Power Quality Blog

10 Actions Companies Can Take Right Now to Reduce Energy Costs and Carbon Emissions

Published by Reuters Events^TM, The industrial energy efficiency playbook, November 2022.
A report in conjunction with ABB and Energy Efficiency Movement^TM .

Executive summary

The world’s industries stand at an energy crossroads in 2022. The urgency of climate change demands action on all sides—from industry, governments and civil society. Energy shortages, brought on by the loss of Russian oil and gas supplies following the invasion of Ukraine in February 2022, have led to inflationary pressures and new energy security challenges that only add to this urgency.

Improving energy efficiency is an under-exploited opportunity to reduce both costs and emissions. While there has been a lot of discussion about how individuals can contribute to saving energy and how consumers can take steps to reduce their bills, the significant potential for energy efficiency and cost improvements in industry has received less attention.

Industry is the world’s largest consumer of electricity, natural gas and coal, according to IEA figures.ⁱ The sector accounts for 42% of electricity demand, equal to more than 34 exajoules of energy. The iron, steel, chemical and petrochemical industries are the largest consumers of energy among the world’s top-five energy-consuming countries—China, United States, India, Russia and Japan.ⁱⁱ

This energy consumption carries high costs in the current inflationary environment. It also created nine gigatonnes of CO_2, equal to 45% of total direct emissions from end-use sectors in 2021, according to the IEA.ⁱⁱⁱ

The importance of energy efficiency in this context cannot be overstated. “We call it the ‘first fuel,’” says the IEA’s senior programme manager for energy efficiency, Kevin Lane. “It needs to be front-loaded across all the sectors.”

This report, developed with ABB and other members of the Energy Efficiency Movement, provides a playbook for executives to address energy efficiency to help mitigate climate change and rising costs. It details 10 actions that industrial leaders should consider for their organisations (see Fig. 1) beyond basic energy efficiency “hygiene” measures like switching equipment off when not in use, converting fluorescent or halogen lighting to LED or insulating walls and piping.

The 10 actions have had to meet some essential criteria: 1) they rely on mature – secure, widely available – technologies; 2) they are material enough to have a meaningful impact on both energy costs and emissions; and 3) they can be deployed quickly without complex or expensive integrations. One of the most attractive aspects of energy efficiency for industry is that in many cases companies can enjoy significant improvements with little or no capital spending.

This list is by no means exhaustive and should be viewed as an inventory of short and medium-term opportunities for industry, as well as an invitation to discuss and document the solutions, use cases and best practices in energy efficiency. Readers are invited to engage with #energyefficiencymovement to share their energy challenges and discuss ideas and lessons learned.

10 key energy efficiency actions for industrial leaders

Action #1: audit operations for energy efficiency

One of the quickest and easiest sources of energy efficiency gains in industry can come from optimising the way assets and processes function. An energy efficiency audit creates an important baseline for a business to make improvements and identify improvement opportunities. Audits can be carried out by established energy service companies (ESCos) and will provide a benchmark to identify potential areas for efficiencies, develop an action plan and measure progress.

What’s involved?
An initial energy audit usually involves an analysis of historical energy consumption and the efficiency of equipment that is powered with electricity or fossil fuels, along with costs and operating characteristics. The ESCo will provide a catalogue of energy-using equipment, along with features such as load factors and demand profiles, to identify areas where savings could be realised. Once a baseline has been established, it may be possible—using sensor technologies and automation—to make auditing an ongoing process that yields continuous improvements. Audits may be part of a broader energy management certification process, such as ISO 50001.

What are the impacts?
While the audit itself does not directly create efficiencies, the measures that can be identified as contributing to efficiency could have a significant impact on costs and energy use.

How much does it cost?
Commercial and industrial energy audits can cost on the order of $0.25 per square foot ($2.70 per square meter).^iv

How complex is it?
Audits are easy to carry out since the ESCo, which should be certified by a body such as the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), takes care of the process. The ASHRAE, as an example, provides standards for three types of audits:^v

Level 1: walk-through survey.
Level 2: energy survey and analysis.
Level 3: detailed analysis of capital-intensive modifications.

How quickly do you get results?
According to published sources,v even a level 1 commercial energy audit can help identify zero-cost efficiency measures that can immediately reduce energy use and costs by between 5% and 10%. More thorough audits can typically reveal up to 20% savings in buildings that have not been subjected to efficiency measures for a decade or more, with up to 40% reductions in cost and consumption being seen in some cases.

What are the critical success factors?
Getting an accurate audit result depends on having access to as much information as possible, which may involve deploying sensors and tracking energy consumption over a period of months.

What do the experts say?
“Before you can make any decisions, you need to get the facts straight,” says Morten Wierod, president of the electrification business area at ABB. “Companies can do that with an energy audit, which in practice involves installing sensors to measure all the points of consumption. The first 50% reduction, with today’s energy prices, will pay for itself within the first year—far faster than installing solar panels on the roof, for example.”

Action #2: right-size industrial assets and processes

Detailed analysis of industrial assets often reveals that equipment tends to be bigger than needed for the job it is doing, according to Adrian Guggisberg, president of the motion services division at ABB. This is because a margin of error is usually allowed for as part of plant designs or simply because the operating conditions have changed over time. Cumulatively, the oversizing of many components can result in excessive energy use and inefficient device loading. Matching equipment capacities to loads more accurately leads to more efficient energy and asset use.

What’s involved?
Right-sizing industrial equipment for the task at hand requires detailed understanding of operational requirements, device efficiency and loading profiles. Depending on the equipment in question, it may be possible to improve loading by adjusting settings, upgrading or re-designing the asset, but if not then it may be necessary to swap out the machine for one more accurately sized for the process involved.

What are the impacts?
Swapping out motors so that they run with 95% loads will improve the efficiency of operations, ABB’s Guggisberg notes.

Re-designing and upgrading plate heat exchangers to fit with the operating conditions also has a high impact on the overall efficiency of the industrial process. A heat exchanger is designed for a specific process at the time of purchase and design parameters are normally not the same as actual operating conditions.

A few years down the line, most plants have changed their operating conditions and they will not give the same outlet temperatures as before. An upgrade of the heat exchanger will be needed, which can easily be done in gasketed plate heat exchangers by adapting the number of plates.

How much does it cost?
Since the wholesale replacement of oversized assets is unlikely to be cost-effective, right sizing can be carried out gradually as part of a plant’s ongoing asset lifecycle management. When done in this way, it may be possible to realize immediate capital expenditure savings through the procurement of smaller, less expensive assets.

How complex is it?
Most of the complexity involved in right-sizing lies in the need to get accurate information on load profiles. This can be obtained from an analysis of operating modes and device specification, potentially facilitated through sensor data.

How quickly do you get results?
Reducing the power requirements of industrial assets yields immediate results in terms of energy and emissions reduction. If right-sizing is introduced as part of the standard replacement cycle, the speed and scale of results will depend on the lifecycle of assets.

What are the critical success factors?
Margins of error are built into industrial processes for good reason: to prevent failures that can compromise
safety and production. In right-sizing, therefore, it will be key to address the following questions:

• How large is the oversizing of an asset?
• What are the chances of the asset being used to its full capacity?
• How meaningful will be the savings that can be achieved by right-sizing?

Another success factor is to ensure that process design and procurement teams are aligned on efficiency targets. Identifying an oversized asset is of little practical use if guidance to reduce the size is ignored when the asset is later replaced.

What do the experts say?
“As an example, most of the electrical motors in industry are oversized, because when the thing was designed it went through different hands—and everybody puts a margin on top,” says ABB’s Guggisberg. “Running an electric motor at 65% load is just not good efficiency.”

Action #3: bring connectivity to physical assets

Many industrial leaders do not have a clear view of where energy is used in their operations. By connecting physical assets using the industrial Internet of Things (IoT), companies can better understand how assets are used, enabling smarter, leaner operations. Recent research from ABB, however, reveals that just 35% of industrial organisations globally have implemented IoT technologies at scale.^vii

What’s involved?
Industrial IoT technologies can track energy flows through a plant and show areas where energy is being needlessly used. This could be down to energy use in ancillary systems, oversized assets (see above), faulty equipment, heat losses or where no electricity should be needed, such as illuminating an unoccupied room.

What are the impacts?
There are losses in all industrial processes, with up to 95% of primary energy being lost on the way to carrying out the work for which it is needed.^viii The point of connecting devices is to uncover previously unseen sources of waste. Although it is obviously impossible to know how significant these will be, a better understanding of how assets and workflows consume energy will almost certainly yield areas for improvement.

How much does it cost?
It is possible to connect physical assets even with limited sensor deployments. If the sensors are installed as part of wider moves to digitisation, then the efficiency gains and energy cost reductions they provide can contribute to the digital program’s overall business case.

How complex is it?
IoT technologies are increasingly mature and simple to implement, although some integration work may be needed to obtain meaningful results from the data. Once the data is obtained, the complexity of what can be done with it is almost limitless. For example, companies are increasingly relying on detailed ‘digital twins’ of real-life operations to study the impact of process changes without affecting actual production. These digital twins can be used for a wide range of simulations, including efficiency studies. Using digital twins for modelling, testing and commissioning in a virtual environment instead of a physical setting —in effect, moving bits rather than atoms—also uses far less energy.

How quickly do you get results?
If sensor technologies reveal the presence of ghost assets—devices that are drawing power without doing any useful work—then these can be switched off or decommissioned straight away, delivering immediate cost and emissions benefits to the enterprise. Elsewhere, the exercise may reveal malfunctioning or incorrectly configured assets that require maintenance, adjusting or replacing. In these cases, the time taken to see results will depend on the remedial work needed.

What are the critical success factors?
Integrating data sources into visualisation and analysis software is key to ensuring efficiency gains can be identified with ease. And domain expertise is needed to tune the algorithms and analytics that go into making better decisions about electricity use. Without this, there is a risk that results could be disappointing.

What do the experts say?
“You can do a lot with sensor data that is already available,” says Paul Röhrs, senior global digital advisor at Microsoft. “You just bring it all into the same place and let the data speak to other silos of data. The critical part is to get the data from the machine.”

Action #4: install high-efficiency motors

In industry, powertrains are used in countless applications to convert electrical energy into motion. The main elements of an industrial electric powertrain are the motor, variable speed drive and the application itself, such as a pump, fan or compressor.

The potential for powertrain efficiency is vast, says Professor Johann Kolar, head of the Power Electronic Systems Laboratory team at ETH Zürich, the Swiss federal institute of technology. It is estimated that an astonishing 46% of the world’s electricity is used to produce mechanical energy through electric motor-driven systems. In industry, the consumption rises to two-thirds of total electricity.^ix

The International Electrotechnical Commission establishes a range of international efficiency (IE) standards for motors, ranging from IE1 (‘standard’) to IE4 (‘super-premium‘).^x There are moves to introduce an even more advanced standard, IE5. More efficient motors tend to be more expensive but can yield important efficiency gains. Given the pervasiveness of motors in industry, a widespread transition to more efficient machines can yield major energy and emissions reductions.

What’s involved?
Installing high-efficiency motors simply involves replacing older machines with ones that have a higher efficiency rating. Roughly 75% of the industrial motors in operation are used to run pumps, fans and compressors, a category of machinery that is highly susceptible to major efficiency improvements.^xi

What are the impacts?
It has been estimated that if the more than 300 million industrial electric motor-driven systems currently in operation were replaced with optimised, high-efficiency equipment, global electricity consumption could be reduced by up to 10%.^xii

How much does it cost?
Upgrading to more efficient models will require capital investment as there can be up to a 40% price differential.^xiii However, expenditure on motors is often an attractive proposition because of their ease of installation. They can generally be installed without any modification of industrial systems.^xiv

How complex is it?
Wholesale motor replacements might not be worthwhile in all cases but most of the electric power consumed by motors is used by mid-sized machines.^xv

How quickly do you get results?
Energy-efficient motors deliver immediate results in terms of energy and emissions reduction—and can pay for themselves in less than a year.^xvi

What are the critical success factors?
To maximise the efficiency gains from newer electric motors, it obviously helps to transition to the most efficient models available. That has cost implications which should be considered in the context of shorter payback times.

What do the experts say?
“Electric motors have been in use for 150 years and they are the steady workhorses behind our everyday lives,” says ABB’s Tarak Mehta, president of the motion business area at ABB. “Yet over the past decade, they have undergone a period of exceptionally rapid technological advancement. Some of the very latest motors specify energy losses about 15% lower than those delivered by earlier models.”

Action #5: use variable speed drives

Today, most industrial electric motors operate at a steady speed and the motion they impart is regulated through valves (for fluids), dampers (for air) and brakes (for material). However, this way of controlling the motion, says ABB’s Guggisberg, is akin to controlling the speed of your car with the brake while pushing the accelerator to the floor: energy is just wasted.

Variable speed drives are technologies used to control the speed of motors and the amount of torque produced—a crucial part of managing the energy consumed by motor-driven systems. Energy consumption is intelligently calibrated to match the amount of work that needs to be done. “Variable speed is like using the accelerator to control the speed of your car,” says Guggisberg.

What’s involved?
Introducing variable speed drives to electric motor-driven systems is easy and straight forward. A technology provider or an ESCo can support in identifying which motors currently in use could and should be equipped with a drive to improve energy efficiency.

What are the impacts?
Installing variable speed drives can improve the energy efficiency of a motor-driven system by up to 30%, yielding immediate cost and emissions benefits.

How much does it cost?
The payback time of a variable speed drive in terms of energy savings is short (typically 1-2 years) in relation to its expected lifetime. High energy prices obviously shorten it further.

How complex is it?
As with the transition to more efficient motors, the introduction of variable speed drives does not require any changes to industrial processes.

How quickly do you get results?
The financial benefits accrue from the moment the variable speed drive enters operation and continue throughout its lifetime.

What are the critical success factors?
In common with many other efficiency improvements, leadership will need to decide whether the benefits of introducing variable speed drives warrant the immediate investment. This in turn will depend on the number, size and usage profile of installed motors – and the price of electricity.

It is worth noting that developments in motor and drive efficiency are increasingly being motivated by regulation, so investments in more efficient machines can help address compliance needs as well.

What do the experts say?
“Not every motor will benefit from a variable speed drive,” says Mehta at ABB, but even assuming around 50% of today’s motors would be upgraded, “we’re talking about a major global improvement in energy efficiency.”

Action #6: electrify industrial fleets

Growing momentum behind vehicle electrification is bringing down the cost of batteries and electric drivetrains. This, coupled with sustained high oil prices, is making electric powertrains an increasingly attractive proposition for industrial vehicles such as forklifts, mining vehicles, trucks and delivery vans.

What’s involved?
Transitioning to electric fleets will happen alongside development of charging infrastructure and low-cost, low-carbon electricity.

What are the impacts?
The efficiency gains apparent in industrial powertrains are also significant in mobility, where there is a shift from internal combustion engines (ICEs) to electric propulsion. Electric motors can achieve more than 95% efficiency, while diesel engines only reach 45% efficiency in the optimum load range.^xviiLevels of electrification are negligible in air and maritime transport at this time, yet in road transport and in industrial mobility, fleet electrification is gaining rapid traction.

Replacing the diesel engine in a 24-tonne excavator with an electric powertrain, which combines battery power with a high-efficiency electric motor and drive, can eliminate 48 tonnes of CO2 emissions per year, according to ABB.^xviii In addition, regenerative braking for industrial vehicles can reduce fuel consumption by up to 30%.

How much does it cost?
Electric vehicles in the transportation of goods today are more expensive than traditional models, making the business case for switching on financial grounds alone more complex.^xix Beyond capital expenditure considerations, however, on average electric vehicles have operating costs as much as 60% lower than equivalent vehicles powered by a diesel engine, mainly due to improved efficiency, reduced fuel consumption and reduced maintenance needs.^xx

How quickly do you get results?
Vehicle electrification yields immediate efficiency gains although, as noted, the capital cost of moving to an electric fleet means financial benefits may be modest in the short term. Linked to vehicle electrification, industrial fleet owners can realise further efficiency gains through digital management of fleets, including optimising charge schedules.

What do the experts say?
“Fleet optimisation is the first thing to do to reduce emissions, especially as green technologies are not yet available in all regions and countries,” says Florence Noblot, head of environmental, social and corporate governance at DHL Supply Chain, a division of Deutsche Post DHL, the global logistics company.

Action #7: use efficient, well-maintained heat exchangers

Heat transfer is crucial when it comes to making an industrial process energy efficient and heat exchangers are used for heating and cooling in almost any industry worldwide. Heat exchangers, as static equipment, are often not subject to proactive maintenance and optimisation, instead experiencing run-to-failure operation without realising the environmental and cost impact of lost heat transfer.

Keeping a heat exchanger at its optimal performance level over time is crucial to ensure energy-efficient processes. Up to 2.5% of the world’s CO₂ emissions come from unmaintained heat exchangers. This can be prevented simply by cleaning heat exchangers on a regular basis.

Selecting the right heat exchanger technology is another important step in optimising energy efficiency for a given application. An innovative and compact plate heat exchanger, for example, can be 25% more efficient than a shell-and-tube heat exchanger.

Furthermore, 20% to 50% of industrial energy input is lost as waste heat, for example in the form of hot exhaust gases or cooling waters. Recovering and reusing that heat in further processes is an important step in improving overall efficiency and decreasing carbon emissions. Solutions could be to reintegrate the heat in the process itself, or to provide the heat for use elsewhere, for example in district heating, electricity generation and so on.

What’s involved?
A review of thermal losses in heat exchangers can be carried out by an ESCo or a specialist service provider. This can help form the basis of a thermal efficiency strategy with maintenance on site or costed technology upgrades.

What are the impacts?
“By first of all securing proper maintenance of heat exchangers, and then also make sure to select the right equipment for new installation, or by upgrading poorly performing heat exchangers, this has a large impact on the energy consumption,” says Julien Gennetier, vice president of Alfa Laval’s energy division.

How much does it cost?
Plate heat exchangers may have lower capital costs than shell-and-tube models because they are a sixteenth of the weight and use a tenth of the floor space, offering savings on shipping, handling and installation.^xxi Plate heat exchangers also offer a reduced operational cost due to higher thermal efficiency.

How complex is it?
Cleaning or upgrading heat exchangers is an easy process that can be carried out as part of planned maintenance. Converting existing technology to a more efficient solution needs some re-piping but for many processes offers direct effects on operational cost.

How quickly do you get results?
Results start to accrue directly upon installation and with regular maintenance.

What are the critical success factors?
The significant efficiency gains available through heat exchanger equipment upgrades makes them highly desirable where possible, although expert advice is needed on the sizing and process integration.

What do the experts say?
“2.5% of the world’s carbon dioxide emissions come from inefficient heat transfer in heat exchangers, due to them not being cleaned and maintained properly,” says Kajsa Dahlberg, cleantech business developer at the heat exchanger manufacturer Alfa Laval. “By simple measures, the energy consumption can be immediately reduced.”

Action #8: switch gas boilers to heat pumps

Heat pumps are seen as key for global decarbonisation as a replacement for fuel-fired boilers. The IEA forecasts the technology will allow more than 50% of homes to use electricity for heating by 2050.^xxii In industry, the technology can have similar benefits for space heating and can also be used for process heat of up to 356 F (180 C).^xxiii Industrial heat pumps make it possible to reuse excess heat from a process for other purposes, such as industrial process heating or space heating, avoiding the need for fuel-fired boilers.

What’s involved?
Heat pumps take advantage of thermal gradients to improve the efficiency of electricity-to-heat generation processes, so should be considered wherever there is a need for low-to moderate process heat or space heating.

What are the impacts?
Heat pumps are by far the most efficient way of obtaining low-to-moderate heat from electricity.

How much does it cost?
Industrial heat pumps can cost up to $90,000, according to published sources.^xxiv

How complex is it?
Upgrading thermal equipment is not a trivial affair and in the case of heat pumps there may be constraints on the suitability of the environment for installation. However, the clear financial and ecological benefits of reducing energy demand for heating can provide a sound basis for upgrade programs.

How quickly do you get results?
The financial benefits start to accrue from installation. Heat pumps last up to 25 years, with a payback time of five years or less.^xxv

What are the critical success factors?
As in other areas, it is important to choose the right heat pump technology for a given application and consider whether it would be advisable to add thermal heat storage to the setup.^xxvi

What do the experts say?
With heat pumps, “You use one unit of electricity and you’re getting three units of heat coming out the other end, which is an astonishing piece of magic,” says Lane of the IEA.

Action #9: deploy smart building management systems

The built environment accounts for some 40% of total energy use and 30% of global greenhouse gas emissions, according to the United Nations Environment Programme.^xxvii For industry, this is perhaps unsurprising in that buildings and associated infrastructure are seldom designed from the bottom up with energy efficiency in mind.

