Lightening strikes a transmission line. A breaker opens. The fault clears within a few cycles. The line is re-energized. No customers lose power. Repeat this series of events 25 times in a month normally void of anything but sunshine. It is certainly not a typical month for the Salt River Project (SRP, Phoenix, Arizona, US). But, the atypical happens, and this would constitute a really bad month in terms of power quality, especially for certain voltage-sag-sensitive customers.

Interestingly enough, this unusual string of events does not impact the traditional reliability indices such as the System Average Interruption Duration Index (SAIDI) and the System Average Interruption Frequency Index (SAIFI). So, how do we explain sudden customer dissatisfaction when corporate reliability metrics indicate good system performance? Do we need additional metrics?

SRP found reliability indices to be a great tool for measuring distribution system performance, but they do not reflect the “pain” customers experience as a result of voltage sags. Taking into consideration that EPRI estimates the annual cost of voltage sags in the United States to be $24 billion — nearly three times more than that of outages — SRP realized something critical was missing. The utility needed an additional performance indicator to capture the impact of sags on customers.

A RESPONSE

SAIDI and SAIFI are relatively simple measures of the average duration and frequency of outages. Depending on your corporate philosophy and the dictates of your regulators, an outage is defined as zero voltage in excess of 1 or 5 minutes. SAIDI is a good indicator of a utility's ability to dispatch crews and repair the system to restore power after an outage. It also reflects the redundancy or automation designed into the grid for the purpose of quick outage recovery. SAIFI reflects a menagerie of performance drivers including the grid's resilience, performance of protection systems, weather, animals, age of components and anything else that might contribute to causing or preventing faults.

Metrics for classifying voltage sags are two-dimensional and thus more complex. Every sag has a magnitude and a duration component, both of which contribute to the pain inflicted on sensitive industrial processes. Therefore, in 2000, SRP began capturing a new metric referred to as sag energy, which is the undelivered energy, or the power the customer should have gotten, but didn't as a result of the sag.

Sag energy is used to measure the significance of a single voltage disturbance, and it can be summed over a period of time such as a month. Tracking it over months will expose performance trends. Most importantly, compounded customer aggravation is an unofficial, unscientific metric. SRP also tracks the number of sags occurring each month, or sag count, as a performance metric.

Power-quality indices are reported monthly to SRP executive management. The reports include a system average Sag Energy Index (SEI) and Sag Count Index (SCI), as well as the same indices delineated for a statistically valid sample of about 80 individual substations. These metrics identify sag performance trends and potential problems at specific 69/12-kV substations or the lines connecting them. More importantly, they send up red flags that customers served by those substations may need some proactive customer care, or at least some explanations.

VALIDATION

One of SRP's earliest validations of the SEI was found in the results of its ongoing customer satisfaction survey. During a two-month period, SRP saw a statistically significant drop in customer satisfaction, followed by a complete rebound during the next survey cycle. The first response was to review the distribution reliability indices, all of which looked good. Surprisingly, SRP found that the SEI and SCI for the corresponding time period showed significant voltage sag activity. Further research showed that an uncommon (for winter in Arizona) series of lightning storms had rattled the grid without causing sustained outages. These disturbances did not count in the SAIFI and SAIDI calculations. SRP reliability indices missed this series of events, but the utility's customers certainly did not.

Another learning experience occurred two years ago, when the power-quality indices indicated severe sag activity across the entire service territory. Large customers were complaining of high production losses. Yet, SRP's reliability indices were showing no signs of excessive breaker activity. Follow-up study showed that the sags were the result of faults at interconnected independent power plants. As a result, SRP and its neighboring independent power plants started meeting regularly to resolve repeating problems and to share best practices for extra-high-voltage maintenance.

The value of tracking power-quality indices is just unfolding. As the utility gains experience in correlating anomalies in SEI and SCI to real grid events, new and valuable information is emerging. In fact, data that once was used only by the power-quality department is now available on the intranet to engineering groups throughout SRP. System protection, transmission operations, transmission planning, line maintenance and customer service all use SEI.

Power-quality metrics will never replace the traditional reliability indices, but rather will supplement them. Power-quality metrics measure something different: customer pain. And after all, you have to be able to measure it to manage it and, ultimately, to satisfy your customers.

---

John Blevins, manager of Power Quality Services at Salt River Project, has 21 years of experience in energy, power delivery and retail services. Blevins oversees a department that provides power-quality consulting and investigation services to SRP customers, and participates in corporate projects and initiatives to improve quality on the power-delivery system. He holds a BSME degree from the University of Kentucky and a technology MBA degree from Arizona State University. He is a registered professional engineer in Arizona, a certified energy manager and a member of IEEE. John.Blevins@srpnet.com

PERMANENT SUBSTATION MONITORS

SRP operates a system of 80 power-quality monitors installed on the 12-kV bus within 69/12-kV substations. This equates to a sample of about 25% of SRP's distribution substations, with at least one monitor on each of SRP's 69-kV subtransmission loops. The result is a statistically valid sample of the grid's performance with a 90% confidence in the average SEI.

All substation monitors are permanently installed high-resolution monitors manufactured by Schneider Electric (Rueil-Malmaison, France). Though most of the monitors installed over the past 10 years are ION 7700 models, SRP now installs the newer ION 7650 meters. The majority of the meters have high-speed Ethernet connections, accommodating the vast quantity of data transfer associated with high-resolution oscillography.

The system also includes multiple servers and communication systems that allow data to be served to users over the web, provide paging to notify SRP staff and customers of power disturbances, and produce reports and, of course, the power-quality indices.