Energy Theft Detection Technical Brief

By the Bidgely R&D Team

Non-technical loss (NTL) continues to be a growing financial concern for electricity distribution companies, especially in developing countries where energy theft is one of the major contributors to NTL. But many existing broad-scale theft detection methodologies lack precision, require significant manual confirmation processes, can send false-positives, and even fail outright to detect some forms of theft.

Bidgely’s UtilityAI™ Energy Theft Detection solution, however, uses advanced AI-based data science to drill down to the premises level, discovering theft patterns in behind-the-meter energy consumption to find tariff misuse, direct theft, meter tampering, and other non-technical loss.

This technical brief by the Bidgely R&D team explains our data science approach, which has led to theft detection successes such as:

  • 70-95% for correctly detecting anomalies
  • Bookings of 50-60% in energy theft cases
  • Strike rates of more than 90% in tariff misuse cases