Expanding Metering Data’s Value

By Guerry Waters, Oracle Utilities

Smart grid investments have inspired both regulatory concern and negative public reactions in recent months. As a result, many utilities are delaying smart metering rollouts and spinning out pilot programs in hopes of clarifying and quantifying benefits.

Abnormal meter reads may indicate poor installation. BI applied to meter data can detect the problem while installation crews are still in the field.

Vendors are now introducing tools that may help. Prepackaged, productized business intelligence (BI) and engineered hardware/software BI appliances are pointing the way for utilities toward easier and faster quantifications of smart metering’s value.

Most utilities today have some experience with BI. Those experiences vary, however, not only among utilities but also within utility departments. Some utilities have developed sophisticated geographic information system (GIS) tools with which to explore spatially-oriented operations like field service or outage management. Others have integrated utility-oriented BI applications into several utility operations. Still others have implemented enterprisewide BI.

The one common denominator of these systems is their high cost. The largest systems have generally been built by outside consultants and other high-cost experts who must be retained whenever the utility requires changes. Internal experts may have developed tools and reports that only they know how to run. Staff may spend considerable time reworking and maintaining integrations and data extractions. These ongoing expenditures and risks are not acceptable for a utility attempting to deliver electricity at the lowest, most reasonable cost.

Fortunately, concurrent with the rise in data volume from smart metering and the advent of meter data management, vendors have begun to introduce new approaches to BI that can dramatically reduce long-term BI costs while also helping utilities maximize the value of their data. They include productized BI, Bi appliances and unstructured analytics.

The application-specific tools of productized BI frequently link to a larger enterprise- or utility-specific BI engine application. For the utility currently focused solely on extracting value from meter data, however, it is generally not necessary to implement the underlying engine for anything other than the immediate application.

Productized BI generally includes:

  • Extractors and schema designed to move application-specific data into a data warehouse.
  • Pre-built dashboards—generated by the BI engine from the extracted and formatted data—designed to answer the most pressing and frequent staff questions and to identify specific issues that require further investigation and possible correction.

While some utility-specific tailoring is generally required, productized BI takes most of the work out of BI system implementation. Utilities are generally able to get the products up and running within a few days rather than the months or years required for typical BI implementations of the past.

With BI appliances, hardware devices with pre-installed BI software reduce the cost and time of a BI implementation. These applications can handle extremely large data volumes (including those typical of interval meter data) and speed up the analytic process—often by using technologies like NoSQL databases, in-memory analytics, R, or Hadoop.

Traditional BI, with data formatted into a schema, provides outstanding results when you know what questions to ask and the answers to expect. Unstructured analytics has a limitation, however: It may not be helpful when you don’t know the questions to ask in advance.

A lack of right questions clearly characterizes the field of interval consumption analysis today. Data will have relevance for projects currently only in the planning stages, such as integrating large amounts of distributed renewable resources and building microgrids. Until utilities are farther along in those projects, however, they will not have the clarity to define all the answers they need. Querying data before it goes into a structured format will help to address those issues. All the above BI tools help utilities maximize the value of the huge volumes of meter data made available by smart metering. They spring from two facts about data analysis:

  • Pattern identification is a direct road to problem solution.
  • The more data a utility can handle quickly, the faster a utility can discern the pattern.

Uses for these patterns are virtually endless and may have immediate bottom-line effects on tampering, AMI performance, supply costs and overbuilding.

Estimates of the cost of meter tampering vary widely. U.S. utilities may lose as much as $6 billion annually; worldwide losses are clearly far greater. Tamper alerts can help. They are far more effective when backed by BI that detects unusual patterns of low usage, however. BI that permits staff to drill into device and customer histories helps field investigators.

In addition, utilities can lose both revenue and customer satisfaction when faulty meters or head-ends result in incorrect bills. BI applied to only a few weeks of meter data, however, can identify the percent of normal readings received on time. Drilling down to compare the late or abnormal intervals by head-end, device, location or supplier can provide a fast track to the source of the problem.

Utilities almost invariably generate or contract for significantly more electricity than customers actually consume. BI can reduce that overage by helping utilities:

  • Refine consumption by degree-day. This permits utilities to vary supply based on short- and long-term temperature predictions.
  • Compare interval data from residential customers with and without electric vehicles or from homes before and after the purchase of an electric vehicle. This helps utilities anticipate demand growth as electric vehicles come online.

Spatial BI also helps pinpoint the location and timing of potential grid overloads and bottlenecks, helping utilities better identify the size and location of needed remedies. This can keep a utility from overbuilding

BI holds real smart grid promise for utilities.

MORE INSIGHT AT http://power-grid.com

Find more information about the subject of this article online by going to the website and typing “business intelligence” into the search engine. You’ll find:

  • Contract details for Bermuda Electric Light Co.’s October choice of BI partner,
  • Berg Insights’ discussion of BI and smart meters worldwide,
  • Pike Research data on Bi and the upcoming “data tsunami,”
  • And more.

Visit us online at http://power-grid.com for all the details.

Guerry Waters is vice president of industry strategy at Oracle Utilities.

More PowerGrid International Issue Articles
View Power Generation Articles on PennEnergy.com
Previous articleRisky Business—How Power Contracts Allocate Uncertainty
Next articleHow Does the Industry Create True Interoperability?

No posts to display