An Energy Central Expert Community Intelligent Utility

Article Post

OGE's three-tiered architecture aids data analysis

Big Data is inundating businesses as well as social interactions, flowing from and into these sources, so utilities are in good company when it comes to making sense of it or extracting business value from it.

We've run a few columns on the topic and on utility strategy around it, which you might find useful. Here are some links: "Data Analytics: Start 'In the Middle'," "Data Analytics for Smarter Grids," "Harnessing Big Data for Business Value" and "Data Analysis: What's It Telling Us?"

I looked for further insights during last week's webcast, "Deep Data: Navigating from Raw Data to Enterprise Information Management," by the Utility Analytics Institute, a sister enterprise to Intelligent Utility. (For the slide deck and audio portion of that webcast, click on the title.)

According to Mike Smith, vice president, UAI, we're in the early years of the smart grid "value phase" in which utilities leverage data for value. Thus, utility spending on analytics is slated to grow rapidly from the low- to mid-two hundred million dollar range annually last year to well over a billion dollars within five years.

UAI surveys show that money and talent are the top hurdles in this quest. 

Brian Eakin, lead forecast analyst in the strategy and business planning group at OGE Energy Corp., said that OGE has nearly 800,000 customers over 30,000 square miles of Oklahoma and Arkansas and it operates nine fossil fuel plants and nearly 800 megawatts of wind capacity.

The main driver for analytics now, Eakin said, is to avoid building additional fossil fuel generation until at least 2020. That will mean leveraging demand response for commercial and residential customers, further investments in renewable energy (largely wind), distributed generation and geothermal heat pumps, while reducing peak demand by 500 MW, of which 70 MW is slated for this year.

A snapshot of Big Data at OGE would include 52 million meter reads per day, which eventually will more than double, and an integrated operations center, which is forecast to receive up to two million event messages per day from wide area networks, AMI, meter alarms and outage management systems. The installation of a distribution management system (DMS) also will send mega-streams of data into the integrated operations center. An outage management system and integrated volt/VAR control program adds yet more data to the mix.

As Eakin explained it, OGE has a three-tiered information architecture that includes a performance-based warehouse, expanded data integration and new analytical and presentation capabilities. The data warehouse is designed to expand to hundreds of terabytes of data by 2014, so headroom obviously is desired for these systems. Data integration will incorporate real-time messaging in addition to traditional extract-transform-load, batch-based processes. Outputs are configured so that all analysts see a "single version of the truth," a "master data information web service" that solves the dilemma of analysts using separate systems with variations of "the truth." 

When it comes to data visualization, Eakin said, it's more effective to project data on a map rather than to read a table or graph, thus geospatial-enabled presentations. Finally, a culture of collaboration at OGE has different business units comparing notes on data analytics successes and challenges.

"Governance of analytics is an important thing here at OGE," Eakin noted. "We let people know what tools should be used, how to use them and what's already been done, so we can prevent people from reinventing the wheel."

A quick look at outcomes: OGE's meter interval data, for instance, allows demand response measurement and verification, it allows dynamic segmentation methodologies; and it provides new research on load. Load forecasting allows for revenue forecasting and, therefore, more accurate budgeting. The integrated operations center provides insights into asset performance and condition-based maintenance rather than modeling.

Roque Marinho, director of business enterprise intelligence for the Cobb Electric Membership Cooperative, a large co-op outside Atlanta with about 460,000 electric and natural gas accounts, has a handy definition: "Business intelligence is information that promotes action. The data needs to tell you something."

Cobb EMC is finishing smart meter deployment in its service territory and Marinho pointed out how the effective presentation of data from an advanced metering infrastructure (AMI) project can pinpoint problems with any single household for resolution.

Here you have a glance at how a couple leading utilities are plunging into data analytics. The webcast in particular has some insightful questions and answers, so replay it at your convenience.

You'll hear leading practitioners, your peers and others explain the nuances around unlocking value from data at the Utility Analytics Institute's first summit, taking place next week, Feb. 15-16, in Orlando, Florida. 

Phil Carson
Intelligent Utility Daily