Medium Voltage Switching Transient Induced Potential Transformer Failures; Prediction, Measurement and Practical Solutions

Published by Daniel McDermit1, David D. Shipp2, Thomas J. Dionise3, Visuth Lorch4,
1Sr. Project Manager Turner Construction Company Chicago, IL
2PE Eaton Electrical Group Power Systems Engineering Warrendale, PA Fellow, IEEE
3PE Eaton Electrical Group Power Systems Engineering Warrendale, PA Senior Member, IEEE
4Eaton Electrical Group Power Systems Engineering Warrendale, PA

Abstract— During commissioning of a large data center, while switching medium voltage circuit breakers without any appreciable load, several potential transformers failed catastrophically. A detailed investigation, including computer simulation was performed. Ferroresonance produced by switching transients associated with opening and closing the vacuum breakers, was determined to be the cause. The analysis also determined that the close-coupled power transformers were also in jeopardy. Field inspections involving grounding improvements coupled with solution simulations were made. High speed switching transient measurements were performed to verify the analysis and the surge protective devices solution (arresters and snubbers). This paper walks the reader through problem recognition, simulation, field measurements and solution implementation. Special focus will be made on the field measurements verification.

Keywords- switching transients, vacuum breakers, potential transformers, ferroresonance, EMTP simulations, surge arresters, RC snubbers.

I. INTRODUCTION

A. Facility Description

The ultimate build out of this facility strategically located in the midwest is approximately 400,000 sq. ft. of colocation (CoLo) data center white space and containers that will be supported by 250,000 sq. ft. of Central Utility Building (CUB) and approximately 25,000 sq. ft. of office and support space. The data center will ultimately support approximately 59.4 MW of critical load.

The first completed phase of the facility, Phase I, is approximately 200,000 sq. ft. of data center space, which consisting of 95,000 sq. ft. of standard CoLo type white space and 96,000 sq. ft. of container based data facility, as well as 125,000 sq. ft. of utility/mechanical plant space. The utility/mechanical plant consists of 1) 12 chillers and cooling towers and associated equipment, 2) thermal storage tanks, 3) air handlers, and 4) critical electrical support equipment including 12 generators, UPSs and PDUs. This equipment supports four CoLo rooms of data center white space. The mechanical and electrical infrastructure also supports the equivalent of seven CoLo rooms of critical load in approximately 50 to 60 containers (each container is capable of supporting 350 kW of critical load).

B. Commissioning History

On February 18, 2009 the construction team, commissioning team, testing engineers and contractors were performing manual open transition testing from one of the data centers three main utility feeds to Generator power. The Appendix gives an excerpt from the Sequence of Operation testing that was being performed when the first Unit Sub Station (USS) had an event where a Potential Transformer (PT) failed. The Appendix describes each step in detail.

The technician performing the aforementioned sequence of operations, specifically Step #32, noticed that there was an issue with the generator kW reading on the digital feeder protection relay which caused him to stop and investigate with the NM1 (Normal Main – Primary Unit Supply) closed, the GM (Generator Main) open and the water-cooled load banks at 600 kW. During his investigation in the Unit Substation Room, smoke began to appear from the bus PT truck, Phase A. The PT was easily de-energized by opening NM1 in manual mode. A quick investigation indicated that there were no primary or secondary fuses blown on the PT.

A similar incident occurred two days later while the technician was setting-up to perform the same testing sequences on another USS. The technician had the generator tags cleared and started the associated USS generator to verify parameters. In manual mode the technician open-transitioned the source back to Utility A. Within a matter of minutes the technician heard something “pop” and then noticed smoke appearing through the vents in the NMI (Utility A) section. Upon investigation of the USS, it was noted that there was a PT failure in NM1.

C. Recognition of a Problem

The project team experienced a series of PT failures that occurred while the commissioning team and equipment vendors were performing open transition sequence testing while utilizing the feed from UTS-A and the feed from UTS-C. The stress on the project team became heightened and everyone from the end user down to the technicians working on the equipment were driven to discover the reason for the PT failures. The question arose, had anyone on the project experienced such an event or had knowledge of similar events? A few of the team members recalled similar events while working on their last data center project which reminded them of the PT failures experienced here. In summary, on the previous data center project, the authors performed an engineering analysis at their facility to simulate the problem observed at the present jobsite.

In the previous report, the authors stated “We are very much aware of how vacuum breaker induced switching transients can cause transformer failure. We have been doing many computer simulated switching transient studies recently to quantify the problem and to verify the solution. A key element to the failure mechanism is short cables between the vacuum breaker and the transformer”. This immediately sent up a red flag to the team electrical manager and others on the team, since all of the PT failures occurred around the time of open transition (switching) operations. Additionally, the distance between the NM1 breakers and the primary side of the transformers is relatively short (contained within the substation itself). We discovered that this facility was potentially experiencing a phenomenon called “Transient Voltage Restrike”. In transient voltage restrike, the combination of variables that can cause transformer failures usually involve a vacuum circuit breaker interrupting inductive loads that are supplied by cast resin power class transformers.

D. Failures of PTs During Vacuum Breaker Switching

Fig. 1 shows the PTs which failed during vacuum breaker switching. The photo shows the PT sustained the damage. The fuses did not blow and remained intact. Typically, a PT may have on the order of 8000 turns per winding. Close examination of the PT shows the damage consisted of a series arc. Such damage is indicative of ferroresonance. When the breaker opens, a DC charge is trapped on the stray capacitance of the cable, which is imposed on the primary winding of the PT. The DC trapped charge saturates the iron of the PT, which fatigues the winding insulation. The frequency of the PT ferroresonance observed at this facility was about 20 Hz. In this special case of ferroresonance called PT saturation, the PT may draw only 0.1A which is not enough to blow the fuse on the PT primary. Consequently, the series arc could last for hours or up to weeks until the insulation breaks down, at which point ionized gases are produced, and a complete fault occurs blowing the fuse. In the worst case, the ionized gasses contained in this confined space develop an arc phase-to-phaseto- phase, i.e. a three-phase fault. This three-phase fault would cause significant damage to the switchgear. Such a failure would result in significant downtime, reducing the reliability of the power delivered to the mission critical loads.

Figure 1. Damage to PT but Fuses Remained Intact

Fortunately, the PT failure mode at this facility was not catastrophic. Instead, after vacuum breaker switching, the following was observed on two separate occasions: 1) the fuse did not blow, but smoke came out of the PT compartment, and 2) the fuses blew before the ionized gasses could take out the entire cell.

E. End User Response

Although the PT failures did not result in massive failure, the end user was concerned the PT failure could have escalated into severe damage of the switchgear. The end users response was to investigate the event, develop a test procedure to investigate the root cause of the failed PTs, and to ultimately recommend a solidly engineered correction to the sequence of operation or a re-engineering of the electrical gear itself.

II. SWITCHING TRANSIENT THEORY

A. Decision to do a Study

Fig. 2 summarizes the PT failures at Substations USS1B and USS8B on the line-side and load side of the 1200A vacuum circuit breakers. At USS1B, there were both line-side and load-side PT failures. At USS8B, there was one line-side PT failure and three load-side PT failures. The figure calls attention to the issues: 1) a large number of PT failures at two different substations, 2) the failures involved both line-side and load-side failures, and 3) the failures occurred with regularity during switching of the 13.2 kV feeders when transitioning from utility to generator. For these reasons, it was decided to create a model of the power system and conduct a switching transients study to determine the root cause of the PT failures. The decision to conduct a switching transients study permitted simulation of a variety of switching events, determination of sensitivity of parameters contributing to the failures and demonstration of the effectiveness of proposed solutions. Further field testing was placed on-hold until the study was completed and the system response understood.

Figure 2. Summary of PT Failures at USS1B and USS8B

B. Background on Ferroresonance

The primary focus of this study was ferroresonance. ANSI/IEEE Std 100-1984 defines ferroresonance as “a phenomenon usually characterized by overvoltages and irregular wave shapes and associated with the excitation of one or more saturable inductors through a capacitance in series with the inductor.” The key elements are saturable inductors in series with capacitance. The nonlinear inductance (XL) is usually associated with the core of a transformer. The transformer core will saturate with flux as voltage increases. The transformer has a saturation curve, which gives flux as a function of voltage. XL has high value for nonsaturation, and XL has low value when the core saturates. The saturation curve has a “knee” where the change takes place. Transformers are designed to operate near the “knee”. In the ferroresonant circuit, the capacitance (XC) can be the capacitance of cable, overhead line, or stray capacitance of transformer windings or bushings.

Under normal operation, XC is smaller than XL. However, if some switching event causes the voltage to increase, then the transformer core may be pushed into saturation and XL lowered. It is possible at some higher voltage this lower saturated value of XL may equal XC, forming a series resonant circuit called ferroresonance. As in the normal series resonant circuit, the source voltage (VS) does not change much, but the voltage for nonlinear inductance (VL) and voltage for stray capacitance (VC) increase and oppose each other. Since ZL is nonlinear, the voltages VL and VC become distorted or irregular.

Some type of system disturbance is needed to “jolt” the transformer XL into a lower saturated value equal to the system capacitance XC. This “jolt” allows XL = XC and ferroresonance starts. Ferroresonance can continue for a long time (minutes, hours or even days) since little resistance (R) is in the circuit to damp the oscillations.

In the case of this medium voltage distribution system, the nonlinear inductance is either the 14,400/120 V PT or 13.2/0.48 kV, 3000/4500 kVA power transformers. The capacitance is dominated by the stray capacitance of the cables. The “jolt” needed to initiate the ferroresonance is the opening of the vacuum circuit breaker which feeds downstream PTs and the power transformer.

C. Modeling Ferroresonance

In modeling this medium voltage distribution system for such ferroresonance analysis, it was important to accurately represent the opening of the primary vacuum breakers, stray capacitance of the cable, and nonlinear inductance of the transformer being switched, i.e. transformer saturation. The authors modeled these critical circuit components in the Electromagnetic Transients Program (EMTP).

The authors’ study approach was to first model the steadystate conditions with the transformer (PT or power transformer) energized. In this way, it was possible to show the normal excitation current drawn by the transformer to magnetize the nonlinear inductance. Next, the authors simulated the actual switching conditions which produced ferroresonance during transition from utility source to generator source. These opening conditions produced ferroresonance, as evidenced by erratic voltage and current waveforms shown later in the paper. Finally, the authors added mitigation in the form of RC snubbers to provide damping and mitigate the PT ferroresonance.

The actual switching conditions consisted of opening the 13.2 kV utility feeder to USS8B, and closing the 13.2 kV generator feeder breaker. The worst case for ferroresonance occurs with an unloaded transformer. Either a PT or power transformer may experience ferroresonance. On this basis, Eaton examined the electrical distribution system and selected the worst case conditions to check for PT and power transformer ferroresonance. Snubbers sized for this worst case will protect the PT and power transformers during less severe switching operations.

D. Computer Simulations of Actual Conditions

Switching transients simulations were conducted in the electromagnetic transients program (EMTP) to investigate the possible failure of the PT due to transient overvoltages during the circuit switching of the vacuum circuit breakers. The circuit model developed in EMTP consisted of the source, breaker, cable, PT, and Transformer T-8B. The cable was represented by a Pi model consisting of the series impedance and half of the cable charging at each end. (In some cases, multiple Pi models are used to represent the cable.) The vacuum breaker was represented by a switch with different models for opening (current chop of 5A), re-strike (excessive magnitude of TRV), re-ignition (excessive frequency of TRV) and closing (pre-strike). The three phase transformer model consisted of the leakage impedance, magnetizing branch, winding capacitances from high-to-ground and low-to-ground. The PT model included saturation effects. Actual switching scenarios that resulted in PT failures were simulated, and the results of these simulations are described below.

a) Open utility-side 13.2 kV feeder to USS8B followed by closing generator-side feeder (open transition to generator)

Case N6 simulates the actual case during the testing even though the transition time in the simulation is much shorter than the actual time. Case N7 is the same as Case N6 but with a snubber installed. Fig. 3 compares the study results for Cases N6 and N7 for the primary voltage at the 3,000 kVA dry type transformer T-8B. The application of the snubber circuit (Case N7), greatly reduced the transient overvoltage (TOV) magnitude at the 13.2 kV bus and the oscillation frequency. The oscillation frequency of roughly 22,000 Hz (Case N6) can be reduced to 2,000 Hz (Case N7) and the resistor in the snubber circuit will damp the oscillation within 3 milliseconds.

A closer examination of the transient overvoltage is given in Fig. 4. Fig. 4 compares the study results for Cases N6 and N7 for the primary voltage at the 3,000 kVA dry type transformer T-8B zoomed from 0 to 30 milliseconds. The figure illustrates the opening of the utility side circuit breaker.

Figure 3. TOV at USS8B During Transition from Utility to Generator with and without Snubber Protection
Figure 4. TOV at USS8B During Transition from Utility to Generator with and without Snubber Protection (Zoom of Utility Side Breaker Opening)

The application of the snubber circuit (Case N7) reduced the transient overvoltage from 170 kV peak with an oscillation frequency of 1,594 Hz to 26.6 kV peak with an oscillation frequency of 215 Hz. The study also shows a high DC offset for Case N6 due to the energy transfer between the stray capacitor and inductance in the Transformer T-8B and PT. The magnitude of the transient overvoltage may be smaller due to the operation of the surge arrester, however the oscillation frequency will remain the same.

Similarly, Fig. 5 compares the study results for Cases N6 and N7 for the primary voltage at the 3,000 kVA dry type transformer T-8B zoom from 90 to 120 milliseconds. The figure illustrates closing of the generator side circuit breaker. The application of the snubber circuit (Case N7) reduced the oscillation frequency from 20,000 Hz (Case N6) to 2,000 Hz (Case N7) and the period of transient can be was reduced from oscillatory down to 3 milliseconds. The oscillatory condition may cause the PT failure over a long period.

2) Close generator-side 13.2 kV feeder to USS8B (enhanced model of open transition to generator)

Case N8 simulates closing the generator-side feeder breaker with an enhanced model of open-transition. Case N9 is the same as Case N8 but with the snubber installed. Fig. 6 compares the study results for Cases N8 and N9 for the primary voltage at the 3,000 kVA dry type transformer T-8B. The application of the snubber circuit (Case N9), reduced the oscillation frequency from 22,700 Hz (Case N8) to 1,486 Hz (Case N9) and the resistor in the snubber circuit will damp the oscillation within 3 milliseconds. Again, the oscillatory condition may cause the PT failure over a long period.

Fig. 7 illustrates the study results for Case N8 for the primary voltage at the 3,000 kVA dry type transformer T-8B. The figure was zoomed from 5 milliseconds to 10 milliseconds and the resonance condition is clearly illustrated in the figure. The simulated ferroresonance condition in Fig. 7 is a near match for the ferroresonance condition captured with the highspeed power quality meter in Fig. 9.

Fig. 8 compares the study results for Cases N8 and N9 for the bus voltage at USS-1B. Since this location is only 10 feet from Bus USS-8B, the results will be similar to the primary voltage at the 3,000 kVA dry type transformer T-8B. The application of the snubber circuit (Case N9), reduced the oscillation frequency from 22,700 Hz (Case N8) to 1,485 Hz (Case N9) and the resistor in the snubber circuit will damp the oscillation within 3 milliseconds. Again, the oscillatory condition may cause the PT failure over a long period.

3) Comparison of PT connections – Open utility-side 13.2 kV feeder to USS8B

The PTs that failed were connected wye-grounded wyegrounded (Yg-Yg). The analysis was expanded to consider the benefit, if any, of the open delta PT connection. In Case T3, there is a nominal load of 100 kW with 0.97 lagging power factor on Transformer T-8B. The 13.2 kV circuit breaker is opened on the utility side feeding transformer T-8B. First, the Yg – Yg PT connection was considered with no snubber. Fig. 9a shows the Transformer T-8B primary voltage. After breaker opening, the transient overvoltage was as high as 103.61 kV with an oscillation frequency of 1,378 Hz. Such a TOV will result in re-ignition of the breaker and a higher TOV.

Figure 5. TOV at USS8B During Transition from Utility to Generator with and without Snubber Protection (Zoom of Generator Feeder Breaker Closing)
Figure 6. TOV at USS8B During Energization of Transformer T-8B from the Generator with and without Snubber Protection

Next, in Case T4 the Yg – Yg PT connection with a snubber was considered. Fig. 9b shows the transformer T-8B primary voltage. After breaker opening, the TOV was reduced from 103.61 kV (Case T3) to 17.413 kV (Case T4). The oscillation frequency was reduced from 1,378 Hz (Case T3) to 214 Hz (Case T4). However, the PT Ferroresonance is not eliminated due to the PT connection of Yg-Yg in the temporarily ungrounded circuit during open transition.

Finally, in Case T5 the open-delta PT connection with a snubber was considered. Fig. 9c shows transformer T-8B primary voltage. After breaker opening, the TOV was reduced from 103.61 kV (Case T3) to 17.416 kV (Case T5). The oscillation frequency was reduced from 1,378 Hz (Case T3) to 214 Hz (Case T5). With the PT connected open delta, the PT Ferroresonance is eliminated. Adding a resistor to the PT equal to 50% of the burden will assist in damping the ferroresonance.

E. Solutions involving snubbers, grounding and ferroresonance elimination

1) Snubbers

Once the need for snubbers was determined, the surge capacitor was selected. The surge capacitorwas selected based on system voltage. Years of experience has shown the industry that certain values work well for given applications. For example, 0.25 micro-Farad surge capacitorworks well for most 15 kV class applications. This value does a good job of slowing down fast transients to be within the dV/dt rating of 15 kV class transformers, switchgear, motors, generators, etc. The capacitor industry makes these surge caps with standard available values. Alternatively, one could carefully evaluate the ring waves and fine tuned the natural frequency to meet a special value, but such an fine tuned approach generally requires a custom build, i.e. special capacitance values with long delivery times. Such custom values are not required, and the long lead times did not meet the schedule of the project. Once the need for snubbers was determined, the first attempt was to use standard available surge caps for that voltage level.

First, the resistor was sized to match the surge impedance of the source cable feeding the circuit. In this case, two (2) cables were address. The surge impedance of the utility source cables was about 30 ohms while the generator cables was about 20 ohms. Since the snubber must take care of both conditions, 25 ohms was selected as a compromise. The authors have found that a perfect match is not required, but rather a few ohms difference is effective. By matching the resistor to the cable surge impedance, the high frequency waveform impinging on the circuit will not double due to reflection. The resistor cancels the reflection as well as provides damping, especially for the DC offset voltage, that causes ferroresonance.

Next, the snubber values were entered into the computer model in EMTP. (EMTP is recognized world-wide as the premier tool for transient analysis of power systems.) In ETMP, the switching transients were simulated to ensure that the selected snubber components actually accomplished their objective. The discharge current through the resistor was evaluated for three criteria: 1) the Joule (or equivalent watt rating) must be well within thermal ratings, 2) the switching transient voltages must be well below BIL and 3) the dV/dt must be well below transformer limits. All three criteria must be satisfied. For 15 kV transformers, if the ring wave frequency can be reduced to between 250 Hz and 1000 Hz, then there will be no problem. This final simulation also proves that the natural frequency is not too low to cause other problems. Once the surge capacitor was added, its capacitance dominated over the other capacitances in the entire circuit, and reduced the oscillation frequency to an acceptable value.

The specifications for the RC snubber circuit are given in Fig. 10. The resistor average power rating is 750 W at 40 °C, and the peak energy rating is 16,000 Joules. This snubber was applied at other locations in the system. Although cable surge impedance varies somewhat from location to location, this snubber circuit provides adequate mitigation of the voltage transients produced by the vacuum breaker at this location. Selection of one set of RC snubber values simplifies the design and installation, i.e., only one type of resistor and capacitor must be used. The 13.2 kV RC snubber circuit is required at every 13.2 /0.48 kV transformer.

Figure 7. TOV at USS8B During Energization of Transformer T-8B from the Generator with and without Snubber Protection (Zoom of ferroresonance)
Figure 8. TOV at USS1B During Energization of Transformer T-8B from the Generator with and without Snubber Protection

2) Grounding

The wiring methods and grounding means for the 13.2 kV system were inspected. The evaluation included the method of grounding, effectiveness of all connections, presence of ground loops, and compliance with the NEC, FIPS 94, and other applicable standards. The evaluation determined both the utility substation and the generation plant had well designed ground mats that provided effective grounding for their respective electrical equipment. However, it was noted that the ground mats were not tied together. Conduit was relied upon to make this connection, but the use of PVC in several locations broke the continuity.

Figure 9. Open utility-side 13.2 kV feeder to USS8B with a) Yg – Yg PT, b) Yg – Yg PT with snubber, and c) open-delta PT with snubber

As a result, ground mats would be exposed to transferred earth potentials during switching and lightning, imposing a potential difference on such components as surge arresters, cable shields, and Yg-Yg PTs. Transferred earth potentials refers to the phenomena of the earth potential of one location appearing at another location where there is a contrasting earth potential. The transfer of potentials may occur over conductors that have been purposely placed between the two locations. For example, consider the situation of Fig. 11. A ground fault occurs at the remote substation between the primary transformer busing and its respective ground mat. The return path for the ground fault current is through the earth back to the main substation ground mat and its respective system ground.

Figure 10. Snubber Specifications and Surge Arrester Arrangement for the Transformer Protection
Figure 11. Transferred Earth Potentials a) Typical substation, b) fault current path, and c) fall of potential distribution.

This earth return path will have a finite amount of impedance Zg = Rg + j Xg where Rg is the earth resistance and Xg is the earth reactance between the two substation ground mats. The potential difference between the mats is given by:

ΔEg = IgZg (1)

Any electrical conductors, which may extend from within one substation to within the other, may be subject to this potential difference ΔEg. When ΔEg is high, due to excessively high ground fault currents or excessive earth impedance Zg, that serious voltage stresses may be imposed upon various system components never designed to withstand them. For this facility with independent ground mats, high frequency impulses on vacuum circuit breaker closing may cause momentary voltage differences between the ground mats. This imposes a potential difference on the PTs well above the rated line-to-ground voltage, forcing the PT into saturation.

III. MEASUREMENTS

A. Before Solutions

Prior to the installation of the snubbers, several transients were captured on one of the Unit Sub Stations. One of these is shown in Fig. 12. It was noted that the lower load levels resulted in higher over-potential and higher dV/dt signals being developed from the breaker operations. Before installing snubbers, the captured screen shot of Fig. 12 was obtained when utility feeder NM-1 was opened at 300 kW showing a high level of dV/dt signal. It was noted that the duration and peak voltage of the transient had increased from previous test sequences at higher load levels. The peak voltage measured was 49.6 kV, and the duration was 876 μsecs.

Figure 12. Switching Transient Capture at PT
Figure 13. Voltage Dividers and PQ Meter for High-Speed Transient Capture

B. Test Equipment Selection

The system under test consisted of a 4.5 MVA dry transformer connected delta-wye with PTs connected Yg-Yg. The medium voltage power transfers were performed by vacuum breakers. Three (3) 13.2 kV sources power the substation: 1) utility source UTS-A to the NM1 breaker, 2) UTS-C to NM2, and 3) GEN11B through the generator output breaker (GM1) to the substation breaker GM. Test probes were installed on each phase of the line side of each substation breaker, and one set at the line side of the transformer. Each test set consisted of a laptop, a power quality meter, three voltage divider (100/1 ratio) probes, and connecting wires as shown in Fig. 13 and described further in Table I.

C. Personnel Safety Issues at 13.2 kV

To perform this test, the specialized test equipment was installed at USS-11B. The installation of these devices significantly changed the hazard in the area. Specialized pretest briefings were implemented to verify all participants were aware of their roles, as well as the hazards of the test procedure. Additionally, steps were implemented to isolate the USS-11B from USS-7B. This simplified the test and the model. Table II gives the most important prerequisites and precautions.

TABLE I. EQUIPMENT FOR SWITCHING TRANSIENTS MEASUREMENTS

.

TABLE II. TEST PREREQUISITES AND PRECAUTIONS

.

D. Test Procedures to Assure Testing will not Fail the PT or 3 MVA transformer (sneak up on problem)

A testing scheme was developed to create individual breaker closing and opening electrical transients at four discrete load levels on USS-11B: 1200, 600, 300 and 100 kW. These events were evaluated by all parties involved, and concurrence was given to proceed to the next test step. This test was designed to approach the potentially more damaging conditions in a deliberate manner. Prior to conducting the actual tests, each test condition was simulated in EMTP. Table III summarizes the results of the simulations of the test conditions. As can be seen in the results, as the load level on the transformer was reduced, the severity of the switching transient overvoltage, ring frequency and ferroresonance condition increased. This advance knowledge of the system response enabled careful tracking of the field test results. Should the field results depart dramatically from the predicted response, the test could be stopped. In this way, it was possible to “sneak-up” on the problem.

E. Confirmed Problem Existed

Testing began with no snubbers and the 1200 kW load and proceeded down to the 300 kW load. The test sequences prior to the snubber installation were terminated at this point at the concurrence of the design engineers, facility operators and the authors because the symptoms of TOVs and dV/dt were worsening with each lower load increment. Fig. 12 shows the worst of transient overvoltage captured at the PT. All parties were satisfied the root cause of the PT failure was determined without actually failing a PT. Next, the snubbers were installed, and the captured screen shot of Fig. 14 was obtained when utility feeder NM-1 was opened at 300 kW load. The snubbers eliminated the dV/dt transient response, and no TOVs occurred at the PTs as shown in the voltage waveforms of Fig. 14. All parties involved were satisfied the snubber was effective in mitigating both the dV/dt and TOV problem. The client was confident that system reliability would not be compromised.

IV. SOLUTIONS

TABLE III. SUMMARY OF TEST CONDITION SIMULATIONS

.

Given that the facility was in the final stages of commissioning, a solution had to be both economical and timely. Using this criteria, three technically sound solutions were proposed and evaluated. Ultimately two of three proposed solutions were implemented as explained below.

A. Change to Different PT Connections (rejected)

The PTs that failed were connected Yg-Yg. In fact, all of the PTs on the 13.2 kV circuits were of the same configuration. The analysis has shown the benefit of the open-delta PT over the Yg-Yg PT, even when the snubber is applied in both configurations. With the open delta PT, the ferroresonance was completely eliminated. However, with the Yg-Yg PT ferroresonance of about 20 Hz persists. The authors discussed replacing the existing the Yg-Yg PTs with open delta PTs. However, it became apparent this would be cost prohibitive for several reasons: 1) the relays would have to be changed, 2) the metering would have to be changed, 3) the commissioning would have to be repeated because the commissioning was in the final stages, and 4), the project was nearly completed and such a change-out would add significant delay. For these reasons, the possibility of changing to different PT connections was rejected.

