PPL Electric Utilities reduces revenue losses with AMI


By Bernie Molchany, Michele Pierzga and Jackie Lemmerhirt

Advanced metering infrastructure (AMI) can be a powerful tool to combat energy loss by allowing examination and analysis of more frequent and detailed data. PPL Electric Utilities uses AMI in conjunction with meter data management (MDM) software to analyse consumption and account data, identifying where power is being lost or stolen.

Revenue loss from energy theft, meter tampering, and equipment error is generally believed to be substantial – estimates range from one to four percent of total revenues for most utilities. In the past, utilities relied on meter readers, or “feet on the street,” to mitigate this problem by observing meters that were tampered with or out of order.

Today, utility companies that have implemented AMI systems no longer have meter readers physically checking equipment. These utilities must devise new ways of detecting theft or equipment problems. By creatively using the rich data set provided by AMI systems regarding meter reads and meter status, utilities can outsmart thieves and identify damaged meters remotely.

Using reverse rotation flags, outage count indicators, interval data, and metered usage on previously cut meters using AMI data can help utilities pinpoint candidates for further field investigations. AMI indicators are often corroborated by on-site investigations that indicate theft is occurring at a specific premise.

Often, the potential to perform theft detection and reduce meter tampering significantly contributes to the justification for a full-scale rollout of AMI. Addressing meter tampering and theft provide enormous benefits to utilities, including increased revenue due to more accurate billing of lost revenue, and reduced service order costs by prioritising revenue protection leads for field personnel.

But collection of higher frequency data and meter status is just the beginning of the revenue protection solution. MDM software helps utilities analyse AMI data, providing knowledge about customer energy use. In-depth analysis helps pinpoint where and by whom power is being diverted, making it easier to identify cases of theft. For example, analysis allows the utility to discover when there is energy use on non-paying accounts and when there is no use for specific time periods on an active account.

ppl elec1

Table 1 – Revenue protection workflows at PPL Electric Utilities

Revenue protection at PPL Electric Utilities
Meter tampering and equipment problems such as broken meters that result in lost revenue are significant issues for PPL Electric Utilities, and the Revenue Protection team has in the past used various strategies to identify specific target accounts for investigation. Most of these strategies involved manual analysis of large quantities of data, a labour intensive exercise.

The team began looking at MDM software as the means to help them simplify the process for identifying possible cases of theft, meter tampering, or equipment problems. An optimum solution would automatically notify team members of anomalies around usage patterns. It would also allow users to create rules and logic, manage the list of outputs, tweak logic for better results, and group the results by geographic location to make it easier to assign work to field investigators.

The Revenue Assurance team believed that an AMI infrastructure with a robust MDM system would take a significant amount of guesswork out of identifying possible theft cases. Rather than make assumptions of the cause of a reduction in consumption, the granularity of data available from MDM can provide a pattern that can be used to identify theft or failing equipment with a high degree of confidence so that the site visit to confirm will be fruitful.

ppl elec2

Figure 1 – Usage pattern indicating abnormal usage

PPL Electric Utilities chose Aclara Software™’s revenue management solution, Revenue Vision, to improve analyses of large volumes of interval, daily, and meter data collected by its AMI system. By combining various meter, premise, and account data, Revenue Vision makes it easier to identify problem meters. PPL is using the Aclara utility applications as well as Aclara Software’s customer service applications.

The revenue protection software identifies suspicious consumption patterns by applying specific, utility-defined screening tests to a targeted population of accounts, meters, or other entities. At PPL Electric Utilities the goal was to define tests narrowly enough so that the data combination would yield a true “hot list” of manageably sized accounts for investigation.

The Revenue Protection team began its project by evaluating existing tests that it already used to assess monthly meter readings. During the course of the review, the company’s biggest revenue loss issues such as equipment malfunctions, installation issues, and potential theft were determined, and the usage patterns identified that were indicative of each problem as well as the customer class or attribute that would be tested. Upon completion of this exercise, the team came up with eight logic tests to implement within the Revenue Vision application.

The next steps in implementing the logic tests required that a workflow be set up for each of the tests (Table 1). The workflows consist of a name, brief description, the group of entities that would be included in the test, and the filters necessary to identify the attributes of the entities that were included. Once the workflows were completed, the team determined how often to run the test.

