Maximizing the benefits of AMR for theft detection through data analytics


Conference: Metering, Billing/CRM India
Location: New Delhi, India
Presenter: Rajesh M. Bansal
Abstract: Presented by Rajesh M. Bansal at Metering, Billing/CRM India

Turnaround of Delhi discoms has been a big success story in Indian power sector & has delivered a message to the entire world that public private partnership in electricity distribution sector would work. The country (India) looses more than Rs. 280 Billion due to theft annually.

Almost 2/3rd of Delhi electricity distribution is catered by BSES. The reduction of AT&C losses in Delhi owes a lot to having the right metering strategies and choosing the right technologies to improve billing & containing energy theft.

While the obvious choice was to migrate from electromechanical meters to electronic meters, a need was felt to

  • Choose the right metering system at all levels (grid, distribution transformer & consumers).
  • Choose meters with enough tamper evident features ,
  • Adopt efficient way of downloading the data from the meter (Handheld/AMR etc) and
  • Analyze smartly the downloaded data to conclude theft.

Right metering system with tamper evident features was implemented & electronic data was downloaded thru’ handheld MRI / AMR.  The critical part was to detect theft and other under-recording anomalies.

This paper discusses on AMR technology used in BSES & on leveraging AMR for theft detection  through data analytics tool.

BSES has installed AMR modems for all premium & high value consumers totaling around 12,000 installations.  Also all the meters in the grid feeders have been AMR enabled.  Through GSM based AMR ,we were not only able to add value in terms of reduction in operating cost but also use AMR as a tool for theft control in  premium and high value consumers’ segment.

Data analytics has helped us a lot in deriving maximum benefits out of AMR.

Data analytics initiative which was pioneered in BSES Delhi basically covers two main streams viz,

  1. Tamper analysis &
  2. Consumption analysis.

Tamper analysis:

The tampers in electronic meters can be broadly classified as:

  • Voltage circuit  related tampers
  • Current circuit related tampers
  • Power circuit related tampers.
  • External interferences like ESD, magnet, high frequency signals, DC pulses etc.

All the above methods leave some symptoms on the data stored in the meter. The tamper analysis is all about looking for symptoms on the stored data like phasor diagrams, event records etc.

The data received from meter through AMR are analyzed through different logics to detect theft.

Consumption analysis:

This analysis does not necessitate data from meter. Monthly energy consumption recorded in the meter would serve the purpose.

Benchmarks have been developed against major segments of consumer groups. Consumption of each consumer is mapped against the established benchmark. Anomalies in the consumption are matched with the consumer’s attributes (like house area, house income etc) to conclude theft.

Theft which escapes the meter or the consumer whose meter data could not be downloaded would get caught through this methodology. The suspects short-listed based on above two methods are subsequently raided for booking theft.

Data analytics was initially started with the high value consumers. Due to its extremely encouraging results and high hit rate, it has been extended to cover mass segment of consumers as well.

Future Plan
In order to further improve the Enforcement strike rate, it is planned to add few new features in meter and Intelligent MODEM based AMR – SMS communication system is to be installed at consumer end.

The normal SMS based system which flashes SMS, incase meter indicates tamper, normally generate pseudo messages and has lost its credibility. BSES has developed a model based on which, SMS will be generated with very high probability of actual theft.