On the contrary, factories, warehouses and other industrial structures, along with auxiliary assets such as lighting and heating, ventilation, and air conditioning (HVAC) systems, are usually specified to minimise capital outlays, often at the expense of efficiency. This means there is plenty of scope for cost savings and efficiency gains with relatively simple, quick-payback interventions. Properly insulating buildings is perhaps one of the fastest and most cost-effective ways to save on energy.

Industrial facilities can save energy and costs by installing digital systems to control HVAC systems, lighting, blinds and so on. These systems typically sense when people are no longer present in the environment and respond accordingly, dimming or switching off lights and closing windows and blinds to minimise wasted energy.

HVAC systems are responsible for almost 50% of commercial building energy consumption, 35% of which is usually wasted, according to ABB partner BrainBox AI.^xxviii

What’s involved?
The purpose of a computer-based building management (or automation) system (BMS) is to monitor and regulate a building’s electrical and mechanical assets, such as power systems, lighting and ventilation.^xxix

Artificial intelligence (AI) can be used to analyse building use patterns and adjust temperatures, requiring almost nothing in the way of intervention. Similar impacts can be achieved using smart meters and intelligent thermostats to adjust building conditions to the real-time needs of workers, rather than relying on wasteful always-on heating, cooling and ventilation.

What are the impacts?
A BMS might typically control around 40% of a commercial building’s energy loads, rising to 70% if the system covers lighting as well.^xxx In an industrial setting, the impact of a BMS will depend on the extent to which building and industrial process loads are managed separately.

Combining artificial intelligence with industrial IoT can cut HVAC emissions by as much as 40% and reduce energy costs 25% while increasing occupancy comfort 60% and extending equipment life 50%.^xxxi

How much does it cost?
Published costs for BMSs in the United States range from $2.50 to $7.50 per square foot ($26.91 to $80.73 per square meter), depending on factors such as the usage and age of the building and whether the installation is new or is an upgrade.^xxxiiMost industrial facilities built after 2000 will have included a BMS as standard, although in older buildings this may need updating.^xxxiii

HVAC optimisation systems can effectively pay for themselves. In New York, for example, the application of these systems is expected to largely help avoid the need for retrofitting energy-inefficient buildings at a cost of upwards of $20 billion.^xxxivSome solution providers also offer flexible consumption models and “zero capex” arrangements, marketing smart building systems on a subscription-based “as-a-service” basis wherein end users pay only from a portion of the cost savings realised.

How complex is it?
The complexity of a BMS installation will depend on the number of building subsystems it covers, with advanced deployments potentially extending to applications such as fire safety and access control as well as HVAC and lighting. At the level of the user, the point of a BMS is to make life easier, for example by switching off lights and HVAC systems when nobody is around.

How quickly do you get results?
Although BMSs are not cheap, automation can deliver immediate results upon commissioning, with research showing the payback times for building energy management systems in commercial buildings dropped from 5.4 to 0.7 years between 1977 and 2017.^xxxv

AI takes a while to learn usage patterns in a building and then apply that knowledge to temperature optimisation, so efficiency gains may not fully kick in for a few months.

What are the critical success factors?
A common problem with BMSs is that the full functionality of the system is not utilised, leading to sub-optimal performance and a reduced return on investment. To overcome this problem, researchers say it is important to consider commissioning, user involvement in the BMS specification and perceptions of vendor performance.^xxxvi

What do the experts say?
“Sizing an HVAC system in an industrial environment is not that simple because of the heat loads you have,” says Guggisberg of ABB. “Often you see people have made mistakes and they have the cooling system running full blast and the doors are open because they can’t get enough air into the room.”

The level of impact that can be achieved with such technologies will vary in each building but could be significant. “No system is fully optimised at the moment,” says Kolar of ETH Zürich.

Action #10: move data to the cloud

Demand for digital services is growing rapidly. Many of the energy efficiency opportunities highlighted in this report rely on massive data storage and computing power to derive insights from operational information. But storing and using data takes energy. Global data centre electricity use in 2021 was 220-320 TWh, or around 0.9-1.3% of global final electricity demand. This excludes energy used for cryptocurrency mining, which was a further 100-140 TWh in 2021.^xxxvii

The global demand for data processing has been growing fast, and it isn’t likely to decrease any time soon—quite the contrary, in fact. For this reason, the technology industry has had a strong focus on energy efficiency. It has achieved impressive gains through concepts such as server virtualisation and cloud computing.

Further efficiencies have come through energy-efficient HVAC systems, motors, variable speed drives and the utilisation of waste heat from data centres—all opportunities highlighted in this report. Industrial organisations looking for greater energy efficiency can tap into many of these gains associated with more intelligent use of data and cloud-based management.

What’s involved?
Moving data to the cloud is a key aspect of harnessing information about industrial assets and processes and applying analytics to optimise how systems are operated and how much electricity they consume. Research suggests, moreover, that cloud-based data centres are themselves upwards of 90% more energy efficient than on-premise computing approaches.^xxxviii

The concept of “multitenancy” allows a single compute resource to serve multiple clients simultaneously, dramatically increasing utilisation rates. Local, privately-owned servers often lack the intelligence and hypervisors characteristic of advanced cloud systems and consequently are frequently kept in an “always-on”state that wastes electricity, even when an application is only occasionally used.

Virtualisation means producing (and buying) fewer servers to handle the same processing burden. Further, many data centre operators, especially hyperscale cloud players and colocation service providers, are among the most advanced users of renewable energy sources in their power mix.

What are the impacts?
Moving from on-premises, privately managed data processing to a cloud-based data centre presents a range of energy efficiency opportunities. The economies of scale associated with data centre consolidation can be significant in terms of cost. Cloud providers also recognise data centre cooling as one of—if not the single biggest—operating expense in their business and are aggressive about seeking energy efficiency improvements.

How much does it cost?
The ability to consume storage and compute functionality on an as-a-service basis means enterprises can shift capital spending to a variable operating expense and enjoy the flexibility and financial benefits that go along with that. While initial costs may increase (for example, when an organisation takes a “hybrid” approach, maintaining both a private data centre and working with a third-party provider), cloud-based arrangements are consistently shown to offer lower total cost of ownership over time.^xxxix

How complex is it?
Transitioning compute workloads and data management to the cloud is a relatively straightforward proposition for most enterprises. Many compute workloads can be shifted with a simple transaction subscribing to a cloud service, while others require a more expensive and involved set-up, usually with a technology partner.

How quickly do you get results?
Most enterprises can expect to see immediate energy efficiency benefits. The extent of gains will depend on factors such as the number of servers used, the overall volume of data and the compute intensity of applications.

What are the critical success factors?
Power usage effectiveness (PUE)—the amount of energy that ends up being used for computation—is perhaps the most important metric in data centre design, highlighting the importance of efficiency for the industry. Ways PUE can be improved include switching off idle technology equipment, consolidating and virtualising servers and storage, running power distribution at higher voltages, using energy-efficient chipsets and management features and installing high-efficiency uninterruptible power supplies (UPS).^xl

Improvements in data centre design can drastically improve efficiency, with industry figures showing that computing power rose 550% between 2010 and 2018 with no more than a 6% energy gain.^xli But even companies that have the resources to implement smart data centre design features such as intelligent cooling might struggle to match the efficiency of server and network virtualisation that comes with hyperscale cloud services.

What do the experts say?
“Our public cloud platform, Microsoft Azure, can help you save up to 93% of your energy consumption and is up to 98% more carbon efficient than on-premises solutions,” says Christoph Pawlowski, industry advocate for sustainability at Microsoft, based on a study.^xlii

Outlook and conclusions

While the overall impact of efficiency gains will vary greatly from one industrial sector to another, the opportunity for cost and emissions reduction is significant and remains largely overlooked. Many governments also offer incentive schemes for industrial energy efficiency that can help accelerate adoption of relevant innovations.

According to the IEA, the industrial world has work to do; energy efficiency is not on track to meet the goal of net-zero emissions by 2050^xliiias defined in the 2015 Paris agreement. “Now is the time to invest in energy efficiency” says the IEA’s Lane.

As industry considers how best to tackle the twin challenges of decarbonisation and energy affordability, it should be clear that efficiency deserves a much more prominent place on the industrial corporate agenda. An important element in realising gains from energy efficiency is empowering the industrial workforce to utilise relevant innovations.

Training and incentivising employees to make energy efficiency a priority and to use available technologies therefore must be part of any systematic approach to reducing energy consumption. More generally, as organisations consider their options to address climate change and rising energy costs, they are confronted with five possible paths:

• Reduce energy consumption by producing less. In most developed economies, however, this would likely lead to reduced economic activity and lower standards of living, an outcome that is likely to be socially and politically challenging.

• Switch to renewable energy sources. This is already happening, albeit at a rate that is unlikely to achieve global climate goals in time. Within hard-to-abate industries, the energy transition is likely to take decades.

• Increase circular business models. This could reduce the emissions from raw material acquisition and preserve the world’s resources, but will not address those that come from energy use—and the timeframe for introducing circularity is also likely to be measured in decades.

• Create carbon sinks to offset industrial emissions. This can be achieved at low cost, for example by planting trees, but with uncertain results and longer time horizons. Quicker, more effective measures involve immature technologies and are costly.

• Improve energy efficiency. This allows industry to operate much like it has until now, maintaining productivity and profits but with lower costs and markedly reduced emissions. As this report has demonstrated, many efficiency measures can be implemented rapidly, with immediate results.

Mitigating climate change will undoubtedly require industry to pursue all these strategies to varying degrees, but energy efficiency stands out as the one that can produce the greatest results in the shortest time and with the fewest downsides. The technologies required for greater efficiency are available today and current high energy pricing conditions in some locations make the business case for their application more pressing than ever.

“We at times think we have done enough on efficiency,” concludes ABB’s Morten Wierod. “But technology has developed a lot over the last 10 years. It’s opened a new door for energy efficiency. The technology we need is already available – and the time to implement it is now.”

Acknowledgements: This report draws on insights from the following industry experts:

Adrian Guggisberg, President, motion services division, ABB
Tarak Mehta, President, motion business area, ABB
Morten Wierod, President, electrification business area, ABB
Julien Gennetier, Vice president, energy division, Alfa Laval
Kajsa Dahlberg, Cleantech business development, Alfa Laval
Florence Noblot, Head of environmental, social and corporate governance, DHL Supply Chain
Johann Kolar, Professor, head of Power Electronic, Systems Laboratory, ETH Zürich
Kevin Lane, Senior programme manager, energy efficiency, IEA
Christoph Pawlowski, Industry advocate for sustainability, Microsoft
Paul Röhrs, Senior global digital advisor, Microsoft

References

ⁱIEA Key World Energy Statistics 2021: Final consumption. Available at https://www.iea.org/reports/key-world-energy-statistics-2021/final-consumption.
ⁱⁱIbid.
ⁱⁱⁱIEA, World Energy Outlook 2022.
^ivThe R Group website: ENERGY AUDITING. Available at https://thergroupllc.com/services/energy-consulting/energy-auditing-service/.
^vASHRAE: Technical FAQ. Available at https://www.ashrae.org/File%20Library/Technical%20Resources/
Technical%20FAQs/TC-07.06-FAQ-95.pdf.
^viThe R Group. https://new.abb.com/news/detail/87544/new-abb-study-on-industrial-transformation-unveils-critical-relationship-between-digitalization-and-sustainability
^viiBob Shively, Enerdynamics, 2017: How Much Primary Energy Is Wasted Before Consumers See Value from Electricity? Available at https://www.enerdynamics.com/Energy-Curents_Blog/How-Much-Primary-
Energy-Is-Wasted-Before-Consumers-See-Value-from-Electricity.aspx.
^ixFernando Ferreira and Aníbal de Almeida, IEEE Industry Applications Magazine, January/February 2018: Reducing Energy Costs in Electric-Motor-Driven Systems.
^xDanielle Collins, Motion Control Tips, March 9, 2020: https://www.motioncontroltips.com/what-are-international-efficiency-standards-for-motors-and-gearmotors/.
^xiOmdia, 2020: Motor-driven Equipment Research Package.
^xiiP. Waide and C.U. Brunner, International Energy Agency working paper, Paris, 2011: Energy-Efficiency Policy Opportunities for Electric Motor-Driven Systems.
^xiiiAustralian Government Department of Climate Change, Energy, the Environment and Water, 2022: Motors and variable speed
drives. Available at https://www.energy.gov.au/business/equipment-and-technology-guides/motors-and-variable-speed-drives.
^xivABB, January 2021.
^xvABB, January 2021.
^xviHoney Electric website: Are Energy Efficient Motors A Good Investment? Available at http://hoveyelectric.com/hovey-electric-power-blog/bid/64122/How-to-Determine-if-Your-Motor-is-Energy-Efficient.
^xviiIbid.
^xviiiABB, June 2022.
^xixJason Deign, Foresight Climate & Energy, June 28, 2022: A new direction for transport electrification. Available at https://foresightdk.com/transports-new-direction/.
^xxABB, June 2022.
^xxiAlfa Laval, 2022: 5 Reasons to use plate-and-frame heat exchangers instead of shell-and-tube. Available at https://www.alfalaval.com/microsites/gphe/tools/gphe-vs-shell-and-tube/.
^xxiiIEA World Energy Outlook.
^xxiiiReuters Events, June 2022: The Next Frontier: Decarbonising Industrial Heat.
^xxivMade in China: Industrial Heat Pump Price. Available at https://www.made-in-china.com/price/industrial-heat-pump-price.html.
^xxvTermo-plus, May 4, 2019: What is the life expectancy of heat pumps? Available at https://termo-plus.com/blog/life-expectancy-of-heat-pumps/.
^xxviAraner, 2021: Heat pumps key success factors. Available at https://www.araner.com/blog/heat-pumps-key-success-factors.
^xxviihttps://www.unep.org/explore-topics/resource-efficiency/what-we-do/cities/sustainable-buildings
^xxviiiBrainBox AI website, 2022: Making buildings smarter, greener, and more efficient. Available at https://brainboxai.com/en.
^xxixMd. Faruque Hossain, Chapter Seven – Best Management Practices, Sustainable Design and Build, Butterworth-Heinemann, 2019, Pages 419-431, ISBN 9780128167229. Available at https://doi.org/10.1016/B978-0-12-816722-9.00007-0.
^xxxIbid.
^xxxiBrainBox AI.
^xxxiiMid-Atlantic Controls, July 13, 2021: How Much Does a Building Automation System Cost? Available at https://info.midatlanticcontrols.com/blog/how-much-does-a-building-automation-system-cost.
^xxxiiiHossain, 2019.
^xxxivBloomberg, July 13, 2022: BrainBox AI Named a Qualified Vendor by NYSERDA for its Real-Time Energy Management + Tenants Program. Available at https://www.bloomberg.com/press-releases/2022-07-13/brainbox-ai-named-a-qualified-vendor-by-nyserda-for-its-real-time-energy-management-tenants-program.
^xxxvChin-Chi Cheng and Dasheng Lee, International Journal of Energy Research 42(1), July 2018: Return on investment of building energy management system: A review. Available at https://www.researchgate.net/publication/326686419_Return_on_investment_of_building_energy_management_system_A_review.
^xxxviGordon Lowry, Building Services Engineering Research and Technology 23(1):57-66, April 2002: Factors affecting the success of building management system installations. Available at https://www.researchgate.net/publication/239405511_Factors_affecting_the_success_of_building_management_system_installations.
^xxxviihttps://www.iea.org/reports/data-centres-and-data-transmission-networks
^xxxviiihttps://www.spiceworks.com/tech/cloud/guest-article/how-the-cloud-drives-sustainability/
^xxxixhttps://blogs.gartner.com/marco-meinardi/2018/11/30/public-cloud-cheaper-than-running-your-data-center/
^xlChris Loeffler, Buildings, May 1, 2008: 10 Ways to Save Energy in Your Data Center. Available at https://www.buildings.com/feature/article/10192816/10-ways-to-save-energy-in-your-data-center.
^xliSebastian Moss, Data Center Dynamics, February 27, 2020: Huge data center efficiency gains stave off energy surge – for now. Available at https://www.datacenterdynamics.com/en/analysis/huge-data-center-efficiency-gains-stave-energy-surge-now/.
^xliiMicrosoft, 2018: The Carbon Benefits of Cloud Computing: a Study of the Microsoft Cloud. Available at https://www.microsoft.com/en-us/download/details.aspx?id=56950.
^xliiiIEA tracking report, September 2022: Energy Efficiency. Available at https://www.iea.org/reports/energy-efficiency.

ABOUT THE ENERGY EFFICIENCY MOVEMENT

The Energy Efficiency Movement is a forum that brings together like-minded stakeholders to innovate and act for a more energy-efficient world. Through innovation, the sharing of knowledge and insights, investments and the right regulations and incentives, we can optimise energy efficiency and accelerate progress toward a decarbonised future for all.

The Movement was launched by ABB in 2021 and it has received a positive reaction from throughout industry, with around 200 companies joining by November 2022. Among them are Microsoft, Alfa Laval and DHL Group, leaders in their industries and contributors to this report.

join.energyefficiencymovement.com

Source URL: https://search.abb.com/library/download.aspx?documentid=9akk108467a5406&utm_source=abbcom&utm_medium=website&utm_campaign=eeplaybook

Analysis of Measured Data at the Point of Complaints to Power Quality

Published by Petr ROZEHNAL, Jan UNGER, Petr KREJCI, VŠB-Technical University of Ostrava

Abstract. Electricity is one of the most important energy that uses both in the home and in industry. For good and seamless use of electricity are important its quality. Deteriorating power quality is increasingly becoming an important problem for the industry and service companies. The poor quality of electricity to distribution companies must also contend with complaints to power quality. A significant increase in the production of energy from renewable sources such as wind energy, leads to the need to explore new ways of energy systems and their potential effects on the quality of electrical energy. It must be known reliability and responsiveness to changes in the network. Power quality and respect for the parameters given standard ČSN EN 50160th Renewable electricity has to design, connect and operate the power system in place, which will have negative feedback effects on the distribution network and to be reliable.

Streszczenie. Pogarszająca się jakość energii staje się coraz bardziej istotnym problemem dla przemysłu i usług. Spółki dystrybucyjne muszą również zmagać się z roszczeniem do jakości zasilania. Znaczący wzrost produkcji energii ze źródeł odnawialnych, takich jak energia wiatru, prowadzi do konieczności poszukiwania nowych sposobów systemów energetycznych i ich potencjalnych wpływów na jakość energii elektrycznej. (Analiza danych pomiarowych z punktu widzenia skarg dotyczących jakości energii)

Keywords: complaints, supply, wind power, power quality, distribution system.
Słowa kluczowe: skarga, zasilanie, wiatr, jakość energii, system dystrybucji.

Introduction

Electricity generation using winds no modern affair, but just is the opposite. This method of obtaining and transformation of energy goes deep into the history of mankind. However, due to the progress and the gradual electrification of the wind turbines have become a vehicle for the conversion of wind energy into electrical energy. Further development of a major surge in electricity consumption occurred in making the rules, regulations and standards governing, inter alia, the qualitative characteristics of electricity. Nowadays, power producers complies with standard EN 50160, according to which I will analyze the measured data from the turbine Vestas V90 – 2.0 MW, depending on the operation of wind power plants. On 16 3rd 2013 reported distribution companies’ complaint to power quality due to alleged voltage unbalance in the location of the plant. This analysis should serve to clarify the possibility of such a complaint from the impact of wind farms and also to illuminate the issues of the impact of wind power to the grid and the selected parameters electricity. The analyzed wind turbine consists of four-pole asynchronous generator rotor windings brought out the rings.

The causes of complaints on the quality of electricity

The cause of complaints electricity by electricity consumers may be several. Among the reasons that may lead to complaints for power quality, harmonics include the creation of a network, fluctuations in voltage, voltage unbalance and power interruptions to customers.

All these causes that lead to the emergence of complaints about the quality of electricity, can lead to large losses for both the electric power customer and as a supplier of electricity.

All complaints arising from the power quality must be verified energy supplier, who will determine whether there is a legitimate reason for complaint on power quality. The legitimacy of the claim to power quality electricity supplier informs the customer of electricity and in the event that a claim is justified and must take corrective action. All power quality parameters, which are compared in reclaiming the power quality are given in EN 50160^th. [4]

Number of complaints on the quality of electricity in the first half of 2013

Number of complaints on the quality of electricity in each year varies. In the first half of 2013 were a total of 233 complaints about power quality. All of these complaints were dealt electricity distributor.

Table 1. Number of complaints in Northern Moravia in the first half of 2013

Month	January	February	March	April	May	June
Number	76	43	39	34	22	19

Table 1 shows the number of complaints about the quality of electricity in the first half of 2013. The table shows the decreasing number of complaints. For the following measurement and evaluation, were used reported complaints on the quality of electric power, which was recorded in March 2013.

