Part 7 Valuing Reliability

Published by

• John D. Kueck and Brendan J. Kirby, Oak Ridge National Laboratory
• Philip N. Overholt, U.S. Department of Energy
• Lawrence C. Markel, Sentech, Inc.

Published in Measurement Practices for Reliability and Power Quality: A Toolkit of Reliability Measurement Practices, 2004

Prepared by Oak Ridge National Laboratory Oak Ridge, Tennessee 37831-6285 managed by UT-BATTELLE, LLC for the U.S. Department of Energy under contract DE-AC05-00OR22725

---

Although reliability metrics do have their shortcomings, as discussed in Section 6, they are useful for trending when applied consistently with an unchanging set of calculation procedures. For example, Pacific Power has developed several performance standards that are on file with the applicable state commissions. Failure to meet these standards results in financial penalties for the utility. Four of these standards are as follows:

1. In each state, annual SAIDI will be improved by 10% between 2003 and 2005.
2. In each state, annual SAIFI will be improved by 10% between 2003 and 2005.
3. In each state, annual MAIFI will be improved by 5% between 2003 and 2005.
4. In each state, the five worst-performing distribution circuits will be improved by 20% over a 2-year period. Five new circuits will be selected in each state each year for a 5-year period.
   In addition to these metrics, which are part of the agreement with the public utility commissions, Pacific Power also has customer guarantees, again tied to financial penalties if the utility fails to deliver. Three of these are summarized as follows:⁵

1. After an interruption, power will be restored within 24 hours, barring damage due to extreme weather. If this condition is not met, the residential customer receives $50, and the commercial customer, $100.
2. Power will be switched on within 24 hours of the customer’s request or the customer receives $50.
3. Customers will be notified two days prior to a planned interruption; if not, each customer receives $50.

These standards of reliability are part of a program to establish a high level of customer service as part of a fundamental business philosophy. The guarantees provide a concrete example of the value of reliability.

The worst-performing distribution circuits are chosen for upgrade by Pacific Power using the customer hours interrupted, or the numerator of SAIDI.⁶ Experienced judgment can then be used to implement known improvements based on reliable rules of thumb and without extensive analysis.

On the other hand, some utilities such as Commonwealth Edison in Chicago are using large-scale reliability modeling to analyze circuits and choose optimal improvements based on cost and benefit.

After a series of major distribution outages in Commonwealth Edison territory in 1999, ComEd launched a comprehensive investigation looking at equipment, design, personnel, and operations. The corrective actions included substation and feeder inspections, installation of new feeders, feeder upgrades, substation expansions, building of new substations, and the implementation of a new maintenance program. ComEd also developed a predictive reliability model consisting of more than 3,300 feeders. The model provided an intelligent system to automatically identify potential reliability problems and to recommend reliability improvement projects based on expected benefits and costs.⁷ A reliability assessment model quantifies reliability characteristics based on system topology and component reliability data. The model identifies areas of inherently good or poor reliability, and also identifies overloaded and undersized equipment that degrade system reliability. Some typical improvements that a predictive reliability model can explore include

• load transfers between feeders,
• building of new substations and substation expansions,
• addition of line reclosers,
• sectionalizing switches,
• adding new feeder tie points,
• automating feeders,
• undergrounding of circuits, and
• replacement of aging equipment.

The ComEd model uses a simulation that assesses each contingency, determines the impact, and weights the impact by the contingency’s probability of occurrence. A sample reliability assessment is shown in Figure 1. Areas with relatively low reliability are shaded in red. Potential problem areas can be quickly identified. If a red area is adjacent to a blue area, it may be desirable to transfer some customers through reconfiguration to improve the reliability of the transferred customers and to help equalize the reliability of the two areas.

Figure 1. Reliability assessment results. Components are shaded based on expected annual outage hours, a primary driver of SAIDI. Source: Ref. 7.

Computer-generated reliability improvements can be evaluated and different approaches can be compared from a cost-benefit perspective. Interestingly, the cost-effectiveness of reliability improvement projects varies widely from area to area. After the reliability model was completed, an intelligent system was used to automatically identify potential reliability problems and recommend reliability improvement projects based on benefits and costs. Figure 2 shows that the highest ranked project for the Northeast region is more than three times as cost-effective as the highest ranked project in the Southern region. The cost benefit varies widely, different types of projects tend to be more effective for different regions, and the best allocation of money will require flexibility in both the types of projects that are funded and the level of funding for each geographic region. This graph shows that the most cost-effective reliability gains can be made in the Northeast.

Figure 2. Plot of recommendations vs reliability score. Source: Ref. 7.

Some of the recommendations for reliability improvement projects included the following:

• Transfer path upgrade: A transfer path is an alternate path to serve load after a fault occurs. If a transfer path is constrained due to small conductor sizes, reconductoring may be cost effective.
• New tie points: A tie point is a normally open switch that allows a feeder to be connected to an adjacent feeder. Adding new tie points increases the number of possible transfer paths.
• Increased line sectionalizing: Increased line sectionalizing is accomplished by placing normally closed switching devices on a feeder. Adding switches improves reliability by allowing more flexibility during post fault system reconfiguration.
• Feeder automation: Adding SCADA-controlled switches on feeders will allow automated post-fault system reconfiguration.

The entire ComEd distribution modeling effort required less than one year. The results found that the most cost-effective approaches to improving reliability are not always obvious and can vary by feeder and region.⁷

Evaluating reliability improvements can thus require a range of techniques from detailed probability modeling of the entire system to using proven rules of thumb to plan upgrades. Interestingly, the utility discussed here that is simply using tried-and-true methods to plan improvements is meeting its reliability standards every year and posting the results on its website.

Conversely, the utility that developed the comprehensive reliability analysis model and the intelligent system to automatically identify potential problems and recommend improvements, made the news with three major outages in one summer. The subsequent effort modeled, calibrated, and assessed the reliability of more than 3300 feeders.

References

5. Pacific Power, “Customer Service Commitments: Annual Report,” May 2003, available at http://www.pacific-power.com/File/File28086.pdf.
6. Dennis Hansen, A Methodology for Maintaining and Improving Reliability, IEEE 0-7803-7285-9/01, Institute of Electrical and Electronic Engineers, Piscataway, N.J., 2001.
7. Richard E. Brown, Distribution Reliability Modeling at Commonwealth Edison, IEEE 0-7803-7285-9/01, Piscataway, N.J., 2001.