B. Tie Ground Mats Together (implemented)

As explained, the different ground mats, i.e. utility substation ground grid and generation plant ground grid were not tied together resulting in transferred earth potentials. Such transferred earth potentials would occur during switching and lightning, imposing a potential difference on such components as surge arresters, cable shields and Yg-Yg PTs. One solution was to increase the insulation strength of the affected equipment. Another option was to tie the ground mats together. The former option would require replacement of equipment which was undesirable for many of the same reasons mentioned in regard to changing the PT connections. The latter option required minimal investment in an equipment grounding conductor (EGC) and the labor to install the EGC. For these reasons, the authors recommended tying the ground mats together to minimize the transferred earth potential. An EGC was installed outside the conduit runs between the utility substation and the generation plant, and the EGC was bonded to the corresponding ground mats.

Figure 14. Switching Transient Mitigated by RC Snubber

C. Add Snubbers as Study Dictated (implemented)

The study showed that the application of the snubber circuit greatly reduced the transient overvoltage magnitude and oscillation frequency at the 13.2 kV PT and Transformer T-8B primary winding. The resistor in the snubber circuit damped the oscillation and reduced the DC offset of the transient overvoltage to within acceptable levels. Further, the study showed the snubber reduced the ferroresonance of the Yg-Yg PT by providing a source of damping that would otherwise not be present. The snubber, in combination with the existing surge arrester, provided the maximum surge protection for the PT and the Transformer T-8B. The snubber components were of relatively low cost and readily available. The snubbers did not require any modifications to the PT circuit, relaying or meters. The performance of the snubbers was proved by measurements and did not require repeating the commissioning tests. For these reasons, it was decided to install the snubbers.

D. Prove Solution with After-the-Fact Measurements

Following installation of the snubbers, power quality measurements were taken to ensure the proper operation of the snubbers. A high speed power quality meter and capacitive voltage dividers were used to measure the transient overvoltage waveforms at the PT and transformer primary produced during switching of the primary vacuum circuit breaker. The voltage dividers were made of capacitive and resistive components with a bandwidth of 10 MHz. The power quality meter was capable of transient voltage waveshape sampling, 8000 Vpeak full scale, 200 nsec sample resolution (5 MHz sampling). This test equipment ensured accurate capture of the high frequency transients. The measurements verified the waveforms did not exhibit excessive high frequency transients (magnitude, rate-ofrise and frequency). The measurements also verified the PT ferroresonance condition was damped-out.

Figure 15. RC Snubber (Typical Installation for this Facility)

E. Typical Snubber Installation

Fig. 15 shows the side view of the custom snubber circuit. The custom design was required to fit the snubber in the bottom of the transition section, between the No. 3 breaker section and the transformer. The snubber was field installed in the bottom of the transition section. The surge capacitor serves as the base. The resistor is bonded to the surge capacitor bushing and the bus bar, where the cables to the transformer are secured. The flanges on each end of the resistor provide a solid support for the fragile resistors. The capacitor was grounded to the ground bus through a flat-braided highly-stranded strap. The snubber was treated like a lightning arrester, i.e., no fuse was installed.

V. CONCLUSIONS

This paper provided a detailed investigation of several PTs that failed catastrophically during switching medium voltage circuit breakers to transition from utility to generator. Switching transient simulations and field measurements determined the root cause of PT failure was ferroresonance as well as switching transient overvoltages associated with opening and closing the primary vacuum breakers. The analysis determined that the close coupled power transformers were also in jeopardy. The study showed that the application of the snubber circuit greatly reduced the transient overvoltage magnitude and oscillation frequency at the PT and power transformer primary winding. The resistor in the snubber circuit damped the oscillation and reduced the DC offset of the transient overvoltage to within acceptable levels. Further, the study showed the snubber reduced the ferroresonance of the Yg-Yg PT by providing a source of damping that would otherwise not be present. Also, grounding improvements were made by tying together the ground mats in the utility substation and generator plant, to minimize the transferred earth potentials imposed on the PT primary winding. High speed switching transient measurements verified the analysis and proved the surge protective devices solution (arresters and snubbers) were effective.

VI. ACKNOWLEDGMENTS The authors wish to thank Turner Facilities Management Solutions as well as Eaton Power Systems Engineering for their collaboration in the presentation of this work: Dan Wilder, Tim McConnell, and Rashad Hammoudeh of Turner as well as Bill Vilcheck of Eaton. Additionally, we thank Scott Seaton.

REFERENCES

[1] IEEE Standard Dictionary of Electrical and Electronics Terms, ANSI/IEEE Std 100-1984.
[2] Hopkinson, R.H., “Ferroresonant Overvoltage’s Due to Open Conductors,” General Electric, 1967, pp. 3 – 6.
[3] Westinghouse Distribution Transformer Guide, Westinghouse Electric Corp., Distribution Transformer Division, Athens, GA, June 1979, revised April 1986, Chapter 4 Ferroresonance, pp. 36 – 40.
[4] IEEE Guide for Application of Transformers, ANSI/IEEE C57.105-1978, Chapter 7 Ferroresonance, pp. 22 – 28.
[5] Distribution Technical Guide, Ontario Hydro, Ontario, Canada, May1999, original issue May 1978, pp. 72.1-1 – 72.1-10.
[6] Greenwood, A., “ Electrical Transients in Power Systems”, Wiley & Sons, 1971, pp. 91-93.
[7] Kojovic, L., Bonner, A., “Ferroresonance – Culprit and Scapegoat”, Cooper Power Systems, The Line, December 1998.
[8] Shipp, Dionise, Lorch and MacFarlane, “Transformer Failure Due to Circuit Breaker Induced Switching Transients”, IEEE Transactions on Industry Applications, April/May 2011.
[9] ANSI/IEEE, A Guide to Describe the Occurrence and Mitigation of Switching Transients Induced By Transformer And Switching Device Interaction, C57.142-Draft.
[10] D. Shipp, R. Hoerauf, “Characteristics and Applications of Various Arc Interrupting Methods,” IEEE Transactions Industry Applications, vol 27, pp 849-861, Sep/Oct 1991.
[11] ANSI/IEEE, Standard for AC High-Voltage Generator Circuit Breakers on a Symmetrical Current Basis, C37.013-1997.
[12] ANSI/IEEE, Application Guide for Transient Recovery Voltage for AC High-Voltage Circuit Breakers, C37.011-2005.
[13] D. Durocher, “Considerations in Unit Substation Design to Optimize Reliability and Electrical Workplace Safety”, ESW2010-3, 2010 IEEE IAS Electrical Safety Workshop, Memphis.
[14] D. Shipp, N. Nichols “Designing to Avoid Hazardous Transferred Earth Potentials,” IEEE Transactions Industry Applications, July/August 1982.
[15] Shipp, Dionise, Lorch and MacFarlane, “Vacuum Circuit Breaker Switching Transients During Switching of an LMF Transformer”, IEEE Transactions on Industry Applications, January/February 2012.

APPENDIX Table IV gives an excerpt from the Sequence of Operation testing that was being performed when the first Unit Sub Station had an event where a Potential Transformer (PT) had a failure. The PT failure occurred during the performance of Step #32 described in the table below.

TABLE IV. SEQUENCE OF OPERATION TESTING

.

Source & Publisher Item Identifier: 48th IEEE Industrial & Commercial Power Systems Conference. DOI: 10.1109/ICPS.2012.6229608

Impact of Electric Vehicle Charging Station on Power Quality

Published by M. S. Arjun1, N. Mohan1, K. R. Sathish2, Arunkumar Patil3, G. Thanmayi4.
1Department of Electrical and Electronics Engineering, JSS Science and Technology University, Mysore, India
2Department of Electrical and Electronics Engineering, ATME College of Engineering, Mysore, India
3Department of Electrical Engineering, Central University of Karnataka, Kalaburagi (Gulbarga), India
4M. Tech-Energy System and Management, Department of Electrical and Electronics Engineering, JSS Science and Technology University, Mysore, India

ABSTRACT: Global warming has led to the widespread adoption of electric vehicles (EV). With the increasing use of electric vehicles, it is very important to understand the impact of electric vehicle charging. Electric vehicle charging station has a serious effect on the power quality of the local power distribution network, and it cannot be ignored. The electric vehicle charger is a type of non-linear load. This non-linearity introduces harmonics into the charging station. Therefore, a high-efficiency charger in the power grid is required. This research work aims to build a charging station model to analyze the effect of EV chargers on power quality and then shunt active power filter (SAPF) based on P-Q theory and synchronous reference frame (SRF). Theory is implemented in the system to suppress harmonics. The simulation will be carried out under two cases, without active power filter (APF) and with APF when number of chargers associated to the charging station. The simulation results of both the methods will be compared and verify the effectiveness of proposed method. The simulation will be done using the MATLAB/Simulink software.

Keywords: Charging station, Electric vehicle, Power quality, P-Q theory, SAPF, SRF theory

1. INTRODUCTION

Due to huge development in transportation sector by considering environment protection and depletion of fossil fuel, traditional vehicles shift to electric vehicle. Electric vehicles (EVs) are less environment pollution, cheaper mode of transportation and growth in battery technology it gaining much attention in the market. There are three main types of electric vehicles within the worldwide [1].

The continuous development of EVs can have adverse effect on distribution network. EV chargers contain electronic devices similar to non-linear load which causes harmonic problem and affecting power quality in the system [2], [3]. Huge EV charging load’s impact on the three-phase power system is analyzed for different parameter on chargers, number of chargers. It is concluded in that APF+PI repetitive control can efficiently suppress the current harmonics and improve power quality [4], [5]. In the concept of vehicle-to-grid (V2G) is studied and analyzed parameters like frequency stability, voltage stability, and power quality are discussed for different penetration level of plug-in electric vehicles (PEV) [6], [7].

Fujun et al. [8] analyzed the influence of harmonic on load by fast charging station, where it is concluded that when the fast-charging station has a large capacity, harmonics have a greater impact on load. The researchers investigated the impact of harmonic distortion on EV charging station based on the measured data and actual test, result was simulated in MATLAB/Simulink software [9]–[11].

Woodman et al. [12] simulated the harmonic caused by non-linear load in IEEE 13 node test feeder that used to verify the VHDL-AMS model and their effects. In the switched filter compensator (SFC) are used to solve the problem of harmonic distortion network [13], [14]. It compensates reactive power to improve power factor, the results were analyzed and compared with and without SFC. This study examines how more EV charging stations affect power quality is examined using MATLAB/Simulink software. Active power filter (APF) based P-Q theory and synchronous reference frame (SRF) theory is adopted to reduce the harmonic in the system for different load conditions.

This paper is organized as:

i) Section 1 provides introduction;
ii) Section 2 detailed problem formulation for EV charging station is explained;
iii) Section 3 presents proposed methodology and brief review of shunt active power filter (SAPF);
iv) Further analysis of simulation results is discussed in section 4; and
v) Followed by conclusion in section 5.

2. PROBLEM FORMULATION

With the higher penetration of electric vehicle day-by-day it has so many disadvantages. The EV charger is a non-linear load, that introduces harmonics into the system. This will influence the power quality of the framework. The research work evaluates the harmonic contamination in the electric system due to EV chargers. The charging station model connected to shunt active power filter [P-Q] theory and synchronous reference frame theory is implemented and simulated in MATLAB/Simulink software. Here the current harmonics are determined for different load conditions.

3. PROPOSED METHODOLOGY

The proposed work is carried out on shunt active power filter using two methods i) synchronous reference frame (D-Q) theory and ii) instantaneous reactive power (P-Q) theory, to suppress the harmonic distortion and improve the power quality in the distribution system. In the modelling phase following are the assumptions, three-phase source with the rating of 400 V 50 Hz AC and lithium-ion battery which is used in almost all the battery EVs.

3.1. Charging station using P-Q theory-based shunt active power filter

The most reason of the shunt active power filter is to generate compensation currents of the same magnitude but in opposite phases to suppress current harmonics [15]. In this research work P-Q theory method is implemented, also known as instantaneous reactive power theory, and creates a simulation model using MATLAB/Simulink. Figure 1 shows a block diagram of an electric vehicle charging station connected to SAPF. To generate compensating reference current, Clarke transformation is used to convert the, source voltage and current from a-b-c coordinates to alpha-beta coordinates, as shown in [16]–[18].

.

Here zero sequence component is absent as α and β axes make no contribution to zero component. Hence it can be easily eliminated from the system.

.

To generate harmonic reference current active power component (𝑝̅𝑎𝑐), total reactive component (q) and real power (𝑝̅𝑙𝑜𝑠𝑠) from 3Φ AC source is required given as (4).

.

Generated compensating current are then transformed from α-β-0 to 3Φ a-b-c frame by inverse Clarke transformation as shown in (5).

.
Figure 1. Block diagram of P-Q theory

3.2. Charging station using synchronous reference frame theory-based active power filter

A rotating coordinate system’s ability to cause harmonics to change frequency is the foundation of the synchronous reference approach. The reference current has nothing to do with the supply voltage and it is directly derived from the actual load current [19], [20]. An electric vehicle charging station connected to an SRF-based active power filter is shown in block diagram from Figure 2.

Figure 2. Block diagram of SRF theory

To generate the compensating reference current, the 3-phase supply current ia, ib, and ic are transformed into 2Φ (α-β) current in the stationary frame [21]–[25].

.

Next the α-β plane converted into D-Q rotating frame, the phase locked loop (PLL) circuit is applied to generate sine and cosine signals to properly synchronize the current with the voltage. The current expression is given by (7).

.

Here the produced d-axis and q-axis component consist of both dc and ac component. The AC component goes through the low pass filter to remove the harmonic component, and the PI controller is used to reduce each harmonic component’s steady state error.

.

Once the desired harmonic elements are removed from the distorted load current, the D-Q rotating frame is converts back to the a-b-c stationary frame.

.

The extracted reference current is transferred for generation of switching pulses for the inverter.

4. RESULTS OF THE SIMULATION

In this study, a simulation model is created in the simulation toolbox MATLAB2018b to analyze the harmonics produced during the charger’s charging stage. The simulation parameters utilized in this model are displayed in Table 1. By analyzing the total harmonic distortion (THD) content of the two charging stations, comparing the THD values of the two charging stations and analyze which charging station has the lowest THD content. According to the simulation results, the current control approach can deliver the optimum results. Figure 3(a) displays the source side current waveforms before the current is compensated and Figure 3(b) displays the source side current waveforms after the current is compensated.

Table 1. Simulation parameters used for design

.
Figure 3. Source current Is (a) before compensation and (b) after compensation

4.1. Charging station using SAPF

Figure 4 shows the generated compensation currents required for elimination of harmonic distortion present in the system. The algorithm is carried out MATLAB/Simulink software. The fast Fourier transform analysis in Figure 5 fast Fourier transform (FFT) analysis of source current Is, Figures 5(a) and 5(b) shows the source current Is is before compensation, after compensation and total harmonic distortion reduced from 28.24% to 6.23% using APF based on P-Q theory.

Figure 4. Waveform of the compensation current
Figure 5. FFT analysis of source current Is (a) before compensation and (b) after compensation

4.2. Charging station using SRF theory

Figure 6 shows the output waveform of the sine and cosine signals used for synchronization and generate the required reference current signal. Figure 7 shows that the total harmonic distortion is only 2.92%, which is an acceptable value compared to the IEEE standard of 5% limit. Table 2 shows the comparison based on the connection of filter under different load condition.

Table 2. Percentage THD comparison

.
Figure 6. Sine and cosine function from the PLL circuit
Figure 7. FFT analysis of SRF theory
5. CONCLUSION

This work proposes a solution to mitigate the effect of the increase in EV charging stations on power quality. The simulation model of the electric vehicle charging station is created in the MATLAB/Simulink software. To analyze the harmonic contamination, charging station with P-Q theory based shunt active power filter and SRF theory-based active power filter was simulated. Both the charging station with the same parameters and the batteries are used as a load during charging. The charging station with SRF theory has a lower THD value than the charging station with a P-Q theory-based power filter and without a filter. Synchronous reference theory model can viably relieve the harmonic distortion. The generated reference signal is not distorted in the SRF approach, which improves performance and it has good compensation result for harmonic currents.

REFERENCES

[1] A. K. Karmaker, S. Roy, and M. R. Ahmed, “Analysis of the impact of electric vehicle charging station on power quality issues,” in 2019 International Conference on Electrical, Computer and Communication Engineering (ECCE), Feb. 2019, pp. 1–6. doi: 10.1109/ECACE.2019.8679164.
[2] S. Deb, K. Kalita, and P. Mahanta, “Impact of electric vehicle charging stations on reliability of distribution network,” in 2017 International Conference on Technological Advancements in Power and Energy (TAP Energy), Dec. 2017, pp. 1–6. doi: 10.1109/TAPENERGY.2017.8397272.
[3] L. Durante, M. Nielsen, and P. Ghosh, “Analysis of non-sinusoidal wave generation during electric vehicle charging and their impacts on the power system,” International Journal of Process Systems Engineering, vol. 4, no. 2–3, pp. 138–150, 2017, doi: 10.1504/IJPSE.2017.084751.
[4] F. Chen, L. Ruijie, and L. Guanhua, “Research on harmonic analysis and harmonic suppression measures of electric vehicle
charging station,” in 2019 IEEE 2nd International Conference on Automation, Electronics and Electrical Engineering (AUTEEE), Nov. 2019, pp. 71–75. doi: 10.1109/AUTEEE48671.2019.9033199.
[5] M. Falahi, H.-M. Chou, M. Ehsani, L. Xie, and K. L. Butler-Purry, “Potential power quality benefits of electric vehicles,” Sustainable Energy, IEEE Transactions on, vol. 4, no. 4, pp. 1016–1023, 2013, doi: 10.1109/TSTE.2013.2263848.
[6] E. Alghsoon, A. Harb, and M. Hamdan, “On stability and power quality impacts of vehicle to grid (V2G),” Applied Solar Energy, vol. 53, pp. 100–108, Apr. 2017, doi: 10.3103/S0003701X17020037.
[7] S. Shao, M. Pipattanasomporn, and S. Rahman, “Demand response as a load shaping tool in an intelligent grid with electric vehicles,” IEEE Transactions on Smart Grid, vol. 2, no. 4, pp. 624–631, Dec. 2011, doi: 10.1109/TSG.2011.2164583.
[8] T. Fujun, X. Ruiheng, C. Dong, R. Lijia, Y. Quanning, and Z. Yan, “Research on the harmonic characteristics of electric vehicle fast charging stations,” in 2017 2nd International Conference on Power and Renewable Energy (ICPRE), Sep. 2017, pp. 805–809. doi: 10.1109/ICPRE.2017.8390645.
[9] Y. Zhang, D. Yu, G. Zhang, H. Wang, and J. Zhuang, “Harmonic analysis of EV charging station based on measured data,” in 2020 IEEE/IAS Industrial and Commercial Power System Asia (I&CPS Asia), Jul. 2020, pp. 475–480. doi: 10.1109/ICPSAsia48933.2020.9208498.
[10] A. Lucas, F. Bonavitacola, E. Kotsakis, and G. Fulli, “Grid harmonic impact of multiple electric vehicle fast charging,” Electric Power Systems Research, vol. 127, pp. 13–21, Oct. 2015, doi: 10.1016/j.epsr.2015.05.012.
[11] A. Pan, Y. Zhu, L. Ren, T. Chen, S. Wen, and W. Yun, “Harmonic research of electric vehicle fast chargers,” in 2016 IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC), Oct. 2016, pp. 2545–2549. doi: 10.1109/APPEEC.2016.7779947.
[12] N. Woodman, R. B. Bass, and M. Donnelly, “Modeling harmonic impacts of electric vehicle chargers on distribution networks,” in 2018 IEEE Energy Conversion Congress and Exposition (ECCE), Sep. 2018, pp. 2774–2781. doi: 10.1109/ECCE.2018.8558207.
[13] M. H. Mohamed, M. N. Abdelwahab, and A. A. Abdelsalam, “Mitigation of electric vehicle distortion impact on distribution networks,” in 2019 21st International Middle East Power Systems Conference (MEPCON), Dec. 2019, pp. 921–926. doi: 10.1109/MEPCON47431.2019.9008040.
[14] J. C. Gomez and M. M. Morcos, “Impact of EV battery chargers on the power quality of distribution systems,” IEEE Transactions on Power Delivery, vol. 18, no. 3, pp. 975–981, Jul. 2003, doi: 10.1109/TPWRD.2003.813873.
[15] A. Panchbhai, S. Parmar, and N. Prajapati, “Shunt active filter for harmonic and reactive power compensation using p-q theory,” in 2017 International Conference on Power and Embedded Drive Control (ICPEDC), Mar. 2017, pp. 260–264. doi: 10.1109/ICPEDC.2017.8081097.
[16] A. A. Imam, R. S. Kumar, and Y. A. Al-Turki, “Modeling and simulation of a PI controlled shunt active power filter for power quality enhancement based on P-Q theory,” Electronics, vol. 9, no. 4, p. 637, Apr. 2020, doi: 10.3390/electronics9040637.
[17] J. L. Afonso, C. Couto, and J. S. Martins, “Active filters with control based on the p-q theory,” Institute of Electrical and Electronics Engineers (IEEE), vol. 47, no. 3, pp. 5–10, 2000, [Online]. Available: https://hdl.handle.net/1822/1921
[18] S. Round, H. Laird, R. Duke, and C. Tuck, “An improved three-level shunt active filter,” in 1998 International Conference on Power Electronic Drives and Energy Systems for Industrial Growth, 1998. Proceedings., 1998, vol. 1, pp. 87–92. doi: 10.1109/PEDES.1998.1329995.
[19] P. Dey and S. Mekhilef, “Synchronous reference frame based control technique for shunt hybrid active power filter under nonideal voltage,” in 2014 IEEE Innovative Smart Grid Technologies – Asia (ISGT ASIA), May 2014, pp. 481–486. doi: 10.1109/ISGT-Asia.2014.6873839.
[20] C. Jiang, R. Torquato, D. Salles, and W. Xu, “Method to assess the power-quality impact of plug-in electric vehicles,” IEEE Transactions on Power Delivery, vol. 29, no. 2, pp. 958–965, Apr. 2014, doi: 10.1109/TPWRD.2013.2283598.
[21] S. Musa, M. Radzi, H. Hizam, N. Wahab, Y. Hoon, and M. Zainuri, “Modified synchronous reference frame based shunt active power filter with fuzzy logic control pulse width modulation inverter,” Energies, vol. 10, no. 6, p. 758, May 2017, doi: 10.3390/en10060758.
[22] M. U. Kumar and R. S. Rao, “Simulation of SRF control based shunt active power filter and application to BLDC drive,” International Journal of Latest Technology in Engineering & Management (IJLTEM), vol. 1, no. 1, pp. 14–22, 2016.
[23] N. S. Rao and H. J. Jayatheertha, “Modeling and simulation of various SRF methods for shunt active power filter and application to BLDC drive,” International Journal of Advanced Engineering Research and Studies, vol. 1, no. 4, pp. 18–22, 2012.
[24] A. Shah and N. Vaghela, “Shunt active power filter for power quality improvement in distribution systems,” Shunt active power filter for power quality improvement in distribution systems, pp. 22–26, 2005, [Online]. Available: https://www.ijedr.org/papers/IJEDR1302005.pdf
[25] V. Khadkikar and A. Chandra, “Three-phase and single-phase p-q theories applied to three-phase shunt active power filter under different operating conditions: a comparative evaluation,” International Journal of Emerging Electric Power Systems, vol. 11, no. 2, Feb. 2010, doi: 10.2202/1553-779X.2308.

BIOGRAPHIES OF AUTHORS

M. S. Arjun received the B.E., Master of Technology and pursuing Ph.D. degree in the field of electrical and electronics engineering. He is currently working as Assistant Professor, Department of Electrical and Electronics Engineering, JSS Science and Technology University (SJCE) Mysuru, Karnataka, India. His area of interest is on electric vehicles, power electronics, and renewable energy sources. He can be contacted at email: arjunms@jssstuniv.in.

Dr. N. Mohan received the B.E, Master of Technology and Ph.D. degree in the field of electrical and electronics engineering. He is currently working as Associate Professor, Department of Electrical and Electronics Engineering, JSS Science and Technology University (SJCE) Mysuru, Karnataka, India. His area of interest is on power system network, and electric vehicles. He can be contacted at email: mohan.eee@jssstuniv.in.

Dr. K. R. Sathish received the B.E, Master of Technology and Ph.D. degree in the field of electrical and electronics engineering. He is currently working as Assistant Professor, Department of Electrical and Electronics Engineering, ATME College of Engineering, Mysuru, Karnataka, India. His area of interest is on power system and power electronics. He can be contacted at email: dr.sathish21@gmail.com.

Dr. Arunkumar Patil has received the B.E. Master of Technology and Ph.D. degree in the field of electrical and electronics engineering. He is currently working as Assistant Professor, Department of Electrical engineering, Central University of Karnataka, Kalaburagi, Gulbarga, Karnataka, India. His area of interest is on power systems, renewable energy, WAMS, AI and data science. He can be contacted at email: arun201085@gmail.com.

G. Thanmayi has received the B.E and Master of Technology degree in the field of electrical and electronics engineering. Master of Technology Specialization is Energy System and Management. Her area of interest is on power systems, renewable energy, energy system and management, energy and environment. She can be contacted at email: thanmayigrani98@gmail.com.

Source & Publisher Item Identifier: International Journal of Applied Power Engineering (IJAPE). Vol. 13, No. 1, March 2024, pp. 186~193. ISSN: 2252-8792, DOI: 10.11591/ijape.v13.i1.pp186-193

EDP Distribuição’s Development of Support Tools and Platforms for Power Quality Management and Analysis

Published by Fabrice GONÇALVES, António LEBRE, Pedro VELOSO, Fernando BASTIÃO, Nuno MELO, DP Distribuição – Portugal. Emails: fabrice.goncalves@edp.pt, antoniojose.lebrecardoso@edp.pt, pedro.veloso@edp.pt, fernando.bastiao@edp.pt, nuno.melo@edp.pt

24th International Conference & Exhibition on Electricity Distribution (CIRED). 12-15 June 2017.

ABSTRACT EDP Distribuição, as distribution system operator (DSO), has been developing an ambitious Power Quality (PQ) monitoring program.

This paper aims to present the multiple efforts that have been made by a DSO developing support tools and platforms in order to promote the continuous improvement of the PQ monitoring program. The main goals are to produce fast, robust and reliable analysis, transforming PQ data into suitable information, and improve the power quality management process and its transparency, adding value to the network operation and all network users.