Fig.1.Number of complaints in the first half of 2013

Figure 1 shows the decreasing number of complaints about the quality of electricity in the first half of 2013. The largest number of complaints was in the month of January, which saw a total of 76 complaints. In the month of March there were only 39 complaints about the quality of electric power, which is half the number. In the month of June there were only 19 complaints about the quality of electrical energy.

One of the recorded complaints about the quality of electric power was complaints that occurred at the location of wind turbines Vestas v90 – 2 MW. After reporting this complaints were placed portable analyzers BK-Elcom in the substation of wind turbines. The measured power quality parameters were recorded and were used to assess whether a given wind turbine could affect, or could have been the cause of this complaints on power quality.

Specifications turbines Vestas v90 – 2 MW

Windward power tilt mechanism, active routing wind and three-leaf rotor diameter of 90 m OptiSpeed^TM use the technology, which allows the device to operate at optimum speed and optimize its performance. All the plants of this type are also fitted OptiTip ®, which is a system developed for optimizing the rush angle. OptiTip ® has always set the blades to an angle that is specific wind conditions ideal. This contributes to an increase in electricity production and to minimize noise emissions.

The main shaft transfers the energy transfer through the generator. The transmission is a combination of a planetary gear transmission with front helical teeth. Since the transmission of energy, is transmitted through the composite connector on the generator. Generator facility is designed as a four-pole asynchronous generator rotor windings brought out the rings. Middle-voltage step-up transformer is placed in a special room at the end of the engine room. It is a structure with the use of dry resin, which was developed specifically for use in wind turbines. Systems OptiTip ® and OptiSpeed^TM optimize performance for each of the wind speed, independent of the temperature and the air density. At high wind speeds, ensures that energy production does not exceed the rated power.

Wind turbine is equipped with a braking system that, if necessary, stops the rotation. The system adjusts the rotor blades and activated while the hydraulic parking brake. The parking brake is located on the high speed shaft. All functions of the power control and regulate the microprocessor control unit.

The control system is equipped with sensors that ensure safety and optimal operation. Drive mechanism for making blades is by three hydraulic cylinders – one for each rotor blade one. The hydraulic unit in the engine room supplies hydraulic pressure tilting mechanism and braking system. Both systems are equipped with hydraulic accumulators that provide network outages regulated and safe shutdown. Four electric rotary drive mechanisms ensure the rotation of the engine room at the top of the tower. Engine cover made of fibreglass protects all components from rain, snow, dust, sun, etc. Access to the engine room of the tower allows the central hole. The engine is installed service crane system with capacity of 800 kg. Crane can be extended to carry up to 9500 kg.

Readings wind is evaluated computer which automatically controls and monitors the operation of VE and passes both remote and local reporting of operating and fault conditions. Life wind electricity is designed and constructed for 20 years. [3]

Description of measurement

During March and April, were installed portable analyzers BK-Elcom three measuring points. Measurements were carried out simultaneously at the place where the wind turbine is connected to the distribution network, as well as in the distribution substation transformer and high voltage. This allows us to determine whether the operation of wind turbines affects the surrounding system and how. Due to interrupt measurement, weather conditions, etc., were selected period, from which all available data from all measuring points. This is the period from 20.3. to April 3, 2013. Were evaluated waveforms, long-term and short-term flicker, and harmonic distortion and compared these with allowable values in EN 50160th.

To determine the effects of rear-connected wind power were realized measurement performance balance beam distribution network, which is controlled wind turbine is connected. Simplified wiring diagram wind electricity is shown in Figure 2.

Measuring point 1 was at the outlet of the reference beam distribution network in the substation 110/22 kV. Measuring point No. 2 was placed in a distribution transformer (DTS) between wind electricity and MV and measuring point No. 3 was in the connection of wind power plants to the distribution network.

Fig.2. Simplified diagram of the measurement

Selected parameters of electric energy

As mentioned above, the operation of wind power plants can be expected to influence the parameters of the distribution system. Interaction between the distribution system and analyzed wind turbine is defined in the point of common coupling.

The DSO is a priority to ensure a stable supply of electricity if possible with constant system parameters. In terms of quality of the supplied energy is to be monitored, in particular:

1. First Voltage changes.
2. Second Flicker – voltage fluctuations.
3. Total Harmonic Distortion

Ad1) Voltage changes

Fig.3. Relationship of high voltage for operation of Wind electricity

From the previous graphs 1 and 2 it is obvious that voltage waveforms are identical to the low voltage and the high voltage level. At the same time, we can notice that in all measuring points of the voltage is stable and there are no variations in time or not delivered at the time the wind power plant is running at full capacity 2 MW.

Generally, the period of operations in the growing and voltage and the voltage follows the change in power output.

We can therefore claim that wind power has adverse retroactive effect in terms of voltage.

Fig.4. Dependence on the low voltage operation of Wind electricity

Ad2) Flicker – voltage fluctuations

Flicker is defined as the human eye perceptible variation of flux of light sources as a result of periodic dips in sub harmonic frequencies. These voltage changes are generally caused by changes in customer load or changes in generation capacity.

If we analyze the theoretical possibility of flicker that accompanies the operation of wind power, it is possible to identify two basic causes of its origin: the effect of wind gusts and wind power tube effect.

Fig.5. Dependence of short-term flicker severity level of the operation of wind electricity

Fig.6. Dependence of long-term flicker severity level of the operation of wind electricity

Effect of wind gusts in the short-term variations of wind speed from its mean value eliminates the inherent inertia of rotating parts of the wind turbine, due to stronger gusts of more or less eliminates the power turbine control. Effect of wind power tube (mast) suppresses much worse. Tube for flowing a wind barrier that slows him. As a parameter determining the flicker is not applied directly to the voltage drop caused by the flicker, but variable called issue of flicker or a flicker severity. Distinguish between short-term (short term) flicker emission P_st, measured or computed at intervals of ten minutes long (long term) emission flicker P_lt, determined the interval of two hours.

Generally, the more leaves the wind turbine, the emission is less flicker. Systems with frequency converter in most cases have lower emissions than systems with asynchronous generator connected directly.

Rules for the operation of the distribution system define the maximum allowed value long-term rate of flicker severity P_lt and so it must not exceed 0.46. [1], [2]

Rotor inertia is so large that the value P_st is practically negligible and short-term change in speed or direction of the wind does not affect the voltage fluctuations in the distribution network around wind electricity.

Rules for the operation of distribution networks dictate that long-term flicker severity shall not exceed 0.46, which was fulfilled in the entire measurement period. Only the first April increased value P_lt to 0.4 and only in two stages over a period of 2 hours, which was during the peak hours of wind electricity. The MV was a period of lowest value P_lt.

Ad3) Total Harmonic Distortion

Fig. 7. Dependence THD_Uthe operation of wind electricity

Total harmonic distortion, or if THD_Uis shown in Figure 5. It has a characteristic waveform that does not matter too much on the production of wind power, but rather the switching power supply, such as television sets. In this graph, we can see some regularity at intervals. These increases are in the afternoon until the evening, when the switch on these characteristic appliances. Nowadays, manufacturers indicate the value of the total harmonic factor not exceeding 5%. It should be sufficient to avoid adversely affecting other devices connected to the network. Size THD_U in this case does not exceed 2%

Conclusion

Complaints about the quality of electric power are a situation in which the distribution companies must contend. In the first half of 2013 was in Northern Moravia reported a total of 233 complaints about power quality. Highest Number of complaints in January, when it was suddenly Wind power with frequency inverters are currently the optimal solution to achieve a balance between the needs of distribution networks and operators of wind turbines. Distribution system to which it is connected modern wind turbine with a synchronous cascade is not burdened by excessive adverse effects as was the case with older types of power squirrel cage induction generators. The wind turbines have become machines which are, in capital costs, payback period, durability, efficiency and utilization of wind ideal solution.

Size of input power has a direct influence on the voltage at the connection point. However, according to this measurement, there is no direct link between the supplied power and voltage fluctuations in the parent substation 110/22 kV, in which the reference beam is connected. The value of the maximum permissible level of long-term flicker severity is 0.46. In our case the increased value P_lt when averaging the values 1, 2, and third faze 0.35. This deviation can be understood rather as a network background.

When analyzing the total harmonic voltage distortion THD_U No relation was found with changes in the power output of wind power factor and total harmonic distortion ranged from 0.5 – 2%.

In view of the reported complaints on the quality of electricity that the amount by measuring power quality parameters on wind power proves that the plant at the time of measurement does not affect the possibility of a claim. This was the conclusion of a distribution company that has identified this claim on power quality as unfounded complaint. It could be a possible local failure at the customer’s electricity.

Acknowledgements: This work was supported by research project GA CR 102/09/1842 and Student Grant Competition SP2013/137

REFERENCES

[1] ČEZ Distribuce [online]. 2011 – Příloha č. 4. Available from http://www.cezdistribuce.cz/edee/content/file-other/distribuce/energeticka-legislativa/ppds/20011/ppds-2011-priloha-4_def.pdf
[2] MIŠÁK, Stanislav, PROKOP Lukáš, KREJČÍ Petr, SIKORA Tadeusz: Větrné elektrárny s asynchronními generátory v sítích VN, Elektrorevue. [online]. 11.12.2008, 47
[3] Větrná energie [online]. 2010 available from http://www.vetrna-energie.cz/projekty/vetrneelektrarny_7/bantice_17
[4] Rozehnal.P, Unger J., Krejci P., Measurement of selected quality parameters of electricity in places complaints to power quality, Konference EPE2013, ISBN 978-80-248-2988-3, Kouty nad Desnou 2013,

Authors: Ing. Petr Rozehnal, VŠB-TU Ostrava, Department of Electrical Power Engineering, ul. 17. Listopadu 17, 708 33 Ostrava, E-mail: petr.rozehnal1@vsb.cz; Ing. Jan Unger, VŠB-TU Ostrava, Department of Electrical Power Engineering, ul. 17. Listopadu 17, 708 33 Ostrava, E-mail: jan.unger@vsb.cz; doc. Ing. Petr Krejčí, Ph.D., VŠB-TU Ostrava, Department of Electrical Power Engineering, ul. 17. Listopadu 17, 708 33 Ostrava, E-mail: petr.krejci@vsb.cz.

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 89 NR 11/2013

Calculation of the Transient Currents in Transformers using Field-Circuits Methods

Published by Bronisław TOMCZUK, Dariusz KOTERAS, Jan ZIMON, Andrzej WAINDOK
Opole University of Technology, Department of Industrial Electrical Engineering

Abstract. The Field-Circuit Method (FCM) and the new Modified Time Stepping Finite Element Method (MTSF) as well as classical Time Stepping Finite Element Method (TSF) have been used for the transient current calculations in transformers. The execution times for MTSF and TSF methods have been compared. The novel method (MTSF) is about two times faster than the classical (TSF) method. Two constructions of the 1-phase transformer have been studied. The first one is an amorphous modular transformer (T1) and the second one is the transformer with laminated carbon steel core (T2). The calculated inrush currents were compared with the measured ones.

Streszczenie Do obliczeń przebiegów czasowych prądów w transformatorach wykorzystano: Metodę Polowo-Obwodową (MPO), Zmodyfikowaną Metodę Kroku Czasowego (ZMKC) oraz klasyczną Metodę Kroku Czasowego. Porównano czasy obliczeń dla klasycznej Metody Kroku Czasowego (MKC) oraz dla ZMKC wykorzystującej dwuwymiarową analizę polową. ZMKC jest około dwa razy szybsza od MKC. Analizowano dwa 1-fazowe transformatory. Pierwszy to transformator amorficzny budowy modułowej T1, natomiast drugi to transformator tradycyjny z rdzeniem z blachy elektrotechnicznej T2. Otrzymano dobrą zgodność obliczeń z pomiarami. (Obliczenia przebiegów czasowych prądów w transformatorach przy użyciu metod polowo-obwodowych).

Słowa kluczowe: Metoda Kroku Czasowego, Metoda Polowo-Obwodowa, prąd załączania transformatora.
Keywords: Modified Time Stepping Method, Field-Circuit Method, inrush current of the transformer.

Introduction

The small power transformers are switch on/off in their operation. In the switching on moments a non-sinusoidal transient inrush current arises. In some cases, the magnitude of the inrush current is several times higher than the operational load current. Its value depends mostly on the voltage magnitude of the supplying source and the residual flux in the transformer core, as well as the current flux derivative called dynamic inductance L_d [5]. The benefits of the computer simulations and the current determination are well recognized. Knowledge of their values is also important for correct determination of the shelters parameters [1]. A accurate approximation of the inrush current requires detailed information regarding the transformer parameters [2]. If its physical model is approachable, the equivalent circuit parameters can be simply obtained from its measurements. However, during the transformer designing they can be obtained from computer simulations, e.g. from magnetic field calculations [3, 4].

The transients of transformers were analyzed in many works e.g. [3, 5]. Not always the residual (remanent) flux is considered for the transformer soft magnetic material core. In some cases, the hysteresis effects were also taken into account [3]. In this work we carried out the calculations using the equivalent circuit parameters, which have been obtained from numerical analysis. We included different values of the residual flux.

Using magnetic field analysis, the non-linear characteristic of the dynamic inductance, as a function of magnetizing current, has been determined. Also, the leakage inductances have been computed. In the computations the material characteristics have been included and the magnetic residual flux has been indirectly taken into account for the initial value of the magnetizing current fixing.

In this work have been compared commercial programs with program created by authors as part of the grant of Polish Ministry of Science and Higher Education. In this algorithm is used modified Time Stepping Method. This modification consists in using calculated magnetic flux values in every step to determine inductance.

Analyzed objects

To investigate the inrush currents, the new construction of 1-phase transformer with amorphous modular core (T1) and a conventional construction of 1-phase small power transformer (T2) have been chosen [5] (Fig. 1). Each column of the amorphous transformer consists of two hollow cylinders (toroids). Its yokes have rectangular shape with two rounded thinner sides (Fig. 1a) [6, 7]. Contrary to the T1, the T2 transformer core is assembled from packages of thin sheets high-grade steel. In Fig. 1, the main dimensions of the transformers and the assumed Cartesian coordination systems are shown. For the T1 transformer, the turn number of the windings is N=232, while for the conventional one (T2), the winding is wounded with N=182 turns.

Fig.1. Outline of the analyzed transformers with: a) amorphous core – T1, b) traditional core – T2.

Mathematical models Field-Circuit Method (FCM)

Generally, the inrush current of the transformer was analyzed by using Field-Circuit Method (FCM). In this model the calculations of the transients are based on the transformer equivalent circuit (Fig. 2), which is described by the following system of the differential equations [3, 8]:

In the expressions (1), the leakage inductance value is L_s=732 μH for T1 and L_s=177 μH for T2, the RMS value of the excitation voltage U=220 V, the coil resistance value – R=0.24 Ω for T1 and R=0.136 Ω for T2. The core losses resistances of R_Fe=1913 Ω (for T1) and R_Fe=2186 Ω (for T2) have been determined. The currents i and i_μ are unknown functions. The nonlinear dynamic inductance L_d(i_μ) should be determined by the finite element (FEM) calculations.

Fig.2. Equivalent circuit of the transformer under no load state.

As was mentioned in the section 1, the residual flux value Br and the flux magnetic path l_μ, influence on the current value i_μ

Fig.3. Dynamic inductance L_d vs. coil excitation current: a) T1 transformer, b) T2 construction.

The leakage inductance L_svalue of the transformer winding has been determined by using the field analysis under short-circuit state. Resistance R_Fe has been determined from measurements under no-load state, whereas the dynamic inductance L_d curve has been created using the field models. The assumed excitation current values in the models has changed from 0.2 to 100 A. In Fig.3 the dynamic inductance verso excitation current is presented.

Time Stepping FEM (TSF) and its modification –MTSF method

In the second mathematical model (called MTSF) the governing expressions for the magnetic field is represented by Maxwell’s equation with the magnetic vector potential ^→A . If the eddy currents in the iron core are neglected, the magnetic field can be expressed by the partial differential equation (PDE)

where _μ(^→B) is the nonlinear permeability of the material.

In the area of the windings, the magnetic field can be governed by the equation

where ^→J is the total current density.

The Galerkin’s approach is the most popular method for matrix of elements formulation. In this proposed model, the weighting functions are the same as the shape functions for this particular weighted residual method. According to the Galerkin’s method, the magnetic vector potential can be expressed as

where N_j are the element shape functions and the A_j are the approximations of the vector potential at the nodes of the elements. Thus, the formulation for the field, in the current currying regions, is expressed by:

For the other subregions of the transformer the functional is expressed

Taking into account the average length path l of the flux in the coil and its cross-section S, the electric circuit can be described by the:

The integrals in (8) refer the region S₁ with the positive direction of the excitation current and the region S₂ with the negative current direction.

Contrary to the commercial computational methods of the time variable FEM [9], we have solved the equation (8) with our software. For the field calculation with the discretized functionals we use the FEMM software. In this modeling method the 2D field calculations have been done. In the case of amorphous transformer, according to its geometry, it is difficult to obtain real magnetic field distribution in 2D. Therefore, depth of this object was fitted for magnetic flux value.

Our method characterizes simplicity and multitask system. Thus the computational platform don’t need to execute so much iterations like in the classic TSF algorithm. Contrary to the TSF method, the values of the dynamic inductance, which concern the field values in the step “i”, are stored in the separate matrix, which is located in the RAM memory. Thus, we don’t need to compute integrals within the eq. 8 in each step of the computational process. It is a great advantage of the MTSF, because only at several time steps the problem must be solved.

Calculation results | Amorphous transformer (T1) – inrush currents

In this paper the supplying voltage for the primary winding under the no load state of the transformer has been assumed. However, the magnetic flux density distribution has been calculated for many values of the current excitation. For example, in Fig. 4 the flux density is presented. The field analysis is devoted to the dynamic inductance determination.

Fig.4. Flux density distribution for I=1.5 A.

Fig.5. Inrush current waves for T1 transformer (φ=0): a) FCM, b) MTSF

Fig.6. Inrush current waves for T1 transformer (φ=90°): a) FCM, b) MTSF

In Figs. 5 and 6 the comparison of the calculated and measured inrush current waves for T1 transformer, under excitation phase φ=0 and φ=90° have been presented. The FCM method and the MTSF one give similar results. However, the first one is about 1.5 times faster. In the case of the φ=90°, both models give almost the same inrush current waves (Fig. 6).

For our investigations, in the analyzed amorphous core the maximal value of the residual magnetic flux density B_r reaches only 0.2 T. In this paper are presented transient inrush current waves for two values of the residual flux density: B_r=0.1 T and B_r=0.2 T (Figs. 7a and 7b). Increasing the residual flux density value causes increasing of the inrush current peak value. In the case of B_r=0.1 T the current peak value is 50% higher than for the net value of the B_r. In the second case (B_r=0.2 T), the current peak reaches i=65 A.

Fig.7. Inrush current waves for T1 transformer: a) for B_r=0.1 T, b) B_r=0.2 T

Inrush currents for the conventional transformer (T2)

We also calculated the inrush currents for the conventional construction of the transformer T2. The calculations have been executed for three different values of the core residual flux density. The assumed switching on phase φ=0 has been chosen. The comparison of the calculation results for two computing methods shows, that the MTSF method is more adequate for transient calculations, (Figs. 8 and 9). It can be observed a fine attenuation of the current waves in the case B_r=1.2 T. The MTSF calculation method gives more accurate results comparing to the simplified field-circuit method (Fig. 8).

Fig.8. Inrush current waves for T2 transformer for B_r=1.2 T: a) FCM, b) MTSF

Fig.9. Inrush currents for T2 transformer: a) B_r=0 T, b) B_r=0.8 T

Fig. 9 shows the calculated current waves for two values of the residual flux density: B_r=0 T and B_r=0.8 T. We can observe that the residual flux density strongly influenced on the inrush current. The current values in the first times period of the current wave are about one hundred times greater than those simulated without the residual magnetism.

The calculations of the currents have been executed with a TSF method, as well. The method is included in many commercial FEM applications. To compare the calculation times for all the methods, we also computed the problem using FCM model. The MTSF method was two times faster than the TSF one for the inrush current calculations (Tab. 1). The computations have been done with using the AMD64 3200+ processor and 3GB of RAM.

Table 1. Compared CPU times for analyzed T2 transformer

	FCM	TSF	MTSF
CPU time [s]	543	1782	846

Conclusions

The field-circuit method (FCM) and time stepping FEM (MTSF) for simulation of the inrush current waves in the two transformers have been studied. In the case of amorphous transformer T1 both methods give similar results. However, for T2 transformer analysis, the more accurate results arise from the MTSF method. First of all its is due to time discetization within numerical process.

