Anomaly detection refers to detecting objects whose behavior deviate significantly from what is expected or standard. It may be detecting theft which is intentionally caused by some consumers in the electrical network, or it can be detecting any other abnormal behavior which may be caused by some mechanical damage in the network. The invention uses multi-layered statistical and machine learning techniques to detect anomalies using smart meter data and some knowledge about the electrical distribution network. The main advantage of this technique is that it is data driven. The learning model proposed in the solution will improve over time.


UBD Ref No: U38    Status: Granted

Inventors: Sambaran Bandyopadhyay, Zainul Charbiwala, Tanuja Ganu, Mohamad Iskandar Petra

Documents: U38-G-US, BN/N/2016/0012, US 14/862,487, US 2017/0082665, US 10,379,146

  • United States

    Filed: Sep 23, 2015

    Granted: Aug 13, 2019