Article Post

Grid data and analytics

Electricity, Electric Equipment, Transformer

As utilities complete the roll out of smart grid technologies, including smart meters and other intelligent distribution network devices, they are finding tangible examples of how data and analytics can support their grid operations. In this article, you’ll meet two analytics professionals to learn about their work to leverage data and analytics from their smart grid deployments. These stories come from presentations at our Utility Analytics Week conference in September and are told by:

  • Derl Rhoades, Southern Company / Alabama Power
  • Jonathan Pettit, Oncor Electric Delivery

You can also download a white paper -- essentially the outtake version of this article -- that features 11 pages of commentary by Rhoades, Pettit and several other utility analytics professionals, by clicking on this link.

Alabama Power: Outage analytics for distribution operations

Alabama Power has 4.4 million retail customers, of which the company has automated metering for 4.3 million. “Our smart meter projects are pretty much done,” said Rhoades. “We do meter alerts for values other than billing. We do voltage. We do amps, phase angles and momentary outages.”

Rhoades said that 10 or 15 years ago, many people in distribution didn’t see much value in AMI, but today, most of them can’t live without it, especially on the outage management side. He said just 10 percent of customers actually call with an outage when they have one.

However, when a utility has an AMI system integrated with an outage system, many people call every day. “Our system basically assimilated a phone call,” he said. “When a meter comes in, it goes in as if the customer called, so it goes into the system. We can predict the outages. The beauty about it is, now we have so many people calling that the outage prediction model is perfect.”

Then when customers call in to restore and the meters report that they’re restored, the Alabama Power system automatically takes them off the outage list. “If the meter was reported out or is predicted out, it’s in the list,” Rhoades said. “So it comes back, and the meters start coming off the list, and if a meter does not report back in, the system automatically pings the meter and determines if the outage is cleared or not. If the outage is not cleared, naturally operations would roll somebody there. We can do it all in a matter of 10 minutes versus waiting until somebody makes a phone call. It has improved our outage management system and restoration times. It was one of the first things we did.”

Rhoades said that to make it all work, a utility has to have its interfaces right, and it has to monitor those interfaces. At Southern Company, everything runs on a bus system, so it is essential to make sure the bus systems are operational, “Because once you start depending on all these outages coming in from your meters, you need to make sure that systems are up and operational. It’s taken us a while to get all the right monitoring points to make sure they’re flowing properly,” Rhoades said, “But we got all that done.”

Alabama Power had some problems with the ping response time from its OMS system that reduced the success rate below an acceptable level, but setting the priority messages correctly fixed the problem. “Everything works fairly well, and it’s a good project,” he said. “It really helps our distribution people in managing outages.”

Oncor: Theft detection and outages

Oncor has 175,000 smart meters left to install to reach its goal of 3.2 million installed. Jonathan Pettit, advanced meter system (AMS) manager at Oncor Electric Delivery, said the key at Oncor is AMS for a variety of reasons, including analytics and automation.

“Currently, we have about 250,000 automated operations a day for which, in the past, we would have had to roll a truck. To date, the automation has saved between 4 million and 4.5 million truck rolls, along with the associated fuel and man hours.”

Analytics also help Oncor protect its revenue from loss through theft. The protection involves the use of manual queries. Pettit said that although the queries are automated, Oncor doesn’t kick out service orders automatically. He said the utility wants to reach the point of 90 to 95 percent assurance of true theft or revenue protection. “The processes have an average 80 percent hit rate in a range from 60 to 90 percent,” Pettit said. “The one that’s 90 percent is about ready to be automated. There are many ways to steal electricity, and it’s amazing what people will do. You’ll find that theft is probably twice as high as you ever thought it was, and people are twice as smart as you thought they were, too.”

According to Pettit, there is an urban myth that meters give false indications of outages. “They actually are telling you there’s something wrong with your secondary,” he said. “Even our meter vendor hadn’t thought about this. But in 100 percent of the cases where we had a false indication, we had a loose terminal or a problem with the transformer. With predictive analytics, you can learn that there’s something wrong out there. So we have a team that does nothing but go around fixing these potential problems before they occur.”

Pettit described the integration various systems, including OMS and DMS, as a “big find,” but something the utility should roll out in bits and pieces. However, Oncor’s distribution operation center said, “This is great!” and all of a sudden, Oncor turned on all of the integration. “They’ve been happy,” he said. “And now it’s as though they can’t live without it.”

Pettit said that utilities will likely find that most OMS systems are not ready to deal with a high quantity of messages. Oncor had an outage that affected a half-million people. He said the systems are not scaled for millions of outage messages. “You have to find a way to filter, modify or otherwise limit the amount of messages,” he said.

The future of analytics

Regardless of the project or challenges, utilities have their eyes open for the next killer app in analytics.

Rhoades said as Alabama Power integrates its AMI and SCADA data, it would like to improve light feeders. “Why is one feeder two percent more inefficient than the next feeder when it should be just like it with the same miles of line?” he asked. “That’s something in the future we’re looking at.”

Pettit said much of the future for data analytics in Oncor’s projects is unknown. He said although the smart grid and smart meter programs have been going on for a long time, they represent the biggest change in the electric utility industry in 100 years.

H. Christine Richards is the director of knowledge services for the Utility Analytics Institute.


No discussions yet. Start a discussion below.