The MTSF has been compared with the commercial TSF, [10]. The CPU time is about two times shorter for the MTSF, which validates the algorithm [11].

The influence of the residual flux on the maximum values of the current waves is higher in the case of conventional transformer, compared to the amorphous one (Figs. 6 and 7). It is mostly due to the air gaps length in the modular construction. Thus, the saturation of the transformer T1 core is considerable lower than the saturation of the T2 transformer core.

The calculation method presented in this work has been validated by measurements of the single-phase transformers. We observed relatively good conformity between computed and measured current waves [Figs. 5, 6, 8]. The differences between calculation and measurement results arise from the simplifications in the field analysis and measurement errors. The main difficulty is precisely determination of residual flux value which has significant influence on inrush current. For example, the residual flux is difficult for testing and contributes to the errors of our method, as well. Additionally, it is impossible to determine exactly the air gap length in the prototype tests. Moreover, we can see that the inrush current for amorphous modular transformer has maximal value several times lower compared to the conventional one.

This paper is partially supported by the Polish Ministry of Science and Higher Education under grant no. N N510 533739.

REFERENCES

[1] Jezierski E.: Transformers, WNT Warszawa, (1983) (in Polish).
[2] Grigsby L.L. (Ed.): The Electric Power Engineering Handbook, CRC Press LLC, Boca Raton, (2001).
[3] Zakrzewski K, Tomczuk B., Magnetic Field Analysis and Leakage Inductance Calculations in Current Transformers by Means of 3-D Integral Methods, IEEE Trans. on Magn., Vol. 32, (1996), no. 3, 1637-1640.
[4] Zakrzewski K.: Power effect in magnetic lamination taking into account elliptical hysteresis approach, ISEF’07, Prague, Czech Republic, 13-15 IX, 2007, no. 024 (on CD).
[5] Tomczuk B., Koteras D., Zimon J., Waindok A.: Polowa analiza siłowników elektromagnetycznych i transformatorów, Pomiary Automatyka Kontrola,(2011), no. 3, 264-268, (in Polish)
[6] Tomczuk B, Koteras D.: Eddy current losses in magnetic circuit with solid ferromagnetics. Electrical Review, (2010), no 4, 180-183.
[7] Tomczuk B., Koteras D.: Magnetic Flux Distribution in the Amorphous Modular Transformers, Journal of Magnetism and Magnetic Materials , vol. 323, (2011), no. 12, 1611-1615
[8] Koteras D.: Magnetic field analysis in modular design transformers with amorphous cores, doctoral thesis, Opole University of Technology, (2006).
[9] Meeker D.: FEMM 4.2 User’s Guide, Charlotteville, USA, (2009).
[10] Maxwell 2D v12 User’s Guide, Piscataway, USA, (2009).
[11] Tomczuk B, Koteras D., Zimon J., Waindok A.: Comparison of numerical methods for current determination under no-load transformers, Zeszyty Problemowe-Maszyny Elektryczne, Branżowy Ośrodek Badawczo-Rozwojowy Maszyn Elektrycznych Komel, Katowice, (2011), no 92, 145-150.

Authors: Prof. Bronisław Tomczuk, (Ph.D., D.Sc.) Opole University of Technology, Industrial Electrical Engineering Chair, ul. Luboszycka 7, 45-036 Opole, E-mail: b.tomczuk@po.opole.pl, Dariusz Koteras, (Ph.D., Eng.), address as above, d.koteras@po.opole.pl, Jan Zimon (Ph.D., Eng.), address as above, j.zimon@po.opole.pl,, Andrzej Waindok (Ph.D., Eng.), address as above, a.waindok@po.opole.pl,.

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY (Electrical Review), ISSN 0033-2097, R. 87 NR 11/2011

Effect of a Wind Generator on the Optimal Location of FACTS

Published by Messaoud ZOBEIDI*¹, Fatiha LAKDJA², Yamina Ahlem GHERBI³, Fatima Zohra GHERBI¹ ,
Department of electrical Engineering, laboratory of ICEPS, Djillali Liabes Universiy, Sidi-Bel-Abbes, Algeria(1), Department of electrical Engineering, laboratory of ICEPS ,Saida Universiy,Saida , Algeria (2), Department of electrical Engineering, USTO-MB Oran , Durable Development of Electric power laboratory, Algeria(3)

Abstract. The development has contributed to an increase in the consumption of electric power, which increases the generation and transport of electrical power. Consequently, electric power systems are becoming more complicated and, hence, interest is to find ways to exploit them effectively and economically. The solution to these problems through improved control of power systems already in place. The proposed elements that control and improve power system are the FACTS devices (Flexible Alternating Current Transmission System). The object of this paper is used new methods to find the optimal location of Thyristor Controlled Series Capacitor and Static Var compensator with and without wind turbine generator, the proposed method as testing in systems of IEEE 14 bus, using Power world simulator software version 18 Education.

Streszczenie. Celem artykułu jest oprawa efektywności systemu farm wiatrowych przez optymalizację lokalizacji kondensatorów (Thyristor Controled Series Capacitors) i kompensatorów mocy biernej. Analizowano system połączeń zgodny z IEEE 14 z wykorzystaniem oprogramowania Power world simulator. Optymalizacja lokalizacji FACTS w syastemie farm wiatrowych

Keywords: Power system, FACTS, Wind generator, Power transfer distribution factors, sensitivity of voltage.
Słowa kluczowe: FACTS, fary wiatrowe, kompensacja mocy biernej.

Introduction

Due to the augmentation of electrical energy and the complicated of the electrical grids. Which result in many problems as overloading or contingencies, where the system does not become secure. The main objective of the engineers is to enhance power system safety [1]. FACTS controllers such as Thyristor Controlled Series Compensator can be help to reduce the flows in heavily loaded lines, low system loss, enhanced the stability of the network, reduced cost of production, improve power system security [2], [3].

Today, Most of the electrical networks content the renewable energy (wind, solar …), however, we can find many type of electrical generators depending on the type of energy for this reason that we are studying the impact of wind generation on the optimal location of TCSC and SVC . In 2016, the global total of electricity generation capacity from wind power amounted to 486,790 MW, an augmentation of 12.5% contrasted to the previous year. Installations augmented by 54,642 MW, 63,330 MW, 51,675 MW and 36,023 MW in 2016, 2015, 2014 and 2013 respectively. [4]

These papers [5],[6],[7],[8] talk about different methods for optimal location of TCSC , the Power transfer distribution factor is suggested method for an optimal place of TCSC, the PTDF can be calculated by using DC power flow system parameters, for calculating the PTDF in faster way [9],[10],[11].

The sensitivity of voltage is method used for optimal location of SVC, it is developed by Newton Raphson [12] The main object of this paper is optimal location of TCSC and SVC, and studying the impact of wind generation on the optimal location of TCSC and SVC with and without wind generator, Results obtained through the emulation on IEEE 14 bus, using Power world simulator software version 18 Educational.

Materials and Methods Modelling of the series facts device TCSC

The TCSC is series type of FACTS, it consists of an inductance in series with a thyristor valve, shunted by capacitor, as shown on figure 1 [13],[14].

The impedance of TCSC can be given as following:

where: α : The firing angle, X_L: The reactance of the inductor, X_C:is the reactance of the capacitor, X_l is the effective reactance of the inductor at firing

Static Var compensator

SVCs are a shunted type of FACTS controller and is a Static Var absorber or generator whose output is adjusted to exchange inductive or capacitive current to maintain the bus voltage. SVCs consist shunt reactors and capacitors,[15] which are controlled by thyristors as shown in the figure 2.

The main objectives of SVC are to increase the stability limit of the power system, to decrease voltage fluctuations during load variations and to limit over voltages due to large disturbances.[16] [17]

The SVC equivalent susceptance is,

Suppose the SVC is installed at bus k , the reactive power which injected by SVC can be describe as equation following :

Wind turbine generator

Wind power is the use of air flow through wind turbines, which convert mechanical energy to electrical energy. Today, there are several different concepts of wind turbine generator as shown in figure 3, this classification based on their connection with network, the most commonly of the new models are types 3 and 4 (USA and oversea). However, many users of type 1 and 2 in-services around the world. [18],[19].

The Power Transfer Distribution Factor

The Power transfer distribution factor shows the sensitivity of the flow on line to a transfer of power between two buses i to j, it means the change of real power in a branch flow for a 1 MW exchange between two buses [21],[22]. It shows by the following expression:

This method used for optimal location of TCSC and the TCSC must be placed in most sensitivity line.

where: m– line index, k – bus where power is injected, l – bus where power is taken out, Δf_m– change in megawatt power flow on line m when a power transfer of ΔP_{k to l} is made between k and l. ΔP_{k to l} – power transferred from bus k to bus l.

The sensitivity of voltage

The sensitivity of voltage model was developed by Newton-Raphson method, power flow equation is given as

ΔP: the change in the real power, ΔQ: the change in reactive power, ΔV and Δδ are the deviations in bus Voltage magnitude and angle.

To get voltage expressed dV/dQ , the ΔP must assumed to be zero, the final expressed of dV/dQ can be written as:

The equation 7 is a sensitivity of voltage to an injection of reactive power at a bus has on various parameters. The SVC device can enhance the voltage stability by injected or absorbed reactive power, the voltage sensitivity used for optimal location of the SVC devices [12].

Results and discussion

In this part, we are used electrical network transmission ( IEEE 14 bus system [23].) for applied the two precedent methods of optimal location of SVC and TCSC ,in different cases of studies .

The steps of simulation are:

• Find the optimal location of FACTS ( TCSC and SVC) .
• Placed FACTS in optimal location and compared the results with and without FACTS.
• Injected a wind generation and compared the results with and without FACTS.

The parameters of wind generator are as following:

Type of Turbine: type 3 (machine model: GEWTG, Exciter model: EXWTGE, Governor model: WNTDGE) . The generator at bus 3 will be modelled using a GEWTG Machine, which models a 30 MW aggregation of GE 1.5 MW DFIGs (Doubly-fed induction generators). It will be equivalent 20 generators of 1,5 MW DFIGs. We replaced the generator of the bus 3 by wind turbine. Then we will find the best location of the TCSC.

Optimal location of TCSC

The IEEE 14 bus system drew by power world simulator, the lines don’t have limits, it observed different limit for each line , such as the highest limit could be seen in the line 4-2 (103 %) , this value indicated the line couldn’t supported transferred power (MW/MVAR) , it mean the line 2-4 overloaded. As show in the figure 4.

Suppose the line 2-4 is overloaded (103%). The main goal of this simulation is solved the overloaded. Using the optimal location of TCSC.

We Suppose the line 2-4 is overloaded (103%) as shown in figure 3, we use the PTDF to find the optimal location of TCSC, for solving the overloaded problems. the results of PTDF show at flowing table:

Table 1. PTDF when line 2-4 overloaded

We use the PTDF to find the optimal location of TCSC, for solving the overloaded problems. The results of PTDF are in the table 1

The PTDF of each line when line 2-4 was overloaded. it presented on the table 1, it observed that the most sensitivity lines were the lines 2-4 and 2-5, 37.4% , 27.68% respectively the lines 3-4 and 2-3 had the same sensitivity with 17,87 % it mean same parameters.

Before we place TCSC for each case, we must determinate the total impedance of line. Suppose the compensation is 70% , we can find the values of total impedance (X__(total)) of X_line by equation 8.

X_TCSC = 70% X_line or X_TCSC = 0.7 X_line
So X_total= X_line – 0.7 X_line = 0.3X_line

New impedance of each line can be given as following:

X_total(2-5) = 0.173880 x 0.3 = 0.052164 pu
X_total(3-4) = 0.171030 x 0.3 = 0.051309 pu
X_total(2-3) = 0.197970 x 0.3 = 0.059391 pu

Table 2. The overloaded of line 2-4 after installing TCSC and wind generator

From the table 2, it is indicated the overloaded of line 2-4, after placed TCSC, and TCSC with wind generator together, the two cases explain as following:

Case one, optimal location of TCSC without wind generator:

1) TCSC in line 2-5 : After installed the TCSC in the line 2-5, it observed that, the overload of the line 2-4 decreased from 103% to 87% and creased from 75% to 80 % on the line 1-2. The TCSC reduced the overload on the line 2-4.

2) TCSC in line 3-4: When installed the TCSC in the line 3-4, it observed that the overload of the line 2-4 increased from 103% to 113 %. it indicated that not optimal location of TCSC.

3) TCSC in line 2-3: The TCSC installed on the line 2-3, it observed that the overload of the line 2-4 decreased from 103% to 84% . The TCSC reduced the overload on the line 2-4.

Case two, optimal location of TCSC with wind generator:

1)TCSC in line 2-5 and wind generator at bus 3: When we place the TCSC in the line 2-5 in presence of Wind generator placed on bus 3, the overload is decreased from 103% to 80%, it was better result compared with to install WG. The Wind Generator could be playing function of generator and system of protection.

2)TCSC in line 2-3 and wind generator at bus 3: if the TCSC is placed in the line 2-3, it removed the overload present in line 2-4 from 103% to 80%, the Wind generator was produced the active and reactive power which missed by the power system, the produced power relieved the stability of the power system.

3) TCSC in line 3-4 and wind generator at bus 3: if the TCSC is placed in the line 3-4 and wind generator installed on bus 3 , it noticed that the overload present in line 2-4 increased to 103%, according to the previous result, this line was not the optimal location of TCSC, but compared with last result the overload decreased by 10 % , from 113% to 103 % .

We can conclude that the TCSC and the wind generator can improve the power flow in the network. In our example, they suppress the overload in the line 2-4.

Table 3. The power loss with and without generator wind

The table 3 shows the active and reactive power loss with and without generator wind.

1)The total of active power loss with wind generator decreased for three cases, if TCSC placed in line 2-5 , it noticed that the total active loss minimized almost 3,73 MW , When TCSC placed in line 2-3 as shown in figure 11 the total of active power loss minimized by 3.69 Mw . Finally, The TCSC placed in line 3-4 the active power loss decreased by 3.74 Mw .

From the results, it observed the most reduced loss active power when TCSC installed in 3-4 . But this line was not the optimal location, because the overload did not removed on the line 2-4 .

The wind generator could be reduced the active power loss.

2)The total of reactive power loss with wind generator decreased for three cases, if TCSC placed in line 2-5 , it noticed that the total active loss minimized almost 9,52 Mvar , When TCSC placed in line 2-3 the total of reactive power loss minimized by 5,43 Mvar . Finally, The TCSC placed in line 3-4 the reactive power loss decreased by 17,17 Mvar .

From the results show in the table 2, it observed the most reduced loss reactive power when TCSC installed in 3-4 . But this line was not the optimal location, because the overload did not removed on the line 2-4 . The wind generator could be reduced the reactive power loss.

Optimal location of SVC

In this section, we used the same power system (14 IEEE bus system) , in this case we try used shunt type of FACTS ( SVC ) for controlling the voltage , we remember that the limit of voltage used in this simulation is ±10%.

For example:

If bus 1 has more or less 10% than it’s voltage (drop voltage or overvoltage) , we can say , that is risk ,we should use protection devices.

Table 4. Voltage profile of IEEE 14 bus in de base case

From the result of voltage profile which indicates in table 4 , it is noticed that most voltage of buses is between 1.06 pu and 0.92962 pu , however there are three critical buses, which are bus 12 (0.84036 pu) , bus 13(0.85519 pu) and bus 14 (0.8926 pu) .

We have just drop voltage in three buses, there is not overvoltage.

The voltage sensitivity factor used to install the SVC in the optimal placement in the power system, where the SVC should be placed in most sensitivity bus.

Table 5. Voltage profile of IEEE 14 bus in de base case

Table 5 shows the voltage sensitivity when we injected a reactive power. We can see different values, 0 in the slack bus (because the slack-bus has constant voltage) . The most sensitivity bus is bus 12 (0.00718792). If the SVC is installed in this bus, it will inject a reactive power which can be help to improve the voltage at the critical buses.

Table 6. Compared voltage magnitude for three cases

From the result after installing SVC at bus 12 as shown in table 6 , the voltage is improved for all buses than the base case , almost the voltage is more than 1 pu just for the critical buses improve from 0.84036 pu to 0,99703 pu at bus 12 , from 0,85519 pu to 0.9779pu at bus 13 and from 0.8926 pu to 0.98684 pu at bus 14. The slack-bus always has constant voltage.

We are noticed that the reactive power injected by SVC in bus 12 is 29.6 MVar .

The SVC is used to limit the transfer of reactive power for reduce the drop voltage. The optimal location of SVC is very important for improving the security of power system. When the SVC placed at bus 12 and changed the generator at bus 3 by wind generator, it can see that the reduction of voltage with small value of each bus , because of the variation of frequency due to the variation of speed of turbine generator.

Conclusion

This paper presents new methods for the optimal location of FACTS. The proposed method is the power transfer distribution factor used for TCSC and sensitivity voltage for SVC. We conclude, the wind generator permits to enhance the function of TCSC and SVC. As result, the active and reactive loss is decreased after injected the wind generator. The install of FACTS (TCSC, SVC) and wind generator simultaneously can be economically.

REFERENCES

[1] A. N. L. Sayyed., P. M. Gadge , Sheikh R.U., “Contingency Analysis and Improvement of Power System Security by locating Series FACTS Devices TCSC and TCPAR at Optimal Location, ” IOSR-JEEE, 2014 , p. 19-27.
[2] J. Navani, M. Goyal, , S. Sapra, “Optimal Placement of TCSC and UPFC for Enhancement of Steady State Security in Power System, ” International Journal of Advances in Engineering Science and Technology, 1, 2013 , pp. 122-129, 2013.
[3] S. Singh, “Location of FACTS devices for enhancing power systems’ security, ” in Power Engineering, 2001. LESCOPE’01. 2001 Large Engineering Systems Conference on, 2001.
[4] Wind power by country, [online], Available at: https://en.wikipedia.org
[5] Thanh Long Duong, Yao JianGang , VietAnh Truong , A new method for secured optimal power flow under normal and network contingencies via optimal location of TCSC, Electrical Power and Energy Systems journal , 53, 2013.
[6] Madhura Gad, Prachi Shinde, S.U.Kulkarni ,“ Optimal ocation of TCSC by Sensitivity Methods”, International Journal Of Computational Engineering Research, Vol. 2 Issue. 6,pp 162-168,2013.
[7] Ghamgeen I. Rashed , Yuanzhang Sun , H. I. Shaheen, Optimal Location and Parameter Setting of TCSC for Loss Minimization Based on Differential Evolution and Genetic Algorithm , International Conference on Medical Physics and Biomedical Engineering, Physics Procedia 33 1864 – 1878. (2012)
[8] P. S. Vaidya and V. P. Rajderkar, Enhancing Power System Security by Proper Placement of Thyristor Controlled Series Compensator (TCSC), International Journal of Engineering and Technology, 4, 5, October 2012.
[9] Darko Šošić, Ivan Škokljev, Nemanja Pokimica, Features of Power Transfer Distribution Coefficients in power System Networks , INFOTEH-JAHORINA ,13, pp. 86 – 90 (March 2014).
[10] Chong Suk Song, Chang Hyun Park, Minhan Yoon & Gilsoo Jang , Implementation of PTDFs and LODFs for Power System Security , Journal of International Council on Electrical Engineering, 1, 1, pp. 49-53, 2011.
[11] Henrik Ronellenfitsch, Marc Timme, Dirk Witthaut, A Dual Method for Computing Power Transfer Distribution Factors, JOURNAL OF L ATEX CLASS FILES, 13, 9, SEPTEMBER 2014.
[12] Jitendra Singh Bhadoriya , Chandra Kumar Daheriya,“An Analysis of Different Methodology for Evaluating Voltage Sensitivity”, International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, vol.3, no. 9,pp 12239 – 12246,September 2014
[13] S.V. Jethani ,V.P. Rajderkar, “Sensitivity based optimal location of TCSC for improvement of power system security, ” International Journal of Research in Engineering and Technology, vol. 3, 2014, pp.121-124.
[14] J. Srinivasa Rao , J. Amarnath , “Enhancement of Transient Stability in a Deregulated Power System using Facts Devices, ” Global Journal of Research In Engineering, vol. 14, 2014.
[15] A. Edris, R. Adapa, M.H. Baker, L. Bohmann, K. Clark, K. Habashi, L. Gyugyi, J. Lemay, A. Mehraban, A.K. Myers, J. Reeve, F. Sener, D.R. Torgerson, R.R. Wood,“Proposed Terms and Definitions for Flexible AC Transmission System (FACTS)”, IEEE Transactions on Power Delivery, vol.12, no. 4,pp 1848-1853, October 1997.
[16] Ryan.M, High-Voltage Engineering and Testing ,3rd ed, Institution of Engineering and Technology. 2013.
[17] Padmavathi S.V., Sahu S.K., Todoran G., Jayalaxmi A., “Modeling and simulation of static var compensator to enhance the power system security”, in: Proceedings of the International Conference “ Postgraduate Research in Microelectronics and Electronics (PrimeAsia)”, Visakhapatnam, India, 19-21 Dec. 2013 , IEEE,06 February 2014, pp. 52-55.
[18] Working Group Joint Report – WECC Renewable Energy Modeling Task Force & IEEE Working Group on Dynamic Performance of Wind Power Generation, Generic Stability Models for Type 3 & 4 Wind Turbine Generators for WECC, International conference of IEEE Power & Energy Society General Meeting, Vancouver, BC, Canada, 2013.
[19] Jens Fortmann , Modeling of Wind Turbines with Doubly Fed Generator System , Springer Fachmedien Wiesbaden, Germany,2014, pp 3-5.
[20] EPRI, “Proposed Changes to the WECC WT4 Generic Model for Type 4 Wind Turbine Generators”, Prepared under Subcontract No. NFT-1-11342-01 with NREL, Issued to WECC REMTF and IEC TC88 WG27 12/16/11; (last revised 1/23/13).[Online].Available:https://www.wecc.biz/Reliability/Report_on_WT4_Model_Description_PP012313.pdf.
[21] Ravi Kumar, S. C. Gupta and Baseem Khan , Power Transfer Distribution Factor Estimate Using DC Load Flow Method, International Journal of Advanced Electrical and Electronics Engineering (IJAEEE),2, 6, pp. 155-159( 2013).
[22] Allen J. Wood, Bruce F. Wollenberg, and Gerald B, Power Generation, Operation and Control 3 thrd edition, John Wiley & Sons, Inc., Hoboken, New Jersey . 2013.
[23] http://icseg.iti.illinois.edu/ieee-14-bus-system/ last update 7 may 2014.

