[4M3-GS-13] AI application: Electric power

[4M3-GS-13-04] Comparison of Detection Performance of Anomaly Power Consumption Patterns by Outlier Detection Methods Focusing on Daily Household Power Consumption Patterns

〇Miho Hashimoto¹, Takahiro Nishigaki¹, Takashi Onoda¹ (1. Aoyama Gakuin University)

Keywords:Outlier detection, Anomaly detection, Household power consumption

With increasing energy-saving consciousness in Japan, Home Energy Management System(HEMS) attracts attention. HEMS is the system managing household power comsumption and is the tool supporting power saving. However, HEMS is at risk of a cyber attack through the Internet. So, it is necessary to detect a cyber attack. We focus on daily power consumption patterns of each household and detect anomalies by outlier detection methods. The outlier detection methods are `Hotelling', `k-NN', `Local Outlier Factor', `One Class SVM' and `SVDD'. As a result, it was shown that it is possible to detect anomalies focusing on daily power consumption pattern.