Identifying non-technical losses with machine learning

Identifying fraud and malfunctions by applying machine learning techniques to data collected by meters.
Lightning Talk 9 Oct 2018 9:35 am Park Hyatt Vienna, Austria

Presented At:


You need to be logged as a Meets Attendee to view this content. Login

In this lightning talk you will hear results and key findings from Enel’s South America Revenue Protection project to identify fraud and malfunctions by applying machine learning techniques to data collected by meters, including:

  • Locating and isoltating non-technical losses by combining granular grid and consumer data, to optimise revenue protection
  • Applying machine learning to smart meter data to better predict where losses are located
  • Automatically identifying drops in consumption and obtaining other features related to demand
  • Providing estimates for the amount of energy which will be recovered, helping to guide future inspections
  • Creating a more directed approach in guiding investigations into energy theft
  • Protecting revenue, and reducing bills for customers

Presented By:

Mario Namtao Shianti Larcher

Data Scientist - Infrastructure and Networks Digital Hub

Mario Namtao Shianti Larcher

Data Scientist - Infrastructure and Networks Digital Hub




Mario Namtao Shianti Larcher

Data Scientist - Infrastructure and Networks Digital Hub