INTRODUCTION

EDP Distribuição (EDP D) is a EDP Group Energias de Portugal company. As the main Distribution System Operator (DSO) in Portugal mainland, EDP D has about 6,1 million of network customers.

To ensure a high level of Quality of Service (QoS), EDP D has been developing systematically, since 2001, a program for Power Quality (PQ) monitoring, according to NP EN 50160 standard and the national QoS Regulation Code [1]. Globally, this program allows to characterize PQ at distribution level, to improve the operation and maintenance of the distribution network, to support the network users and to report PQ information, namely to regulatory authorities and online.

To support this challenging PQ program, EDP D has been developing a suitable PQ monitoring technical platform. With the evolution of monitoring devices and tools and subsequent learning of them by EDP D, this program has had a substantial development. In 2010, the introduction of PQ permanent monitoring in some HV/MV substations was an important step.

EDP D is facing multiple and diverse challenges in the management and development of its PQ monitoring program. Besides the technological evolution taking place, regulatory and standardization changes, demands and recommendations are an important issue to take into account. Sometimes the available technology, hardware or software level doesn’t allow for an immediate answer to that, leading to an onerous and long-term development in partnership with manufacturers.

The program implementation has allowed to obtain a wealth of PQ data and results, building up a very substantial and interesting repository of PQ information and knowledge. From this point of view, several challenges are arising. Some related with the quantity of data (big data), the guarantee of its quality and compatibility with all sources. Others related to the improvement of data analysis in order to allow a full and efficient use of data, and to provide required and reliable information to all stakeholders.

Simultaneously, the dimension and complexity growth drive to big challenges at management level and in guaranteeing of security and transparency. The systematization of PQ processes could be a useful and helpful tool. To face all these challenges, EDP D has been developing support tools and its PQ monitoring platform.

PQ MONITORING PLATFORM CURRENT STATEMENT

EDP D has been progressively developing a PQ monitoring platform in order to fulfil the PQ monitoring program to answer to the increasing demands and challenges. The PQ monitoring platform comprises the PQ monitoring recorders, communication infrastructures and PQ management centre (databases, management and analysis’ software and its related applications). The present global topology is shown in Figure 1. With this architecture it is possible to assure the requirements of the Portuguese QoS Regulation Code, including the most recent changes from the last edition and all other internal requirements. Therefore, systematic monitoring campaigns in substations are performed, as well as permanent monitoring in HV/MV substations.

Portable PQ recorders are used for temporary monitoring campaigns and voltage measurements are performed in MV busbars of HV/MV substations and LV busbars of MV/LV secondary substations.

Figure 1. EDP D’s PQ monitoring platform.

Since 2010, fixed class A PQ recorders according to the IEC 61000-4-30 standard have been installed in all new HV/MV substations and also in those submitted to a major refurbishment, in order to perform permanent PQ monitoring. Voltage measurements are performed in MV busbars, as well as current measurements.

Some customers are supported by a dedicated PQ monitoring in order to perform an accurate characterization of the PQ supplied and to identify improvement actions. Typically, a portable class A PQ recorder is installed for monitoring for at least one month.

About 75% of fixed PQ recorders installed in HV/MV substations are provided with Ethernet communications supported by a dedicated internal communications network. All other PQ recorders (the remaining 25% of the fixed and all the portables) are equipped with a 3G modem, supported by a mobile communications UMTS network.

Thus, all PQ recorders are provided with fast and reliable remote communications to a central server, regardless of their type and location of installation.

All the collected data is stored automatically in a bulk SQL database and processed in order to create PQ reports according to the requirements of the Portuguese regulation and for other types of analysis. These specific reports are produced using a dedicated web based application, which reads data directly from the SQL database.

In order to reinforce the proper storage and the security of PQ data, the PQ management centre was transferred to a Data Protection Centre, allowing backup of the data and an easily expansion because of their virtualized structure.

Since 2014, the PQ plan and results from the monitoring program activities have been published on EDP D’s website.

PQ MONITORING PLATFORM DEVELOPMENTS

In order to guaranty PQ monitoring platform evolution, according to EDP D PQ’s monitoring strategy, some new features are being implemented or under analysis.

Warning system on data quality and PQ limits violation

To have an awareness of the issues that require attention in PQ monitoring platform, EDP D felt the need to develop a warning system that allows, in an automatic way, fast identification of problems in data quality and PQ limits violation analysis, according to parametrized EDP D’s PQ requirements.

Each PQ recorder has a specific software manufacturer, which ensures, among other functionalities, the configuration of some kind of alerts. However, the available alerts are different from one device to another, depending on its manufacturer.

Regarding this limitation, and to ensure the same level of alerts independently of used PQ recorder, the warning system will be developed over the existing web based application where PQ data is processed, in order to provide PQ overview reports.

The specification for this tool includes the following functionalities:

Data Quality – Communication faults with PQ recorders, errors in data collection, percentage of data’s coverage for each PQ recorder, proper recording of voltage events, errors in weekly-based automatic process of data compilation;

PQ analysis – Use of continuous phenomena data and voltage events, compared with definable limits for daily analysis and information of Power Quality Index for analysis according to NP EN 50160 standard;

Daily Automatic Report – Possibility to set the hour for the process’s start, to report on the 24 hours of the previous day and to send by email or to use the web based application for report access and search.

Application of PQ indexes

The growth of monitoring points and consequent increase of available data offers a comprehensive history, allowing analysis with a fairly representative time base, which becomes representative of the network. It is therefore important to have a systematization approach to transform PQ data into relevant information for network management and operation. Thus, it is important to choose and/or develop adequate PQ indexes capable of translating PQ data into necessary PQ information.

Regarding continuous phenomena, the recorded data from each monitoring point is currently transformed into a continuous phenomena Power Quality Index by the web based application. This index translates in numerical terms and for each of the continuous phenomena, how far we are from nonconformity, providing a clear perception of the performance of each monitored point in relation to NP EN 50160 standard, unequivocally, identifying situations of nonconformity. However, and despite the relevant information that this index already produces, it needs to evolve in order to make an easier comparison in historical terms, i.e. the same monitoring point at different periods of time, but also in comparative terms, i.e. from one monitoring point to another.

Regarding voltage events, particularly voltage dips, EDP D has, for many years, consistently followed the evolution of the number of voltage dips per MV busbar at HV/MV substations within the scope of PQ monitoring program. The classification of the voltage dips has been made according to the criteria of NP EN 50160 standard. However, the number of voltage dips per MV busbar does not reflect the number of customers affected by these voltage events, i.e. a voltage dip has the same impact if the MV busbar feeds only one customer or ten thousand.

Table 1. Classification of dips in terms of residual voltage and duration according to the NP EN 50160 standard.

.

In order to overcome this limitation, EDP D is developing the use of SARFI (System Average RMS Variation Frequency Index) concept in MV network. In this case, EDP D uses the concept of SARFIMV considering MV delivery points (MV/LV secondary substations and MV customers).

.

where
a is the number of voltage dips per MV busbar,
b is the number MV/LV substations supplied by the MV busbar,
c is the number of MV/LV substations supplied by MV busbars,
n is the number of MV busbars under analysis

It is important to develop flexible tools to achieve automatic calculation, particularly in the selection of criteria used to calculate the SARFIMV, enabling to perform some sensitivity studies.

This tool for automatic SARFIMV calculation is intended to be developed over the existing web based application. The main functionality of this tool enables users to choose the criteria that will be used for calculation. This means that users can easily calculate a SARFIMV depending on voltage dips residual voltage u or voltage dips duration t, or even more depending on both (curve).

This functionality allows sensitivity studies to be carried out, in order to understand which information is possible to get from the index changing the selected criteria.

PQ processes systematization

The clear definition and systematization of PQ processes was identified as an important tool to help in the operation, management and development of the PQ monitoring program.

As a starting point, a core process was defined – Manage Power Quality, based in the regulatory requirements to characterize the PQ in the distribution network operated by EDP D, and the respective sub-processes as shown in Figure 2.

Therefore, the main goal and scope of this process is to manage the verification of PQ in the distribution network, in accordance with the Portuguese Regulation and respective procedures to PQ monitoring and information reporting.

This PQ core process is part of EDP D’s QoS Management Process, which also includes continuity of supply.

Figure 2. Overview of the core process and subprocesses.

In addition, another management tool was introduced with the definition of compliance controls inside each process.

FUTURE TRENDS

The role of PQ in the current distribution networks is becoming increasingly important and EDP D wants to remain at the forefront of its potentialities.

As such, EDP D is focused on some future projects that aim to be a step forward from the current use and management of PQ data.

Unified Data Model (MDU)

There is a permanent need to get extra value from PQ data. In this sense, a project (MDU) is being developed in order to correlate PQ data and other network data sources, such as continuity of supply and SCADA data.

This project is being designed so that any user with the right permissions can get information in an automated way from different sources and, thus, better understand some occurrences in the network. With this functionality, manual manipulation as well as possible errors arising from any misuse are prevented.

This project will also help dealing with the constantly growth of recorded PQ data due to a continuously increase of the number of devices used in permanent PQ monitoring. In this sense, the application of big data and data analytics technologies will be of great value.

Online PQ information for operation support

As stated before, the increasing of recorded PQ data implies a similar increase of automation of its analysis. Besides that, it is important that PQ analysts and the operators that manage network online can perform faster actions in case of potential network PQ problems, anticipating, in most cases, their resolution.

Taking into account the difficulties related to slightly time lagged PQ analysis, like the ones on warning system, also under development, working on a time base, it is important to provide the operators with automatic and online information in case of customized PQ inputs for just-in-time actions.

Taking this in consideration, a future project will be developed in order to, as currently featured with the SCADA system, provide the operators with the relevant information that helps their decision-making to ensure a better QoS in global terms and a better PQ in particular.

When a PQ recorder registers a violation of a predefined limit, it triggers an automatic communication to the database, which in turn, sends it to a system available in the operation centre.

This information can be related to PQ continuous phenomena, such as voltage r.m.s., harmonics, flicker, etc. or related with voltage events, mainly voltage dips and swells.

CONCLUSIONS

For many years, EDP D has been developing a PQ monitoring program and platform in order to fulfil several requirements and challenges.

Year after year, more requirements and new challenges appear, some related with the increasing quantity of data, but, more than that, with the necessity to identify and work this data in order to improve the reliability and availability of the network that EDP D operates and manages.

As such, EDP D has some ongoing work, developing tools for PQ monitoring platform to produce fast, robust and reliable analysis, transforming PQ data into suitable information, developing its warning system on data quality and PQ limits violations as well as applying PQ indexes, but also improving power quality management process and its transparency with processes systematization and compliance control.

Despite the relevant work already developed in PQ monitoring field, EDP D is also aware of near future challenges, already having a portfolio of future projects that will allow a step forward from the current use and management of PQ data, improving it with Unified Data Model (MDU) and online PQ information for operation support.

Acknowledgments-The authors thank the availability and collaboration from Teresa Couceiro, Flávio Cação, António Margalho and Nuno Pinho.

REFERENCES

[1] Energy Services Regulatory Authority (ERSE), 2013, “Regulamento de Qualidade de Serviço do Setor Elétrico”, Diário da Republica, Regulamento nº455/2013 (2ª série).

Source & Publisher Item Identifier: ISSN 2515-0855. doi: 10.1049/oap-cired.2017.0421. http://www.ietdl.org

Harmonic Measurement and Analysis during Electric Vehicle Charging

Published by Mohd Zamri Che Wanik1, Mohd FadzilMohd Siam1, Afida Ayob2, Subiyanto2, Azah Mohamed2, Abu HanifahAzit3, SaharuddinSulaiman3, Mohamed Azrin Mohamed Ali4, Zahrul Faizi Hussein1, Ahmad Kamil MatHussin1,

1TNB Research Malaysia, Malaysia, 2Power System Research Group, Universiti Kebangsaan Malaysia, Malaysia, 3TNB Distribution Malaysia, Malaysia, 4Malaysian Green Technology Corporation, Bandar BaruBangi, Malaysia. Email: mzamri@tnbr.com.my

ABSTRACT This paper presents and describes harmonic measurement and analysis of studying harmonic propagation during electric vehicle (EV) charging. The measurement study is performed on a golf cart and two modern type of EVs. Harmonics from a single EV charging and a group of EV charging was measured. The voltage and current waveform during the charging was captured and analyzed to investigate the harmonic components that exist in the electrical system. Total current harmonic distortion (THDi) and total voltage harmonic distortion (THDi) were both calculated. Modern electric vehicles are found to release low THDi but higher THDv compared to a golf cart. On the measurement study during charging of a group of EV, it is found that the summation of THD is not linear with a number of vehicles. The finding of the study reveals that harmonic contamination from EV charging on electrical grid is not as critical as thought by most of power system researchers and engineers.

Keywords: Electric Vehicle; Total Harmonic Distortion; Electric Vehicle Charging; Harmonic; Power Quality

1. Introduction

In light of high energy usage, environmental pollution and rising fossil fuel prices, current dependence on internal combustion engine (ICE) technology employed in vehicles should be reduced and the widespread use of electric vehicle (EV) as the transportation of choice in 20 to 30 years time should be increased. It is estimated that EV vehicle penetration will increase gradually where 35% is projected at 2020 and will reach 50% by the year of 2024 [1-2].

The general effect on distribution systems caused by the spread of EV will be substantial load increase and large increment of system voltage and harmonic distortion. Another issue that should be considered is the coincidence between the charging start time and the eventual evening load peak period, which varies with customer and country. For charging, EV batteries need DC current so the grid AC current will be converted to DC by battery charger. The charger is basically the rectifier/inverter with controller integrated with protection circuit. This is where the concern rises because inverter/rectifier is known as a harmonic source. So one of the concerns with electric vehicle charging is the harmonic contamination to the electrical grid.

There is no agreement however on how much the total harmonic distortion (THD) can be released to the network during charging. In one of the published report, total current harmonic distortion (THDi) is reported between 2.36% to 5.26% at the beginning of charging and reaching up to 28% at the end of charging. However total voltage harmonic distortion (THDv) is claimed only to range between 1 and 2% with power factor close to unity. For commercial chargers, THDi from measurement recording values are between 60% to 70% [1,3-4].

The concern of the engineers and researchers are when a large number of EVs charging simultaneously to the power system grid. What is the sum of THD when EV multiplies in numbers? Many believe that THD will increase with the number of vehicles. It is however difficult to find a report that discusses this issue. Moreover the values indicate in the previous paragraph was reported [3-4] more than 10 years ago so the values do not represent the actual amount of harmonic generated from state-of-the-art charger technologies that is used in modern electric vehicles.

The study reported in this paper tries to investigate the harmonic distortion from a single EV and a group of EVs. The measurement was performed on an older type of EV and two modern types of EV. The results of the study is useful in understanding the harmonic distortion contribution from EV connected to the grid for charging.

This paper is organized as follow. First EV technology is briefly reviewed. The harmonic distortion limit then highlighted. Harmonic measurement during a single vehicle charging is then presented. Finally the harmonic distortion during charging of a group of EV is discussed and commented.

2. Electric Vehicle Technology

Electric vehicle that will be seen creating issues on the power system grid will be of two types. First type is plug in hybrid where there is a combination of ICE and battery. The second type will be all electric vehicle where this vehicle depends solely on battery. For both types of EV, electrical power is needed from the grid for charging.

Example of typical PHEV hardware arrangement is depicted in Figure 1. This is for parallel type of PHEV. There is also series type of PHEV but the focus of this paper is on the grid connection issue so it is not important if the EV is of parallel or series type. For all electric vehicle, the components in the light green dash lines are not needed and removed leaving only batteries, charger, power electronic drive and propulsion electric motor.

When EV is charging, the only active components are charger and battery because EV is not moving. For grid interfacing studies, the EV can be represented electrically as depicted in Figure 2. Battery charger is basically a rectifier which converts AC current from the grid to DC current to charge the battery. The rectifier is normally of active type where the power electronic devices are employed for switching devices. In this diagram power electronic devices are IGBT.

Figure 1. PHEV typical layout [5].
Figure 2. Battery charger circuit diagram [6].

EV charger is normally of conductive type even though inductive type is available. Conductive chargers have direct plug-in connection to the supply e.g. using an extension power cord to plug form the wall outlet into the EV. Inductive charger on the other hand use magnetic coupling as a mode of energy transfer. Comparing these two types, conductive charger is simpler to design, having higher efficiency and more popular.

3. Harmonics Distortion Limit

Harmonics distortion has detrimental effects on electrical equipment inside power system. Severity of harmonic is determined by the percentage of total harmonic voltage distortion. Harmonic voltage is referred to as a sinusoidal voltage having frequency equal to an integer multiple of the fundamental frequency of 50/60 Hz supply. Total harmonic distortion is calculated as follows[ 7]:

.

In Malaysia, the main utility company, Tenaga Nasional Berhad, has set a limit on total harmonic voltage distortion according to voltage level as tabulated in Table 1. For current there is no distortion limit set.

4. Harmonic Measurement

A few field measurement activities were conducted investigate the harmonic during EV charging. The measurement is performed using Fluke power quality meter. The picture of the meter is portrayed in Figure 3. The meter has the capability to directly display the voltage and current waveforms in real time. In addition, the measurement data can be saved into the memory of the meter for further analysis.

For individual harmonic, three types of electric vehicle is compared. The first type of EV is a commercial type of modern EV ( EV1). Second type is modern EV prototype (EV 2). Third type is a golf cart. Due to classified and sensitive issues, the name of the manufacturers and the brand of these EVs are not disclosed.

Table 1. THDv limit in TNB Distribution system [7].

.
Figure 3. Fluke power quality meter.
Figure 4. Modern EV 1 voltage and current waveform.

Figure 4 shows the voltage and current waveforms of EV 1. Figure 5 and 6 show THDv and THDi for EV 1 respectively. THDv and THDi after analysis are 1.5% and 11.6% respectively. For EV 2, the voltage and current waveform are depicted in Figure 7. THDv and THDi for EV 2 is 1.2% (Figure 8) and 9.2% (Figure 9) respectively. Figure 10 shows the voltage and current waveform for the golf cart. Figure 11 and 12 shows THDv and THDi for the golf cart which are 1.1% and 34.4% respectively.

Figure 5. Total voltage harmonic distortion from EV 1.
Figure 6. Total current harmonic distortion from EV 1.
Figure 7. Modern EV 2 voltage and current waveform.
Figure 8. Modern EV 2 total voltage harmonic distortion.
Figure 9. Modern EV 2 total voltage harmonic distortion.
Figure 10. Golf cart voltage and current waveform.

The measurement results reveal that the THDi from chargers is low for modern EV but for THDv, modern EV is releasing higher distortion. The lower THDv from golf cart can be explained due to lower current magnitude drawn during the charging. For modern EV, even though THDi is lower, THDv is higher than measured for golf cart. This is due to higher current magnitude drawn by modern EV.

5. Harmonic From a Group of EV

In the previous section, THDv measured from a single EV was presented. The value was recorded lower than limit set by utility. But the concern on harmonic is when a group of EVs connected to the grid simultaneously for charging. There are many engineers who thought that the summation of THD from a group of EVs is linear with the sum of EV. To find out the answer, a measurement is performed to measure THD from a group of EVs.

In Malaysia, modern EV is not yet commercialized so the study cannot be performed on the commercial EVs. But there are many golf carts available which can be used in the study. The results will not represent the modern EV harmonic but the main objective is to investigate the sum of THD from a single EV and a group of [EV].

Figure 11. Golf cart total voltage harmonic distortion
Figure 12. Golf cart total current harmonic distortion.

The measurement was performed at a golf club at Universiti Kebangsaan Malaysia. The measurement was carried out on a working day. On that day, it was raining from early morning until 10:00 am. But the golfing activities were active where all golf carts left the charging point by 9.00 am. The first golf cart returned to charging station after 11 am. The time, number of vehicle, and the line where they are connected are detailed in Table 2. The measured quantities are voltage and current harmonics distortion. The measurement data was recorded initially every one hour interval until 6:00 pm. Starting at 7:00 pm, the golf cart start to arrive more often so the measurement was taken for every 15 minutes.

The data are visualized through a graph as shown in Figure 14 for THDv and Figure 15 for THDi. Instead of time, the number of vehicles is used for x-axis to facilitate the pattern of consumption from these EV. The line where these golf carts are connected were also identified. The consumption starts to increase substantially starting from 6:30 pm (18:30 hrs) when golfers start to return the golf cart to the pick-up point.

Figure 13. Golf Club distribution Circuit.

Table 2. Time, number of vehicle and connection.

.
Figure 14. Measured THDv during monitoring.
Figure 15. Measured THDi during charging.

The THD is also measured both for voltage and current as shown respectively in Figure 14 and Figure 15. Eventhough THDv for each single vehicle was measured 1.1% but a PCC, the value is only 0.7% for onevehicle. With the increaseof vehicles connected, the THDv steadily increase. For eight vehicles on L1, the recorded THDv is only 1.9%.

For THDi of one vehicle, THDi 34.4% is recorded but at PCC, THDi is recorded 47.8%. With the increment of golf cart charging the THDi value at PCC decrease to 16.6% for 3 vehicles but increase again for12 vehicles. Surprisingly THDi do not increase much when vehicle increase from 16 to 22 on L1.

6. Conclusion

In this paper, the harmonic measurement study and analysis during electric vehicle charging is presented. The measurement is performed on golf carts and two modern types of EV. The measurement result shows that modern EV release lower THDi compared to a golf cart which is expected. Unfortunately for THDv, modern EV cause higher percentage of distortion. For a group of EV charging, THD for both voltage and current are found not to be the direct summation of the THD from a single vehicle. The results of the study areuseful in understanding the harmonic distortion contribution from EV connected to the grid for charging.

7. Acknowledgements

This investigation study is part of a project jointly carried out by TNB Research Malaysia, Universiti Kebangsaan Malaysia, TNB Distribution Malaysia and Malaysian Green Technology Corporation. The project is funded by TNB Research Malaysia through funding TNBR/RD55/ 2012 and partly supported by Universiti Kebangsaan Malaysia via research grant GGPM-2011-071.

*TNB is an abbreviation for Tenaga Nasional Berhad which is a main electrical utility company in Malaysia. TNB research is a research arm of TNB.

REFERENCES

[1] R. Liu, L. Dow and E. Liu, “A survey of PEV Impact on Electric Utilities,”IEEE PES Innovative Smart Grid
Technologies Conference, January, 17-19, 2011, Anaheim, CA, USA
[2] EPRI, “Environmental Assessment of Plug-In Hybrid Electric Vehicles,” Technical Report 1015325, July 2007.
[3] J. Orr, et al, “Current Harmonics, Voltage Distortion, and Powers Associated with Electric Vehicle Battery Chargers Distributed on the Residential Power System,” IEEE Trans. On Industry Applications, Volume: IA-20, Issue 4,1984
[4] J. Gomez and M. Morcos, “Impact of EV Battery Chargers on the Power Quality of Distribution Systems,” IEEE Power Engineering Review, October 2002.
[5] Emadi, A., “Transportation 2.0.,” IEEE Power and Energy, Vol. 9, No. 4, July/ Aug. 2011.
[6] W.Kramer, S. Chakraborty, B. Kroposki and H. Thomas, “Advanced power electronic interfaces for distributed energy systems – part 1: system and topologies,” National Renewable Energy Laboratory, Golden, Colorado, Tech. Rep. NREL/TP-581-42672, Mar.2008.
[7] Malaysian Energy Commision, The Malaysian Distribution Code, http://www.tnb.com.my/tnb/application/uploads/uploaded/the%20malaysian%20grid%20code.pdf

Source & Publisher Item Identifier: Engineering, 2013, 5, 215-220. doi:10.4236/eng.2013.51b039 Published Online January 2013 (http://www.SciRP.org/journal/eng)

A Review of the Harmonic and Unbalance Effects in Electrical Distribution Networks due to EV Charging

Published by Lauri Kütt, Eero Saarijärvi, Matti Lehtonen, Department of Electrical Engineering Aalto University School of Electrical Engineering, Espoo, Finland. lauri.kutt@aalto.fi, eero.saarijarvi@aalto.fi
Heigo Mõlder, Jaan Niitsoo, Department of Electrical Engineering, Tallinn University of Technology Tallinn, Estonia

Abstract—For wide use of electric vehicles (EVs), there are different aspects of the electric power system to consider for making it ready for the increased load by battery charging. The topics include power production, peak load management, distribution transmission capacity but also distribution network power quality and many more. This paper presents an overview on the likely power quality impacts in the distribution networks associated with EV charging. Based on a literature review, focus is especially put on harmonics and load unbalance in the network. Most relevant papers observing these topics are presented summarizing their contribution. The power quality aspects in distribution networks discussed here are not often presented within analysis of permissible EV penetration levels. Harmonics or voltage unbalance and effects associated with these could introduce additional limits to the EV charging capacity for the distribution networks. Therefore the analysis on the EV charging influence on these power quality topics requires also high-priority discussions before drawing conclusions on the distribution networks capabilities.

Keywords-electric vehicles, battery charger, power quality, harmonics, harmonic distortion, voltage unbalance

I. INTRODUCTION

Electrical vehicles are considered everyday use commuter vehicles with significant on-board energy storage and can use the mains power supply for charging the energy storage. Typical such vehicles are battery electric vehicles (BEVs) and plug-in hybrid vehicles (PHEVs). Due to similarity of the process of the on-board battery charging, there would be only slight differences in the impact for the power networks depending on the vehicle type. Therefore, within this paper, all such vehicles are referred to only as electric vehicles (EVs).

Typical EVs are provided with batteries having energy storage capacities from some kWh up to several tens of kWh [1]. Accounting losses, charging such batteries means even higher energy usage from distribution network. It has to be kept in mind that for providing acceptable convenience level for the EV owner the battery would require full recharge in time of just limited hours. To accomplish this, power of the chargers is expected to be high, starting from 1.6 kW for single phase onboard chargers for home use [2] reaching into hundreds of kW for ultrafast charging [3]. As the most probable location for EV charging is home [4] the likelihood of addition of powerful single-phase loads to the residential network is very high.

The increasing use of EVs is being promoted actively for several benefits in environmental aspects and energy efficiency. Assuming thermal power plant origin of power, overall EV efficiency is at least 23,1% [5] while a vehicle with internal combustion engine utilizes 12,5% of fuel primary energy. Besides higher efficiency, the EVs can offer additional cost effectiveness and other benefits. Discussions in public are however not regarding the effects to electrical networks and it is often presumed that the distribution networks provide the necessary overhead and are ready to accept the EV charging loads [6]. The distribution network transfer capacity availability has recently seen quite much discussion, with results presenting clearly that the distribution networks can have limitations in EV charging [7] [8] support even for a relatively low EV penetration levels. As a remedy, several control and moreover smart charging scenarios have been proposed to increase the charging capabilities to the network [9].