Authors: PhD student Messaoud Zobeidi , Department of electrical Engineering laboratory of ICEPS, Djillali Liabes Universiy, Sidi-Bel-Abbes, messoud91@yahoo.fr ; Pr Fatiha Lakdja , Department of Electrical Engineering, University Saida , laboratory of ICEPS, Djillali Liabes Universiy, Sidi-Bel-Abbes , flakdja@yahoo.fr ; Dr Yamina Ahlem Gherbi , Department of electrical Engineering, USTO-MB Oran , Durable Development of Electric Power laboratory, Algeria aygherbi@yahoo.fr ; Pr Fatima Zohra Gherbi, Department of Electrical Engineering laboratories of ICEPS, Djillali Liabes Universiy , Sidi-Bel-Abbes,Algeria ,fzgherbi@yahoo.fr

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 96 NR 9/2020. doi:10.15199/48.2020.09.09

Fluid Interaction in a Complex Terrain Wind Farm

Published by Bukurije HOXHA¹, Risto V.FILKOSKI² ,
University of Prishtina “Hasan Prishtina” (1), Ss. Cyril and Methodius University, Skopje (2)
ORCID: 1. 0000-0002-8890-2054; 2. 0000-0002-3743-318X

Abstract. Optimisation of the placement of wind turbines in a farm is an important stage in wind farm construction. Koznica, a mountainous terrain available for wind farms, is considered in the present work, with the aim to optimise the configuration of a park of 10 turbines. The one-year measurements carried out in the specified place of the Koznica mountain have confirmed the wind energy potential. The present work focuses on analysing how the distance between wind turbines affects the energy produced in configurations of 2, 3 and 5 diameters distance between the turbines, using a particular turbine type with predefined technical characteristics. Interaction analysis is conducted in terms of wake effect, affecting the annual output energy and wind farm efficiency, depending on the farm configuration. The wake effect here is shown as the wind speed deficit. That deficit intensity is removed from the current and previous intensity of each respective turbine. Finally, the difference between the farm organisation’s optimised form and the previous configurations is shown, emphasising the annual energy produced depending on the capacity factor.

Streszczenie. Optymalizacja rozmieszczenia turbin wiatrowych na farmie jest ważnym etapem budowy farmy wiatrowej. W niniejszej pracy uwzględniono Koznicę, górzysty teren dostępny pod farmy wiatrowe, w celu optymalizacji konfiguracji parku 10 turbin. Roczne pomiary przeprowadzone we wskazanym miejscu góry Koznica potwierdziły potencjał energetyki wiatrowej. W niniejszej pracy skupiono się na analizie wpływu odległości między turbinami wiatrowymi na energię wytwarzaną w konfiguracjach 2, 3 i 5 średnic odległości między turbinami, z wykorzystaniem określonego typu turbiny o określonych parametrach technicznych. Analiza interakcji prowadzona jest pod kątem efektu czuwania, mającego wpływ na roczną moc wyjściową i sprawność farmy wiatrowej, w zależności od konfiguracji farmy. Efekt kilwateru jest tutaj pokazany jako deficyt prędkości wiatru. Ta intensywność deficytu jest usuwana z aktualnej i poprzedniej intensywności poszczególnych turbin. Na koniec pokazano różnicę między zoptymalizowaną formą organizacji gospodarstwa a poprzednimi konfiguracjami, podkreślając roczną produkcję energii w zależności od współczynnika wydajności. (Oddziaływanie miedzy turbinami w złożonej terenowej farmie wiatrowej)

Keywords: wind turbine, capacity factor, efficiency, energy yield, complex terrain.
Słowa kluczowe: farmy wiatrowe, turbiny, oddziaływanie między turbinami

Introduction

Airflow as a potential energy source is free. Air is used as the main source of energy production in wind farms and the initial assumption is that it, as an interaction fluid, will have identical energy potential for all turbines on the farm under consideration. In order for the flow of this interaction fluid not to be affected from the previous turbine to the next, it is necessary to have an optimal possible distance between the turbines. Objectively, the optimal distance is impossible to achieve to a large extent in the case of terrains such as this one that is the subject of analysis in this case. This is due to the frequent ups and downs of the hilly terrain for which the analysis is performed. Energy is a vital key to socio-economic development and economic growth [1]. Wind energy constitutes a clean and economical source to generate electricity and is suitable for countries with moderate to high wind potential [2]. Based on Global Wind Energy Council (GWEC) data’s 2018, the worldwide newly installed wind energy capacity exceeds 60 GW in 2020, distributed over about 100 countries, and it is estimated that the globally installed wind energy capacity will exceed 800 GW [3]. Wind energy has the potential to produce power for each hour of the day and in different capacities due to different wind speeds [4]. To make the most of wind energy, between the other, improvements in the design, construction and materials must be made to have the highest energy conversion efficiency [5]. Today, wind power is considered a booming sector, and wind turbine manufacturers face difficulties meeting the market demand [6]. They are manifested in different forms, such as in the construction of the rotor part of the turbine, with the addition of other units, i.e., mini turbines [7 – 9], thus creating multiple rotors. Such a method is not yet widely used and what is known is that it contributes to a significant reduction in the total cost of a wind farm. During planning a wind farm, it is necessary to assess the wind resources of the potential site [10]. The element that requires further study is the fatigue of the structure because of the turbulence generated and the problems with reduced overall efficiency for operating in this form if it is not preceded by static side analysis [11]. According to the work [12], significant problems which must be solved to support the development of wind energy are the insurance of power transmission capacity of the electricity network, balancing the energy generated by wind power plants related to the error control of forecasting, and transition of the energy system from conventional to renewable generation. In this study, a detailed description of the parameters that characterise the placement of wind turbines in a farm according to the software and analytical method is realised. Wind measurements must be as accurate as possible to achieve optimal placement of turbines in a wind farm [13]. Greater accuracy of measured data means a more extended period of measurement and collection of wind parameters at the place where the wind farm will be built. Calculations regarding the optimisation of the placement of wind turbines on the wind farm in Koznica were made through WAsP software where the change of the interaction of the turbines between themselves can be simulated and clearly understood [14], in addition to the numerical method WEST, which is more appropriate. Studies have shown a very good correlation between the data obtained from the use of the software in question and the numerical form of the simulation [15]. Installed wind energy capacity has grown from less than 20 gigawatts (GW) in 2000 to 590 GW by the end of 2018 and already provides 6% of the electricity consumed in the world [16]. By knowing the important role that renewable-energy sources play in this context, identifying suitable sites for installing renewable-energy facilities is a crucial task [17, 18]. The wind farm layout optimisation (WFLO) problem consists of finding the turbine positions that maximise the expected power production in terms of coordinates and altitude of hub height [19]. However, the effect of wind veer causing a partial yaw error over the rotor span is rarely considered. Analysis of such effects becomes increasingly important as the dimensions of the wind turbine rotors increase [20]. To estimate the energy potential when it comes to wind energy, measurements must be made on the source side, and in this case, the key data is the wind speed. Measurements are a significant part of the cost of investing in a wind farm, and it is usually attempted to simulate the interactions of the individual turbines in a wind farm through some software such as CFD [21, 22]. To help in understanding the wind farm processes, some of the widely used wind energy application programs are being used, such as WAsP, WindPro, WindFarmer, WindSim, and Windographer [23 – 25]. However, because WAsP was developed for linear, steady flow analyses, its applicability for analyses of wind flow over complex terrain is very limited. In recent years, nonlinear, unsteady flow analyses have become possible because of the rapid improvement of computers [26 – 28]. In the last part of this work, further investigations have been carried out to compare the WEST results with linear simulation WAsP. Finally, considering commercial models of a wind turbine, the wind potential has been estimated for each turbine in this park. It is important to remark that the scope of the WEST method is not to yield more accurate or exact results concerning WAsP but to be a cheaper and faster alternative for simulations, to have good results in less time. The paper takes into consideration a very detailed and unrealized analysis of the complex mountain terrain. The more contributing is the fact that in the case of installing wind turbines in our case we have a significant change in terms of altitude. The impact of such a change is negative. This is because in such a case some “fictitious tunnels” are created which are negative loads for the turbines.

Material and Methods

The wind farm considered in this study is in Koznice, a mountainous terrain without obstacles such as tall buildings, rocks, or trees. The considered terrain of 20 km – 20 km is considered. The determination of the coordinates is done based on the map of wind density and speed, of course, for each configuration looking to meet that condition. Figure 1 shows the specified location of Koznica, specified by its met mast.

The two most essential elements that are important when analysing a potentially potential place in the field of wind energy are the intensity of the average wind speed and its direction. The wind direction in Koznica is shown in figure 2.

Regarding the interaction of air masses with special emphasis, it is important to take the interaction between the height referred to build the wind park and the complex terrain represented with the degree of roughness. In detail, in table 1 are presented the average speeds at heights 40, 60 and 84m.

Table. 1. Wind speed intensity

Measurement height, m	84	60	40
Average wind speed, m/s	6.16	5.85	5.64

Then for energy calculation are considered 3 type of wind turbines and their number is 10. The average annual wind speed of Koznica at 84 m height above ground level is shown in Fig. 3. Maps are created using the WAsP software based on annual data from one-year measurement of met mast of location Koznica and formed digital maps of terrain orography and roughness.

Fig. 3. Average annual wind speed map of candidate region Koznica at 84 m height above ground level

Types of wind turbines considered are presented in table 2 with their technical data.

Table. 2. Main data for wind turbines used in the study, thrust coefficient, C_EF, and wake decay coefficient.

WT Type	D	P_f	Z_h	W_r	Wind speed, m/s	Thrust Coefficient	α
Siemens SWT-130-3.3	130	3.3	10	12.5	6.6	0.80	0.0609
Vestas V-126-3.45	126	3.45	87	20	6.9	0.785	0.0627
GE Wind GE-130-3.4	130	3.4	85	18	6.5	0.83	0.0628

Power production modelling

To estimate the power production of WF under the wake effect, we need at the first stage to determine the power generated by each WT. There are many expressions to approximate the power curve of WT that are elaborated in detail by [29, 30]. Thus, the power production of WT is estimated as follows:

where C_EF represents the efficiency factor expressed as follows:

In the present study, C_EF is assumed to be equal to 50%. The total power generated by WTs operating under wake effect is:

Wind farm efficiency is obtained using this equation:

According to Jensen wake model, the w_df – wind velocity deficit is expressed as follows:

Based on the Jensen model as described in figure 4 graphically, it is said that the near wake region is for 2D, Intermediate is 2-3D, and far wake > 5D distance. To see this effect then it is used for a real wind farm with real wind speed, direction and standard deviation data [31-33].

Fig. 4. Illustration of the Jensen wake model

As stated initially, the difference in energy produced is not high for the best case according to Jensen and optimisation according to software and terrain taken in study [34-36]. The expression for electricity calculation is multiplied by 0.98, as a correction coefficient that brings innovation in the field of energy calculation. Difference in energy yield for each case is described in figure 5. We see that the huge amount of energy yield is for the highest distance between wind turbines, 5D.

Fig. 5. The difference in energy generation, in MWh/yr.

Fig. 6. Applying the Jensen model to determine the efficiency for 2D, 3D and 5D and their comparison.

The following figure shows the ratio of the largest amount of energy generated during the year for the placement of turbines at 5D distance compared to 3D and 2D cases and 3D/2D cases.

After analysing the change in the intensity of the annual energy produced within the distances then the optimisation model is formed so that we can have a clear reflection of the speed in each position of the turbines and the organising of the park, presented in figure 7.

Fig. 7. Organised wind park after optimisation of the turbines position

To carry out the analysis of the produced energy and efficiency it is necessary to present the data of their speed and coordinates. The data are valid for the measuring height of 110m as the average data for all turbines. The distance shown is in order from 1-2, 2-3, onwards.

Fig 8. Wind speed in m/s for each wind turbine and their distance in optimised placement in meters

From the previous figure in the optimisation configuration, we have in each turbine the speed above 6 m/s which is an indicator for the wake effect.

Conclusions

In this paper, the Jensen model has been used to study and evaluate the wind energy potential of a wind farm in Kosovo, exactly Koznica. This model allowed reconstructing the distribution of wind fields of complex territories, providing helpful information about wind farm layout optimisation. The interaction between the wind turbines is explained in terms of the wake effect created by the previous turbine in the next one. Finally, in the last part of this paper, all turbines in that windfarm have been investigated by WAsP simulations, confirming the new proposed methodology as an acceptable way for windfarm analysis. It has been shown in the paper that the wind speed deficits obtained from the Jensen wake model for a wind turbine, as a function of wind direction, depend on how we observe wakes. Results are different when considering that the deficits are observed by a point measurement, met mast, compared with those observed by a second turbine due to partial wake interaction. Moreover, the remarkable correlation between the approximated method through many calculations and that achieved by the WAsP software can be noticed, due to the linearity that accompanies them. When comparing the use of such a method with different software, we can see that we have a higher correlation with WAsP software because the modelling is linear. Future research will focus on the investigation of the structure and characteristics of the wakes originated from large wind farms under different ABL flow cases involving different land surface covers and different atmospheric stability conditions.

REFERENCES

[1] Joselin Herbert G.M., Iniyan S., Sreevalsan E., Rajapandian S., A review of wind energy technologies. Renewable and Sustainable Energy Reviews. Volume 11, Issue 6, August 2007.
[2] NA. Kallioras, N.D. Lagaros, M.G. Karlaftis, P. Pachy, Optimum layout design of onshore wind farms considering stochastic loading. Adv. Eng. Softw. 88, 8–20 (2015). https://doi.org/10.1016/j.advengsoft.2015.05.002
[3] Wu, X.; Hu, W.; Huang, Q.; Chen, C.; Chen, Z.; Blaabjerg, F. Optimized Placement of Onshore Wind Farms Considering Topography. Energies 2019, 12, 2944.
[4] Demolli H., Sakir A., Ecemis A., and Gokcek M., “Wind power forecasting based on daily wind speed data using machine learning algorithms,” Energy Convers. Manag., vol. 198, 2019.
[5] Glassbrook, K.A., Carr, A.H., Drosnes, M.L., Oakley, T.R., Kamens, R.M., Gheewala, SH, 496 2014. Life cycle assessment and feasibility study of small wind power in Thailand. Energy for 497 Sustainable Development 22, 66–73.
[6] Abbes M., Belhadj J. Development of a methodology for wind energy estimation and wind park design. JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY 6, 053103 (2014)
[7] Watson, S., Moro, A., Reis, V., Baniotopoulos, C., Barth, S., Bartoli, G., Bauer, F., Boelman, E., Bosse, D., Cherubini, A., Croce, A., Fagiano, L., Fontana, M., Gambier, A., Gkoumas, K., Golightly, C., Latour, M., Jamieson, P., Kaldellis, J., and Wiser, R., Future emerging technologies in the wind power sector: A European perspective,Renewable and Sustainable Energy Reviews, Vol. 113, 10 2019, pp. 109-270.
[8] Flynn, D., Rather, Z., Ardal, A., D’Arco, S., Hansen, A.D., Cutululis, N.A., Sorensen, P., Estanquiero, A., Gomez, E., Menemenlis, N. and Smith, C., 2017. Technical impacts of high penetration levels of
wind power on power system stability. Wiley Interdisciplinary Reviews: Energy and Environment, 6(2), p.e216.
[9] Cherubini A, Papini A, Vertechy R, Fontana M. Airborne wind energy systems: a review of the technologies. RenewSustainEnergyRev2015; 51:1461–76.
[10] M. Ayala, J. Maldonado, E. Paccha, and C. Riba. Wind Power Resource Assessment in Complex Terrain: Villonaco Case-study Using Computational Fluid Dynamics Analysis. Energy Procedia, 107(September 2016):41–48, 2017.
[11] Porté-Agel, F., Bastankhah, M. & Shamsoddin, S. Wind-Turbine and Wind-Farm Flows: A Review. Boundary-Layer Meteorol 174, 1– 59 (2020). https://doi.org/10.1007/s10546-019-00473-0.
[12] Jonaitis, A.; Gudzius, S.; Morkvenas, A.; Azubalis, M.; Konstantinaviciute, I.; Baranauskas, A.; Ticka, V. Challenges of IntegratingWind Power Plants into the Electric Power System: Lithuanian Case. Renew. Sustain. Energy Rev. 2018, 94, 468–475.
[13] Al-Abadi A, Youjin K., Ertunç. Ö, Delgado A. (2016). Turbulence impact on wind turbines: experimental investigations on a wind turbine model. Journal of Physics Conference Series 753(3). https://doi.org/10.1088/1742-6596/753/3/032046.
[14] Mortensen, N. G. (2018). Wind resource assessment using the WAsP software. DTU Wind Energy. DTU Wind Energy E, No. 174
[15] Milanese M, Congedo PM, Colangelo G, et al. (2019) Numerical method for wind energy analysis in WTG sitting. Renewable Energy 136: 202–210.
[16] Yusta J., Arantegui R., Measuring the internationalisation of the wind energy industry. Renewable Energy 157 (2020). Volume 157, September 2020, Pages 593-604.
[17] Díaz-Cuevas, P. GIS-Based Methodology for Evaluating the Wind-Energy Potential of Territories: A Case Study from Andalusia (Spain). Energies 2018, 11, 2789.
[18] Asnaz M.S.K., Yuksel B., Ergun K. (2020) Optimal Siting of Wind Turbines in a Wind Farm. In: Machado J., Özdemir N., Baleanu D. (eds) Mathematical Modelling and Optimisation of Engineering Problems. Nonlinear Systems and Complexity, vol.30 Springer, Cham. https://doi.org/10.1007/978-3-030-37062-6_6.
[19] Gu H., Wang J., Irregular-shape wind farm micro-siting optimisation. Energy 57, 535-544 (2013).
https://doi.org/10.1016/j.energy.2013.05.066.
[20] Bardal, L. M., Sætran, L. R., and Wangsness, E.: Performance Test of a 3 MW Wind Turbine – Effects of Shear and Turbulence, Energy Proced., 80, 83–91, https://doi.org/10.1016/j.egypro.2015.11.410, 2015.
[21] Tabas, D.; Fang, J.; Porté-Agel, F. Wind Energy Prediction in Highly Complex Terrain by Computational Fluid Dynamics. Energies 2019, 12, 1311.
[22] Richmond, M. Antoniadis, A. Wang, L. Kolios, A. Al-Sanad, S. Parol, J. Evaluation of an offshore wind farm computational fluid dynamics model against operational site data. Ocean Engineering Volume 193, 1 December 2019, 106579.
[23] Yılmaz U., Balo F., and Sua S. Simulation Framework for Wind Energy Attributes with WAsP. Procedia Computer Science. Volume 158, 2019, Pages 458-465
[24] Carvalho, D. & Rocha, A. & Santos, C. Silva & Pereira, R., 2013. “Wind resource modelling in complex terrain using different mesoscale–microscale coupling techniques,” Applied Energy, Elsevier, vol. 108(C), pages 493-504.
[25] Gasset, N.; Landry, M.; Gagnon, Y. A Comparison of Wind Flow Models for Wind Resource Assessment in Wind Energy Applications. Energies 2012, 5, 4288–4322.
[26] Uchida, T.; Takakuwa, S. A Large-Eddy Simulation-Based Assessment of the Risk of Wind Turbine Failures Due to Terrain-Induced Turbulence over a Wind Farm in Complex Terrain. Energies 2019, 12, 1925.
[27] Mohamed B, Fadela B, Mounir K (2015) Optimisation of the wind turbines location in Kaberten wind farm in Algeria. Energy
[28] CL Archer, A. Vasel-Be-Hagh, C. Yan, S. Wu, Y. Pan, J.F. Brodie, A.E. Maguire, Review and evaluation of wake loss models for wind energy applications, Appl. Energy 226 (2018) 1187e1207.
[29] J. Scire, F. Robe, M. Fernaua, R. Yamartino, A User’s Guide for the CALMET Meteorological Model, Technical Report Version 5, Earth Tech Inc., Concord, MA, USA, 2000.
[30] M. Milanese, L. Tornese, G. Colangelo, D. Laforgia, A. de Risi, Numerical method for wind energy analysis applied to Apulia Region, Italy, Energy 128 (2017) 1e10.
[31] F. González-Longatt, P. Wall, and V. Terzija, “Wake effect in wind farm performance: Steady-state and dynamic behavior,” Renewable Energy, vol. 39, no. 1, pp. 329–338, Mar. 2012.
[32] CL Archer, A. Vasel-Be-Hagh, C. Yan, S. Wu, Y. Pan, J.F. Brodie, A.E. Maguire, Review and evaluation of wake loss models for wind energy applications, Appl. Energy 226 (2018) 1187e1207.
[33] Bachhal A., Vogstad K., Lal Kolhe M., Chougule A., Beyer G.H. Wake and Turbulence Analysis for Wind Turbine Layouts in an Island. ICPRE 2018. E3S Web of Conferences 64, 0 (2018).
[34] Lopez-Villalobos, C.A.; Rodriguez-Hernandez, O.; Campos-Amezcua, R.; Hernandez-Cruz, G.; Jaramillo, O.A.; Mendoza, J.L. Wind Turbulence Intensity at La Ventosa, Mexico: A Comparative Study with the IEC61400 Standards. Energies 2018, 11, 3007.
[35] Emejeamara, F.; Tomlin, A.; Millward-Hopkins, J. Urban wind: Characterisation of useful gust and energy capture. Renew. Energy 2015, 81, 162–172.
[36] W. Miao, C. Li, G. Pavesi, J. Yang, X. Xie, Investigation of wake characteristics of a yawed HAWT and its impacts on the inline downstream wind turbine using unsteady CFD, J. Wind Eng. Ind. Aerodyn. 168 (2017) 60e71.

