Article Post

BGE: analytics pilot and lessons learned

Orange customer service

In a pre-conference workshop at Utility Analytics Week, Thomas Humphreys, Baltimore Gas & Electric's project manager for advanced meter capabilities, described how the investor-owned utility tackled the prospect of using data analytics. 

In his introduction, Rick Brakken, CEO of DataRaker, BGE's vendor for its pilot analytics program, said that BGE had been well-organized for a pilot evaluation in that the utility leadership saw data analytics "as an outcome, not an afterthought" of its advanced metering infrastructure (AMI) initiative.

I found Humphreys' presentation the most interesting of five made on Tuesday in part because we'd followed the winding path BGE followed to get regulatory approval to proceed with its AMI initiative. 

You may recall that Maryland regulators insisted that the utility establish clear customer value for an AMI project that would depend on 50 percent funding under the American Recovery and Reinvestment Act. One outcome: BGE customers would not be charged for the project until they'd received a smart meter and, once a new meter was installed, they would gain access to their energy use data via an online tool.

(We covered the bumpy regulatory process, which at the time drew widespread industry attention, in a series of columns: "Maryland PSC Linked Cost Recovery to Consumer Benefits," "Readers Applaud Maryland Commission" and "BGE Wins Smart Meter Approval, With Conditions.") 

Humphreys' presentation proved to be something of a blow-by-blow description of the process BGE undertook to move a data analytics concept to a pilot project with well-defined outcomes. 

BGE serves 2,300 square miles of central Maryland, with 1.2 million electric customers, Humphreys' told those assembled in Arlington, Texas, for the Utility Analytics Week conference sponsored by the Utility Analytics Institute, a division of Energy Central, parent of Intelligent Utility. 

Beginning this spring, BGE has managed to install about 90,000 interval meters, with an ambitious schedule to finish by 2014. (That will require a rate of 20,000 meters installed each week from now until then.) January 2012 marked the installation of a meter data management system (MDM) and a new "customer care billing system" that replaces a 20-year old customer information system (CIS). 

Next month the utility will introduce a "smart energy management program," which will provide Web-based access to energy use information for customers who've received their interval meter. Meter pinging for outages and remote connect/disconnect will soon go live and, in July 2013, customers can opt for peak time rebates, according to Humphreys. 

While the AMI effort moved forward, Humphreys' department set up a pilot program to help BGE business units define "value cases" (i.e., use cases) for how business operations and practices could be improved by using data analytics from advanced meter data. BGE has proceeded with a pilot project focused on the use cases with the "highest value potential," he said. 

Humphreys' project team will monitor the performance of analytics in achieving value during the pilot but, he noted, "we really want the business [unit] to take ownership of their own analytic capability and, ultimately, define a business case for long-term [analytics] deployment." 

"We want each business unit to show enough value, enough savings, to absorb the shared cost of [data analytics] implementation," Humphreys said. 

Several steps in the process were valuable, according to Humphreys. 

A new vice president of smart grid technologies took an interest and extended an invitation to the utility's various business units to develop use cases for data analytics. A consultant (Accenture) helped the business units develop and define value from data to improve business operations. DataRaker, eventually chosen to help implement the BGE pilot study, provided a case study focused on Florida Power & Light that seemed to parallel BGE's circumstances and goals.

Then, Humphreys' team held intensive meetings with business units to understand their "biggest needs and frustrations," including existing data limitations. The project team catalogued the emerging use cases, which included, for instance, simple reports of 24/7 load profiles of a particular customer, aggregation of those load profiles at the transformer level, voltage profiles for a feeder, revenue protection actions, the list grew. 

"We started to get more ideas just by sitting down and having long, focused discussions with the business units," Humphreys told the Utility Analytics Week audience. 

The project team winnowed the ideas to those that could be put into action quickly, noted the use cases that would require tweaking the available tools or developing new sources of data and listed longer-term objectives that would not be implemented until after the pilot had run its course.

"Value themes" included revenue protection, AMI operations (particularly security), distribution system operations and planning, voltage management, engineering standards, targeting customers for energy efficiency programs, meter to cash efficiencies, Humphreys said.

The BGE team arrived at the training stage just last week, using actual BGE data with the vendor's tools. That produced some "energetic responses" from the business units as they recognized the possibilities. One example: an output that caused the utility's revenue protection chief to initiate an immediate action item for a field crew. 

Now that the pilot is underway, Humphrey's team will hold monthly forums for each business unit in the pilot to see what's working, what's not and determine the real business value of the analytics that have been tested. What costs can be taken out of business processes? What new revenue might be produced? What would a full-scale, future deployment look like? 

"If we can get ten, fifteen business units to show the benefits of full deployment, this should be an easy sell," Humphreys said. "From what I've heard so far, I'm optimistic that we'll build a positive business case." 

According to the pilot's timeline, conclusions around a business case for analytics should be possible by the end of the year.

BGE's lessons learned appeared to be several-fold. Data analytics should be viewed as a planned outcome, not an afterthought. A pilot study to test early wins that could validate the vendor and technology while establishing the use case made sense. Executive sponsorship to raise the project's profile and articulate the potential value is crucial. And fostering business unit ownership to develop the use case and enthusiasm for outcomes has paid dividends.

More to come; stay tuned.

Phil Carson


No discussions yet. Start a discussion below.