Topics for the analysis of EV charging impacts to distribution networks can be listed as thermal loading, voltage regulation, harmonic distortion levels, unbalances, losses and transformers loss of life [10]. This paper aims to summarize the investigation presented until now about the aspects of harmonics and load unbalance associated with EV charging. For an introduction, a summary is provided about the harmonic currents produced by the EV chargers. Here the expected harmonic distortion and power levels for chargers are described, also case measurement results are provided. This is followed by a discussion about the summation of harmonic currents and the harmonic voltages seen in the distribution networks. Some remedies are discussed to decrease the harmonic currents levels. In the last part the studies of network voltage unbalance are discussed.

The focus of the distribution networks in this paper is put on the residential networks, where the EV charging could bring severe addition of power electronic load and associated power quality issues. Industrial networks are not observed at this time as usually the industrial networks have their own power supply network and rather company specific loads. Residential networks however supply a large number of customers and failures to meet supply standards may result in high customer dissatisfaction.

II. REQUIREMENTS FOR THE EV CHARGERS AND NETWORKS

In Europe, most common standard for public power supply is EN 50160 [11], which sets conditions for

• voltage magnitude variation,
• voltage harmonics,
• interharmonic voltage,
• voltage unbalance

and many more. All loads that are connected to the power network must provide so low effect on the network that it does not lead to violation of the power supply conditions stated in this standard. This means also that the EV charger, once connected to public network, must not influence the network operation to the extent that can cause deviation from the standard. It should be kept in mind that parameter variations beyond standard limits can cause malfunction and failure of different devices connected to network.

The requirements for the EV chargers specifically are not standardized at this time. In general, the EV chargers have to fulfill requirements for loads that can be connected to electric power network described by electromagnetic compatibility IEC 61000 series standards (and similar IEEE standards in USA). These standards set the emission levels, including the harmonic currents, power factor etc. that a charger is allowed to have. The standards applying for the low-power EV chargers are IEC 61000-3-2 [12] and IEC 61000-3-4 [13], which set limits to the harmonic emissions generated by the charger.

Limits for voltage unbalance in LV networks are provided in standard IEC 61000-2-2 [14]. In general, the permissible level in the networks is 2%, but the question is somewhat more complex. The voltage unbalance is a result of unequal phase currents and the control of the phase currents is possible generally for the three-phase chargers. A greater problem would arise from the use of large number of single-phase chargers in the network, which can still provide presence of significant unbalance.

III. CHARGER HARMONICS

Every non-linear load is expected to provide a harmonic pattern, meaning specific harmonic magnitude and phase values. The quantitative amount of current waveform harmonics can be expressed as the total current harmonic distortion (THDi)

.

where h is the harmonic order number, H is the highest number of harmonic observed, Ih is the RMS-value of the current h-th harmonic component and I1 is the mains frequency current component RMS value. Similarly, the total voltage harmonic distortion THDv is calculated, but with the current values in (1) replaced with respective voltage values.

EV battery requires DC for charging and the conversion from AC as well as charging control is provided by the power electronics converters, presenting non-linear load. The harmonics associated with the EV charging are closely related to the charger circuit topology that is providing an interface towards the AC network. The simplest single-phase full-bridge rectifier, or for higher power ratings three-phase diode rectifier, (similar to all uncontrolled rectifiers) provide the highest current harmonics to the AC power network. The circuits and control strategies have evolved rapidly to include more network friendly features, such as power factor correction (PFC) and current waveform shaping. Evolution of the rectifiers’ topologies, the associated waveforms and the current harmonics levels have been laid out in [15]. Authors have relied on real measurements of different chargers, presenting actual waveforms that provide good information on possible charger topology. First chargers measurements from 1993 revealed use of uncontrolled or low-control rectifiers, revealing average THDi of 50%. Measurements of 1994 commercially offered EVs present average THDi of 20%, while the EVs tested in 1995 presented near-sinusoidal current waveforms and THDi below 7.5%. It should be pointed out that the chargers with high THDi present the older topologies, not likely present for the modern vehicles.

More recent analysis, presented in [16] for different controlled battery charger front ends using modern power semiconductors, shows that there is a variety of topologies offering THDi well below 5% at load of 50…100% of rated power. The harmonics levels would remain lower than the limitations by applicable standards. The power factor at the same is time greater than 0.99. However, there seems to be slight dependence between the harmonic levels and input voltage levels. With higher charger input voltage, the lower harmonics (below 13th) are at slightly higher level, while for the lower mains voltage the higher harmonics (above 15th) present higher values. The practical measurements of modern commercial EVs in [17] present charging THDi of 11.6%, similar to results in [18]. There is also a rather good example of a commercial EV with a charger providing around 4.5% THDi presented in [19]. A calculation for the standard limits of the harmonic emissions has been presented in [20]. Based on the limits in IEC 61000-3-4 [13] for individual harmonic emission levels, the highest THDi allowed for a charger would be 17.3%. Comparing only THDi, the modern EVs could fit nicely in the standard limits. It has to be pointed out that the THDi does not reveal the levels of the individual harmonics, which could still exceed the limits regardless of THDi value.

A comparison of the charger harmonics typically indicated is presented in Table I.

TABLE I. COMPARISON OF EV CHARGERS HARMONIC DISTORTION LEVELS

.
IV. CHARGING PROCESS

Different battery types and their charging cycles have been presented in [23] comparing lead-acid batteries and Li-ion batteries charging cycles and charger operation. Lead-acid, not very widely used in modern EVs due to heavy mass, has 4 different charging stages, with low, medium, topping and trickle modes. Li-ion type batteries would follow charging patterns fast-topping-trickle. Although the lead-acid battery charger shown is three-phase and Li-ion single-phase, it might be reasonable to assume that the chargers would present similar current waveforms regardless of the battery chemistry. It has been presented that the THDi during the heavier charging load is lower than during the low (trickle) load, discussed also in [24]. The conclusion, that the current waveform degradation in the low-power mode would be more harmful considering losses and distortions in the distribution networks, cannot be agreed with. It does not take into account the fact, that in low-power operation load current as well as harmonics have smaller magnitudes, which means that the effect to the network, on the contrary, is weaker. This has been presented with practical measurements of another commercial EV using the built-in charger [18]. Results show rather constant THDi value of 11% over the entire range of charging power, with a slight increase for the near-full capacity. However, it can be seen from the associated voltage THD measurements, that when the charging current decreases and THDi increases, THDv instead decreases. Also in the same paper a three-phase more powerful charging has been observed, with clearly observable relation of THDi increasing as the charging current decreases.

In [25] a charging profile has been proposed from measurements of EV charger, also used in many other papers, presenting constant current and then constant voltage charging operation. The particular measured data shows difference in waveforms considering different state of charge (SOC). Again, the THDi is higher when the battery is approaching the complete charge and the output power of the charger drops. Different charging power profiles with measurements of 5 different EVs have been presented [26]. The charger control methods for different EVs are likely very different, as the charging times vary. Similarly the charging current profiles follow different patterns. The chargers observed are modern chargers, with PFC likely included, as the true power factor is specified in most cases as very close to unity.

V. HARMONIC CURRENTS FOR MULTIPLE CHARGERS

A typical distribution network has a large number of different non-linear loads connected to it. Adding different manufacturers’ EV chargers, it is likely that there are a variety of different harmonic patterns present. The diversity of the patterns may lead to notable harmonic cancellation. This effect occurs when harmonics with different phase angles provide a sum in the magnitude that is smaller than the individual harmonics magnitudes. It is still rather complicated to evaluate this effect. Cancellation is more probable as the number of consumers increases [24]. It has also been indicated that harmonic cancellation is more expected at higher harmonic orders, which can then account for the relatively minor THDi decrease observed in different studies. In most papers, it is rather common, that only the harmonics levels are observed, as the utilities are required to keep the harmonics levels under a given limit. If diversity of chargers is not taken into account, the harmonic problems could be overestimated [27].

For the correct estimation of the harmonics levels and cancellation, phase angle values of individual harmonics are also required besides magnitudes. An example of a more sophisticated model allowing observing harmonics cancellation due to diversity in magnitudes and phase angles has been presented in [27]. One of the pilot papers in this area is [28] where multiple different EV chargers in the network have been observed. There are 5 different rather simple charger topologies described, assigned for samples of EVs. Several probabilistic parameters are included such as distribution of charging times and SOC. Monte Carlo simulation method with sample size of 100 is used for the analysis of the complex system. It is reported, that 10% smaller harmonic current magnitudes were seen compared to the simple summing of magnitudes.

In [25] the measurement results of charging process have been presented, with analysis performed by dividing the harmonics into real and imaginary parts, observed up to 15th harmonic. Identical chargers are observed, however with waveform and THDi dependent on the battery SOC, which could also be observed as different chargers in the network. Similar to [28], the battery state and also charging start time have been observed as stochastic parameters. A model has been defined taking into account the probability density functions and solution has been found using analytical and the Monte Carlo method. A good convergence has been achieved with sample size of 7 chargers. Monte Carlo method has also been used in [29] where convergence is presented with a slightly larger sample size of 25. This would mean that this amount of different chargers could provide the harmonic cancellation already near the maximum possible value. The THDi decrease from 27.5% down to 25.1% due to diversity has been stated, which is almost 10% improvement.

Analysis with real and imaginary components for each harmonic has been described in [30]. The paper presents rather powerful chargers with THDi over 40% at connection point. The 11 kV medium-voltage network has been simulated with 36 chargers, each at power level of 8.2 kVA, which makes it difficult to witness the total cancellation effect. Similar approach in [31] uses real measurements of 30 kW charger currents. Charger current waveform indicates significant high frequency harmonics content. The analysis aims to develop a probabilistic method for evaluation of the current waveform. However as the chargers are assumed all the same and other simplifications are introduced, the result does not provide good indications of benefits of the probabilistic approach.

In conclusion, many studies have presented that the harmonic cancellation between different chargers and loads brings the expected EV charging harmonic current magnitude lower. The view on the network has been rather simplified and ideal, with networks only having EV charging load. However, there is still a lack of overview on the matter to what extent actually the harmonic cancellation could reach, assuming background with real everyday loads connected. These loads have their own specific harmonic patterns that can contribute to the harmonic cancellation.

VI. HARMONIC VOLTAGE LEVELS

The harmonic currents can cause presence of significant harmonic voltages. Prediction of levels of voltage distortions has been analyzed in [27], taking into account the probabilistic characteristic of the EV charging currents. The worst case scenario with least residential linear load available during nighttime is targeted. EV penetration of 50% is assumed and voltage distortion is calculated for individual buses. The method presented gives a probabilistic output, revealing the expected probability of a voltage distortion above allowed value. Seasonal varieties have also been accounted as two scenarios are presented. This method has its virtue in the fact, that the modern power supply standards [11] require the operation within stricter limits for a specified portion of time. The probabilistic method can be used to determine the probability of each bus to exceed or meet the requirements.

Worst-case scenario of residential network harmonic voltages has been presented in [20] where a smaller LV network with 15 kVA supply transformer is observed. In this investigation, a worst-case waveforms were constructed, using highest permissible limits for a commercial charger set by standard [13], resulting in charger with THDi of 17.3%. The network load also included typical household appliances for balance. Model of the network is comprehensive and takes into account also lengths and types of cables feeding the customers. It is reported that the worst-case scenario with all households having EV presented the THDv from 2.6% to 5.2%. However, for another case with less EV penetration, THDv stable value at 1.5% level has been shown regardless of addition of EV. Measurements of THDv have also been presented for a test site on a real network, indicating THDv over 2% also without the EV charging. More suggestions for a worst-case simulation have been presented in [32] for a larger network, where also the network is observed not as a lumped model but the effects of cables have been included. THDv is also dependent on the harmonic voltage drop on the cables, due to harmonic current present in the cable. This way, the THDv can be estimated for the consumer furthest away from the supplying transformer.

VII. STRESS ON THE NETWORK COMPONENTS

Harmonic effects in distribution networks cause additional stress on the network components due to the increased harmonic currents and voltages. The components to suffer the most are the distribution transformers, cables and fuses [24]. The most direct effect would be seen with the distribution transformers, which could encounter much greater stress and heating due to the increased harmonic losses but also increased load due to EV charging. The direct approach to the harmonic losses has been presented in [22] where a three-leg distribution transformer is observed using electromagnetic model of the transformer. The 2 MVA transformer is observed under a load of 8 rapid chargers, each with maximum power of 250 kVA. Charging stations present ideal 3-phase rectifier load, with THDi at 24% level. The losses in the transformer are analyzed throughout the day, with greatest variability of losses due to winding losses. The core losses are presented rather stable regardless of load. The comparison is provided in reference to the same transformer being loaded with equal sinusoidal current. In general, operation with high harmonics would bring 6% higher losses than in the case of sinusoidal current. It has been estimated in [24] that total harmonic current distortions level of 70% could bring increase in transformer lifetime consumption by up to 10%. A quadratic relation to between harmonic level and transformer loss of life is proposed.

For maintaining the network components operation or lifetime with presence of harmonics the allowed load is decreased. Higher derating is needed to mostly for the transformers. It has been concluded that the cables do not require as large derating [24]. Another problem emerging from presence of harmonics and voltage unbalance is the neutral wire current [33], which can become large and pose risks if overloaded. Approach on neutral current evaluation due to the EV charging has been presented in [21].

VIII. REMEDIES AGAINST HARMONICS

As a remedy to the high THDi, introduction of some filtering components is one of the solutions. In [34] authors have been investigating single and three-phase chargers with the addition of series reactors. The simulations presented show that for single-phase rectifier there is little effect by the series reactor, while for the three-phase charger the reactor can provide THDi improvement of 50%. In the same paper, some investigation for shunt active filters for a larger parking facility has been presented. In addition to THDi decrease, the reduction of harmonic currents is indicated to provide a decrease of the neutral currents from 56 A down to 5 A. However, neither network parameters nor configuration have been presented for deeper insight. Options to use dedicated harmonic suppression filters have been discussed in [35]. The harmonic current source in the studied case is a EV three-phase charger. For the three-phase rectifier topology is has been presented that 5th and 7th harmonic frequency single-tuned filters are providing a good remedy. This filtering option has been presented for the large-power charging stations, while it is mentioned that generally the small-power charging there would be no need for any filters.

The uncontrolled charging can be regarded for providing the worst-case charging load currents. The analysis of current harmonic distortion during uncontrolled charging and the possible decrease during controlled charging has been analyzed in [36]. The simulation observes an IEEE 19 bus system with transformer having 100 kVA power rating. Charger used has rather high THDi of 31.9%. Uncoordinated charging at low EV penetration results in THDi of 12%, while for coordinated charging this is at 6%. For the implementation of charging control for the real networks, however, some discussions on the targets would be necessary. Next step in charging control is controlled charging and an approach to use smart charging control for asset protection has been presented in [37]. The power quality improvement strategy is proposed for protecting the components most prone to bad power quality. Derating of distribution transformer operating power in case of high harmonic currents is evaluated with the K-factor method. Smart load management with in-house priority system is proposed, accompanied by a smart meter for the control. The rescheduling of EV charging using the smart load management has been shown to provide significant THDi reduction, for the benefit of transformer health.

IX. PHASE UNBALANCE

The unbalanced voltage in three-phase network can be observed as a sum of positive, negative and zero sequence voltage components. The voltage unbalance is commonly observed as negative sequence unbalance, a relation of negative sequence voltage to positive sequence voltage

.

where U1 is the positive sequence voltage and U2 is the negative sequence voltage of frequency 50 or 60 Hz.

The EV chargers most likely to be deployed would be the single-phase chargers found onboard the EVs. It is likely that for a large number of customers and three-phase system, the charging current would be distributed uniformly among the phases. There is still a possibility that given the stochastic characteristic of the charging, there could be some time instances when the charging current is loading the phases unequally. The current imbalance will in turn cause voltage unbalance on the distribution transformer.

During simulation presented in [38], the current and resulting voltage unbalance have been observed for a car park. Relatively high power transformer (1 MVA) was employed only for the supply of 140 units EV single-phase chargers with rated current of 10 A. For this configuration the expected phase current imbalance expected was up to 13%. The voltage unbalance limit was set at 1.3% and it was presented that current imbalance over 11% level would cause the voltage unbalance that would exceed the limit. In a comprehensive analysis in [26] the authors have presented also the typical LV network background voltage unbalance. For some periods, the voltage unbalance even without EVs shows values above 1% level. Comparing to standards for compatibility levels of 2% in LV networks [14], there is still some headroom, though not very much. The analysis has been presented, using random single-phase load placement on the network with 200 customers supplied by 400 kVA transformer with 1000 variations introduced. Different case simulation results are provided, with conditions set on maximum phase unbalance power of 100 kVA or voltage drop of 10% in the network. It is presented that while for the 5 A currents the expected unbalance is quite low, multiple single-phase loads with 20 A load current are more likely to cause unbalance over 1%.

Voltage unbalance analysis in [39] uses the voltage unbalance sensitivity definition. This is a function of EV charging location in the grid and the charging current. Network normal load is modeled as each household on the LV feeder is assumed random power from 0 up to 5 kW. The first case of phase current unbalance is presented with constant uneven load rather than probabilistic approach. At this load configuration, it is reported that there is almost 0.9% voltage unbalance in the beginning and over 1.8% unbalance in the end of the feeder, without the charging load. Adding some 20 A chargers, the near-limit 1.96% value is reached (tolerable 2%). For similar stochastic scenario with Monte Carlo method used, a likely distribution of voltage unbalance in the furthest feeder is presented. It is shown that for 34% of time, it is likely that the 2% voltage unbalance limit is exceeded.

X. DISCUSSION

The EV chargers used in the distribution network would need to fulfill the standard conditions for all loads and it can be shown [40] that if the IEC 61000 standards are met then also it would be ensured that the network operation is meeting the EN 50160 conditions. From the analysis presented, several aspects should still be brought forward. The EV charging differs from other loads, as to long charging times mean long lasting load with high and constant power level, and high coincidence [26]. The EV charger harmonic patterns, and if single-phase chargers are used, the voltage unbalance, could have significant levels for extended period of time.

As the EV charging is more of a stochastic phenomenon, a vast number of different charging and thus load scenarios are available that are possible at any time instance. Therefore, probabilistic approach would be appropriate. The power supply standards themselves, such as EN 50160 [11] are setting the parameters for a specific portion of time (95%) and refer to the probabilistic evaluation of load. The probabilistic methods therefore could be used more directly to evaluate meeting the operation criteria in the power supply standards.

The analysis on the expected harmonic levels is comprehensive; however 2 aspects in the analysis should be discussed. First, the total current harmonic distortion has been referred in several cases as the factor determining the power supply quality. Instead, it should be considered that the voltage harmonic distortion could be a problem. The current distortion as by definition (1) refers to the proportionality of harmonic currents in relation to the mains frequency current component. For determining the impact on the network and the voltage harmonics that are caused by the current harmonics, the actual harmonic current values would have to be used. Second, the modern household devices tend to use more power electronic regulators and converters and therefore an increase of harmonic currents can also be expected even without the EVs charging. It would be interesting to see in the analysis how the actual residential load harmonics and the EV harmonics cancellation occurs.

The voltage unbalance due to heavy single-phase loading, added with high harmonics levels could bring failures due to exceeding power supply standard limits, but for example also due to neutral wire overload. It is therefore necessary that the need for essential reinforcements to the network and sub– stations would be analyzed prior to the large-scale EV deployment. Utilities should consider also the power quality aspects associated with high charging loads with high priority.

REFERENCES

[1] Gan Li, Xiao-Ping Zhang, Modeling of Plug-in Hybrid Electric Vehicle Charging Demand in Probabilistic Power Flow Calculations, IEEE Transactions on Smart Grid, Vol. 3, No. 1, Mar 2012, pp. 492 – 499.
[2] Shengnan Shao, Manisa Pipattanasomporn, Demand Response as a Load Shaping Tool in an Intelligent Grid With Electric Vehicles, IEEE Transactions on Smart Grid, Vol. 2, No. 4, Dec 2011, pp. 624 – 631.
[3] H. Hõimoja, M. Vasiladiotis, S. Grioni, M. Capezzali, A. Rufer, H.B. Püttgen, Toward Ultrafast Charging Solutions of Electric Vehicles, CIGRE 2012, paper C6-207.
[4] L. Zhao, S. Prousch, M. Hübner, A. Moser, Simulation Methods for Assessing Electric Vehicle Impact on Distribution Grids, IEEE PES Transmission and Distribution Conference and Exposition, 2010, 7 p.
[5] C. Roe, J. Meisel, A.P. Meliopoulos, F. Evangelos, T. Overbye, Power System Level Impacts of PHEVs, 42nd Hawaii International Conference on System Sciences, HICSS, 2009, 10 p.
[6] R.C. Green II, L. Wang, M. Alam, The impact of plug-in hybrid electric vehicles on distribution networks: A review and outlook, Renewable and Sustainable Energy Reviews, Vol. 15, Iss. 1, Jan 2011, pp. 544 – 553.
[7] L. Zhao, S. Prousch, M. Hubner, A. Moser, Impact Assessment of Varying Penetrations of Electric Vehicles on Low Voltage Distribution Systems, IEEE PES Transmission and Distribution Conference and Exposition, 2010, 7 p.
[8] V. Tikka, J. Lassila, J. Haakana, J. Partanen, Case Study of the Effects of Electric Vehicle Charging on Grid Loads in an Urban Area, IEEE PES International Conference and Exhibition on Innovative Smart Grid Technologies (ISGT Europe), 2011, 7 p.
[9] Leonor Trovão and Humberto M. Jorge, Power Demand Impacts of the Charging of Electric Vehicles on the Power Distribution Network in a Residential Area, International Youth Conference on Energetics (IYCE), 2011, 6 p.
[10] J. Taylor, A. Maitra, M. Alexander, D. Brooks, M. Duvall, Evaluation of the Impact of Plug-in Electric Vehicle Loading on Distribution System Operations, IEEE Power & Energy Society General Meeting, 2009, 6 p.
[11] EN 50160:2010, Voltage Characteristics in Public Distribution Systems. European standard.
[12] IEC 61000-3-2 ed3.0:2005, Electromagnetic compatibility (EMC) – Part 3-2: Limits – Limits for harmonic current emissions (equipment input current ≤ 16 A per phase).
[13] IEC/TS 61000-3-4 ed1.0:1998, Electromagnetic compatibility (EMC) – Part 3-4: Limits – Limitation of emission of harmonic currents in low voltage power supply systems for equipment with rated current greater than 16 A.
[14] IEC 61000-2-2 ed2.0:2002 Electromagnetic compatibility (EMC) – Part 2-2: Environment – Compatibility levels for low-frequency conducted disturbances and signalling in public low-voltage power supply systems.
[15] S.H. Berisha, G.G. Karady, R. Ahmad, R. Hobbs, D. Karner, Current Harmonics Generated By Electric Vehicle Battery Chargers, International Conference on Power Electronics, Drives and Energy Systems for Industrial Growth, 1996, Vol. 1, pp. 584 – 589.
[16] F. Musavi, M. Edington, W. Eberle, W.G. Dunford, Evaluation and Efficiency Comparison of Front End AC-DC Plug-in Hybrid Charger Topologies, IEEE Transactions on Smart Grid, Vol. 3, No. 1, Mar 2012, pp. 413 – 421.
[17] M. Zamri, C. Wanik, M. Fadzil, M. Siam, A. Ayob, S. Yanto, M. Azah, A. H. Azit, S. Sulaiman, M. Azrin, M. Ali, Z. F. Hussein, A. Kamil, M. Hussin, Harmonic Measurement and Analysis during Electric Vehicle Charging, Engineering, 2013, Vol. 5, No. 1B, pp. 215 – 220.
[18] N. Melo, F. Mira, A. de Almeida, J. Delgado, Integration of PEV in Portuguese Distribution Grid. Analysis of harmonic current emissions in charging points. Electrical Power Quality and Utilisation (EPQU) 2011, 6 p.
[19] C. Wenge, M. Stotzer, T. Winkler, P. Komarnicki, Power quality measurements of electric vehicles in the low voltage power grid, Electrical Power Quality and Utilisation (EPQU), 2011, 5 p.
[20] R. Bass, R. Harley, F. Lambert, V. Rajasekaran, J. Pierce, Residential Harmonic Loads and EV Charging, IEEE Power Engineering Society Winter Meeting, 2001, Vol. 2, pp. 803 – 808.
[21] A.J. Collin, S.Z. Djokic, H.F. Thomas, J. Meyer, Modelling of Electric Vehicle Chargers for Power System Analysis, Electrical Power Quality and Utilisation (EPQU), 2011, 6 p.
[22] P.S. Moses, M.A.S. Masoum, K.M. Smedley, Harmonic Losses and Stresses of Nonlinear Three-Phase Distribution Transformers Serving Plug-In Electric Vehicle Charging Stations, IEEE PES Innovative Smart Grid Technologies (ISGT), 2011, 6 p.
[23] J.P. Trovao, P.G. Pereirinha, L. Trovao, H.M. Jorge, Electric Vehicles Chargers Characterization: Load Demand and Harmonic Distortion, Electrical Power Quality and Utilisation (EPQU), 2011, 7 p.
[24] J.C. Gomez, M.M. Morcos, Impact of EV Battery Chargers on the Power Quality of Distribution Systems, IEEE Transactions on Power Delivery, Vol. 18, No. 3, Jul 2003, pp. 975 – 981.
[25] P.T. Staats, W.M. Grady, A. Arapostathis, R.S. Thallam, A Statistical Method for Predicting the Net Harmonic Currents Generated by a Concentration of Electric Vehicle Battery Chargers, IEEE Transactions on Power Delivery, Vol. 12, No. 3, Jul 1997 , pp. 1258 – 1266.
[26] J. Meyer, S. Hahle, P. Schegner, C. Wald, Impact of electrical car charging on unbalance in public low voltage grids, Electrical Power Quality and Utilisation (EPQU), 2011, 6 p.
[27] P. T. Staats, W.M. Grady, A. Arapostathis, R. S. Thallam, A Statistical Analysis of the Effect of Electric Vehicle Battery Charging on Distribution System Harmonic Voltages, IEEE Transactions on Power Delivery, Vol. 13, No. 2, Apr 1998, pp. 640 – 646.
[28] J.A. Orr, A.E. Emanuel, D.J. Pileggi, Current Harmonics, Voltage Distortion, and Powers Associated with Electric Vehicle Battery Chargers Distributed on the Residential Power System, IEEE Transactions on Industry Applications, Vol. Ia-20, No. 4, Jul/Aug 1984, pp. 727 – 734.
[29] M.S.W. Chan, K-T.T. Chau and C-C.C. Chan, Modeling of Electric Vehicle Chargers, Annual Conference of the IEEE Industrial Electronics Society, IECON 1998, Vol. 1, pp. 433 – 438.
[30] E.W.C. Lo, D. Sustanto, C.C. Fok, Harmonic Load Flow Study for Electric Vehicle Chargers, IEEE International Conference on Power Electronics and Drive Systems, PEDS 1999, pp. 495 – 500.
[31] Lu Yanxia, Jiang Jiuchun, Harmonic Study of Electric Vehicle Chargers, International Conference on Electrical Machines and Systems, ICEMS, 2005, Vol. 3. pp. 2404 – 2407.
[32] V. Monteiro, H. Goncalves, J.L. Afonso, Impact of Electric Vehicles on Power Quality in a Smart Grid Context, Electrical Power Quality and Utilisation (EPQU), 2011, 6 p.
[33] J.J.M. Desmet, I. Sweertvaegher, G. Vanalme, K. Stockman, R.J.M. Belmans, Analysis of the Neutral Conductor Current in a Three- Phase Supplied Network With Nonlinear Single-Phase Loads, IEEE Transactions on Industry Applications, Vol. 39, No. 3, May/Jun 2003, pp. 587 – 593.
[34] J. Balcells, J. Garcia, Impact of Plug-in Electric Vehicles on the Supply Grid, IEEE Vehicle Power and Propulsion Conference (VPPC), 2010, 4 p.
[35] Wenhai Yang, Zhanlong Zhang, Jingmin Wang, Yajing Gao, Simulation of Electric Vehicle Charging Station and Harmonic Treatment, International Conference on Systems and Informatics (ICSAI), 2012, pp. 609 – 613.
[36] S. Deilami, A.S. Masoum, P.S. Moses, M.A.S. Masoum, Voltage Profile and THD Distortion of Residential Network with High Penetration of Plug-in Electrical Vehicles, IEEE PES Innovative Smart Grid Technologies Europe (ISGT Europe), 2010, 6 p.
[37] M.A.S. Masoum, P.S. Moses, S. Deilami, Load Management in Smart Grids Considering Harmonic Distortion and Transformer Derating, IEEE PES Innovative Smart Grid Technologies Europe (ISGT Europe), 2010, 7 p.
[38] G.A. Putrus, P. Suwanapingkarl, D. Johnston, E.C. Bentley, M. Narayana, Impact of Electric Vehicles on Power Distribution Networks, IEEE Vehicle Power and Propulsion Conference, VPPC, 2009, pp. 827 – 831 .
[39] F. Shahnia, A. Ghosh, G. Ledwich, F. Zare, Voltage Unbalance Sensitivity Analysis of Plug-in Electric Vehicles in Distribution Networks, Australasian Universities Power Engineering Conference (AUPEC), 2011, 6 p.
[40] M.N. Moschakis, E.L. Karfopoulos, E.I. Zountouridou, S.A. Papathanassiou, On Adaptation of Electric Vehicle and Microgrid Issues to EMC-Power Quality Standards, Electrical and Electronic Engineering, Vol. 2, No. 5, 2012, pp. 249 – 257.