Authors. First author is Msc. Ass. Bukurije Hoxha, Faculty of Mechanical Engineering, University of Prishtina, street “Sunny Hill”, nn, 10000, Prishtina, e-mail: bukurije.hoxha@uni-pr.edu. Second author is Prof. Dr. Risto V. Filkoski, Faculty of Mechanical Engineering, Ss Cyril and Methodius, Rudjer Boshkovic Str. 18, 1000 Skopje, R.N. Macedonia, e-mail: risto.filkoski@mf.edu.m

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 98 NR 4/2022. doi:10.15199/48.2022.04.02

Some Results of Research into Harmonics in the High Voltage Networks with Distributed Nonlinear Loads

Published by Lidiia KOVERNIKOVA,
Energy Systems Institute of the Siberian Branch of the Russian Academy of Sciences

Abstract. The paper presents the results of the analysis of harmonic parameters that were measured at the nodes of connecting railway substations to the feeding network. Traction load is nonlinear, unbalanced and stochastic. The 3-rd and 5-th harmonics are dominant in the current of the traction load. The paper shows some properties and specific features of the measured parameters of 3-rd and 5-th harmonics.

Streszczenie. W artykule zaprezentowano rezultaty analizy harmonicznych zmierzonych na węzłach połączeń podstacji kolejowych z siecią zasilającą. Sieć trakcyjna jest nieliniowa, niezrównoważona i stochastyczna. Dominują trzecia i piąta harmoniczna. (Rezultaty badania harmonicznych w sieci trakcyjnej z nieliniowym obciążeniem)

Keywords: harmonic, measurement, railway substation, traction load.
Słowa kluczowe: sieć trakcyjna, harmoniczne

Introduction

Researchers have conducted very many experimental studies on harmonic conditions in the electric networks. The results of their analysis are presented in [1-11]. The studies aimed to analyse the properties and specific features of the harmonic conditions.

The presented paper gives some results of the measurements and analysis of the 3-rd and 5-th harmonic parameters at railway substations. The railway substations receive power from the 110-220 kV public networks with a frequency of 50 Hz. The feeding network is almost radial. The railway substations are distributed along the feeding network. The traction loads affect one another.

The objective of the research was to study the properties of harmonic behaviour in the network with distributed traction loads for identification of regularities. The obtained information will be used further to develop models of nodal loads for the calculation of nonsinusoidal conditions in the high voltage network.

Object of research

Electrified railway in Russia has been a source of harmonics for many years. It is very extended. In East Siberia railway occupies a special place since it is part of the Trans-Siberian railway. The railway runs across the territories with sparse population and small electric loads. At the same time the railway traffic is heavy. The time interval between trains varies from 5 to 20 minutes. The railway substations are as a rule located at a 40-60 km distance from one another, i.e. they are quite evenly distributed along the feeding network. Each section of railway between two substations receives power from two sides. The railway substation has two 40 MVA three-winding transformers. One of them is a reserve transformer. The 25 kV winding of the transformer supplies power to the traction network, whereas 6 (11) kV winding supplies power for auxiliary needs of the substation and to non-traction consumers located near the substation.

Electric locomotives receive power from the traction network. They are driven by DC engines. The engines are powered through single-phase two-pulse rectifier circuits. The rectifier circuits in the traction network cause harmonic currents. The currents of the 3-rd and 5-th harmonics have the largest value. The harmonic currents penetrate through transformers into the 110-220 kV feeding network and cause a distortion of the voltage waveform. Thus, the traction load for the feeding network is nonlinear, unbalanced and distributed. The measurements show that the voltage nonsinusoidality at the nodes of connecting railway substations to the feeding network exceeds the standard limits established in Russia [12].

The considered railway section is situated in East Siberia between the substations Mysovaya and Novoilinsky. The section contains 9 railway substations. Measurements were taken at four substations: Mysovaya, Tataurovo, Zaigraevo, Novoilinsky. For simplicity, we will denote the substations by letters M, T, Z and N respectively. The substations are located as follows: the substation T – in 126 km from the substation M, the substation Z – in 86 km from the substation T, the substation N – in 47 km from the substation Z. Arrangement of the substations relative to each other is presented in Fig. 1, where EPS – electrical power system.

Fig.1. Arrangement of the railway substations

The measurements were carried out at the points of common coupling of railway substations to the feeding network, i.e. on the high voltage side of the transformers. Measurements were performed with the aid of the device “OMSK”, which measures not only the indices of power quality but also currents, powers and other parameters. Each measurement was performed during 24 hours. The parameters were measured mainly in an interval of 1 minute. For voltage measurements the device was connected through voltage transformers to the high voltage buses. For current measurements the device was connected to current transformers installed at the inputs of high voltage transformers.

Measurements were made for three connection schemes of the railway substation:

• scheme I – traction network feeder is connected, i.e. traction network receives electric energy from the 25 kV transformer winding at the railway substation;
• scheme II – traction network feeder is disconnected, i.e. traction network does not receive electric energy from the 25 kV transformer winding at the railway substation;
• scheme III – passive filter in the traction network is disconnected at the railway substation.

The disconnection of the traction network feeder at either of the two substations means that the section of traction network receives electric energy only from one substation. Passive filter is tuned to absorb the 3-rd and 5-th harmonic currents. Fig. 2 shows the scheme that explains connection of transformer and circuit breakers for three schemes during measurements. The number near the circuit breaker corresponds to the number of the scheme in which the circuit breaker was switched on.

Fig.2. Connection scheme of transformer and circuit breakers

Harmonic voltages in terms of the standard requirements

The measurements show that the standard limits [12] are considerably exceeded at 3-rd and 5-th harmonics. The values of measured harmonic voltages U3, U5, that represent the values with a probability of 95% are presented in Table 1. The values exceeding the limits are highlighted in bold.

Table 1. Measured values of U₃, U₅ [%]

Table 1 shows that:

• disconnection of traction network feeder at substation M increased insignificantly the 3-rd harmonic voltage in phase B, but decreased it in phase C, and decreased the 5-th harmonic voltage in all phases;
• disconnection of traction network feeder at substation T decreased voltage at all harmonics in all phases;
• disconnection of feeder at substation Z decreased the 3-rd harmonic voltage in two phases (A, B) and increased it in phase C, increased the 5-th harmonic voltage in all phases; disconnection of passive filter decreased the 3-rd harmonic voltage in all phases but increased the 5-th harmonic voltages;
• disconnection of both traction network feeder and passive filter at substation N increased the voltage of the 3-rd and 5-th harmonics in all phases.

High harmonic voltages at disconnected traction network feeder testify to the fact that they are formed not only by the currents from the traction network but also by the currents drawn in the feeding network from other nonlinear loads. The decrease in the 3-rd harmonic voltage after disconnection of passive filter at substation Z is the evidence of passive filter malfunction.

The obtained results confirm that the harmonic conditions are complex, unpredictable and require thorough research before modelling them.

Active and reactive powers of the fundamental frequency

Fig. 3 shows the curves of active and reactive powers for phase B for schemes I and II for substation N. They demonstrate a typical character of change in the powers at railway substations. The curves of active and reactive powers at all substations are very similar. When the traction network feeder is connected, the powers are highly variable (Fig.3a). When traction load is disconnected, the curves of powers are ordinary (Fig.3b). The highly variable character of powers at connected traction network feeder occurs as a result of summing up the powers of a large amount of electric locomotives that operate simultaneously. In Fig. 3b a long-term decrease in powers corresponds to the night time. The power curves at connected feeder represent the total powers consumed by electric locomotives and non-traction loads. At the same time the powers of traction loads exceed the power of non-traction loads by 2-4 times.

Fig.3. Variation of P and Q for: a) scheme I, b) scheme II

The 3-rd and 5-th harmonic currents

Daily measurements of the harmonic currents represent time-series of values (Fig.4).

Fig.4. Scatter plot of the 5-rd harmonic current as function of time

The analysis of the measured time-series of harmonic currents shows that they are non-stationary. We present the results of stationarity analysis of the time-series for the 5-th harmonic current at phase A for substation Z as an example. The time-series was divided into 4 equal intervals of 360 elements each. The mean value and variance were calculated for each interval and are presented in Table 2. The data of the Table 2 show that the mean values and variances for each interval differ in value, which gives evidence of non-stationary time-series. Analysis of the measured currents of the 3-rd and 5-th harmonics at the other substations has showed that their time-series are also non-stationary.

Table 2. Mean values and variances

The current curves closely resemble the above given power curves, but have a different shape. In the majority of the studied cases the values of correlation coefficients between harmonic powers and currents are low, which testifies to the weak correlation. However, in some cases there is a noticeable and high correlation, even with an opposite sign. The correlation coefficients between the active and reactive powers and the 3-rd and 5-th harmonic currents for substation N are presented in Table 3 for the sake of illustration.

Table 3. Correlation coefficients between I₃, I₅ and P, Q

In scheme II the correlation coefficient in phase С equals -0.88, which is vividly shown in Fig.5. The 5-th harmonic current decreases with the active power increase and increases with its decrease. The curve of the 3-rd harmonic current changes less sharply than the curve of the 5-th harmonic current. At the same time, we clearly see the sections, where with the increase of active power the current value decreases, and vice versa. The correlation coefficient in phase C between the active power and the 3-rd harmonic currents is equal to -0.32.

Fig.5. Variation of the harmonic currents and the active power

The current waveform is much distorted. It changes with time, but in general, it remains typical despite the great variety. Fig. 6 presents the oscillograms of currents for the connected feeder and disconnected feeder at substation M.

The current waveforms are much less distorted, when the feeder is disconnected.

Fig.6. Current oscillograms for: a) scheme I, b) scheme II

The analysis of harmonic composition of the traction load current shows that the value of the 3-rd harmonic current varies from 25% to 30% of the fundamental frequency current, and the value of the 5-th harmonic current is within the range from 8% to 10% of the fundamental frequency current. Table 4 presents the statistical estimates of the 3-rd and 5-th harmonic currents in one of the phases of each substation. The obtained values are of approximately the same order of magnitude at all the substations

Table 4. Statistical estimates of the 3-rd and 5-th harmonic currents

The curves of powers and currents demonstrate a largely probabilistic character of harmonic behaviour. Fig. 7 presents the histogram of the 3-rd harmonic current in phase А, which is measured at substation N in scheme I. This histogram has one peak.

Fig.7. Histogram of the 3-rd harmonic current

The histogram of the 5-th harmonic current in Fig. 8 has two peaks. The histograms are constructed to get an idea of the distribution function form of the measured currents of the 3-rd and 5-th harmonics. Suitable models describing the probability distribution functions of the measured harmonic parameters are to be yet chosen at a later date.

Fig.8. Histogram of the 5-th harmonic current

The properties of active and reactive components of harmonic currents are of particular interest for constructing models of non-linear loads. The histograms in Fig. 9 present the values of active and reactive components of the 5-th harmonic current. The histogram of the active current components has several faint peaks (Fig.9a). The histogram of the reactive current components (Fig.9b) has two peaks as well as the histogram of the current module in Fig. 8. The histograms of the values of active and reactive current components are the probability density functions of different forms.

Fig.9. Histograms of active a) and reactive b) components of the 5-th harmonic current

Fig. 10 presents the currents of the 3-rd harmonics in phase A of substation N for schemes I and II in the form of scatter plots on a complex plane. The diagrams make it possible to evaluate the phase angles of currents. The distributions of phase angles for 3-rd harmonic in schemes I and III differ very little from each other. The phase angles for the 3-rd harmonic are within the range from 0 to π. The phase angles for the 5-th harmonic are within the range from π/2 to 2π. Disconnection of the traction substation feeder considerably changes the phase angles. The changes take place in the scatter plots and ranges of phase angles. The phase angles for the 3-rd harmonic are distributed within the range from 0 to 2π, and for the 5-th harmonic – within the range from –π/2 to π/2.

Fig.10. Scatter plots of the 3-rd harmonic currents for: а) scheme I, b) scheme II

Analysis of the interrelation between voltages and currents of the 3-rd and 5-th harmonics
The values of harmonic voltages at the points of common coupling are largely determined by the values of currents of loads connected to the node. The influence of harmonic currents passing through the substation transformers on the values of corresponding voltages is assessed by the correlation coefficients in Table 5.

Table 5. Correlation coefficients between I₃, U₃ and I₅,U₅

The correlation coefficients are determined for all the schemes given in Table I. The values of correlation coefficients that correspond to the noticeable and high values are shown in bold. There is a considerable linear relationship between the voltages and currents of the 3-rd and 5-th harmonics at substation Z in scheme II and of the 5-th harmonic at substation N in scheme II. In most of the other cases the relationship is weak, which indicates a strong influence of the harmonic currents of other nonlinear loads on the voltage. The harmonic voltage at the substation arises due to the effect of numerous nonlinear loads connected to the feeder.

Resonance conditions at the 3-rd and 5-th harmonics

The measurements at substation T that were made by metering device demonstrated a sharp increase in the 3-rd and 5-th harmonic voltages. Fig. 11 shows a range of measured 5-th harmonic voltages in which resonance conditions are well seen. A sharp increase in voltage occurs after the 15-th measurement. It turned out that at this moment a 44 MVAr capacitor bank was switched on at the substation of power supply organization, which is located in the area of railway substation T, to maintain the fundamental frequency voltage.

Further analysis and calculations showed that after switching the capacitor bank a resonance loop occurred between the capacitor bank and network at the 3-rd and 5-th harmonics. Before the connection of capacitor bank the input conductance at the network node at the 3-rd harmonic was inductive, whereas after the connection its value decreased almost by 4 times. At the 5-th harmonic the input conductance became capacitive but of a very low value. The capacitor bank compensated the inductive conductance of the network node. The 3-rd and 5-th harmonic conditions are unbalanced. Unbalance of voltages and currents increased after the connection of capacitor bank.

Fig.11. Change in the 5-th harmonic voltages after connection of capacitor bank

Conclusions

At all the substations, where the measurements were taken, the standard limits harmonic voltages were exceeded. The limits of the 3-rd and 5-th harmonic voltages were exceeded most frequently and to a greater degree.

Oscillograms of phase currents essentially differ from sinusoidal form when feeders of traction network are switched on. Currents are considerably unbalanced. Traction load introduces a significant probabilistic component to the harmonic behaviour in the network. The harmonic currents in the network with numerous distributed nonlinear loads are conditioned by the effect of numerous loads.

Currents of the 3-rd and 5-th harmonics represent nonstationary time-series. They are weakly correlated with fundamental frequency active powers in the scheme with connected feeders of traction network. A greater extent of correlation is observed in the schemes with disconnected passive filters. Considerable correlation occurs in the schemes with disconnected feeders of traction network. Strong correlation between harmonic currents and voltages is revealed in the schemes with disconnected feeders of traction network at all substations except for T.

Probability distributions of currents of the 3-rd and 5-th harmonics have single- and double-peaked histograms, whose forms usually differ from the normal distribution. Histograms of active and reactive current components have different forms of probability density functions.

In the general case the phase angles of currents of the 3-rd and 5-th harmonics are within the range from 0 to 2π and change at disconnection of the traction network feeder.

Connection of capacitor bank resulted in resonance conditions at the 3-rd and 5-th harmonics, which increased voltages and currents of the 3-rd and 5-th harmonics.

Acknowledgment: The work was supported by the grant of the Leading Scientific School of the RFSS НШ-1507.2012.8.

REFERENCES

[1] T.C. Shuter, H. T. Vollkommer, T.L. Kirkpatrick, A survey of harmonic levels on the American electric power distribution system, IEEE Trans. on Power Delivery, vol. 4, No. 4, October 1989, 2204-2213.
[2] A.E. Emanuel, J.A.Orr, D.Cyganski, E.M.Gulachenski, A survey of harmonic voltages and currents at distribution substantions, IEEE Trans. on Power Delivery, vol. 6, No. 4, October 1991, 1883-1890.
[3] A.E. Emanuel, J.A.Orr, D.Cyganski, E.M.Gulachenski, A survey of harmonic voltages and currents at the customer’s bus, IEEE Trans. on Power Delivery, vol. 8, No. 1, January 1993, 411- 421.
[4] Y.J. Wang, L. Pierrat, L. Wang, Summation of harmonic currents produced by AC/DC static power converter with randomly fluctuating loads, IEEE Trans. on Power Delivery, vol. 9, No. 2, April 1994, 1129-1135.
[5] A. Mansoor, W. M. Grady, A. H. Chowdhury, M. J. Samotyj, An investigation of harmonics attenuation and diversity among distributed single-phase power electronic loads, IEEE Trans. on Power Delivery, vol. 10, No. 1, January 1995, 467-473.
[6] A.Cavallini, G.C.Montanari, M.Cacciari, Stochastic evaluation of harmonics at network buses, IEEE Trans. on Power Delivery, vol. 10, No. 3, July 1995, 1606-1613.
[7] Chung-Hsing Hu, Chi-Jui Wu, Shih-Shong Yen, Yu-Wu Chen, Bor-An Wu, Jan-San Hwang, Survey of harmonic voltage and current at distribution substation in Northern Taiwan, IEEE Trans. on Power Delivery, vol. 12, No. 3, July 1997, 1275-1284.
[8] A.Cavallini, R.Langella, A. Tesla, F. Ruggiero, Gaussian modeling of harmonic vectors in power systems, 8th International conference on Harmonics and Quality of Power, Athens, Greece, 14-16th Oct.1998, Proccedings, vol.II, 1010-1017.
[9] Probabilistic Aspects Task Force of Harmonics working group Subcommittee of the Transmission and Distribution committee, Time-varying harmonics: part I – characterizing measured data, IEEE Trans. on Power Delivery, vol. 13, No. 3, July 1998, 938-944.
[10] I.M.Neidawi, A.E.Emanuel, D.J.Pileggi, M.J.Corridori, R.D. Archambeault, Harmonics trend in NE USA: a preliminary survey, IEEE Trans. on Power Delivery, vol. 14, No. 4, October 1999, 1488-1494.
[11] A. Ardito, S. Malgarotti, A. Prudenzi, A survey of power quality aspects at industrial customers in Italy, 17th International conference on Electricity Distribution, Barcelona, Spain, 12-15 May, 2003, Proccedings, vol.II, 1010-1017.
[12] State standard R 54149-2010. Electric energy. Electromagnetic compatibility of technical equipment. Power quality limits in the public power supply systems. Moskva. Standartinform. 2012.

