Unsupervised Nonintrusive Extraction of Electrical Vehicle Charging Load Patterns

Abstract

Extracting electric vehicle (EV) charging loads is an important aspect that enables smart grid operators to make informed and intelligent decisions about conserving power and promoting the reliability of the electrical grid. This paper presents an unsupervised algorithm to extract the EV charging loads (EVCLs) nonintrusively from the smart meter data. The proposed algorithm can run on low-frequency smart meter sampling data and only requires the real power measurement, which is the type of data communicated and recorded by most smart meters. Validation results on real aggregated household loads have shown that the proposed approach is a promising solution to extract EVCLs and that the approach can effectively mitigate the interference of other appliances that have similar load behaviors as EVs. Furthermore, the extraction of such load behaviors can be aggregated and open further smart grid analyses and studies.