Source & Publisher Item Identifier: 2013 12th International Conference on Environment and Electrical Engineering. DOI: 10.1109/EEEIC.2013.6549577

Measures and Technical Means for Increasing Efficiency and Reliability of Extra High Voltage Transmission Lines

Published by Vladislav KUCHANSKYY1, Paul SATYAM2, Olena RUBANENKO3, Iryna HUNKO4,
Institute of Electrodynamics of NAS of Ukraine (1), Kiev, Ukraine University of the West of England (2), Bristol United Kingdom, Regional Innovational Center at the Faculty of Electrical Engineering University of West Bohemia (3), Plzen, Czech Republic, Vinnytsia National Technical University (4), Vinnytsia, Ukraine

Abstract. Electromagnetic transients are considered in the implementation of three-phase automatic reclose on the transmission line of extra high voltage 750 kV. The influence of automatic shunting of phases and pre-insertion active resistance for limiting the characteristics of the aperiodic component of the current, which obstructs the transition of full current through zero, is evaluated. The paper analyses measures taking into account the effect of changing the degree of compensation of charging power and the angles of switching on an SF6 circuit breaker. Sub-schemes of disconnected undamaged phases of the extra high voltage transmission line for the investigation of the aperiodic current component have been developed. The values of the pre-insertion active resistances of different connection and automatic shunting of the phases are determined at which there is an effective reduction of the characteristics of the aperiodic component of the current. In the software environment, a model was developed and switching transient processes were simulated in the 750 kV transmission line. Operating modes that are potentially dangerous for SF6 circuit breakers are determined and recommendations are given to avoid them. Currently the technical and economic requirements for power transmission lines designed for the transport of electricity from large power plants and for the communication of powerful energy systems are increasing. Today there is importance of reducing specific investment in the construction of new and reconstruction of existing lines. The solution to these issues is associated with the maximum use of power lines by increasing their power transfer capability and controlling modes, especially in operating emergency conditions and post-emergency operation of power systems.

Streszczenie. Stany przejściowe elektromagnetyczne są brane pod uwagę przy wdrażaniu trójfazowego automatycznego ponownego zamykania na linii przesyłowej bardzo wysokiego napięcia 750 kV. W pracy wpływ automatycznego bocznikowania faz i rezystancji wstępnej w celu ograniczenia charakterystyki aperiodycznego składnika prądu, który utrudnia przejście pełnego prądu przez zero. W pracy przeanalizowano parametry uwzględniające wpływ zmiany stopnia kompensacji mocy ładowania i kątów włączenia wyłącznika SF6. W środowisku oprogramowania opracowano model i symulowano procesy przejściowe w linii przesyłowej 750 kV. (Techniki zwiększania skuteczności i niezawodności linii wysokiego napięcia 750 kV)

Keywords: electromagnetic transients, automatic phase shunting, three-phase automatic reclose, aperiodic component of current.
Słowa kluczowe: zjawiska elektromagnetyczne, linie bardzo wysokiego napięcia, procesy załączania

Introduction

Due to the modernization of the switching equipment – replacement of air circuit breakers on SF6 circuit breakers in the bulk electrical networks, there was a need to analyze the switching transients in 750 kV extra high voltage transmission lines. Due to the lacks of experience in operating SF6 circuit breakers at 750 kV extra high voltage (EHV) substations, accidents have occurred which have significantly reduced the reliability of the operation of bulk electrical networks [1-6]. One of the typical examples of switching that causes damage of SF6 switch is the fast on/off cycle [1-3]. On the other hand, the reliability is one of the most crucial subjects of power and electrical energy systems [7-9].

Such a typical switching example is the cycle of a three-phase automatic recloses (TPAR) in the event of a nonliquidated metal short circuit in one of the phases. In case of unsuccessful TPAR, the damaged phase is switched on to the non-liquidated short circuit and the damaged phases are switched-off. After switching on the non-damaged phases, there may be an aperiodic current component iacc (ACC) with characteristics values are significantly higher than the maximum permissible passport data of the SF6. Exceeding the value iacc and duration Tap of the ACC leads to a delay in the transition due to zero full transient current fig. 1. A durable damping process iacc with a value exceeding the maximum allowed prevents the full current from passing through zero, resulting in damage to the gas switch chamber. According to the requirements specified in [10], the ACC value may not exceed 58%.

It is assumed that when the switch is off, arc extinguishing occurs in each phase at the moment when the current passes through a zero value. Since the dispersion in the action of the poles of a SF6 circuit breaker when switched on does not exceed 0.001 s, the contacts are simultaneously closed. So, in the case of an unsuccessful TPAR, when the arc on the overhead line during the TPAR pause did not go out, it is possible to significantly delay the process of arc extinction [1-3].

A significant aperiodic component in the switching currents occurs not only with whole groups of shunt reactors. As a result of a series of transient calculations, the most unfavorable moments of contact closure after a TPAR pause and the moments of disconnection of an uncorrected failure by the value of the aperiodic component were found.

The objective of this work is to identify the conditions for the appearance of an aperiodic component in the current, to analyze the process of damping it using pre-insertion resistors, and also to analyze alternative solutions to the problem as a whole.

Fig.1. Total transient current
The reason for the appearance of the aperiodic component of the current during switching

In studies [2, 10-13], the simultaneous influence of the above measures on the significance of the ACC characteristics was not considered. The studies analyzed the effect of pre-insertion active resistances on the gas switch on magnetization currents and resonant overvoltages when switching on the EHV power line to the unloaded autotransformer. There are also studies [14,15] devoted to the estimation of the influence of pre-insertion active resistances and the moment of switching on the value of the characteristics of the ACC without analyzing the effect of changing the degree of compensation of the charging power of the line.

It should be noted that there are known works on the use of automatic phase shunting (APS) during single phase automatic re-closure (SPAR) to compensate for the recharge arc and elimination of abnormal resonance overvoltages [15-23]. Studies on the evaluation of the effects of ASF when performing the TPAR are also given. There are also no studies to compare the different types of connection of pre-insertion active resistors to the values of the characteristics of the AC current. As shown by previous studies, the application of only reducing the degree of charge power compensation and the use of controlled switching do not always effectively reduce the ACC [1-3, 10, 11].

This paper deals with the switching of switching on of the low-voltage transmission line to short-circuit one of the phases, which leads to the activation of TPAR. Studies have shown that the changes in the switching angle of the gas switch and the change in the degree of compensation of the charging power of the low voltage transmission line cannot effectively reduce the aperiodic component to the passport values [24, 25]. The total current value in the switch iC(t) is determined by the expression:

.

where iinv=Iinvcos(ωt+α–Ψ) is involuntary component current in circuit breaker; ω is angular velocity; α is moment of commutation; Iinv and Ψ is amplitude and phase of involuntary value of current; iap(t)=Iapsin(α–ϕ)e-t/τ i is aperiodical component in circuit breaker; Iap is amplitude of aperiodical component; t is time of electromagnetic transient; τ is the damping constant of the ACC; iosc(t)=Itre-t/τosccos(ωt+α–Ψtr) is decaying current transient component in circuit breaker. Itr and Ψtr is amplitude and phase of decaying transient component; τosc is the damping constant of decaying current transient component.

The initial value of the aperiodic component of the current depends on the moment the circuit breaker closes (for example, if the switch-on occurs when the instantaneous value of the mains voltage is close to zero, then the aperiodic component has the largest value equal to the amplitude of the periodic component of the current). The damping time constant of the aperiodic current is determined by the ratio of the active and inductive resistances in its circuit.

The aperiodic time constant of the aperiodic component depends on the ratio of the inductance and resistance of the circuit:

.

where LΣ is equivalent inductance of an equivalent circuit: RΣ is equivalent resistance of the equivalent circuit.

So, the initial value of the ACC current component depends of values (1). On Fig. 2 are shown AP in phase A in cycle of TPAR. The permissible value of ACC component for SF6 circuit breakers 750 kV is 58 % of total current at the transition process.

Fig.2. Excess of ACC component the maximum permissible value

Measures and technical means of improving the reliability and efficiency of EHV power transmission lines due to switching The main task of existing measures and means used in extra high voltage power transmission lines is to increase the reliability and operational efficiency of the operating modes of the main electric networks that they form. There are innumerable modes of operation of power lines, especially for operational normal operating modes. Normal modes of electrical networks mean those that allow long-term operation without any restrictions for both consumers and equipment of the networks themselves. But when critical deviations of values of parameters of an electric network are reached, it is necessary to speak already about occurrence of an abnormal mode for which use of traditional methods the controls will be inadequate and therefore ineffective. The consequences of uncontrolled abnormal mode can be not only the deterioration of the technical and economic performance electrical network, but also damage to the equipment of responsible consumers, as well as failure of the main equipment of the network itself with the further development of a system crash.

It should be noted that classification of abnormal modes, which occur in the abnormal nonsinusoidal and nonsymmetrical modes of extra high voltage power lines is given in [15-22]. The use of the term of abnormal mode is not accidental, because when working out literary sources [1-6, 10-13] and studies of experimental results [15-17,19- 22], it was concluded that this kind of modes is fundamentally different from traditional ones. The difference and the special characteristics of mode is that they are caused by an abnormal regime, primarily due to the effect of the source of distortion. In [18, 19, 25,26], the division of this type of mode into two main categories, depending on the resonance at a certain frequency, is shown on the basic harmonic and higher harmonic components.

Also, this developed classification of abnormal modes includes the considered form of the TPAR operation mode. An abnormal mode of operation exists during the implementation of the TPAR cycle and until the breaker is possibly damaged.

In extra high voltage power lines, the following measures and technical measures apply, shown Fig. 3:

• the use of pre-insertion resistances in SF6 circuit breakers, which can be connected in series or in parallel Fig. 3 a) and b);

• applications of automatic phase shunting when a phase disconnected is shunted by circuit breakers Fig. 3 c);

• the controlled switching to ensure switching moments at the required time Fig. 3 d);

• disconnecting of groups of shunt reactors, as well as the use of open-phase modes of groups shunt reactors Fig. 3 f).

At Fig. 3 LM, LE – inductance of the shunt reactor which compensate capacitance between phases and Capacitance between phase and earth; LL, LR – inductance and active resistance of the transmission line; LS1, LS2, R1, R2 – equivalent inductance and active resistance of systems; Rpre – active pre-insertion resistance of SF6 circuit breaker.

On Fig. 3 a, b) Q1 indicates the opened main contacts of the SF6 circuit breaker and Q2 denotes the auxiliary circuit breaker of the pre-insertion active resistance, which are closed to reduce the effect of the aperiodic current component.

On Fig. 3 c), Q1 denotes the open main contacts of the SF6 circuit breaker and Q2 denotes the auxiliary contacts of the automatic phase shunting device for phase closure. On Fig. 3 d), Q1 denotes a circuit breaker with a controlled switching device is designated for switching at an exact time.

On Fig. 3 f), Q1 denotes the line phase power switch and Q2 shows the switch of the shunt reactor for the case of studying the influence of the open-phase operation mode of the line on the aperiodic current component.

In Fig. 3 also indicates a short circuit that occurs on an extra high voltage overhead power line.

Fig.3. Equivalent circuits of phase extra high voltage transmission line

The aperiodic time constant of the aperiodic component in case of application pre-insertion connected in series:

.

where A = LLLMLE; B = LS1LS2; C = RLRMRE; D = RS1RS2; E = RS1+RS2; F = RLRM; G = RLRE; I = RMRE; K = LS1+LS2; M = F+G.

The aperiodic time constant of the aperiodic component in case of application pre-insertion resistors connected in parallel:

.

The above measures have proven themselves well in suppressing and limiting the characteristics of abnormal resonant overvoltages in non-sinusoidal and asymmetric operating modes [16-19, 25,27,28]. It should be noted that today there is no work to assess the impact of measures and means on the aperiodic component of the total transient current (1) and especially aperiodic time constant (2).

Switch with pre-insertion active resistor can be used to extinguish large multiple voltages on the second harmonic. As shown by research [25,27-29], overvoltages on the second harmonic can exist for a long time, so the decisive factor in determining the scattering energy is the time of the emergence of an abnormal regime with unloaded autotransformer.

The method described in the article solves the problem of putting the line under voltage, avoiding the danger of damage to electrical equipment by resonant overvoltages on the second harmonic. The use of pre-insertion resistors can reduce the amplitude and duration of this type of overvoltage. It is important to note that the decision to use pre-insertion resistors in each case must be supported by the results of mathematical analysis. Thus, the use of pre-insertion resistances in SF6 circuit breakers has not been verified by influencing the characteristics of the aperiodic component. Initial transition conditions for each phase corresponding to the maximum and minimum of the sine wave voltage [27-3`]

.

The objective of the ASP is to reduce the electrostatic as well as the electromagnetic components of the feed current in order to ensure a successful SPAR [31-34].

It should be noted that the paper considers the worst case when the voltage in the intact phases goes beyond zero and the case when the voltage reaches its maximum value. In the first case, such initial conditions of the electromagnetic switching transient cause the maximum initial value of the aperiodic component of the current. As shown by the damping study of the aperiodic component of the current is determined by the ratio of the circuit’s active resistance to the total inductance [1, 2, 3]. According to the data, the damping process for certain lines continues to 0.003-0.04 s. A controlled switching device for aperiodic suppression can be used in combination with the above.

Fig.4. Sine wave voltage phases A, B, C with designated switching points

As we can see from (1) iap(t) depends on α – moment of commutation. So we can that controlling switch device can reduce this two component. All circuit breakers are equipped controlled switching device Switch Sync F236. The sinusoid with possible moments of commutation is depicted on Fig. 4. To assess the impact of the switching moment on the characteristics of the aperiodic component, a simulation model is developed. The description of which is given in the next section.

When using controlled switching, the time constant will be equal to:

.

In case of switching-off the SR group, the decay time constant will be:

.

In case of application APS decay time constant will be:

.

To perform the time constant calculations, the following extra high voltage power line data were adopted. for the next lines Khmelnytsky Nuclear Power Plant (Ukraine) – Rzeszow (Poland), South-Ukrainian nuclear power plant (Ukraine) – Isaccea (Romania), Western Ukrainian (Ukraine) – Albertirsa (Hungary). The resistance is Rpre-ins= 400 Ohms for both cases Fig. 3 a) and Fig. 3 b). In Table 1 are shown parameters of equivalent systems and shunt reactors.

Table 1. Parameters of systems and shunt reactors

.

The calculations of the aperiodic time constant according to the expressions (3), (4), (6), (7), (8) are given in Table 2.

As can be seen from the results in Table 2, the lowest constant time value is observed when using controlled switching. In the case of ASP, the lowest value will be observed, which will lead to the longest running of the aperiodic component.

Table 2. Results of calculation of aperiodic time constant

.
Fig.5. Reducing the aperiodic time constant at different measures and means
Fig.6. Equivalent scheme for replacement of ultra-high voltage power line

For example, in [1-6, 10-13] it is recommended to disable the shunt reactor group to reduce the characteristics of the aperiodic component. As you can see from the results, such an event reduces the time constant thereby increasing the duration of the transient. In fact, shutting down the bypass reactor group is ineffective. To test the efficiency, the results of which are shown in Fig. 5 were carried out.

As can be seen from Fig. 5, the use of shutdown of the shunt reactor group does not lead to a significant change in the time constant, while further dragging on the value of the impedance resistance does not lead to the required increase in the time constant.

In order to verify the correctness of the formulas and theoretical statements, a simulation model was developed for the analysis of electromagnetic transients in the TPAR cycle.

To study the electromagnetic transients, the following equivalent circuit of the transmission line replacement was used (Fig. 6), Which takes into account the real cycle of transposition. EMF was taken into account for each phase of the ultra-high voltage power transmission line. The single-phase metal short circuit was modeled in the middle of the line phase.

In Fig. 6 the following designations are introduced:

CAE CBE CCE are capacitances between the phase of the line and ground, XAB, XBC, XAC are interphase inductors, CAB,CBC,CAC are interphase capacitors, EA,EB,EC are EMF of phase, QA,QB,QC are SF6 circuit breakers, l1,l2,l3 are lengths of sections of transposition steps, l are length of line.

The list of EHV lines for which research was conducted is listed on Table 2.

Table 2. The list of EHV lines of Integrated electrical power system

.

In order to verify the correctness of the formulas and theoretical statements, a simulation model was developed for the analysis of electromagnetic transients in the TPAR cycle.

Modelling of cycle three phase auto-reclose of extra high voltage transmission line

The model was developed to study the processes at single phase auto-reclose in the environment MATLAB/Simulink which are illustrated on Fig. 7. There were made calculations to find the effective measure to prevent this kind of overvoltages. The three phase power system is simulated by voltage sources with fixed voltage and inductance. The overhead line is simulated by two parts, which are given complex matrices with distributed elements or values on the forward and reverse sequence.

Fig.7. Model of extra high voltage transmission line
Fig.8. Pre-insertion resistors connected in series
Fig.9. Pre-insertion resistors connected in parallel
Fig.10. Controlled switching device
Fig.11. Automatic phase shunting

Using the data in Table 1, we simulate extra high voltage line modes using a specific measure and technique. Each case is illustrated in Fig. 8-10.

As can be seen from Fig. 8-11, the simulation results confirm the theoretical assumptions about the duration of existence of the aperiodic component and the derived equations (3), (4), (6-8). The shortest duration of the aperiodic component is observed when using controlled switching (9). The highest duration is observed with the use of automatic phase shunting (10). The use of pre-insertions witched resistors also does not produce the required result (7-8).

Conclusions

1. When designing extra high voltage power lines, a careful analysis of the electromagnetic transients accompanying the switching during operation is necessary. Quick on-off cycles (symmetric or non-phase) should be avoided in case of TPAR. In particular, in such cases it is necessary to pay attention to the cycle of fast three-phase automatic reconnection, in which the switch may be damaged due to the aperiodic component of the current.

2. The results of the work confirm the theoretical principles, which consist in the inefficiency of using preinsertion resistances and automatic phase shunting to reduce the characteristics of the aperiodic component of the current of the total transient process. This also applies to the disconnection of a group of shunt reactors as measure. The derived formulas for the determination of the aperiodic time constant of the aperiodic component confirm the above. Changing the resistance value of the pre-insertion resistor when disconnecting the group of shunt reactors also confirmed the inefficiency of using the same means even when used together. This analysis was performed without any significant assumptions based on the mathematical models developed in MATLAB which confirm the theoretical propositions.

REFERENCES

[1] Sawi ck A., Modified arc models in a SF6 power circuit breaker, Przegląd Elektrotechniczny, 1 (2018), 93-96
[2] Czaban A., Lis M., Chrzan M., Szafraniec A., Levoniuk V. , Mathematical modelling of transient processes in power supply grid with distributed parameters, Przeglad elektrotechniczny, 2018, nr 5, 17-20.
[3] Machowski J., Kacejko P., Robak S., Mi l l e r P., Wancer z M., Analizysystemu elektroenergetycznego w średniookresowym planowaniurozwoju, Przegląd Elektrotechniczny, R. 89 (2013), nr 6, 234-237
[4] Fl isowsk i , Z. Kosz taluk , R., Methods of overvoltage reduction in powernetworks,Przeglad Elektrotechniczny, v. 77 (2001), nr. 11, 267-271
[5] Chmielewski T., Oramus P., Florkowski M., Modeling of electric arc inanalysis of switching overvoltages during current interruption by LVelectromechanical relays, Przegląd Elektrotechniczny, 2016, nr. 2, 202-206
[6] Liang Ji, Booth C., Dyśko A., Kawano F., Beaumont P., Improved Fault Location Through Analysis of System Parameters During Autoreclose Operations on Transmission Lines, Power Delivery IEEE Transactions on, vol.29 (2014), No. 6, 2430-2438
[7] A.-M. Hari r i , H. Hashemi-Dezaki, and M. A. Hejazi , “A novel generalized analytical reliability assessment method of smart grids including renewable and non-renewable distributed generations and plug-in hybrid electric vehicles,” Reliability Engineering & System Safety, 2020, vol. 196, 1-17.
[8] H. Hashemi -Dezaki, A.-M. Hariri, and M. A. Hejazi , “Impacts of load modeling on generalized analytical reliability assessment of smart grid under various penetration levels of wind/solar/non-renewable distributed generations,” Sustainable Energy, Grids and Networks, 2019, vol. 20, 1-16.
[9] H. Hashemi -Dezaki, H. Askarian-Abyaneh, A. Shams-Ansari, M. Dehghani Sanij, and M. A. Hejazi, “Direct cyber-power interdependencies-based reliability evaluation of smart grids including wind/solar/diesel distributed generations and plug-in hybrid electrical vehicles,” International Journal of Electrical Power & Energy Systems, vol. 93,1-14.
[10] Sili Y.; Gang L.; PeiRen W.; Dui feng Y. ; Xiangxiang G. ; Yuan Y. , Discussion on the Problem About Capacitive Current Switching of EHV and UHV AC Circuit Breaker, 2017 4th International Conference on Electric Power Equipment – Switching Technology (ICEPE-ST), 22-25
[11] Kachesov V. , Kachesov D. , Requirements for switching algorithms of EHV shunt compensated OHL by SF6 circuit breakers, Proceedings of International Conference on power systems Transients (IPST 2011), Delft, Netherlands, 2011, 15-18
[12] Naumkin I., Balabin M., Lavrushenko N., Naumkin R. Simulation of the 500 kV SF6 circuit breaker cutoff process during the unsuccessful three-phase autoreclosing, Proceedings of International Conference on power systems Transients (IPST 2011), Delft, Netherlands, 2011, 5-11
[13] Jian H., Xiaofeng J., Xiaoguang H., Automated monitoring and analysis for high voltage circuit breaker, 2010 5th IEEE Conference on Industrial Electronics and Applications, Taichung, 2010, 560-564
[14] Gong R., Wang S., Luo X., Analysis and Design in Extra High Voltage Circuit Breakers Employing Shunted Capacitors, 2012 Sixth International Conference on Electromagnetic Field Problems and Applications, Dalian, Liaoning, 2012, 1-4
[15] Lazimov T., Saafan E., Babayeva N., Transitional processes at switching-off capacitor banks by circuit-breakers with pre-insertion resistors, 2015 Modern Electric Power Systems (MEPS), Wroclaw, 2015, 1-4
[16] Lazimov T., Imanov S. , Saafan E., Transitional recovery voltages at capacitive currents switching-offs by vacuum and SF6 circuit-breakers, 2010 Modern Electric Power Systems, Wroclaw, 2010, 1-5
[17] Kuchansk yy V. , Application of Controlled Shunt Reactors for Suppression Abnormal Resonance Overvoltages in Assymetric Modes, 2019 IEEE 6th International Conference on Energy Smart Systems (ESS), Kyiv, Ukraine, 2019, 122-125
[18] Kuchansk yy V., The prevention measure of resonance overvoltges in extra high voltage transmission lines, 2017 IEEE First Ukraine Conference on Electrical and Computer Engineering (UKRCON), Kiev, 2017, 436-441
[19] Tugay Y. , The resonance overvoltages in EHV network, Proceedings of IEEE Sponsored Conference EPQU’09 – International Conference on Electrical Power Quality and Utilisation, Poland, Lodz, 2009, 14-18.
[20] V. Kuznetsov, Y. Tugay and V. Kuchanskyy, “Investigation of transposition EHV transmission lines on abnormal overvoltages”, Technical electrodynamics, 51-56, Nov. 2013.
[21] V. Kuznetsov, Y. Tugay and V. Kuchanskyy, ” Influence of corona discharge onthe internal ovevoltages in high way electrical networks”, Technical electrodynamics, 55- 60, Nov. 2017.
[22] Kuznetsov V., Tugay Y., Kuchanskyy V. Overvoltages in open-phase mode. Tekhnichna Elektrodynamika. No 2. pp.40–41.2012
[23] Kuchans k yy V. , The application of controlled switching device for prevention resonance overvoltages in nonsinusoidal modes, 2017 IEEE 37th International Conference on Electronics and Nanotechnology (ELNANO), Kiev, 2017, 394-399.
[24] Vladislav Kuchanskyy, Olena Rubanenko, Influence assesment of autotransformer remanent flux on resonance overvoltage UPB Scientific Bulletin, Series C: Electrical Engineering, 2020, 82(3):233-250.
[25] Live Tank Circuit Breakers. Buyer’s Guide. ABB AB. 2014. 152 p.
[26] T. Oros z , Z. Á. Tamus and I . Vajda, “Modeling the high frequency behavior of the Rogowski-coil passive L/r integrator current transducer with analytical and finite element method,” 2014 49th International Universities Power Engineering Conference (UPEC), Cluj-Napoca, 2014, pp. 1-4, doi: 10.1109/UPEC.2014.6934663.
[27] Su C., Wang P., Liao C., Chou M., Analysis of harmonic overvoltages during transformer energization for mass rapid transit systems, 2016 10th International Conference on Compatibility, Power Electronics and Power Engineering (CPE-POWERENG), Bydgoszcz, 2016,. 33-38
[28] Abdulsalam S. , Xu W. , A sequential phase energization method for transformer inrush current reduction-transient performance and practical considerations, IEEE Transactions on Power Delivery, 2007, vol. 22(1), 208-216
[29] Brunke J . H. , Fröhl ich K. J . , Elimination of transformer inrush currents by controlled switching – part II: application and performance considerations, IEEE Transactions on Power Delivery, 2001, vol. 16(2), 281–285
[30] Cho C. , Lee J. , Min B., Application of controlled switching device for high voltage circuit breaker in KEPCO real power system, 2017 4th International Conference on Electric Power Equipment – Switching Technology (ICEPE-ST), Xi’an, 2017, 492-495
[31] Hasibar R., Legate A., Brunke J., Peterson W., The Application of HighSpeed Grounding Circuit breakers for Single-Pole Reclosing on 500 kV Power Systems, IEEE Transactions on Power Apparatus and Systems, vol. PAS-100, No. 4, 1512-1515
[32] Baina H., Xin L., Jianyuan X., The Analysis of Secondary Arc Extinction Characteristics on UHV Transmission Lines, 2008 International Conference on High Voltage Engineering and Application, Chongqing, 2008, 516-519
[33] Mizoguchi H., Hioki I ., Yokota T., Yamagata Y. Tanaka K., Development of an interrupting chamber for 1000 kV highspeed grounding circuit breakers, IEEE Transactions on Power Delivery, vol. 13, No. 2, 495-502
[34] Chi T., Xin L., Jianyuan X., Zhen-xin G., Comparison and Analysis on Very Fast Transient Overvoltage
Based on 550kV GIS and 800kV GIS, 2008 International Conference on High Voltage Engineering and Application, Chongqing, 2008, 288-291.