Authors: Ph. D. Lidiia Kovernikova, Energy System Institute SB RAS, 130, Lermontov Str., Irkutsk, 664033, Russian Federation, E-mail: kovernikova@isem.sei.irk.ru.

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 89 NR 11/2013

The Coreless Superconducting Fault Current Limiter 15 kV 140 A

Published by Michał MAJKA, Janusz KOZAK, Electrotechnical Institute

Abstract. The superconducting fault current limiter (SFCL) is a device allowing for a more effective use of the existing power network infrastructure. The limitation of short-circuit currents by the SFCL to safe levels will result in the network elements being susceptible to smaller electrodynamic and thermal overloads. This paper presents the electrical scheme, design and numerical model of the 15 kV class SFCL prototype.

Streszczenie. Nadprzewodnikowy ogranicznik prądu zwarciowego (NOPZ) jest urządzeniem pozwalającego na lepsze wykorzystanie istniejącej infrastruktury sieciowej. Ograniczenie przez prądów zwarciowych do bezpiecznego poziomu sprawi, że elementy sieci będą narażone na mniejsze przeciążenia cieplne i elektrodynamiczne. W artykule przedstawiono schemat elektryczny, projekt i model numeryczny prototypu ogranicznika na napięcie 15 kV. (Bezrdzeniowy nadprzewodnikowy ogranicznik prądu 15 kV 140 A).

Keywords: superconductivity, superconducting fault current limiter, SFCL, numerical analysis.
Słowa kluczowe: nadprzewodnictwo, nadprzewodnikowy ogranicznik prądu, SFCL, analiza numeryczna.

Introduction

The superconducting fault current limiter (SFCL) introduces minimal impedance to the power system under normal conditions and high resistance during faults, limiting short circuit current. The main duty of the SFCL is decreasing the fault current to safe level and avoid network instability. The electrodynamic forces occurring during the course of a fault current may damage the devices of the electric power system, such as transformers, generators or busbars in switching stations, within tens of milliseconds. Every such failure of an electric power network entails expensive and time-consuming repairs. Therefore, it is vital that the network’s operation be secured with a reliable protection system. A rapid increase of the resistance of a superconductor on crossing the current critical value I_c makes it possible to build reliable superconducting fault current limiters (SFCLs). SFCLs react very rapidly by limiting the first, the most dangerous, surge current during a current fault condition, thus protecting the devices of the electric power network from the dynamic effects of current faults. The SFCL responds before the first cycle peak and provides an effective means to limit excessive fault currents to safe levels without the disadvantages of conventional fault current mitigation methods.

The SFCLs provide an economic solution for protecting the devices of the electric power system against excessive short circuit currents in case of faults. The application of a SFCL leads to an increase of the allowable short-circuit power at the point of connection of new power generating sources, which is determined by the short-circuit parameters of the power network. This, in turn, will result in an increase of the capability of the power network for connecting distributed generation energy sources based on renewable energy sources. Present researches on current limiters focus on resistive type [5], inductive type [1], [2], [4], [6], [7] and limiters with a saturated core [10]. The drawback of the concept of inductive limiter with shielded core was insertion of a finite impedance in the line even during normal operation, and the large size and weight of the iron core [9]. The presented solution of a coreless construction reduces the weight of the device and the size of the primary copper winding and the voltage on the limiter during the normal operation is negligible [4].

The design of the SFCL

A design of a 1-phase inductive type superconducting limiter is presented in Figures 1-4. The limiter was designed to work in a 15 kV power system. Its main parameters are presented in table 1.

Fig. 1. Design of the SFCL (six identical units connected in series)

A three-winding superconducting current limiter has two primary windings and one secondary winding [1], [4]. The primary winding, placed on the outer ring, is made of a copper wire. The second primary winding, placed in the inner ring, is made of a 2G superconducting tape. The third winding is a shorted secondary winding made of a 2G superconducting tape, placed in the inner ring.

Fig. 4. One unit electrical connections of the SFCL

The primary winding made of 2G tape is connected in parallel with the copper primary winding. All three windings are magnetically coupled. The magnetic coupling between the 2G tape windings in the inner ring is greater than the magnetic coupling between the 2G tape winding and the copper winding in the outer ring. The coupling coefficient between primary HTS and secondary HTS windings is 0.97 and between primary copper winding and secondary HTS winding is about 0.52.

Table 1. Parameters of SFCL

The limiter will be placed in a cryostat with an external vacuum insulation and cooled in a liquid nitrogen bath (Fig. 1). The cryostat of the limiter will be made of GFRP (Glass Fiber Reinforced Polymer). It will be fitted with four copper current leads (Fig. 1) to which the primary, both copper and HTS, windings terminals will be connected. This will allow to record the distribution of currents in these windings during short-circuit tests.

The limiter consists of six identical modules connected in series (Fig. 1 – 4), which allows to lower the voltage of the individual windings. There is 2.5 kV per one module. The superconducting tapes will be insulated with 0.025 mm thick polyimide film with a 0.040 mm silicone adhesive during the winding process. Dielectric strength of DuPont Kapton FN polyimide film is 5.9 kV.

Each module consists of two carcasses of different diameters which are made of composite materials reinforced with fibreglass. The copper winding will be wound onto an external bobbin and the superconducting windings on an internal bobbin. In each of the six modules the primary copper winding has 36 turns and is connected in parallel with two primary superconducting windings. The primary superconducting windings have 12 turns each and are connected in series. The secondary superconducting windings consist of two shorted superconducting windings, each with 12 turns. Both the primary and the secondary superconducting windings are wound onto a single bobbin in such a way that their turns are positioned one on top of the other, which provides a very good magnetic coupling between the windings and this, in turn, reduces the voltage during the SFCL’s performance in nominal conditions.

The primary copper winding will be wound using a 3 mm x 6 mm copper wire. The superconducting windings will be wound using the SF12050 superconducting tape with 2 μm silver layer and a resistance of HTS tape 0.104 Ω/m in resistive state at 77.4 K [3]. The primary and secondary superconducting windings are of the same length and have the same number of turns. A Kapton tape will be used to insulate the superconducting windings. Figure 4 represent the connections of the windings of each of the six modules.

Numerical model of SFCL

The numerical model of the limiter was developed in the “Transient Magnetic” FEM-circuit Flux2D software [8]. The geometry of the actual model of the limiter was substituted with a simplified axially symmetric geometry (Fig. 5).

Fig. 5. Simplified geometry of numerical model in Flux2D for all 6 units

Fig. 6. Electric circuit of numerical model of SFCL in Flux2D

The outer circuit of the numerical model is presented in Fig. 6. The thermal issues which occur in the windings of the limiter are included in the user subroutine written in Fortran. According to this procedure, in every step in the calculations the temperature of the limiter’s winding is determined using the energy balance, based on the present value of the current flowing through the limiter’s windings. The energy balance equation takes into account the transition of the heat from the limiter windings to the cooling liquid. After determining the current temperature of the winding, the resistance of the winding is calculated on the basis of experimentally determined R(I,T) relation for the SF12050 superconducting tape [3].

Simulations were performed for model of limiter whose parameters are presented in table 1. Thanks to the performed simulations, courses of a fault current in the circuit with and without the limiter were obtained (Fig. 7), as well as the changes of resistance and temperature of individual limiter windings during the limitation of the fault current (Fig. 9).

In the stand-by state, i.e. the first 40 ms of calculations, the superconducting windings of the limiter are in the superconducting state and a nominal current of 140 A flows through the limiter (Fig. 8). The voltage value in all models of the limiter is lower than 1 V, which results from a minor leakage reactance.

During a short-circuit lasting from 0.040 sec. to 0.200 sec., a fault current flows through the limiter. The peak value of the current in the shorted circuit i_p = 40 kA was limited to 4.7 kA (Fig. 7). The course of the fault current causes the HTS windings to heat up very rapidly. The temperature of the windings increases from an initial temperature of 77.4 K to a maximum temperature T_max which is reached at the moment of switching off of the short-circuit (Fig. 9c).

Fig. 7. Current waveforms in the circuit with and without SFCL

Fig. 8. Current waveforms in the windings of the limiter in stand-by state

Fig. 9. Numerical model – current waveforms in the windings of the limiter (a), the changes of resistance (b) and temperature of individual limiter windings (c) during the limitation of the fault current (graphs for HTS I and HTS II windings overlap).

The performed simulation shows that the temperature of the superconducting windings increases much faster than the temperature of the copper windings, and it reaches different values at the moment of switching off of the short-circuit. In designing the limiter, it was assumed that the maximum temperature of the limiter’s superconducting windings at the moment of a short-circuit occurrence would not exceed 200 K and the fault current peak value would be below 5 kA

Conclusion

The developed design in which the superconducting windings are wound simultaneously onto a single bobbin allows to obtain a very high coupling factor between the windings and minimize the leakage reactance of the limiter, which minimizes the voltage in the limiter in the stand-by state. In case of a 2-winding design in which the primary copper winding is magnetically coupled with a secondary HTS winding, there always occurs leakage reactance, which causes losses in the stand-by state. The use of a connection in parallel of a copper coil and a superconducting coil in the primary winding protects the short circuit from opening in case when the superconducting tape is damaged. The fault current limiting capability of a 3-winding limiter is determined mostly by the impedance of the copper winding coupled in parallel with the primary superconducting winding.

An analysis of the results of the numerical simulations confirmed that it is possible to build an inductive type coreless superconducting fault current limiter that will effectively limit the peak value of the fault current from 40 kA to 5 kA within 160 ms. The number of turns in the primary copper winding and the superconducting tape length in the superconducting windings must be such that the temperature of the HTS windings does not exceed the maximum allowed temperature of the superconducting tape. Due to a substantial increase of the temperature of the limiter’s HTS windings, the short circuit must be switched off by a conventional circuit breaker before the temperature of the HTS winding reaches the maximum value.

This work was supported in part by the National Centre for Research and Development under Grant UMO2012/05/B/ST8/01837.

REFERENCES

[1] Kozak J., Majka M., Janowski T., Kozak S., Wojtasiewicz G., Kondratowicz-Kucewicz B., “Tests and Performance Analysis of Coreless Inductive HTS Fault Current Limiters”, IEEE Trans. Appl. Supercond., 21 (2011), No. 3, 1303 – 1306.
[2] Naeckel O., Noe M., “Design and Test of an Air Coil Superconducting Fault Current Limiter Demonstrator ”, IEEE Trans. Appl. Supercond., 24 (2014), No. 3, article nr 5601605.
[3] Czerwinski D., Jaroszynski L., Majka M., Kozak J., “Analysis of Alternating Overcurrent Response of 2G HTS Tape for SFCL”, IEEE Trans. Appl. Supercond., 24 (2014), No. 3, 5600104
[4] Kozak J., Majka M., Kozak S., Janowski T., „Design and Tests of Coreless Inductive Superconducting Fault Current Limiter”, IEEE Trans. Appl. Supercond., 22 (2012), No. 3, 5601804
[5] Kozak J., Majka M., Kozak S., Janowski T., „Comparison of Inductive and Resistive SFCL”, IEEE Trans. Appl. Supercond., 23 (2013), No. 3, 5600604
[6] Heydari H., Sharifi R., “Three-Dimensional Pareto-Optimal Design of Inductive Superconducting Fault Current Limiters”, IEEE Trans. Appl. Supercond., 20 (2010), No. 5, 2301 – 2311
[7] Naeckel O., Noe M., “Conceptual Design Study of an Air Coil Fault Current Limiter ”, IEEE Trans. Appl. Supercond., 23 (2013), No. 3, article nr 5602404.
[8] de Sousa W.T.B., Nackel O., Noe M., “Transient Simulations of an Air-Coil SFCL”, IEEE Trans. Appl. Supercond., 24 (2014), article nr 5601807.
[9] Janowski T, Wojtasiewicz G., Kondratowicz-Kucewicz B., Kozak S., Kozak J., Majka M., „Superconducting Winding for Inductive Type SFCL Made of HTS Tape With Increased Resistivity”, IEEE Trans. Appl. Supercond., vol. 19, issue: 3, pp. 1884 – 1887, 2009.
[10] Hong H., Su B., Niu G.J., Cui J.B., B. Tian, Q., Wang L.Z., Wang Z.H., Zhang K., Xin Y., “Design, Fabrication, and Operation of the Cryogenic System for a 220 kV/300 MVA Saturated Iron-Core Superconducting Fault Current Limiter”, IEEE Trans. Appl. Supercond., vol. 24, issue: 5, article nr 9002204, 2014.

Authors: dr inż. Michał Majka, Instytut Elektrotechniki, ul. Pożaryskiego 28, 04-703 Warszawa, E-mail: m.majka@iel.waw.pl; dr hab. inż. Janusz Kozak, Instytut Elektrotechniki, ul. Pożaryskiego 28, 04-703 Warszawa, E-mail: j.kozaki@iel.waw.pl

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY, ISSN 0033-2097, R. 92 NR 7/2016. doi:10.15199/48.2016.07.06

IEEE Standard 519-2022 Update and Changes From Prior Version (-2014)

Published by Mirus International Inc.

APPLICATION NOTE IEEE STD 519 – 2022 UPDATE_HARMONIC STANDARD

IEEE Standard for Harmonic Control in Electric Power Systems (IEEE Std 519-2022), is a highly recognized and referenced standard, and has been updated in 2022 from its prior 2014 version. The purpose of the standard is to establish goals for the design of electrical systems that include both linear and nonlinear loads².

Please note that Mirus is not responsible for any misinterpretation of the standard, and the following notes are Mirus’ interpretation only. The more significant changes are shown in bold:

• Title Change: ‘Recommended Practice and Requirements’ changed to ‘Standard’. Although the title has removed the word ‘recommended’, the fact is that the use of IEEE Std 519 still remains wholly voluntary. Several other verbage simplifications have been made throughout the document as effort to simplify and help avoid misinterpretation of the standard.

• The introduction mentions that users should not add equipment that affects the impedance characteristic in such a way that voltage distortions are increased. The word ‘passive’ has now been removed from this description, so now we can interpret that ALL equipment should be considered to meet this criteria.

• Scope added to include Inverter-Based Resource (IBR) and/or Distributed Energy Resource (DER) installations which direct the User to IEEE Std 1547³ or IEEE Std 2800⁴ for current distortion limits if the combined site rated generation is >=10% of the annual average load demand. A Decision Tree has been provided as Figure 1 in the standard to determine whether IEEE Std 519 current limits apply at the PCC.

Figure 1: Decision tree for applying current distortion limits at PCC²

• Additional note in the scope that the limits given in this document are justifiable only at the PCC, and not intended to be used for the evaluation of equipment. This was always the intent of the standard, but some additional clarification has been included. Although IEEE Std 519 does not specify limits on individual equipment, Mirus accepts that some engineers/specifications may choose to apply specify limits on certain non-linear loads. This strategy can be used to ensure compliance at the PCC provided there is no system resonance conditions at the residual harmonics and/or all nonlinear loads contributing to the distortion at the PCC have been considered.

• Additional clarity provided for the ‘maximum demand load current’ to clarify how it is calculated by the max 15 or 30min demand, and what to do if 12 months of data is not available. If not available, then it shall be based on the projected 15- or 30-min demand. Previously it was not clear what should be done for new installations or when 12 months of data was not available.

• Harmonic measurements for IEEE 519 evaluation are required to be made up to the 50^thusing Class A instruments. Class S would only require measurement up to the 40^thharmonic. This clarification may help those looking to use or purchase a power analyzer.

• Even-order harmonic current limits have been relaxed. Previously all even-order harmonics were limited to 25% of the individual harmonic limits as provided in the current limit tables. Now the current harmonic limits have been increased for even harmonics.

For h<=6 -> Limits have increased from 25% to 50% of the individual limits shown in the tables.

For h>6 -> Limits have increased from 25% to 100% of the individual limits shown in the tables.

Even-order harmonics are not desirable on power systems, and their appearance can indicate asymmetry between the positive and negative halves of the waveform. Perhaps it was felt that these increases were required due to the larger numbers of IBR, AFE drive and active filters, which may result in some levels of even order harmonics that may not have always been meeting the more stringent even-order harmonic limits in the previous 2014 standard.

• Annex A. Significant changes to the Interharmonic Voltage Limits and Rationale. The standard mentions that more detail can be provided by referencing a paper by Drapela, et al [B3] “Issues and Challenges Related to Interharmonic Distortion Limits”, 2020. ⁵

Interharmonics can be difficult to address when dealing with active devices such as IBR, AFE Drives and active filters because they can often generate interharmonics. Passive filters typically treat harmonics over a frequency range so both harmonic and interharmonic frequencies are usually addressed.

The recommended limits (not prescriptive) are based on analysis of the sensitivity of lighting equipment (LEDs) to prevent flicker, and also more strict limits required on sensitive non-lighting equipment. Appropriate limits should be based on specific needs (local, country, conditions, etc.).

References

1. SOLV^TM Software Registration and Download, https://www.mirusinternational.com/register.php
2. IEEE Std 519-2022, IEEE Standard for Harmonic Control in Electric Power Systems.
3. IEEE Std 1547-2018, IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces.
4. IEEE Std 2800-2022, IEEE Standard for Interconnection and Interoperability of Inverter-Based Resources (IBR) Interconnecting with Associated Transmission Electric Power Systems.
5. Drapela, J., M. Halpin, R. Langella, J. Meyer, D. Mueller, H. Sharma, A. Testa, N. Watson and D. Zech, “Issues and Challenges Related to Interharmonic Distortion Limits,” 2020 19th International Conference on Harmonics and Quality of Power (ICHQP), Dubai, United Arab Emirates, 2020, pp. 1–6, doi: 10.1109/ICHQP46026.2020.9177933.

Mirus International Inc. [2022-09-22] 1-888-TO MIRUS www.mirusinternational.com MIRUS-AN003-A

IEEE 519-1992 Compliance

Published by Electrotek Concepts, Inc., PQSoft Case Study: IEEE 519 Compliance, Document ID: PQS0321, Date: October 10, 2003.

Abstract: IEEE Standard 519-1992 is a standard that addresses the need for limiting the harmonic current a customer injects onto the utility system. It also protects the customer by specifying maximum harmonic voltage distortion levels that utilities can supply.

The installation of a 3% choke on each drive and a 150kVAr harmonic filter on the bus reduced the harmonic currents to acceptable values.

This case presents the evaluation of IEEE 519 compliance for an industrial facility supplying adjustable-speed drives.

PROBLEM STATEMENT

A wastewater treatment plant is installing five (5) 100 HP pulse width modulation (PWM) adjustable-speed drives (ASDs). The utility has specified that IEEE 519 current limits must be met.

The combined drive load has the following characteristics:

Drive Rating: 500 HP
Bus Voltage: 480 V
Fundamental Current: 600 A

SYSTEM CONFIGUATION

Figure 1 illustrates the oneline diagram used for the analysis of IEEE 519 compliance.

Figure 1 – Oneline Diagram for IEEE 519 Case

Short circuit and load information:

I_SC(short circuit) = 26750 Amps @ 480 Volts
I_L(maximum average demand load current) = 1200 Amps
I_SC/I_L (low side PCC) = 22

IEEE 519 CURRENT LIMITS

IEEE Standard 519-1992 is a standard that addresses the need for limiting the harmonic current a customer injects onto the utility system. It also protects the customer by specifying maximum harmonic voltage distortion levels that utilities can supply. The standard should be used for guidance in the design of power systems with nonlinear loads. Table 1 summarizes the current requirements.

Table 1 – Current Limits for Individual Customers (120V – 69kV)

where:

SCR: ratio of the short circuit current at the point of common coupling to the maximum average demand load current (I_sc/I_load)

TDD: Total Demand Distortion, current distortion in percent of the maximum average demand load current

IEEE 519 defines the point of common coupling (PCC) as: A point of metering or any point as long as both the utility and the customer can either access the point for direct measurement of the harmonic indices meaningful to both or estimate the harmonic indices at a point of interference (POI) through mutually agreeable methods.

LOAD CURRENT EVALUATION

Table 2 summarizes the current requirements for the initial case with no harmonic current reduction. This evaluation illustrates the need for harmonic current mitigation. As can be seen in the table, most of the individual harmonic currents and the total demand distortion are exceeded for this case. In addition, the bus voltage distortion of 9.8% is higher than the generally accepted limit of 5%.

The transformer impedance can be determined from:

Z_tx(Ω) = (kV² / MVA) * Z_tx(%) = (0.480² / 1.5) * 6% = 0.0092Ω

The harmonic voltages are determined by multiplying the current injection times the impedance at each harmonic. The maximum average demand load current is used to scale the individual harmonic currents for comparison with the specified limits.

