Powering Strong Communities

Finding Value in Utility Data

Utilities have an abundance of data available to them, yet few are unlocking the full value of this data through analysis that can help utilities forge stronger relationships with customers, enhance reliability, foster efficiency, and more. As public power utilities find ways to tap into a wealth of data to make informed decisions – or help customers do so – they must also think critically about how such decisions can alter energy efficiency and demand.

Look Outside the Average

More than half of public power providers in the United States have advanced metering infrastructure that can deliver data in hourly intervals or even smaller timeframes, according to researchers from Guidehouse.

That kind of consumption data – segmented by the hour or more frequent intervals – is powerful in the hands of someone like Pasi Miettinen, CEO of Sagewell, a software company and consultancy specializing in analytics for utilities. Among other things, his company helps utilities create data-driven programs that target specific types of customers for the greatest possible program adoption and utility gain.

“Program design without looking at meter data first is almost always a mistake,” Miettinen said. “The reason for that is that utilities are used to looking at data in averages, such as the average load shape. When you look at individual customers, they are astoundingly volatile. They don’t look like the averages at all.”

Miettinen pointed to a utility his firm works with that discovered 20% of customers made up almost 50% of the coincident peak consumption. These customers were concentrated geographically and were not the biggest, most expensive residential properties. Sagewell analytics identified a group of homes in the 1,200-square-foot range that had an 11-kilowatt coincident peak on average, with very little consumption the rest of the year. “They just happened to use their air conditioning systems and other large end uses a lot during the coincident peak,” he said.

This finding reveals the power of customer segmentation and target marketing. “If you assume that all customers are alike and you design programs based on an average customer, you may overlook an enormous amount of low-hanging fruit,” Miettinen said.

Along with finding the customers who use the most power, utilities can use analytics to disaggregate consumption data and identify large loads within a household that could be schedulable, such as electric vehicle chargers and pool pumps. This is valuable because having those loads operate in off-peak hours is less disruptive for consumers and more likely to be adopted.

Utilities also can use meter data to plan marketing programs that support the utility now and in the future. A case in point is the peak-season flip many southern utilities have seen in the past 20 years. When Miettinen started his career as a load forecaster at a southern utility more than two decades ago, the utility’s peak occurred in the summer, and it was due to air conditioning load. Now, many utilities in the region are winter-peaking.

“Heat pumps installed in the South turned out to be very inefficient for winter heating,” he said. When temperatures drop low enough, those inefficient heat pumps switch to auxiliary resistance heating coils, a big power drain. Utilities in the Northeast, where Miettinen lives now, tend to have performance requirements for heat pumps, and the one in his Boston home is good to minus-17 degrees.

"The equipment that utilities encourage sometimes has significant impacts on the load shapes of the future,” he said, adding that small, ductless systems – used primarily for air conditioning – don’t replace much fossil fuel consumption in the winter. Alternatively, high-efficiency, whole-home heat pump systems are what he calls customer- and utility-friendly. “They’re cost-effective for the customer to operate and also generate good margins for the utility, which lowers the cost of electricity for everyone,” Miettinen said.

“A key insight from data analysis is that there are technologies that are beneficial for both customers and utilities,” he said. “Finding the intersection of the two is good program design.”

Knowledge Brings Savings

For big power users like university campuses or municipal facilities, data can reveal issues with power use and facilitate conservation. In New York, state facilities are under executive orders from the governor’s office to cut back energy use and meet specific efficiency targets. Fortunately, organizations with big loads can learn more about energy consumption at their facilities through the New York Energy Manager, a web-based service offered by the New York Power Authority, the nation’s largest state public power organization.

The Energy Manager, based in Albany, consolidates data from more than 20 different sources, including meters, submeters, and building management systems, said Emilie Bolduc, NYPA’s vice president of distributed energy resources. With the building management systems, arrays of sensors collect data about interior building conditions like humidity, temperature, and carbon dioxide levels.