Authors: Doc., PhD., Vladislav Kuchanskyy, Institute of Electrodynamics of NAS of Ukraine, Kiev, Ukraine kuchanskiyvladislav@gmail.com; PhD., Satyam Paul, Department of Engineering Design and Mathematics, University of the West of England, Bristol, United Kingdom satyam.paul@uwe.ac.uk; Doc., PhD., Olena Rubanenko, Regional Innovational Center at the Faculty of Electrical Engineering University of West Bohemia, Plzen, Czech Republic, rubanenk@rice.zcu.cz; PhD., Iryna Hunko, Department of Electric Stations and Systems, Vinnytsia National Technical University, Khmelnytsky highway 95, 21021, Vinnytsya, Ukraine iryna_hunko@ukr.net;

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 11/2020. doi:10.15199/48.2020.11.28

Negative Sequence Current as a Breaker Failure Protection for Medium Voltage Grids

Published by 1. Tomáš ŠKUMÁT1, 2. Žaneta ELESCHOVÁ2,
Západoslovenská distribučná, a. s. (1). Slovak University of Technology in Bratislava (2)
ORCID. 2. 0000-0002-7596-108X

Abstract. This paper focuses on situations in Medium Voltage (MV) grids where a feeder´s breaker failure occurred. This type of fault is quite severe but well handled in higher voltage systems (high voltage, extra high voltage, etc.), basically in looped grids. However, medium voltage grids were built in a different way even in terms of protection relay in power systems, automatics and backups. Several cases of a breaker failure situation have led us to reconsider the existing protection relay scheme used in Západoslovenská distribučná, a. s. (ZSD) – a distribution system operator. The paper also provides suggestions for a real operation – principle of a breaker failure protection by coordination is improved.

Streszczenie. W niniejszym artykule skupiono się na sytuacjach w sieciach średniego napięcia (SN), w których wystąpiła awaria wyłącznika linii zasilającej. Ten rodzaj błędu jest dość poważny, ale dobrze obsługiwany w systemach wyższego napięcia (wysokie napięcie, bardzo wysokie napięcie itp.), zasadniczo w sieciach zapętlonych. Jednak sieci średniego napięcia były budowane w inny sposób nawet w zakresie zabezpieczeń w układach elektroenergetycznych, automatyki i rezerwowania. Kilka przypadków awarii wyłączników skłoniło nas do ponownego rozważenia istniejącego schematu przekaźnika zabezpieczeniowego stosowanego w Západoslovenská distribučná, a. s. (ZSD) – operator systemu dystrybucyjnego. W artykule zawarto również sugestie dotyczące rzeczywistego działania – udoskonalono zasadę działania zabezpieczenia od uszkodzenia wyłącznika przez koordynację. (Prąd składowej przeciwnej jako zabezpieczenie przed uszkodzeniem wyłącznika w sieciach średniego napięcia)

Keywords: breaker failure (BF), circuit breaker, MV grid, protection relay, busbar, fault, negative sequence, protection relay.
Słowa kluczowe: uszkodzenie wyłącznika, prąd składowej przeciwnej

Introduction

A breaker failure (BF) is considered to be a situation, where a circuit breaker (CB) nearest to a fault (selectivity) does not clear the fault for some reason. There may be several reasons why this happens – CB itself can be damaged (extinguishing medium changed its properties), tripping coil is damaged, auxiliary circuits or secondary circuits encounter a problem – some types of CBs for MV use electronic mainboards), protection relay that trips the CB encounters a failure, etc.

Referring to Fig. 1, let us assume, there is a fault between CB 3 and CB 4. Protective relays associated with CB 3 and CB 4, determined to detect faults on the line between these CBs, operate and command CB 3 and CB 4 to trip. In this example, CB 3 fails to interrupt the fault current. Therefore, all sources that continue to supply the fault current through CB 3 need to be interrupted. Assuming sources at stations A and C, CBs 2, 5, and 7 need to be opened locally, or CBs 1, 6, and 8 would need to be opened remotely [1]. Major disadvantages in relation to a remote backup are wider outages (more customers affected), longer clearing time (also time of voltage dip) but the greatest advantage is independency. Local backup is substantially faster but the primary disadvantage of a local breaker failure protection (BFP) is that it may suffer from a common-mode failure [1].

Fig.1. BF procedure [1]

BF can be caused by a variety of situations – a failure to trip or a failure to clear. In the first case, the breaker contacts do not open after the trip circuit has been energized by the protection (“stuck breaker”). In the latter case, the contacts open but the arc is not extinguished and current continues to flow [1].

Common elements of a BF to interrupt the scheme include the following [1]:

• Scheme initiation by a breaker trip signal such as a protective relay that has operated to trip the breaker.

• Determination that the breaker has tripped successfully by monitoring reset of an overcurrent element (50BF) that responds to each measured phase current (50P) and possibly the sum of these phase currents (50G), monitoring change in state of the circuit breaker auxiliary contact (52a, 52b or 52aa), or a combination of these methods.

• A timer.
• Some means to trip and block closing of adjacent breakers.

• Optional – a separate output contact to issue a re-trip signal to the circuit breaker before issuing a breaker failure output with sufficient margin such that successful opening of the circuit breaker will prevent and undesired BF output.

• Optional – a teleprotection channel to key a direct transfer trip (DTT) and to cancel reclosing of remote circuit breakers. In general, there are several BFP schemes [1]:

• Basic BF scheme.
• Basic BF with re-trip logic.
• BF scheme for dual breaker arrangements.
• BF scheme based on 50 BF pickup time.
• BF scheme with two-step timing arrangement.
• BF initiate seal-in.
• BF minimal current scheme.
• Dual timer BF scheme with a fast breaker auxiliary contact and a current detector reset check.
• Triple timer BF scheme.
• Single-phase tripping, BF, and re-trip logic.
• BF timer bypass scheme.
• Current differential BF protection.
• Ground fault BF on both a live tank circuit breaker and a current transformer column failure.
• Series (tandem) breakers.
• BF protection for generator applications.
• Mechanical indication of breaker status (52a).

BF can be a part of [1]:

• Primary protection for an element,
• Feeder, transformer, motor or transmission line protection devices,
• Centralized bus protection devices,
• Distributed bus protection devices

Merits, advantages and disadvantages of integrating BFP with zone protection relays are discussed in [2]. In the [2] methods to improve the security of BFP are also reviewed.

Fundamentals of a local and remote backup protection in combination with real experience from utilities are discussed in [3]. It is also very important to watch a failure rate and take measures where it is inevitable. Issue of reliability characteristics of the 110 kV / MV station and MV switchgears is presented in [4].

BF Protection by Coordination

Faults occur frequently in MV grids. In general, the number of faults increases downstream from a transmission system to a distribution system. Therefore, there is a higher chance of BF.

Looped grids (HV systems, extra HV systems, etc.) were built with a different approach for a breaker failure situation.

The main difference between looped grids and MV grids may be already apparent. On one hand, there is an operational difference and on the other hand, there is also a technological difference. MV grids are operated radially. Wiring between protection relays is not done like in looped grids. There are also different types of protection relays installed – overcurrent instead of distance protections. Busbar differential protection is also not a common protection in MV grids. Feeder´s breaker failure was historically left just to an overcurrent protection relay installed to a feeding transformer [5, 6] – a so called breaker failure protection by coordination. There is one major drawback in this kind of solution – it may be impossible for the backup protection relay to see all faults.

Referring to Fig. 2, let us assume, there is a fault on the feeder with CB4 that has encountered a BF. Backup for this situation is provided by a protection relay that operates CB1

– in general a transformer´s overcurrent protection relay installed on MV side. Simplicity is a very significant advantage in relation to this solution

– no extra equipment, no risk of mis-operation. It is an ultimate protection, which covers all failures, not just BF (failure of CB, relay, settings, controls and wiring, etc.). The main disadvantage is speed – it is a slow type of BF backup.

Fig.2. BF procedure in MV grids

It is a common practice for distribution, but typically not sufficient for transmission. Solutions for transmission were already mentioned in Introduction. However, the BFP by coordination can be improved by a negative sequence current protection.

MV Feeding Transformer´s Neutral Earthing

There is one major aspect that must be specified, which is HV/MV transformer´s neutral point earthing on the MV side. Basic types are [7]:

• Isolated
• Solid
• Low impedance
• Resonant (compensated)

MV grids in Europe are not operated as solidly grounded [8]. This type of operation is very convenient from economical and maintenance perspective but there is a very severe impact on reliability indicators SAIDI and SAIFI. Other types are common for European MV grids, whereas resonant grounding is very popular indeed. Smaller grids (low capacitive currents of shunt capacitances) can be operated as isolated but with grid expand it becomes meaningless.

Low impedance is mostly used in pure cable grids. Reason for this is very simple – majority of faults are permanent (insulation system breakdown of cables), although there are some pilot projects handling with resonant grounding even in case of pure cable grids – further improvement of reliability indicators. A risk of cross-country fault increases in this case.

Fig. 3 shows how the issue of transformer´s neutral point earthing is approached in ZSD.

There are 3 types of it:

a) Low impedance – low resistance resistor is used.
b) Hybrid – compensated (Petersen coil) and in case of a ground fault a low impedance resistor is temporarily switched in parallel to the coil.
c) Compensated – Petersen coil + auxiliary winding with a secondary resistor.

Type a) is used in pure cable grids.

Type b) is used in grids with majority of cables, but also with some overhead lines – periphery of cities.

Type c) could be used in this case but a primary resistor is kept because a fault location system used in such grids is based on I0 measurement used in ring main units – a primary resistor is a very convenient solution for this method (higher currents during a 1 phase fault in comparison with a secondary resistor). Only the currents are measured – a robust and reliable fault location system used in pure cable grids of ZSD.

There are following basic shunt-type of faults, in relation to Fig. 3:

  • One-phase faults
    o Ground faults
    a) high current
    b) temporary high current
    c) low current (operational)
  • Two-phase faults
    o Two-phase with ground
    o Two-phase
    o Cross-country fault
  • Three-phase faults.

Fig.3. Transformer´s neutral earthing used in ZSD

Negative Sequence – Symmetrical Components

A very common theory and a tool for mathematical and graphical interpretation of asymmetrical phasors in three phase systems is based on symmetrical components. Any three asymmetrical phasors can be replaced by a system of symmetrical phasors (sequences) – positive (index 1), negative (index 2) and zero (index 0) sequence as illustrated in Fig. 4. Phase A is a base.

Fig.4. Graphical interpretation of symmetrical components
.

Equations (1), (2), (3) are derived and considered to be a final product, especially the Equation (2) is important for this paper [9].

Analysis of some basic faults mentioned in Chapter II is provided for the purpose of the paper. A simplified grid shown in Fig. 5 is used for the analysis. No load conditions are assumed in the following theoretical analysis. There is a general impedance connected between the earth and the transformer’s neutral point, which is explained more explicitly later. Let us assume that the fault is at the end of the line. Type of the fault is specified in subchapters.

Fig.5. Simplified MV grid

A. Ground Fault – Low Impedance Grounding

Fig. 6 shows sequence circuits for this case. They are connected in series on the basis of the fundamental theory. For analytical and practical purposes, the following assumptions can be applied – transformer´s sequence impedances ZT,1, ZT,2 and ZT,0 can be neglected. Also, line´s sequence impedances ZLine,1, ZLine,2 and ZLine,0 can be neglected. It´s because their values are several times lower in comparison with resistance of grounding resistor – RG (RG >> ZT and RG >> ZL).

What shouldn´t be neglected in general regarding real grids is grid´s shunt impedance (mainly a line to ground capacitance) shown also in Fig. 5. It is apparent from Fig. 6 that the following aspects have impact on I2 magnitude – low impedance grounding resistor, resistance of the fault itself, shunt capacitance of the grid.

Because the distributed capacitive reactances (impedances) ZShunt,1, ZShunt,2 and ZShunt,0 are very large, while the series impedance values ZT,1, ZT,2, ZT,0, ZLine,1, ZLine,2, ZLine,0 are very small, thus, practically, ZShunt,1 is shorted out by ZT,1 + ZLine,1 in the positive sequence network, and ZShunt,2 is shorted out by ZT,2 + ZLine,2 in the negative sequence network. Since these series impedances are very low (they can be neglected as previously mentioned), Z1 and Z2 approach zero relative to the large values of ZShunt,0 and RG.

Fig.6. Sequence circuits – a) ground fault in the grid

Moreover, Equations (4), (5) and (6) are derived with the assumption of no load conditions [9] in the fundamental theory. Therefore:

.

If RF goes to zero, then magnitude of the fault current is:

.

The real fault recorded by the feeder´s protection relay is shown in Fig. 7 and Fig. 8. The fault was in phase A (in the diagrams L1 marked in green colour). Sequence currents are aligned in phase with the faulted phase A as it is according to the theory. The fault started as a cross-country fault. It is apparent from Fig. 7 – the short circuit current went to several kAs but the protection relay of another feeder cleared it, anyway, after about 60 ms the fault re-appeared, but this time as a single line-to-ground fault. The primary resistor used in this grid limits the current to 600 A.

Fig.7. Real line-to-ground fault in the grid a) – RMS values
Fig.8. Real line-to-ground fault in the grid a) – phasor diagram

Pre-fault and fault current conditions are compared in Table 1, in relation to the real fault depicted in Fig. 7 and Fig. 8. Highly symmetrical load is apparent.

Table 1. Pre-fault and fault conditions

.
B. Ground Fault – Resonant Grounding

Fig. 9. shows sequence circuits for this case, i. e. the same connection as in Fig. 6, just a different device connected to a transformer´s neutral point – adjustable Petersen coil (LP) which compensates line to ground capacitive currents of healthy phases in the location of a fault. Sequence currents are as follows:

.

The same assumptions and neglections made in ground fault with low impedance grounding can be applied also in this case with positive, negative and zero sequence impedances of a transformer and a line. Equation (5) for sequence voltages applies also in this case. If XLp – XShunt,0 = 0, then I1 = I2 = I0 = 0. Situation in compensated grids (resonant grounding) is therefore different. After decay of transient phenomena, the new steady-state currents of either healthy or even faulty feeder have operational values. It is apparent from Fig. 10 and Fig. 11. In symmetrical grids, there is almost no I2 current (Fig. 11). Fig. 10 illustrates behaviour of currents after a fault in phase A occurs – the faulted phase A current decreased a bit, the healthy phase B current increased quite a bit and the healthy phase C current decreased a bit. The reason behind this is the following:

• Phase A system´s capacitance is shortened by a ground fault, so the corresponding capacitive current goes to zero.

• There is a shift of healthy phase voltages, the angle between them is not anymore 120°, but 60° and this fact causes also shift in capacitive currents. There is also a change in magnitude of the currents which is √3 higher than a pre-fault per each phase. Both the angle and the magnitude depend on the value of RF.

• Ratio between a load and a capacitive current is also important.

Fig.9. Sequence circuits – b) ground fault in the grid

This fact is apparent even from Table 1. However, it is less obvious due to the current of the primary resistor. Throughout the system the distributed capacitance XShunt,1 and XShunt,2 is actually parallel with the series reactances XT, XLine and so on, so that in the system I1 and I2 are not quite equal to I0 in the system [10]. If shunt impedances of the system are symmetrical, then capacitive currents of healthy phases do not contribute to I2 current. This is also apparent from Fig. 11.

Fig.10. Real line-to-ground fault in the grid b) – RMS values
Fig.11. Real line-to-ground fault in the grid b) – phasor diagram

However, I2 current would appear temporary – after a primary resistor is switched in parallel with the coil. It is common for the operation that XLp is tuned close to the resonance.

Therefore, in resonant grounded systems, which are symmetrical, I2 would be very low, almost equal to zero. Real fault recorded by a feeder´s protection relay is as follows:

If RF goes to zero, then the fault current is:

.

Essentially, the fault current is equal just to resistive losses of the system´s shunt impedance, however, the coil must be tuned to resonance. In general, a so called detuning current can be assumed and then the magnitude of IF is:

.

Pre-fault and fault current conditions are compared in Table 2, in relation to the real fault depicted in Fig. 10 and Fig. 11. Highly symmetrical load is apparent.

Table 2. Pre-fault and fault conditions

.
C. Line-to-Line

Fault Shunt impedances can be neglected for this type of fault, their contribution to fault current is very small. A fault between phases B-C is considered. The following assumptions applies for the fault – IA = 0, IB = -IC and UfB = UfC (if RF = 0). Therefore, I1 = -I2 and I0 = 0.

Fig.12. Sequence circuits – line-to-line fault

Connection of sequence circuits in Fig. 12 corresponds with the applied assumptions. Equations (11), (12), (13) can be derived from Fig. 12 and the assumptions applied for this type of fault:

.
.

Equation (13) shows the major advantage that comes with negative sequence current protection relay – normal overcurrent relay must be set up to x.IN (x.IB), where x starts roughly at 1,2. Negative sequence current relay is more sensitive by a factor of √3.

Real fault recorded by a feeder´s protection relay is as follows:

Fig.13. Real line-to-line fault – RMS values
Fig.14. Real line-to-line fault – phasor diagram

Pre-fault and fault current conditions are compared in Table 3, regarding the real fault depicted in Fig. 13 and Fig. 14. Highly symmetrical load is apparent.

Table 3. Pre-fault and fault conditions

.

It is apparent that IA increased a bit – this is caused by MV/LV distribution transformers and their windings connection – Dyn or Yzn is used in grids of ZSD. |IB| ≠ |IC| because of load currents.

Load itself can be quite a significant contributor in general regarding the magnitude of I2, however, it is not the case with ZSD´s grids – whole lines are built as three-phase (highly symmetrical shunt impedance / capacitance) and there is no significant 1-phase or 2-phase load connected to the MV grid.

There is I2 during cross-country and a two-phase-to-ground faults. There is no I2 during a three-phase fault according to the theory – a symmetrical fault. It is obvious that I2 protection relay can´t be applied as a general solution according to the theory, but also according to recorded faults.

Aspects Influencing I2 Magnitude

The following aspects and their influence on I2 are analysed for the purpose of this paper – fault´s resistance (RF), line´s length and a special case of asymmetrical load that can occur even in symmetrical grids. Calculations were carried out in Excel based on equations stated in the theoretical chapter. Grid´s line-to-line voltage U = 22 kV, transformer´s SN = 40 MVA. Current´s base for pu system is 1 kA.

The graph in Fig. 15 shows relation between |I2| and RF in grids from Fig. 3 – a) and temporarily in b). Line´s series impedance was neglected in this calculation. When the ratio between a grounding resistor and a cable´s impedance is taken into consideration, then the neglection is acceptable.

It is apparent that I2 magnitude drops relatively fast but one fact must be mentioned – real faults in cable grids and their RF are close to 0. In Fig. 15, there are 3 curves in relation to neutral point resistor´s nominal current. Commonly used values were analysed.

Fig.15. Influence of fault´s resistance on |I2| – low impedance grounding

However, line´s impedance shouldn´t be neglected in case of overhead lines. The length of a line is for sure a limiting factor for I2 magnitude in this case. Commonly used AlFe conductors are included in Table 4, where the reactance per kilometer was calculated for a typical conductor arrangement used in MV grids of ZSD.

Table 4. MV AlFe conductors

.

Relation between the length of a line and I2 magnitude is shown in Fig. 16 for a different type of AlFe conductors from Table 4. Line-to-line fault is assumed at the end of a line.

Fig.16. Influence of the length of a line on |I2|

The graph in Fig. 17 shows relation between |I2| and RF on a line with AlFe 110 – a cross section commonly used in the main part of a MV line. The length of a line starts at 10 km and goes to 100 km with increment of 10 km.

Fig.17. Influence of RF on |I2|

The last analysis deals with the situation which can cause asymmetrical load conditions even in symmetrical grids. It is caused by asynchronous closing of contacts in case of any kind of switchgear used in a grid – a disconnector, a CB or a load-break switch. We have encountered situations, especially with load-break switches mounted to poles, with two contacts closing synchronously and the third one closing after 1s – this time has been the maximum that we have experienced in ZSD grid.

These switches are often used during maintenance in a grid. Therefore, this fact must be taken into consideration regarding time grading of I2 protection relay. A symmetrical grid with IA = 1∠0° pu, IB = 1∠-120° pu and IC = 1∠120° pu is assumed. Closing of a contact in phase C is asynchronous for this analysis and corresponding I2 is analyzed in Fig. 18.

Fig.18. Influence of load asymmetry
Fig.19. Active and reactive power during the fault
Real Fault Before I2 Protection Relay Implementation

The cause of the fault was a direct lightning strike to an overhead line. Direct lightning strikes can be quite easily confirmed thanks to modern technologies and tools provided by meteorological institutes. It was a two-phase fault. The maximum power achieved during the fault was SMAX = 23,16 MVA – Fig. 19 and the supplying HV/MV transformer´s nominal power SN = 25 MVA. It was literally just a heavy load for the machine. The fault was cleared naturally after roughly 40 s because two conductors were broken and fell to the ground – load side (not towards the feeding substation), so the arc was extinguished. This is obviously an unwanted result – the fault itself was there for such a long time that it caused severe damage to the equipment of the distribution grid.

Fig.20. Damage caused to the overhead line after a breaker failure – 2 downed conductors

Implementation Process

Situation in our grids before implementing the current negative sequence protection relay was as follows:

• Backup was handled only by a HV/MV transformer’s overcurrent relay.

• Wiring between the protection relays does not support a similar solution which is used in higher voltage systems. Rewiring the circuits of relays in all our substations would be both time and economically consuming. Implementing the current negative sequence relay required:

• Setting up the function in the existing relays.
• Setting up the signal in a SCADA.
• Testing the relay.
• Testing the communication to a SCADA and a dispatch centre.

All the points mentioned above would be mandatory even in case of rewiring. Thing is that these 4 points were handled in general by our own staff.

A. Negative Sequence Relay Setup

The most important thing that must not be overlooked when we implement the I2 relay is the symmetry of the load throughout the grid. The maximum |I2| in our grids is 5.78 A, the relative value is 5.45% (statistical data are from 91 HV/MV transformers). Therefore, ZSD grids can be considered still very symmetrical. If there is a significant asymmetrical load connected to a HV/MV transformer in a grid, then the I2 protection relay can be setup just to a signalization instead of a standard trip function.

What must be considered in general regarding setup of I2 protection relay is the ratio of feeder´s current instrument transformers and a feeder´s overcurrent protection I> threshold. If there is a situation where some feeders have different ratios of current instrument transformers compared to others, then I2 protection relay must be activated also on feeders with higher ratios. This is a precaution for a situation of a resistive fault that occurred on a feeder with higher overcurrent I> setup. This can trip a feeding transformer and it would be considered a false trip.

B. Implementation to Existing Relays

All protection relays used in the grid are digital. Majority of ZSD MV substations have two main bus-bars. Each busbar has a single supplying transformer (HV/MV). In this case, there is always one bus coupler (BC). Some substations are also equipped with bus sectionalizers (one for each bus), then there are two separate bus couplers. The new standard of our MV substations (GIS – gas insulated substation) is – two main bus-bars, two bus-sectionalizers (with CB) and two bus couplers. Only a few substations have a single bus-bar and no bus sectionalizer (BS). The summary of typical cases in ZSD grids is as follows:

• Two main buses (two HV/MV transformers) and BCs – I2 function is active in protection relays of transformers (MV side) and both BCs.