PPL Electric Utilities generally runs tests weekly, although Revenue Vision allows users the flexibility to change the frequency of test runs. Weekly runs allow better manage of output, adding a security benefit by placing a frequent “electronic eye” on every meter in the field.

ppl elec3

Figure 2 – The meter recorded in
Figure 1 with wires attached
to its potential clip

A new way of doing business
Use of an MDM system fundamentally alters the way revenue protection operations are performed. In the past, significant time was required to perform data queries iteratively and sequentially in a somewhat manual fashion. Also, the data available for such queries was generally limited to daily and monthly consumption. The results were based on an ad hoc process that takes considerable time, with different screening tests being designed and deployed at different times.

MDM changes this paradigm in several ways. Design and implementation of screening tests within MDM are distinctly separate steps. Analyses are designed to fit customer load and data characteristics, allowing PPL Electric Utilities to effectively identify energy theft or tampering. Once an analysis is designed, it is implemented into a regular production process, ensuring that PPL keeps up with the examination of current data, alerting the Revenue Protection team of anomalies as soon as they arise.

The design step involves exploratory analysis of different test schemes by running, reviewing, and comparing different result sets. Hourly data is utilised for these tests and supplemented by external data sources such as weather data, GIS, and customer characteristic data. In the design phase, these tests are run on all or just a sample of customers, with the primary purpose to evaluate the effectiveness of the tests, rather than simply generate customer lists from the tests.

The implementation step is automated. Once logic tests are found to be effective by the analyst, they are put into production by scheduling them as automated runs for whatever period makes sense. All AMI data is initially screened by the validation rules inherent in the MDM system. After validation certain accounts are identified for further review. The revenue protection analyses automatically are run on selected meters. Tests can be nested into a single logic string within a single production run, rather than performed sequentially in multiple runs.

Analysts apply standard tests or test combinations to specific accounts or groups of accounts. Failure of a combination of tests may detect meter tampering. For example, the combination of a loss of power indicator on a meter with a reverse rotation flag is a better indicator of theft than either test alone. No one test determines energy theft or meter tampering, but failures may place an account or meter on the suspicious account list.

ppl elec4

Figure 3 – Toggle
switch controlling
the meter bypass

Looking at the results
PPL Electric Utilities has had positive results from implementation of MDM-based revenue protection software. To date, a total of 40 cases of theft, meter tampering, or equipment issues have been identified through the software and verified through field investigations, with all of those cases resulting in corrective actions. Eighteen cases of equipment problems were identified, which resulted in rebilling of customers. Twenty cases of customer theft were detected, with prosecutions undertaken in all cases.

In an interesting twist, the Revenue Protection team identified two cases of customer-owned generation via windmill and solar panel. These cases were identified through anomalies in blink counts and reverse rotation on the meters. From a utility perspective it is important to identify cases where customers are producing their own electricity so that safety precautions can be addressed and accurate reports of customer load be provided to grid operators. Inaccurate reports can result in penalties to the utility. In addition, the customer may be eligible for a different rate and due a credit for power pushed back into the grid.

Typically, however, reverse rotations and abnormal blink counts are due to meter tampering. For example, in one recent case PPL Electric Utilities was able to identify potential theft by looking at the usage pattern (Figure 1). The graph, taken from reports output from the MDM, indicates a suspicious usage pattern, with the meter going into a reverse rotation a number of times during a single billing cycle. What is more, there are days during the month when the customer is not using any power, while on other days the meter recorded usage. On 20 December, 94 kW of usage was recorded, for example, while on 3 January, when the temperature was 17F, no usage was recorded.

An investigation of the premises based on analysis of the AMI data indicated that the customer had tampered with the meter. Wires were attached to the meter’s potential clip (Figure 2). The bypass was controlled by a simple toggle switch (Figure 3). In this case, PPL Electric Utilities was able to extrapolate usage for rebilling purposes from the periods that were recorded.

Implementing revenue protection software has allowed PPL Electric Utilities to zero in on problem accounts by combining data collected by the AMI system, such as daily readings, interval data, and momentary interruption notifications (blink counts) with other pertinent information such as daily temperatures, meter status, and account status. In the future, the company expects to refine its use of MDM to further maximise the benefits it can derive from the meter data it collects. Even today, in addition to improving its revenue protection activities, it has used the features of its MDM system in customer service, allowing representatives to respond more quickly and accurately to high-bill inquiries. The MDM system also allows PPL Electric Utilities to provide more data to customers about their usage of electricity. This information helps them make informed decisions and take an active role in controlling energy consumption.