Non-Intrusive Load Monitoring using Feed Forward Neural Network

In recent years, electricity demands have increased because of the growing population. In order to reduce energy consumption, several studies have concluded that Non-Intrusive Load Monitoring (NILM) is effective in raising awareness for users to monitor their daily energy consumption which is beneficial for energy conservation. NILM is a technique that monitor and analyze energy usage through load measurements. These load measurements are used for examining appliances power consumption behavior and the data can be used to modify habits of users through utility bills. This paper proposes a feed-forward neural network approach for NILM using magnitude of current harmonics for load identifications. Experiments for steady state and transient state waveform were first conducted to acquire individual signature current harmonics of appliances and the data acquire are being fed into the neural network for training and later used for load identification. The results concluded that data collected from steady state are better and the simulation results reveal that the proposed neural network approach was able to identify the appliances accurately.