• Two main buses (two HV/MV transformers), two BCs and two BSs – I2 function is active in protection relays of transformers (MV side), BCs and BSs (if these are equipped with a CB).

• A single bus-bar (one HV/MV transformer) – I2 function is active only in a protection relay of a transformer (MV side).

• A single bus-bar (one HV/MV transformer) and a BS – I2 function is active in protection relays of a transformer (MV side) and a BS.

I2 protection relays of BCs and BSs are graded from protection relays of transformers – selectivity. So, in case of a BF, there is mostly one of the bus switches (BS or BC) clearing this fault as a backup – only part of a substation would be affected by an outage. In cases with a single bus and no BS, the whole bus would be affected by an outage.

Conclusion

BF is a severe type of fault. Not all MV substations and infrastructure of their protection relays were built in a way similar to higher voltage level substations (wiring mentioned in the introduction, or technological perspective -busbar differential protection). The only backup in case of a breaker failure was historically left to ordinary overcurrent protection relays of feeding transformers and bus sectionalizers – a so called BFP by coordination. The combination of long overhead lines, a fault at the far end of such an overhead line and low load conditions (natural behaviour during nights) can create situations with faulty currents which cannot be identified as a fault for conventional overcurrent protection relays (lower values below a threshold). Feeder´s breaker failure during such conditions is a serious issue.

This paper presents the theoretical analysis combined with real faults. Aspects affecting |I2| are also stated and analysed. Implementation of I2 protection relay considerably improves the situation. BFP by coordination is improved due to this solution – the problem with longer lines is diminished. Response of an operator (at a control room / dispatch centre) cannot be fast enough for breaker failure situations. However, it is not a general solution because there are also symmetrical faults – three-phase faults. These faults would not be identified by the relay. It is also important to say that this type of fault is statistically less occurring. These faults with the conditions already mentioned – the fault is at the far end of a line, the load current is low and the total current is below overcurrent relay´s threshold, would be left just to the operator and remote control of the CB.

Acknowledgment: This publication was created thanks to support under the Operational Program Integrated Infrastructure for the project: International Center of Excellence for Research on Intelligent and Secure Information and Communication Technologies and Systems – II. stage, ITMS code: 313021W404, co-financed by the European Regional Development Fund.

REFERENCES

[1] IEEE Guide for Breaker Failure Protection of Power Circuit Breakers, IEEE C.37.119-2016, 2016.
[2] Kasztenny B., Thompson M., Breaker Failure Protection – Standalone or Integrated With Zone Protection Relays?, Proceedings of the 2nd Annual Protection, Automation and Control World Conference, Dublin, Ireland, June 27-30, 2011.
[3] Xue Y., Thakhar M., Theron J. C. and Erwin D. P., Review of the Breaker Failure Protection practices in Utilities, 2012 65th Annual Conference for Protective Relay Engineers, College Station, TX, USA, 2012, pp. 260-268.
[4] Chojnacki A. Ł., Podstawowe funkcje niezawodnościowe stacji 110kV/SN oraz rozdzielni sieciowych SN, Przegląd Elektrotechniczny, 2016, R. 92, NR 8, pp. 238-241.
[5] Network Protection & Automation Guide, Alstom Grid, May 2011. ISBN: 978-0-9568678-0-3.
[6] Janíček F., Chladný V., Beláň A., Eleschová Ž., Digitálne ochrany v elektrizačnej sústave. Bratislava: Vydavateľstvo STU, 2004 Slovak republic. 360 pages. 1. release, ISBN: 80-227-2135-2.
[7] Procházka K., Vybrané problémy provozu distribučních sítí VN, VÚE Brno, 1992, pp. 1-64.
[8] Grym R., Hochman P., Bermann J., Machoň J., Cichoň B., Chránění II. Havířov: Irena Satinská – IRIS, 2004. 305 pages. ISBN: 80-903540-0-9.
[9] Trojánek Z., Hájek J., Kvasnica P., Přechodné jevy v elektrizačních soustavách. Praha: Nakladatelství technické literatury, 1987 Czech republic. 312 pages. 1. release
[10] Blackburn L.J., Symmetrical Components for Power Systems Engineering. New York: Marcel Dekker, Inc., 1993 USA. 427 pages. 1. Release.

Authors: Ing. Tomáš Škumát, Západoslovenská distribučná a. s., Čulenova 6. 816 47 Bratislava, Slovakia,
E-mail: Tomas.Skumat@zsdis.sk
doc. Ing. Žaneta Eleschová, PhD. Slovak University of Technology, Ilkovičova 3. 812 19 Bratislava, Slovakia,
E-mail: zaneta.eleschova@stuba.sk

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 97 NR 11/2021. doi:10.15199/48.2021.11.25

The Mathematical Model of High Voltage Switch as An Element of a Power System

Published by 1. Andriy CZABAN1, 2. Vitaliy LEVONIUK2, 3. Radosław FIGURA1,
Kazimierz Pulaski University of Technology and Humanities in Radom (1), Lviv National Agrarian University (2) ORCID: 1. – 0000-0002-4620-301X; 2. 0000-0003-2113-107X; ; 3. 0000-0001-8048-5623

Abstract. In this work, on the basis of an interdisciplinary method of modelling, a mathematical model of a power system is presented, main element of which system is an gas switch. For modelling of non-mechanical part, a popular theorem on flickering centre of speed of rotation was used. For a mathematical model of an arc, when the resistance was of non-linear nature, a resistive-capacitive alternative diagram was used. The final system of differential equations was presented in a normal cauche form. Results of simulation were given as drawings and were analysed.

Streszczenie. W pracy na podstawie interdyscyplinarnej metody modelowania zaproponowano model matematyczny układu elektroenergetycznego, głównym elementem którego wyłącznik gazowo-elektryczny typu SF6. Dla modelowania części mechanicznej wyłącznika wykorzystano słynny teoremat o migowym centrum prędkości obracania. Dla modelu matematycznego luku wykorzystano rezystancyjno-pojemnościowy schemat zastępczy, gdy rezystancja rozpatrywała się jako funkcja nieliniowa. Końcowy układ równań różniczkowych przedstawiony w normalnej postaci Causze’go. Wyniki symulacji komputerowej podano w postaci rysunków, które analizują się. (Model matematyczny wyłącznika wysokiego napięcia jako elementu układu elektroenergetycznego).

Keywords: high voltage switch, mathematical modelling, the Hamilton-Ostrogradskii principle, the Euler-Lagrange equation.
Słowa kluczowe: wyłącznik wysokiego napięcia, zasada Hamiltona-Ostrogradskiego, równania Eulera-Lagrange’a, nieustalone procesy.

Introduction

Analysis of unsteady electromagnetic processes of complex power systems is extremely important both at the design stage and during exploitation of the said systems, which are elements of one integrated power system.

It is obvious that such systems are characterized with very sophisticated physical processes, for the purpose of analysis of these it is necessary to use complex mathematical model, based on the theory of common differential equations, and sometimes on common equations and partial derivative equations.

Power systems consist of many various components. In order to provide highly adequate mathematical model it a must to describe each of the operating electric devices in details. Depending on the task, attention should be paid to a required device.

This work concentrates on one of the most crucial element of power systems of high voltage – the gas switch. In order to switch long power supply lines on/off, it is the said device that is used. In the course of analysis of unsteady processes of power supply line of high voltage, an important problem occurs, i.e. the necessity to model electric arc processes in switches of high voltage and high voltage. In a general case when current function curve breaks due to contacts opening, an arc is formed, physical principles of which arc are highly complex. These principles are described by plasma theory, electromagnetic field, thermodynamics theory etc. Nowadays, the models of these devices [1] are quite complex and cumbersome. Therefore they are not always acceptable in the analysis of transients in electrical networks.

Another factor to consider is the mechanical processes that occur in the switch when moving the contacts. These processes take as much time as the electromagnetic ones that occur in the switched elements of electrical networks, and therefore can affect the latter.

It is obvious that providing highly adequate mathematical model of the switch requires both use of complex theory and few experts.

Nowadays, SF6 circuit breakers ABB type LTB 362-800 (T) E4 are widely used for switching elements of ultra-high voltage electrical networks on the territory of European Union and in the CIS countries. These switches consist of two modules, each of which has two pairs of series-connected contacts – two fixed and two movable, which are driven by one mechanism for moving the contacts (see Fig. 1). Capacitors are connected in parallel to the contacts for even voltage distribution during switching.

Figure 2 presents proposed simplified kinematic diagram of right side of the mechanism triggering of switch’s contacts. Left side is symmetrical to the right one.

Fig. 1. The mechanism for moving the contacts of the SF6 circuit breaker ultrahigh voltage company ABB type LTB 362-800 (T) E4

Fig.2. Simplified virtual kinematic scheme of replacement of the mechanism of movement of contacts of the switch.

The mentioned mechanism of movement of contacts (see Fig. 1) consists of two crank mechanisms which are symmetrical concerning a longitudinal axis of the whole mechanism. The mechanism is driven by a spring (not shown). These crank mechanisms have a specific design, because the axis of movement of the spring and the axis of movement of the contacts do not coincide with the center of rotation of the mechanism itself. Such crank mechanisms are called deaxial.

We have already built a mathematical model of such a mechanism [2], however, with the high adequacy of this model, it is too cumbersome and difficult to implement as a computer program for the potential user. Therefore, in the present paper, we propose to use a mathematical model of the switch with a simplified kinematic substitution scheme for the analysis of transients in electrical networks with SF6 switches. A comparative analysis of the work of both models showed the convergence of the obtained results by 92%. This gives grounds for the application of a simplified substitution scheme when the calculations do not require high accuracy. The use of a simplified kinematic substitution scheme does not significantly affect the adequacy of the results, but allows to significantly simplify the model of the switch.

In this article, on the basis of variational approaches, we will build a mathematical model of the electrical network, the main element of which is a switch. This approach makes it possible to avoid the decomposition of a unified dynamic system and to obtain the initial equations of the electromagnetic and mechanical state exclusively from a unified energy approach, by constructing an extended Lagrange function [3].

Recent research analysis

Power system problems are often discussed in scientific articles, there are plenty of alike works. Work [4] provides analysis of problems concerning design of mathematical models and macro models of power supply lines with switches. The mathematical model here is developed in the MatLab/Simulink software package. Obtained results are presented.

In [5], the study of electromagnetic transients during the shutdown of short-circuit currents on the 500 kV transmission line with shunt reactors turned on is presented. Here, the model of the switch and other elements of the electrical network is built in the EMTP-RV software package.

In [6] shows the results of researches of the reasons of accidents of SF6 switches during switchings of the compensated power lines of 750 kV. Based on the simulation model developed in the MatLab/Simulink software package, the study of electromagnetic processes in compensated power lines depending on the switching moments is carried out.

In [7] the influence of transient switching processes in power lines on the operational state of the power system was studied. In [8] the transient electromagnetic processes in the power line with shunt reactors during an emergency shutdown due to a short circuit were analyzed. In these works, the study was also carried out in the software package EMTP-RV.

Analysis of the above mentioned works, and numerous articles, make us judge that their authors in the course of modeling transient electromagnetic states neglect mechanical processes in high voltage switches, which are commensurate in duration with electromagnetic, and therefore can affect them. We also see that in the present time the most popular software packages for the study of switching processes are MatLab and EMTP. However, in these systems, the switches are modeled in such a way that the break in the function of the circuit breaker current occurs exclusively at zero, which is not always true.

Based on the above, the aim of this work is a developing of a mathematical model of the ultra-high voltage switch taking into account the simplified kinematic scheme of replacement of the mechanism of moving its contacts to simplify the switch model and analysis of switching transients in electrical network elements.

Fig.3. Power system diagram

Mathematical model of the system

Figure 3 shows part of power grid of high voltage, which part consists of power system (EMF), internal resistance, inductance, high voltage switch, the latest presented as nonlinear active-capacitive components connection, П-alternative diagram of power supply line with concentrated parameters and resistive-inductive load.

We propose to use variational principles for building of mathematical models of electrical network elements, in particular the modified Hamilton-Ostrogradsky principle [3].

For the system we study, the extended action functional by Hamilton-Ostrogradski and its variation will be the following [3]:

.

where S is the action by Hamilton-Ostrograd, L* is the augmented Lagrange function.

Expanded lagrangian of the integrated system is given as follows [3]:

.

where Т~* – kinetic coenergy, P* – potential energy, Φ* – energy dissipation, D* – energy of outside nonpotential forces.

Next, elements of dispersed expanded non-conservative lagrangian is presented:

.

where LS1, LН, LL – system inductances; RS1, RН, RL – system resistances; eS1 – electromotive force; СL1, CL2 – line’s capacities; gL1, gL2 – line’s conductivities; iS1, iН, iL, igL1, igL2 – currents; CV – capacity of contacts capacitor; rD – non-linear resistance of arc; iD – arc’s current; ∆х – spring’s end shift; Vx – spring’s end speed; k – resilience coefficient; kd – spring diffusion coefficient; m – contacts mass reduced to a spring system; FX – arc displacement force reduced to a spring system; Qj – charge of the j-th element; uj – the voltage of the j-th element.

For example, we present a differential equation that describes the transient process of current in an equivalent power system (System, see Fig. 3) based on the principles of variation.

We write the Euler – Lagrange equation [3]

.

where q is the generalizing coordinate; q̇ – the speed of the generalizing coordinate. As a generalizing coordinate we take the charge – the speed of the generalizing coordinate

qS1 = QS1, q̇S1 = dqS1/dt = dQS1/dt = iS1.

We should note that in the Euler-Lagrange equation (7) we substitute only the components that relate to the element for which we obtain the equation of the electromagnetic state, since the derivatives of other functions (generalized coordinates) are identically zero, because the latter is not differentiated.

We write the Euler – Lagrange equation (7) taking into account (3) – (6) given that

∂Т* / ∂QS1= uV – uCL1.

.

We obtain the equation of the extremals of the Hamilton action functional by changing the order of differentiation (8) and applying the theorem on the derivative of the integral over the upper bound:

.

This way we obtain an equation that describes the current of an equivalent power system:

.

In order to avoid overloading the article with mathematical inferences, we will not reflect the procedure for obtaining the rest of the equations, but only present their final form. You can learn more about the principles of obtaining the equations of such a prolan, for example, in [3, 9]

.

where igL1, igL2 – line leakage; uCL1, uCL2 – line start voltage, line end voltage; uV – voltage between the contacts.

As already mentioned, electric arc resistance rD is nonlinear and it depends on distance between contacts. In order to describe this resistance, it is necessary to know how the resistance is related regarding contacts distance.

On the basis of Fig. 2, the equation takes the following form:

.

Then, referring to the popular theorem on flickering centre of speed, the equation is [10]:

.

On the base (15) and (16) the Vx is calculated by formula:

.

Using (15), angle β is calculated:

.

Next, putting (18) into (17) Vy speed is calculated:

.

Differential equations for finding speed increase will be obtained Δyfakt

.

On the basis of Fig. 2, equations for finding functions Δy and Δxfakt are made

.

In a general case, resistance rD is of exponential nature [11].

Therefore approximation is done using 5-degree polynomial:

.

Dependence (22) initially has a slowly increasing character (simulates arc combustion), and when the contacts diverge by a distance of more than 0.02 m – a sharply increasing nonlinear character (simulates arc attenuation).

Integration is calculated for differential equations: (10) – (12), first equation in (13), (14), (20). This is due to the following second and third equations in (13), (19), (21), (22).

Computer simulation results

Computer simulation of transient states was carried out for power supply line presented in Fig. 3, mechanical processes in high voltage switch of type ABB LTB 362-800 (T) E4 were calculated.

The experience proceeded as follows. The system startup was realised by activating EMF. After reaching transient state (0,18 s) in p. К, symmetrical 3-phase short circuit was triggered. Next, in 0.04 s (time for activation up of relays protection) contacts started to disconnect. The system parameters were: eS1 = 612sin(ωt + 20°) kV, RS1 = 2,032 Ω, LS1 = 0,161 H, RL = 6,85 Ω, LL = 0,333 H, CL1 = CL2 = 0,0000024 F, gL1 = gL2 = 0,0000058 Sm, CV = 400·10-12 F, LH = 0,7 H, RH = 800 Ω, k = 320000 N/m, m = 2,032 kg, kd = 0 Ns/m. The following assumption was adopted: FX = 0 N (arc displacement force not referred to).

Fig.4. Transient switch current

Temporary switch current values are shown at Figure 4. Analysis of the figure let one conclude that for short circuit the surge current is over 7 times higher. Within this time interval, protection system got activated up. Next due to switch operation, the current began to actively drop down and in 0.3 seconds its value was zero. The line got disconnected.

In Fig. 5, voltage between switch’s contacts was presented. It is obvious that in an operating state of the system the voltage value is zero. Next, in the switch activation state, the voltage increases to 750 kV. With contacts disconnection resistance grows, in 0.03 s voltage on contacts equals EMF rated voltage, 630 kV.

Fig.5. Transient voltage between the contacts.

Fig.6. Transient current in power supply line

In Figure 6 transient current in power supply line is presented. It can be observed that in operating state of the system, the maximum current value was 0.9 kА, in short circuit state it was 7 kА. When comparing Figures 4 and 6, a difference is observed, the difference may be a result of complex processes occurring on the line end.

Fig.7. Transient voltage on the right of the switch

Fig.8. Transient voltage on the left of the switch

Figures 7 and 8 are very interesting, they visualize voltage on the right and on the left of the switch. Analysis of the figures leads to the following conclusion: after short circuit on the line end (p. K),the voltage on the line start dropped to 400 kV. After activation/of the switch, voltage on the right dropped down to zero (Fig. 7). As for the second figure (Fig. 8), the case is different. Voltage in a transient state equals system’s EMF.

Fig.9. Switch’s contacts disconnection as a function of time, within time interval [0,2; 0,28] s

Transient disconnection of switch’s contacts within time interval [0,2; 0,28] s shows at Fig. 9. This drawing can be treated as a light theme of this work concept. The approximated curve shows the dependence of arc resistance as a function of the distance of contacts and time. It has been described in the present paper by a mathematical model. This model takes into account the elements that make up the high-voltage circuit breaker.

Conclusions

1. The use of the modified Hamilton-Ostrograd principle gives the possibility to build mathematical models of very complex dynamic objects. Therefore, this method is used to describe electrical power systems.

2. The developed mathematical model of the ultrahigh voltage switch taking into account the virtual simplified kinematic scheme of replacement of the mechanism of movement of contacts, allows to reproduce with a sufficient level of adequacy transient electromagnetic and mechanical processes in the switch taking into account physical principles of switching. In this case, this also applies to the rupture of the current function not only during its transition through zero, but also in a fairly wide time range around zero.

3. The results of the computer simulation presented in this paper fully confirm the theories of transient processes in gas switches, which are complicated high voltage power systems.

REFERENCES

[1] A. Ahmethodžić, M. Kapetanović, and Z. Gajić, «Computer Simulation of High-voltage SF6 Circuit Breakers: Approach to Modeling and Application Results,» IEEE Transactions on Dielectrics and Electrical Insulation, 2011. 18, (4), р. 1314 – 1322.
[2] A. Chaban, A. Szafraniec, and V. Levoniuk, «Mathematical modelling of transient processes in power systems considering effect of high-voltage circuit breakers,» Przeglad elektrotechniczny, 2019. 1, р. 49 – 52.
[3] A. Chaban, «Zasada Hamiltona-Ostrogradskiego w układach elektromechanicznych,» Publisher T. Soroki, Lviv, 2015. (Ukr).
[4] P. Stakhiv, «Discrete mathematical macromodel of electric transmission,» Przegląd Elektrotechniczny, 2013. 4, p. 272-274.
[5] I. Naumkin, M. Balabin, N. Lavrushenko, and R. Naumkin, «Simulation of the 500 kV SF6 circuit breaker cutoff process during the unsuccessful three-phase autoreclosing,» Proceedings of International Conference on power systems Transients, Kyoto, Japan, June 14-17, 2011. р. 5 – 11.
[6] V. Kuchanskyi, «Controlled switching SF6 breakers in the main power electrical grids,» Proceedings of the IED NASU, 2017. 48, р. 38 – 42.
[7] P. Dehghanian, and M. Kezunovic, «Probabilistic impact of transmission line switching on power system operating states,» Proceedings of International Conference on Transmission and Distribution and Exposition (T&D), 2016.
[8] D. Lin, H. Wang, and S. Shen, «An adaptive reclosure scheme for parallel transmission lines with shunt reactors,» Proceedings of International Conference on Transmission and Distribution and Exposition (T&D), 2016.
[9] A. Chaban, V. Levoniuk at all, «Mathematical model of electromagnetic processes in Lehera line at open-circuit operation,» Electrical Engineering & Electromechanics, 2016, 3, p. 30 – 35.
[10] K. Pomorski, «Theoretical mechanics,» Publisher Maria CurieSkłodowska University, Lublin, 2000. (Pol).
[11] V. Sidorets, and I. Pentegov, «Deterministic chaos in nonlinear circuits with an electric arc», Publisher Welding, Kiev, 2013. (Rus).

Authors: dr hab. inż. Andriy Czaban, prof. UTH Rad., University of Technology and Humanities in Radom, Faculty of Transport and Electrical Engineering, ul. Malczewskiego 29, 26-600 Radom, E-mail: atchaban@gmail.com; Ph.D. Vitaliy Levoniyk, Lviv National Agrarian University, Department of Electrical Systems, 1, V. Velykogo str., 80381 Dubliany, Lviv region, Ukraine, E-mail: Bacha1991@ukr.net; assoc. prof., Ph.D Radosław Figura, University of Technology and Humanities in Radom, Faculty of Transport and Electrical Engineering, ul. Malczewskiego 29, 26-600 Radom, E-mail: r.figura@uthrad.pl.

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 97 NR 7/2021. doi:10.15199/48.2021.07.19

Large-Power Synchronous Motor Braking by Field Current

Published by Marian HYLA, Silesian University of Technology, Department of Power Electronics, Electrical Drives and Robotics

Abstract. The paper presents the braking process of a large-power salient-pole synchronous motor forced by field current. The motor control was carried out using a microprocessor block for the excitation of large-power synchronous motors. Cases of free retardation and braking resulting from eddy currents generated in the machine body by the magnetic field of the rotating winding of the rotor with open stator windings were considered. The method proposed was to determine the dependence of braking time on the value of the field current on the basis of the observation of changes in the rotational speed without the knowledge of the parameters of the engine and braking moments of the drive system. The measurement verification of the braking time calculated for the assumed field current has been presented.

Streszczenie. W artykule przedstawiono proces hamowania jawnobiegunowego silnika synchronicznego dużej mocy za pomocą prądu wzbudzenia. Sterowanie silnika zrealizowano za pomocą mikroprocesorowego bloku zasilania wzbudzenia silników synchronicznych dużej mocy. Rozpatrzono przypadek wybiegu swobodnego oraz hamowania na skutek prądów wirowych generowanych w korpusie maszyny przez pole magnetyczne wirującego uzwojenia wirnika przy rozwartych uzwojeniach stojana. Zaproponowano metodę wyznaczania zależności czasu hamowania od wartości prądu wzbudzenia bazującą na obserwacji zmian prędkości obrotowej bez konieczności znajomości parametrów silnika i momentów hamujących układu napędowego. Zaprezentowano weryfikację pomiarową czasu hamowania obliczonego dla założonego prądu wzbudzenia. (Hamowanie silnika synchronicznego dużej mocy prądem wzbudzenia)

Keywords: eddy-current, field current control, synchronous motor braking
Słowa kluczowe: prądy wirowe, regulacja prądu wzbudzenia, hamowanie silnika synchronicznego

Introduction

Large-power synchronous motors are used in industry to drive devices that do not require speed control. Typical applications are the drives of main ventilators in coal mines. The mining regulations in force in Poland require that in each exhaust shaft, in addition to the active main fan or the main fan set, a back-up fan is required , which can be started within 10 minutes [1]. Failure of the main fan and failed start-up of the reserve fan poses a serious threat to the health and life of the employees , and a break lasting more than 20 minutes results in suspension of works and evacuation of workers towards the inspiratory shafts or to the surface [1].

Due to the mass of the fans, reaching several tens of tons, and diameters up to 9 metres, this type of propulsion system is characterized by a high moment of inertia, about 10 times greater than the moment of inertia of the rotor of the propulsion engine. For this reason, coasting times can reach tens of minutes, and the engine can be re-started only after it has been stopped.

It is therefore advisable to shorten the coasting process of motor with braking in order to prepare it for a restart.

Braking methods of large-power synchronous motors

After switching off the motor supply during operation, the rotational speed decreases until it is completely stopped under the influence of the rotary motion resistance forces. The time to stop the motor is strictly dependent on the antitorque moments and the moment of load of the drive system. This method of stopping the motor is often insufficient due to technological reasons as the process lasts relatively long. Therefore other methods are used to accelerate the process of stopping the engine by forced braking.

The motor is braked by converting its kinetic energy into a different kind of energy. During electric braking the kinetic energy is converted into electrical energy and, eventually, into another type of energy, e.g. thermal energy.

In practice, with large-power synchronous motors, dynamic braking is used, in which involves disconnecting the stator windings from the supply network and switching on the resistance into the stator circuit for energy discharging. In order to achieve the most effective braking, relatively large resistances are switched on in the stator circuit, so that the braking from the synchronous speed takes place with the maximum braking torque, and then subsequent stages with appropriately reduced resistance are switched on.

Another method is the use of a mechanical brake placed on the machine shaft and increasing the frictional moment. Mechanical braking is used in solutions where braking reliability is required, even after power failure. This method causes rapid wear of the brake friction covers. For this reason, the mechanical brake is usually used in the final phase of motor braking process at a low shaft speed.

In drive systems with an inverter in the stator circuit, generator braking with the return of energy to the grid is used. By correspondingly reducing the frequency of the motor stator supply voltage, according to the braking ramp that maintains the right angle between the axis of the stator magnetic field and the rotor’s magnetic field, the motor can be braked in a relatively short time. Other electrical braking methods such as counter-current braking, DC braking or single-phase braking are not used in large-power synchronous motors.

Object of the research

The test object was a synchronous motor type GYd-178sp/02 with rated active power 4000 kW, stator voltage 6000 V, stator current 500 A and rotation speed 750 rpm, coupled on a shaft witch two P-1500/10/250/03 type DC generators with rated active power 1750 kW, stator voltage 650 V and stator current 2700 A. During the tests, the synchronous generators were turned off, increasing only the resultant moment of inertia, the moment of friction and the moment of ventilation losses of the entire drive system.

Control of drive operation was carried out by the ProgressPOWER microprocessor block for the excitation of synchronous motors [2] developed in cooperation with the author. The block diagram of the device is shown in Figure 1.