Table 2 – Evaluation of Current Limits for Base Case

Note: I_d is based on the average maximum demand load current

MITIGATION TECHNIQUES

Several techniques for reducing the harmonic current were evaluated:

1. Installation of a 3% choke on each drive (refer to Figure 1 for new current waveform)
2. Installation of a 3% choke and a 30kVAr, 5^th harmonic filter on each drive
3. Installation of a 3% choke on each drive and a 150kVAr, 5^th harmonic filter on the 480 volt bus

Table 3 summarizes the results from the computer simulations.

Table 3 – Evaluation of Current Limits for Various Solutions

SUMMARY

IEEE Standard 519-1992 is a standard that addresses the need for limiting the harmonic current a customer injects onto the utility system. It also protects the customer by specifying maximum harmonic voltage distortion levels that utilities can supply.
The installation of a 3% choke on each drive and a 150kVAr harmonic filter on the bus reduced the harmonic currents to acceptable values.

REFERENCES

IEEE Recommended Practice for Electric Power Distribution for Industrial Plants (IEEE Red Book, Std 141-1986), October 1986, IEEE, ISBN: 0471856878
IEEE Recommended Practice for Industrial and Commercial Power Systems Analysis (IEEE Brown Book, Std 399-1990), December 1990, IEEE, ISBN: 1559370440

RELATED STANDARDS
IEEE Standard 519-1992

GLOSSARY AND ACRONYMS
ASD: Adjustable-Speed Drive
DPF: Displacement Power Factor
PCC: Point of Common Coupling
PF: Power Factor
PWM: Pulse Width Modulation
POI: Point of Interference
SCR: Short Circuit Ration
TDD: Total Demand Distortion
THD: Total Harmonic Distortion
TPF: True Power Factor

For the updated IEEE Standard 519-2022: IEEE Standard 519-2022 Update and Changes From Prior Version (-2014)

PD Pulse Burst Characteristics in Differently Aged Transformer Oils under AC Conditions

Published by JUNHAO LI¹, YANMING LI¹, GUOLI WANG², Martin D. JUDD³,
State Key Laboratory of Electrical Insulation and Power Equipment, Xi’an Jiaotong University (1)
Electric Power Research Institute of China South Grid (2)
Institute for Energy and Environment, Department of Electronic and Electrical Engineering, University of Strathclyde (3)

Abstract. Partial discharge (PD) detection is a technique widely used for high voltage equipment insulation condition assessment. In such applications, an understanding of PD mechanisms, characteristics and development processes is important. In this paper, PD characteristics of three typical transformer oils, aged to different degrees, are examined under AC conditions, using a needle-to-plane electrode system. The PD activity in transformer oil is confirmed as appearing in pulse burst form. Pulse burst characteristics recorded in the study show that the transferred electric charge per PD pulse burst increases with the degree of aging. This increase is accompanied by an increase in the number of discrete PD pulses and by a decrease in cavity formation time. The experimental results of cavity discharge in differently aged transformer oils are discussed and their characteristics are explained in terms of the charge exchange mechanism.

Streszczenie. Przedstawiono charakterystyki wyładowania niezupełnego w trzech typowych olejach transformatorowych o rożnym stopniu starzenia przy zasilaniu napięciem przemiennym. Stwierdzono, że impulsy wyładowania zależą od stopnia starzenia. (Charakterystyki impulsów wyładowania niezupełnego w oleju transformatorowym o różnym stopniu starzenia)

Keywords: partial discharges; oil insulation; pre-breakdown; PD pulse burst; power transformer.
Słowa kluczowe: wyładowanie niezupełne, transformatory mocy, izolacja olejowa.

1. Introduction

Transformers are key and widespread components of the power network. Mineral insulating oil plays a major role in power transformers, acting both as insulation and coolant. The presence of even a minor defect in the insulation structure, under normal operating voltages can create local field enhancement causing partial discharges. The defect may be a protrusion or asperity point on transformer metalwork or winding which was introduced during manufacture/maintenance. Gas cavities can be formed within the oil phase at this defect and cause partial discharge (PD) until their eventual collapse due to either dynamic instability or the diminishment of the sustaining electrical stress enhancement [1~6].

Electrically induced cavities in dielectric liquids are caused by localized injection of current pulses at a high field region where electron avalanches or streamers may develop within the liquid phase. Kattan et al [7~8] demonstrated that most of the electrical energy injected (about 90%) is converted into heat which evaporates the liquid. Pompili, R. Bartnikas et al [9-16] studied the PD pulse burst characteristics in transformer oils with different viscosities and found that PD which occurred in cavities within the liquid appeared in the form of pulse bursts that consist of a series of discrete current pulses. The time interval between discrete pulses can be as low as just a few nanoseconds and is a function of the cavity formation, growth and collapse time in liquid. The first pulse of pulse burst may exhibit a greater magnitude than the second and represents charge injection from the electrode, while subsequent pulses represent PD pulse activity within the expanding cavity.

An oil/paper structure is the typical configuration of transformer insulation and it undergoes long term aging due to gradual physical and chemical degradation subjected to electrical and thermal stress in-service. The decomposed product for insulation aging is solid, liquid and gaseous impurity species such as carbon, water, CO, CO2 and furan products, etc [17]. These impurities will alter the PD pulse burst characteristics in oil. The purpose of the study reported here was to investigate the effect of the degree of aging on the pulse burst characteristics in order to better understand the processes in terms of their effects on PD measurement and implications for insulation diagnostics in electrical plant such as power transformers.

2. Characteristics of the oil specimens

PD characteristics were examined in three differently aged transformer oils using a needle-to-plane electrode system. The first specimen was unused oil; the second was medium aged oil from an in-service 110/35kV, 20MVA transformer and the third was severely aged oil from a 220/110kV, 120MVA transformer which was close to end of life. All oil samples are Karamay 25# transformer oil and from same manufacturer. Their viscosity at 40°C was 13cSt and density at 20°C was about 850kg/m³. The physical and electrical properties of these oil samples are summarized in Table 1. The photo of samples is shown in figure 1.

Table 1 Electrical property of the oil specimens

Property	Unused oil	Medium aged oil	Severely aged oil
AC breakdown voltage, kV (at 2.5mm)	42	39	30
Tan δ at 90°C, 50Hz	0.0025	0.005	0.0184
Neutralization value, mg KOH/g	0.018	0.023	0.039
Water content, ppm	21.4	24.8	32.8
Furan products, ppm	0	0.02	0.25

As is shown in figure 1, with the oil aging degree increase, the color of oil is deeper and deeper.

3. Test system and procedure

The experimental setup used for the study is shown in Figure 2. A needle-to-plane electrode system was used to create a field enhancement site in order that PD was generated. The radius of needle was 40μm and the needle-to-plane gap was 30mm.

AC voltage was generated using a test transformer rated at 50Hz, 10kVA, 0-100kV. The PD pulse bursts were recorded using a 500MHz bandwidth digital oscilloscope, having a sampling rate of 2.5 Gsamples/s. The PD measurement impedance was a wideband resistor, the step response time of which was below 3 ns.

Phase resolved PD (PRPD) patterns represent each PD pulse as a point in a charge-phase diagram and are a well established tool for interpretation of PD activity. In order to study the influence of aging degree on the PRPD measurement, PRPD patterns were acquired using a PD detector that is able to record the complete PD pulse shape as well as its magnitude and phase.

Fig. 2. Circuit of the experimental apparatus

The pulse burst parameters were measured as a function of applied voltage, with the applied voltage increased in steps of 2kVrms and maintaining a voltage constant for about 10 min before continuing to the next step. The PD inception voltage (PDIV) was determined when one or more pulse bursts first appear. The same procedure was continued above the PDIV level by recording the pulse burst activity over 10min. Subsequently, the pulse burst statistical parameters such as the average number of discrete pulses, duration of each burst, time interval between first and second discrete pulses within the PD pulse burst and the maximum amplitude within a pulse burst were computed for each voltage step. PD in oil is a stochastic phenomenon and in order to obtain a statistical law of the parameters, the numbers of PD pulse burst used to compute are above 50 in every applied voltage step.

Results and analysis

PD activity in a needle-to-plane oil-filled structure is concentrated around the peaks of the applied sinusoidal voltage waveform. The duration of a PD burst is much shorter than the power frequency cycle (20ms) so that the externally applied voltage can be regarded as a constant during a pulse burst [11]. The focus of study in this paper is the relatively stable negative PD pulses, since the positive PD pulses are substantially more irregular and erratic.

The PD inception voltage in all three specimens was 20kV. Figure 3 portrays some typical PD pulse burst behavior in unused oil at different voltages. The number and the maximum magnitude of discrete pulses within PD pulse bursts are seen to increase with the applied voltage. The first pulse in the PD pulse burst represents initial charge injection and is larger than the second discrete pulse caused by cavity discharge. A typical PD pulse burst waveform in severely aged oil at 30kV is shown on the same timescale in Figure 4.

Fig. 3. PD pulse bursts in unused oil at (a) 20kV, (b) 22kV, (c) 26kV applied voltage

Fig. 4. Typical PD pulse burst in severely aged oil at 30kV

Differently aged oils show different characteristics with increase of the applied voltage. Figure 5 shows the characteristic variation in number of discrete pulses per PD pulse burst as a function of applied voltage. All three specimens exhibit an increase in the number of pulses but for the severely aged specimen, the increase is more pronounced than for the other two, especially with applied voltage increase.

Fig. 5. Number of pulse within burst as function of applied voltage

Variation in duration time of the pulse burst is delineated in Figure 6. The duration reduces slightly with applied voltage in unused oil and medium aged oil while a sharp reduction with the increase of applied voltage is observed in severely aged oil. This phenomenon can be explained in terms of charge exchange between cavity wall and impurity species. The ions from the cavity discharge are trapped at the cavity interface by electrostatic forces but can be removed through charge exchange with impurity species in the process of cavity growth. The impurity species will diffuse to high electric field point under electric force and make the charge exchange effect more energetic.

Fig. 6. Duration time of per burst as function of applied voltage

The impurity species (especially furans) show a marked increase with degree of oil aging according to Table 1. This will cause the rate of charge de-trapping from the interface between the cavity wall and liquid to increase as well. For this reason, the discharge frequency could be expected to increase with the concentration of impurity species. More frequent discharges in the cavity will contribute to its electrohydrodynamic instability and reduce its lifetime in the oil phase. Hence the duration of pulse bursts decreases with the concentration of impurity species, especially at higher applied voltages. It also can be seen from Figure 6 that the burst duration reduces by about 0.5μs between 20kV and 30kV in the unused and medium aged oil samples but by about 2μs over the same range for the severely aged oil.

The first large discrete pulse represents the initial charge injection to liquid phase, and then the time to form the cavity that supports the subsequent partial discharge process can be determined by computing the time interval between first pulse and second pulse within the pulse burst. Figure 7 portrays the variation in the time interval between the first and second pulse within the pulse burst. The cavity formation time is found to range from 70ns to 350ns. The three test specimens show the same tendency for cavity formation time to decrease with applied voltage. The formation times reported previously in the literature [9] are somewhat longer, ranging from approximately 100ns to 700ns. The one likely reason is the condition of the oil specimens. The samples in literature [9] were new transformer oils. However, the specimens in this study are aged transformer oils which contain impurities. At a given voltage, cavity formation time generally decreases with the degree of aging of the oil sample. The other likely reason is the electric field distribution in the gap, especially in the close vicinity of the needle. The electric field, besides being modified by the charge injected into the liquid at the needle tip, is a function of the electrodes configuration. The needle tip radius is different with literature [9] may lead to such different in unused oil specimen.

Fig. 7. Time interval between first and second discrete pulses within the PD pulse burst as a function of applied voltage

Figure 8 shows the average maximum magnitude of discrete pulses within the burst for all three specimens (The dispersion of maximum magnitude of discrete pulses within the burst is larger, so, the figure 8 and figure 9 only shows the average values of all measured dates). The maximum magnitude increases with the applied voltage in all cases and the magnitude difference is not exceptionally different as a function of ageing. However, we can estimate the approximate transfer charge per PD pulse burst using the formula q = ∫idt. Figure 9 presents the plots of the average charge transferred per PD pulse burst as a function of the applied voltage and reveals a marked increase in the average charge transfer per burst with the applied voltage. Especially for severely aged oils, the charge transfer shows an abrupt increase when applied voltage exceeds 24kV. Comparing Figures 5 and 8 it can be seen that the difference of maximum magnitude of discrete pulse within PD pulse burst is minimal but the number of discrete pulses per PD pulse burst changes significantly across the oil samples. Charge transferred per PD pulse burst increases with oil aging.

Fig. 8. The average maximum amplitude within a pulse burst as function of voltage

Fig. 9. Average charges transferred per PD pulse burst

In addition to the typical phenomena discussed so far, some abnormal PD bursts were also observed during testing, and some results in unused oils were more similar to those of Pompili [9] than the results presented above.

For example, a single pulse often appears in the oil, which may be due to charge injection that fails to generate a cavity and initiate cavity discharge.

In some other instances the growth of the cavity may undergo substantial fluctuation as it continues to discharge as evidenced by the irregular amplitude of the discrete pulses shown in Figure 10. The cavity may divide and create more micro-cavities in the expansion phase, resulting in this phenomenon.

The first large pulse within a PD burst usually corresponds to the charge injection in liquid phase, but in some instances a cluster of extremely small pulses appears at the beginning of the pulse burst, as shown in Figure 11. In such cases it is difficult to identify the “first” pulse. These very small pulses may represent a multi-charge injection process in which it is the total injection energy leads to cavity formation.

Fig. 10. PD pulse burst in unused oil at 24kV

Fig. 11. PD pulse burst in unused oil at 28kV

Fig. 12. PD pulse burst in unused oil at 24kV

Fig. 13. PD pulse burst in severely aged oils at 28kV

There exists another kind of PD pulse burst event that was regularly observed, in which a discrete PD pulse sequence is observed among the successive pulses with a monotonically ascending pulse magnitude, as shown in Figure 12. The interval time between successive pulses increases gradually without a larger initiating pulse. This may be due to cavities having been generated by previous charge injection but which could not sustain PD initially. At a later time, the phase shift of the external AC voltage may lead to re-ignition of PD in one of these “dormant” cavities.

Another instance which was observed in severely aged oil is illustrated in Figure 13. This type of pulse burst may owe to the higher concentration of impurity species in severely aged oil that result in more electrical discharge activity and shorter cavity lifetime. Impurities move to the electrode and increase the electric field aberration on a microscopic scale. Charge injection processes occur simultaneously and initiate separate cavities, thereby leading to relatively independent two-cluster PD pulse burst. This phenomenon was observed only in severely aged oils.

5. Conclusions

The PD pulse burst phenomenon in a 30mm long point-to-plane gap in differently aged transformer oils have been measured under AC conditions. For the three tested transformer oils, the cavity formation times ranged from 80ns to 350ns, compared to the average duration time of per burst of 0.7μs to 3μs. All oils tested exhibited a considerable increase in PD activity with applied voltage. The number of discrete pulses per PD pulse burst ranged from 3 to 25 and increased with the applied voltage and degree of aging. The level of average charge transferred per PD bursts was found to vary from 25pC to 250pC and have the same trend as with the number of discrete pulses per burst.

Impurity species such as moisture, furans, etc., which increase in concentration as the oil ages will tend to increase the number of discrete PD pulses per pulse burst (and hence the charge transferred per pulse burst), while reducing the duration time of pulse bursts and the cavity formation time.

Acknowledgement: The authors appreciate the supported by “The foundamental research funds for the central universities”. The authors appreciate the supported by the “National Natural Science Foundation of China (number: 50977075)”. The authors appreciate the supported by the National Engineering Laboratory for Ultra High Voltage Engineering Technology（Kunming、Guangzhou）.The authors would like to thank Shannxi electric power corporation for supply the oil samples.

REFERENCES

[1] P. K. Watson, W. G. Chadband and M. Sadeghzadhe-Araghi, “The Role of Electrostatic and Hydrodynamic Forces in the Negative-point Breakdown of Liquid Dielectrics”, IEEE Trans Elec. Ins., Vol. 26, no. 4,pp.543-559, 1991.
[2] P. K. Watson, W. G. Chadband, “The Dyanmics of Pre-Breakdown Cavities in Viscous Silicone Fluids in Negative Point-Plane Gaps”, IEEE Trans Elec. Ins., Vol. 23, no. 4, pp.729-738, 1988.
[3] P. K. Watson, “Electrostatic and Hydrodynamic Effects in the Electrical Breakdown of Liquid Dielectrics”, IEEE Trans Elec. Ins., Vol. 20, no. 2, pp. 395-399. 1985.
[4] A. Larsson, M. Bengtsson, “The Sequence of Events in Mid-Gap Laster-Triggered Electrical Breakdown in Liquid Dielectrics”, IEEE Trans. Dielectr. Electr. Insul., Vol. 9, no. 3, pp. 349-352. 2002.
[5] A. A. EI-Sulaiman, A. S. Ahmed and M. I. Qureshi, “Effects of Field Strength and Filtration on Burst Phenomena in aged Transformer Oil under High Non-uniform DC Field”, IEEE Trans Elec. Ins., Vol. 18, no. 2, pp. 163-166. 1983.
[6] N. Bonifaci, A. Denat, and B. Malraison, “Determination of Charge Mobility in He Gas from Current-Voltage Measurements in Point-Plane Geometry”, IEEE Trans. Industry Applications, Vol. 37. no. 6, pp.1634-1640. 2001.
[7] R. Kattan, A. Denat and O. Lesaint, “Generation, growth, and collapse of vapor bubbles in hydrocarbon liquids under a high divergent electric field”, J. Appl. Phys.,Vol. 66. no. 9, pp.4062- 4066, 1989.
[8] R. Kattan, A. Denat and N. Bonifaci, “Formation of Vapor Bubbles in Non-Polar Liquids Initiated by Current Pulses”, IEEE Trans Elec. Ins., Vol.26, no. 4, pp.656-662,1991.
[9] M. Pompili, C. Mazzetti, and R. Bartnikas, “PD Pulse Burst Characteristics of Transformer Oils”. IEEE Trans on .Power. Delivery., Vol. 21, no, 2, pp. 689-698, 2006
[10] M. Pompili, C. Mazzetti, and R. Bartnikas,. “Partial Discharge Pulse Sequence Patterns and Cavity Development Times in Transformer Oils under ac Conditions”. IEEE Trans. Dielectr. Electr. Insul., Vol. 12, no, 2, pp. 395-403, 2005
[11] M. Pompili, C. Mazzetti, and R. Bartnikas, “Phase relationship of PD pulses in dielectrics liquids under ac conditions,” IEEE Trans. Dielectr. Electr. Insul., Vol. 7, no, 1, pp. 113-117, 2002
[12]. M. Pompili, C. Mazzetti, and R. Bartnikas, “Early stages of negative PD development in dielectric liquids,” IEEE Trans. Dielect. Electr. Insul.,Vol. 2, no. 4, pp. 602-613, 1995.
[13]E.O. Forster in Engineering Dielectrics, Vol. III, Electrical Insulating Liquids, R. Bartnikas, Editor, Monograph 2, ASTM, Philadelphia/West Conshohocken, 1994
[14] M.Pompili, C.Mazzetti, and R. Bartnikas, “Comparative PD pulse burst characteristics of transformer type natural and synthetic ester fluids and mineral oils”. IEEE Trans. Dielect. Electr. Insul.,Vol. 16, no. 6, pp. 1511-1517, 2009.
[15] M. Pompili, C. Mazzetti, and R. Bartnikas, “Simultaneous ultrawide and narrowband detection of PD pulses in dielectric liquids”, IEEE Transaction on Dielectric and Electrical Insulation, Vol. 5, pp.402-407, 1998.
[16] M. Pompili, “Partial discharge development and detection in dielectric liquids”, IEEE Transaction on Dielectrics and Electrical Insulation, Vol.16, pp. 1648-1654, 2009.
[17] A.M.Emsley, X.Xiao, R.J.Heywood and M.Ali, “Degradation of cellulosic insulation in power transformers. Part 3: Effects of oxygen and water on aging in oil”, IEE pro-Sci, Meas. Technol. Vol, 147, no.3. pp.115-119. 2000.

Authors: Dr Li Junhao, High voltage department, xi’an jiaotong university China PR, E-mail: xjtuhvljh@gmail.com; Prof Li Yanming, High voltage department, xi’an jiaotong university China PR, E-mail: ymli@mail.xjtu.edu.cn.

Source & Publisher Item Identifier: PRZEGLĄD ELEKTROTECHNICZNY (Electrical Review), ISSN 0033-2097, R. 87 NR 6/2011