All data get analyzed and come back to facility operators in the form of information-rich reporting that shows consumption down to metered appliance levels, usage trends, savings and operational recommendations, and the estimated savings that can be attained from adopting those recommendations. Currently, the Energy Manager platform includes more than 16,000 facilities and has collected some 600 million data points.

“Our customers use NY Energy Manager to identify operational savings,” Bolduc said. “It’s also become a compliance tool because the system demonstrates how much usage was reduced over time.”

At one state university, the system revealed HVAC issues. “They asked us to submeter the air handling units in a recreation center because they were having humidity and odor problems,” Bolduc said. “The building was also consuming way more energy than initially designed.”

The NY Energy Manager team found that mechanical dampers within the unit were not opening to their design settings, causing a ventilation imbalance. “That was one of the problems identified, which saved the university a significant amount of money once the economizers were fixed,” she said. Economizers combine return and outdoor air to cool and ventilate buildings.

The university also learned from NY Energy Manager reports that air handling units were running full force all day even though the building was not occupied 24/7. Once facility managers set up a schedule for ramping the units up and down, they saved an average of 168 kilowatt-hours per day per unit.

Because Energy Manager also consolidates consumption from other utilities – water, wastewater, and gas – Bolduc’s team has been able to help with savings in those areas, too. Using Energy Manager, a bus depot operator found a water leak in a bus-cleaning station. “The operations lead told us they would not have noticed the leak until they got their bill, and it would have been around $6,000 by the time it came,” Bolduc said.

Knocking Opportunity

The value data offers when it comes to targeting or helping customers is matched in the operational arena, where utilities are using data to enhance system performance and reliability. Still, there is some low-hanging fruit that remains untouched by many, said Tao Hong, professor of Systems Engineering and Engineering Management at the University of North Carolina in Charlotte and director of BigDEAL, the Big Data Energy Analytics Laboratory.

Hong is an expert on system forecasting and renewables integration. Asked the key to balancing power supply and demand with intermittent generation, he said: “First, you must understand your demand.”

Doing that effectively requires accurate data, Hong said.

A Guidehouse consulting team estimated that 55% percent of public power utilities have advanced metering infrastructure installed and can access frequent-interval consumption data, which may be recorded every 15 minutes or even every 5 minutes. Even if utilities do not have AMI, most (80% to 90%) have SCADA (supervisory control and data acquisition) all the way down to distribution substations, so they could at least be looking at system data on an area-wide basis, Hong said.

“Most utilities do not use that data for load profiles,” Hong said. “Many are using old meter data that delivers one number per month.” But SCADA data could provide frequent-interval visibility all the way down to distribution system feeders. “That’s way better than what they’re using today.”

Using this type of data for load profiling could help utilities with renewables integration, where system planners must understand demand to have the right supply available for balancing renewable intermittency. Right now, Hong said, many utilities do not have sufficiently accurate load profiles. They also do not know when someone installs rooftop solar because many systems are sold and managed by third parties, and the utility is not involved.

Another area where Hong believes utilities are missing out is in outage prevention. If utilities correlate past outage data with the weather, Hong estimates utility staff could probably figure out the cause for 30% to 50% of outages, then take preventive measures. For instance, simply correlating vegetation-related outages with tree-trimming schedules could have significant impact. The eReliability Tracker, from the American Public Power Association, allows subscribers to track outage causes, but utilities would need to analyze this data against other information to take proactive measures.

So could analysis of voltage and current changes that happen right before and after a fault. Seeing this incipient data – the power surges that accompany lightning strikes – may reveal places where lightning arresters would be of value. These devices give the electricity a path to ground for the over-voltages and power surges that may accompany a lightning strike and cause a fault on the system.

Although Hong sees opportunities that are being missed, he also knows why. “Smaller utilities cannot afford the people who can do this analytics work,” he said. “A town might pay $120,000 in compensation for a mid- or senior-level manager, but that’s not enough for a data scientist.”

Luckily, Hong sees another opportunity public power utilities could embrace. “Ten or 20 utilities could work together and share resources,” he said. “They could hire some data scientists to work for all the utilities together. That’s a business model they might want to try.”