ProgressPOWER excitation power supply block is designed for cooperation with large-power synchronous motors with rated stator voltage 6 kV and excitation current up to 400 A.

The device is managed by a microprocessor system, and implemented algorithms allow to perform asynchronous start-up in classical or starting-choke systems, the control of synchronous operation with the possibility of reactive power or field current regulation, and technological or emergency drive shutdown with energy discharging of the field circuit through inverter operation of the thyristor converter [3], or braking of the motor by field current. The field current control is carried out by a microprocessor system by means of changes in the thyristors switching delay angle of the rectifier that supplies the excitation winding.

Fig.1. Diagram of the synchronous motor control system with microprocessor block for excitation and start-up reactor: M – synchronous motor, WT – thyristor exciter, µP – microprocessor system, PT – thyristor rectifier, UR – start-up system, W – circuit breaker, O – disconnector, WD – choke circuit breaker, DR – inrush choke

In the circuit of the starting resistor, transistor keys were used in the configuration enabling the flow of bidirectional current induced in the field winding during the asynchronous motor start-up. The contactless excitation system allows to increase the reliability and durability of the device.

The microprocessor system, in addition to controlling the current in the excitation circuit, controls the circuit breakers in the 6 kV field, controlling the permissible working area and the state of internal and external protections located, for example, in the switch bay supplying the motor.

A stand-alone operation mode or cooperation with an external, superior control system is available. Cooperation with external devices is carried out through built-in RS-485, USB and Ethernet communication interfaces. Suitable firmware versions enable the device to work with a medium voltage inverter in the synchronous motor stator circuit or control the voltage of the synchronous generator.

Free retardation braking

Figure 2 shows the registered course of the rotational speed of the drive system during free retardation of the synchronous motor. The motor was switched off while running at a synchronous speed, and the braking time was 15 minutes and 26 seconds. The shutdown consisted of opening the stator power switch and quickly discharging the energy of the field circuit by preparing the thyristor bridge to inverter-type work [3].

Fig.2. Speed course during coasting of the motor

The motion equation for the rotating element is as follows

.

where: J – moment of inertia, M – moment affecting the drive system, ω – angular speed.

During the braking process, the load torque M has a negative sign. For an unloaded motor coast, this moment is the sum of the friction torque Mf and the moment of ventilation resistance Mv:

.

where kv – a constant reflecting the influence of rotational speed on the value of the moment of ventilation resistance, hence the equation of motion can be written as

.

and after differentiation, the expression is given for the angular velocity as a function of time

.

After the braking time th the speed ω is 0 (slip s=1), i.e.:

.

Knowing the braking time th, the moment of inertia J can be expressed in the following form

.

and including in (4) the following is obtained

.

simplifying the entry as

.

Equation (7) describes the curve passing through the point A(t=0, s=0) and B(t=th, s=1) of the waveform in Figure 2 (where s – slip corresponding to the speed of rotation) however, such curves are infinitely many depending on the coefficient λ. To determine the coefficient λ, the knowledge of the additional point of ω=f(t) characteristic is required.

Assuming the coordinates of the point C(t=5, s=0.6) for the motor speed curve from Figure 2, the value of the coefficient λ was determined by a numerical method as λ=1.181·10-3 s2, which allows the expression of the curve shape of the braking using the analytical equation without the knowledge of the values of the moment of inertia of the drive system and the moment of friction and ventilation losses.

Equation of the synchronous machine motion

The mathematical model of a salient-pole synchronous machine, due to the magnetic asymmetry of the rotor, is usually presented in the form of a system of equations describing currents, voltages and fluxes expressed in relative units in a d-q system rotating at synchronous speed, supplemented by a motion equation in the following form

.

where: TM – mechanical time constant, ψd, ψq – relative rotor magnetic flux linkages in d and q axis, id, iq – relative stator currents in d and q axis, m0 – relative load torque. According to the presented standard mathematical model, with the stator windings open (id=0, iq=0), the braking torque is equal to the moment of load m0, in which the moment of friction and the moment of ventilation losses can be taken into account.

The spinning rotor of a synchronous machine with forced field current generates a magnetic field permeating the stator’s static structure. The magnetic flux, which is variable in the time and space generates loses in the iron, which can be divided into hysteresis losses and eddy-current losses.

Hysteresis losses over one period are proportional to the hysteresis loop area, and are proportional to the frequency of re-magnetization of the iron. For a synchronous machine with forced current flow into the excitation winding, this frequency is related to the rotor speed.

Eddy-current losses are associated with the induction of iron currents due to the changing magnetic field. With sinusoidal flux variability, based on empirical studies by Richter, they are assumed to be proportional to the square of the frequency of changes in the magnetic field.

In a rotating synchronous machine with open armature winding and forced flow of field current, the magnetic flux is greater than the flux in the rated synchronous operating state due to the lack of reaction of the magneto-motive force of the armature, which causes a much greater impact of eddy-currents on the braking torque than in the synchronous state, not included in the standard mathematical model of the synchronous motor.

Braking the motor by field current

Increasing the effectiveness of the motor braking process is possible thanks to the induction of eddy-currents in the stator magnetic circuit due to the flow of current in the rotating field winding.

The general theory of eddy currents is known, and research on eddy-current brakes has been carried out for over 100 years [4-7], using more and more modern techniques recently, e.g. by means of finite element analysis [8, 9]. The main problem in this type of research is the proper determination of the model parameters and the need to adopt some simplifying assumptions.

Fig.3. The speed curves during motor braking for different field current values

The presented method does not require knowledge of the system parameters, and is based on the observation of the course of the rotational speed change during the motor braking process. In addition, it does not require the use of additional braking devices, using the phenomena caused by the flow of current in the rotating field winding with open stator windings.

Figure 3 shows the recorded waveforms during the deceleration process of the tested motor for different values of the field current. For field current If1=150 A, the overrun time has been reduced to 8 minutes 55 seconds and for field current If2=345 A to 3 minutes 12 seconds.

Assuming that the hysteresis losses are proportional to the speed of the magnetic field’s rotation generated by the field current, and the eddy-current losses are proportional to the square of this speed, the motion equation can be represented as

.

where: A, B – constants with unknown values. Differentializing (10) gives an equation describing the angular speed ω as a function of the field current If and coefficients A and B, and additionally it is difficult to determine the friction torque Mf on the basis of the speed during the coasting, which makes it impossible to determine the numerical values of the unknowns even when knowing additional points of registered speed curve.

There are methods to determine the mechanical losses of the drive system [10-12], but they require measurement of other quantities, e.g. power at the moment when the braking procedure starts.

Assuming that the additional braking moment related with the magnetic field from the field current depends on the rotational speed of the rotor and the rotational speed during the braking process depends on time, a correcting element has been introduced into (7), obtaining a dependence in the following form

.

where C and D are unknown constants, and the t/th member takes into account the influence of speed on the additional braking torque related to the current flow in the field circuit.

For the excitation current If1, the rotational speed ω reaches the value 0 after the braking time th1, which considering in (11) allows to save the coefficient C in the following form

.

Equation (12) means that there are infinitely many coefficients C and D that provide the solution.

Taking into account the braking time of the th2 engine with the If2 excitation current, the numerical value of the D coefficient can be determined

.

and next, coefficient C from (12) can be determined too. For the tested drive system these values were determined as C=2.462·10-4 s-1A-D and D=2.333.

The determined coefficients reflect the influence of various phenomena, such as losses on hysteresis and eddy-currents loses, magnetic circuit nonlinearity, parasitic moments and other phenomena difficult to describe in an analytical manner and requiring the adoption of many simplifying assumptions.

Taking into account the determined coefficients C and D in (11), it is possible to determine analytically the field current causing the motor to stop after a given time tx

.

Based on (14), the dependence of the braking time on the field current shown in Figure 4 was determined.

Fig.4. Dependence of the braking time on the field current on the base (14)

Determining the time after which the motor will stop at the given field current, due to the entanglement of time t in (14), is possible only through a numerical calculation process. For the assumed field current If3=400 A, based on (14), the engine stoppage time was determined as 2 minutes 31 seconds. Figure 5 shows the registered motor speed during braking process in such a case. The real braking time was 2 minutes 23 seconds.

Fig. 5. Recorded motor speed during braking at 400 A field current
Conclusions

Using the braking moments generated from the eddy-currents in the machine body generated by the flow of current in the rotating field winding of the rotor, the motor braking time can be significantly reduced without the use of additional braking devices. It is also possible to quite accurately estimate the field current causing the drive to stop within the set time without knowing the magnetic and mechanical parameters of the system. The presented method of determining the influence of the field current on the course of the synchronous motor braking process does not require knowledge of the motor parameters, and at the same time allows to take into account the influence of real phenomena in the motor stator magnetic circuit, which are associated with the interaction of the rotating magnetic field generated by the current in the field winding.

The considerations concerned the coasting and braking of the synchronous motor non-loaded by the braking torque of driven device. The braking torque of the driven device causes the motor to stop in an even shorter time, but regardless of this, the maximum reduction of the braking time will be achieved at the maximum permissible field current. But it should be taken into account that the current flow through the field winding causes losses in this winding, resulting in an increase of its temperature , simultaneously the reduction of the speed of rotation causes the ventilation and cooling of the winding to become less effective.

The tests carried out with the motor and the microprocessor block for excitation [2] confirmed the possibility of controlling the motor braking time with the help of the field current. The application of the field current control system regardless of the state of the motor operation makes the braking procedure effective even after an emergency power off and opening of the stator windings.

REFERENCES

[1] Decree by the Minister of Economy from 28 June 2002 on occupational health and safety, mining operations and fire protection in underground mines, Republic of Poland Journal of Law from 2002, No 139, item 116 9 and from 2006, No 124, item 863
[2] Hyla M.: Power supply unit for the excitation of a synchronous motor with a reactive power regulator, Mining – Informatics, Automation and Electrical Engineering, vol. 53, no. 1, pp. 17-21, 2015
[3] Hyla M.: Wybrane aspekty sterowania tyrystorową wzbudnicą silnika synchronicznego (in Polish, abstract in English), 5th Conf. Modelling and Simulation, Kościelisko, 2008, pp. 345–348
[4] Morris D. K., Lister G. A.: The eddy current brake for testing motors, Journal of the Institution of Electrical Engineers, vol.35, no. 175, pp. 445-468, 1905
[5] Davies E. J.: General theory of eddy-current couplings and brakes, Proceedings of the Institution of Electrical Engineers, vol. 113, no. 5, pp. 825-837, 1966
[6] Gonen D., Stricker S.: Analysis of an Eddy-Current Brake, IEEE Transactions on Power Apparatus and Systems, vol. 84, no. 5, pp. 357-361, 1965
[7] Venkataratnam K., Kadir M. S. A.: Normalized Force-Speed Curves of Eddy Current Brakes with Ferromagnetic Loss Drums, IEEE Transactions on Power Apparatus and Systems, vol. PAS-104, no. 7, pp. 1789-1796, 1985
[8] Holtmann C., Rinderknecht F., Friedrich H. E.: Simplified model of eddy current brakes and its use for optimization, 10th Int. Conference on Ecological Vehicles and Renewable Energies (EVER), Monte Carlo, 2015, pp. 1-8
[9] Sinmaz A., Gulbahce M. O., Kocabas D. A.: Design and finite element analysis of a radial-flux salient-pole eddy current brake, 9th Int. Conf. on Electrical and Electronics Engineering (ELECO), Bursa, 2015, pp. 590-594
[10] Ilina I. D.: Experimental determination of moment to inertia and mechanical losses vs. speed, in electrical machines, 7th Int. Symposium on Advanced Topics in Electrical Engineering (ATEE), Bucharest, 2011, pp. 1-4
[11] Ćalasan M., Ostojić M., Petrović D.: The retardation method for bearings loss determination, Int. Symposium on Power Electronics Power Electronics, Electrical Drives, Automation and Motion, Sorrento, 2012, pp. 25-29
[12] Ćalasan M. P., Petrović D. S., Ostojić M. M.: Electrical braking of synchronous generators for combined generator and water turbine bearings as well as stray-load losses determination, IET Electric Power Applications, vol. 7, no. 4, pp. 313-320, 2013

Autor: dr inż. Marian Hyla, Silesian University of Technology, Faculty of Electrical Engineering, Department of Power Electronics, Electrical Drives and Robotics, ul. B. Krzywoustego 2, 44-100 Gliwice, Poland, e-mail: marian.hyla@polsl.pl

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 11/2020. doi:10.15199/48.2020.11.04

Influence of Environmental Exposures on Electrical Parameters of Low Voltage Surge Arresters

Published by Piotr PAJĄK1, Bartłomiej SZAFRANIAK1, Anna DĄDA2, AGH University of Science and Technology, Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical Engineering (1), MGGP S.A., Krakow Branch (2)

Abstract. Low voltage surge arresters work in a very different environmental conditions. During the exploitation, under the influence of environmental exposures, the structure of metal oxide surge arresters (MOSA) is gradually degraded. These processes can change their protective properties and lead to a reduction in the effectiveness of surge protection. The aim of the paper is to analyze the influence of environmental exposures on the electrical parameters of low voltage surge arresters.

Streszczenie. Warunki środowiskowe w jakich pracują niskonapięciowe ograniczniki przepięć są bardzo zróżnicowane. Podczas eksploatacji, pod wpływem narażeń środowiskowych, struktura ograniczników przepięć ulega stopniowej degradacji. Procesy te mogą powodować zmiany parametrów elektrycznych ograniczników i prowadzić do zmniejszenia skuteczności ochrony przeciwprzepięciowej. Celem referatu jest analiza wpływu narażeń środowiskowych na parametry elektryczne niskonapięciowych ograniczników przepięć. Analiza wpływu narażeń środowiskowych na parametry elektryczne niskonapięciowych ograniczników przepięć

Keywords: metal oxide surge arresters, impedance spectroscopy, water immersion, leakage current.
Słowa kluczowe: beziskiernikowe ograniczniki przepięć, spektroskopia impedancyjna, degradacja pod wpływem wilgoci, prąd upływowy

Introduction

The main purpose of using surge arresters is to protect the insulation of devices from significant voltage increases that may occur in the power system. Reliable operation of the surge arresters and preservation of reduced voltage values declared by manufacturers within a long time is required in order for this protection to be effective. Surge arresters are exposed to many environmental factors during their exploitation, which directly affect the operating parameters, advance their ageing and in extreme cases may damage them. Current stroke, high temperature, moisture penetration and salt contamination are these negative impact factors. Substantial penetration of moisture in combination with high air pollution, significantly advance the degradation of surge arresters, decreasing the protective voltages of the varistor and with an avalanche way increase in leakage current, they can even lead to its damage [1].

The basic element of sparkless surge arresters are varistors made of zinc oxide (ZnO) with additions of other metal oxides, Bi2O3, CoO, MnO, Sb2O3 among others. The varistors are resistors whose resistance depends on the voltage applied to them. The material from which the varistors are made is properly prepared – first ground and homogenized, and then pressed and sintered at high temperatures. The end result of such a technological process is a polycrystalline structure with unique properties. The basic element of the varistor’s microstructure is grain, which should be characterized by appropriate size, homogeneity and low resistivity. The formation of grain microstructure, occurring at the stage of production processes, determines the creation of current paths in the varistor. The current flowing through the varistor is the sum of partial currents flowing through its structure through various paths. Paths with fewer borders are led by larger currents. Voltage drops in a given varistor sector depend on the number of intergranular borders, which is directly influenced by the amount and size of grains. The total voltage drop on the varistor results from the equalization of voltages by varying the leakage currents [2].

The chemical composition and suitably selected parameters of the varistor manufacturing process allow for the shaping of their strongly nonlinear current-voltage characteristics (U = C · Iβ; V – voltage, I – current, C – constant, β – nonlinearity coefficient) [3, 4].

The mechanism of current conduction on the varistor is related to phenomena occurring at the boundaries between grains. At the interface of neighboring grains there are potential barriers resulting from the electric charge accumulated on the boundary surfaces. In the characteristics of the dependence of the electric field intensity on the value of varistor current density E = f (J), three clearly distinct ranges can be distinguished (Fig. 1): pre-breakdown range (normal operation range), breakdown range (stabilization range) and saturation range [5].

Fig.1. The characteristics of the electric field intensity dependence on the current density [4]

Varistors are usually made in the form of cylindrical disks of various diameters and thicknesses. They are closed in sealed enclosures designed to protect against external factors, in particular against moisture (Fig. 2a, 2b). The geometrical dimensions of the varistor are related to the assumed value of the discharge current (diameter, surface area of the disk), the operating voltage (thickness of the disk) and its ability to absorb energy (disk volume) [6]. Typically, the ZnO grain diameter is a few micrometers (Fig. 2c).

Diagnostic tests of surge arresters are carried out at all stages of their technical life. They are based on various evaluation methods, for example on visual inspection, thermovision studies, or on the measurement of electrical parameters, performed at both constant and alternating voltages. The energy (overvoltage) and environmental (physicochemical) exposures occurring in operation lead to changes in the internal structure and various types of damage to the arresters [6].

Fig.2. A low voltage metal-oxide varistor as the main element of the surge protection device (SPD): a) example of the low voltage SPD design, b) manufactured disk varistors, c) the microscopic level varistor structure

Research Program

This work contains tests and analysis of the results performed on a series of low voltage surge arresters with similar parameters produced by two different manufacturers. Four samples were selected for observing the parameters of the tested arresters, two samples of object A and two samples of object B. The analysis of the results is to determine the permissible dispersion in which the monitored parameters may be contained. The tests were performed on surge arresters designed for installation in low voltage aerial power transmission networks. Their technical parameters are listed in table 1.

Table 1. Selected parameters of tested SPD

.

All presented measurements were made in High Voltage Laboratory of Electrical and Power Engineering Department of AGH University of Science and Technology in Krakow, Poland. Surge arresters were tested with the following procedure:

Impedance spectroscopy

The impedance spectroscopy method was used to observe changes in the dielectric properties of the tested surge arresters. This is a test method used to determine the physical and chemical properties of materials and electrochemical processes. It consists in measuring the linear, electrical response of the tested object as a result of stimulation with a small electromagnetic signal in a wide frequency band. Small voltage induction allows treating the tested element, in this case the surge arrester as a linear element. The measuring instrument depending on the impedance frequency Z(ω) is the frequency response analyzer (FRA). This device generates a stimulation of a specific shape and selects individual points on the frequency scale. An impedance spectrum is created from the measured current responses at selected frequencies. The current flowing through the tested object is combined with two synchronous, orthogonal reference signals (cosωt and sinωt), one of which corresponds to an trigger signal.

Fig.3. Configuration of a laboratory stand for measuring dielectric parameters of tested surge arresters: 1) wideband impedance measurement system (Solartron 1260 + 1296); 2) surge arrester

The characteristics of the relative permittivity εr and the dielectric loss coefficient tgδ were registered depending on the frequency f for all samples, in no way previously operated or subjected to laboratory exposures. The measurements were carried out over a wide range of frequencies, from 10-2 Hz to 104 Hz. The measuring station was equipped with a Solartron 1260A frequency response analyzer, Solartron 1296A dielectric interface and a supervising computer (Fig. 3).

Electrometer – recording of the leakage current waveform

In order to performed detailed diagnostic, it is particularly important to analyze the pre-breakdown range of surge arresters, because this is the state of their normal operation. Then, a leakage current flows through the arrester. Measurement of this current enables the technical condition of arrester to be evaluated. Observation of the value and shape of the leakage current is often used to assess the operation of surge arresters [7-10].

The tested surge arresters were connected in such a way that the measurement only includes the cross-current component of the leakage current. The surface current was shielded – the arresters were placed in a copper band which was grounded.

The leakage currents waveforms were recorded over time, all samples were subjected to the experiment. The measurement took place for one minute at constant voltage 308V, which is the highest value of the operating voltage. The measurements were carried out using an electrometer from the B2987A series of Keysight Technologies.

Exposure to moisture and salt solutions

Surge arresters were combined into two pairs. Each pair is a object A surge arrester and a object B surge arrester. In order to imitate the actual environmental conditions to which devices are subjected during exploitation, each pair of arresters has been drenched in the following liquids: water and 5% NaCl solution. Ageing was executed at an ambient temperature of 23°C to 25°C and lasted 600h. Each time, after another 150 hours of ageing, before the next tests, the surge arresters were dried.

Results and Analysis of Measurements

The article presents the results of testing low voltage surge arresters. Tests were performed using the impedance spectroscopy method. This is one of the nondestructive testing methods that can be used for diagnostic of ZnO varistors. The obtained test results are presented in the following figures (Fig.4 – Fig.7).

Fig.4. The influence of water immersion on ageing of the object A parameters: a) dependence of relative permittivity on frequency εr(f), b) the dependence of the dielectric loss factor on the frequency tgδ(f)

Fig.5. The influence of water immersion on ageing of the object B parameters: a) dependence of relative permittivity on frequency εr(f), b) the dependence of the dielectric loss factor on the frequency tgδ(f)

The analysis of impact of ageing based on the εr and tgδ wideband characteristics shows gradual growth of these parameters’ values. That effect has been observed within the range of low frequencies for all surge arresters. It was also found that the reaction of surge arresters within the same voltage group, immersed in a 5% NaCl solution, is different. Strong growth of object A surge arresters’ tgδ can be observed at the frequency range of kHz’s whereas this parameter does not change for object B surge arresters at the same range of frequency.

Fig.6. The influence of 5% NaCl immersion on ageing of the object A parameters: a) dependence of relative permittivity on frequency εr(f), b) the dependence of the dielectric loss factor on the frequency tgδ(f)

Fig.7. The influence of 5% NaCl immersion on ageing of object B parameters: a) dependence of relative permittivity on frequency εr(f), b) the dependence of the dielectric loss factor on the frequency tgδ(f)

Fig.8. The influence of immersion in water on leakage current of the arresters for 308V: a) object A b) object B

Fig.9. The influence of immersion in 5% NaCl solution on leakage current of the arresters for 308V: a) object A b) object B

Tested surge arresters have reached similar values of leakage currents. That value is influenced by the relatively short ageing time and lack of raised temperature. Nevertheless, the moisture penetration into arresters’ interior increases the leakage current. In the example of object A, it can be observed that the leakage current values grow with time, and the differences between respective waveforms are greater while immersed in a 5% NaCl solution. In the case of object B, these differences are not so significant.

The voltage–current characteristics of metal oxide varistors depend on moisture content at low current, in the vicinity of continuous operating voltage [11]. Greater leakage current of surge arresters immersed in 5% NaCl solution in relation to immersion in water was observed. The changes are more significant for the arresters manufactured by object A than object B, which means that they are much more sensitive to moisture impact. The use of this type of arresters in environments with significant contamination and salinity may cause major difficulties in maintaining the stability of the requested parameters.

Summary and Conclusions

Based on the results, moisture penetration into arresters interior is confirmed. The impact of immersion in water and in a 5% NaCl solution on the surge arresters’ parameters was already noticeable, despite the relatively short ageing period. It points out that a deterioration of their technical condition ensued.

There are visible changes in the shape of wideband characteristics as a result of the surge arresters’ ageing test. Within the range of low frequencies, the εr and tgδ characteristics reach the higher values along with the length of ageing time, as well as for immersion in water and in 5% NaCl solution. Tgδ(f) characteristics indicate an significant increase in the conductivity losses at low frequencies.

The impact of moisture penetration into varistors’ interior is also visible on obtained leakage currents waveforms. Surge arresters aged in 5% NaCl solution show higher values of leakage current compared to an analogous test in water. The values of these currents are different for various series of arresters used within the same voltage group. Analysis of appearances occurring during the ageing processes is an additional source of information on the mechanism of current conduction in the varistor.

REFERENCES

[1] Wanderley Neto E. T., Da Costa E. G., Ferreira T. V., Maia M. J. A., Failure Analysis in ZnO Arresters Using Thermal Images, IEEE PES Transmission and Distribution Conference and Exposition Latin America, Venezuela, (2006)
[2] Mielcarek W., Uwarunkowania technologiczne warystorów tlenkowych, Prace Instytutu Elektrotechniki, Zeszyt nr 212, Warszawa, (2002)
[3] Chrzan K. L., High-Voltage Surge Arresters, Wrocław, Dolnośląskie Wydawnictwo Edukacyjne publishing company, (2003)
[4] Flisowski Z., High Voltage Technology, Warszawa, PWN publishing company, (2017), 226–234
[5] Florkowska B., Furgał J., Zydroń P., Materials Engineering in Electrical Engineering. Laboratory, AGH issues, Kraków, (2012), 72–80
[6] Bonk M., Fusnik Ł., Szafraniak B., Zydron P., Influence of temperature and high-energy stresses on selected wideband characteristics of metal-oxide varistor parameters, Proc. 9th Int. Sci. Symp. Electr. Power Eng. Elektroenergetika 2017, (2017), 306–310
[7] Biglar B., Jayaram S., Cherney E. A., Evaluation of the insulation design of polymer housed surge arresters using saltfog test, 2000 Conference on Electrical Insulation and Dielectric Phenomena, (2000), 385–388
[8] Da Silva D. A., Costa E. C. M., Pissolato J., De Jesus R. C., De Franco J. L., De Abreu S. R., Romano M. A. A., Lahti K., Evaluation of the moisture ingress and the electrical performance on polymeric surge arresters of distribution networks, 18th International Symposium on High Voltage Engineering, (2013), 349–354
[9] Silva D. A., Costa C. M., Franco J. L., Abreu S. R., Jesus R. C., Antonionni M. , Pissolato J., Polymer Surge Arresters: Degradation Versus Electrical Performance, (2012)
[10] Dalai S., Chatterjee B., Study on the effect of moisture ingression into metal oxide surge arrester using leakage current analysis, 3rd International Conference on Condition Asstessment Techniques in Electrical Systems (CATCON), (2017), 330–334
[11] Chrzan K. L., Influence of moisture and partial discharges on the degradation of high-voltage surge arresters, European Transactions on Electrical Power, Euro. Trans. Electr. Power 2004, (2004), No. 14, 175-184

Authors: dr inż. Piotr Pająk, AGH University of Science and Technology, al. Mickiewicza 30, 30-059 Kraków, Poland, E-mail: ppajak@agh.edu.pl, mgr inż. Bartłomiej Szafraniak, AGH University of Science and Technology, al. Mickiewicza 30, 30-059 Kraków, Poland, E-mail: szafrani@agh.edu.pl, mgr inż. Anna Dąda, MGGP S.A., Cracow Branch, e-mail: annadada0709@gmail.com

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 1/2020. doi:10.15199/48.2020.01.14