Electric vehicle charging load filtering by power signature analysis

With the more increase in population which is leading to rapid fossil fuel depletion and adding more pollution into environment, there will be more demand of Electric Vehicle in the future and with the more EV's going to come on road having old power distribution infrastructure it will bring charging problem of EV. Home charging of EV will bring overloading of transformer thus producing adverse effect on smart grid. In such conditions monitoring of EV charging plays a vital role in EV charge Scheduling and to monitor the power loading pattern of different houses under a transformer while EV charging by monitoring the smart meter. In this paper by Non-Intrusive load monitoring (NILM) an algorithm is developed that will disaggregate EV charging load and kWh consumed by EV from the aggregated power signal (real power in watts) of many appliances. This algorithm can also identify EV in presence of other high wattage appliances. Advantage of this algorithm is that it even works with low sampling rate of per second data (1/60 Hz) and can be easily trained with the area specific EV charging known database thus giving high estimation accuracy. This algorithm is checked on a monthly data of a house giving accuracy of more than 90 percentage with the output information of how much kWh energy has been consumed by EV and the mean square error only 0.05 in disaggregating EV charging load.