Low frequency-based energy disaggregation using sliding windows and deep learning

The issue of controlling energy use is becoming extremely important. People’s behavior one the most important elements influencing electric usage in residential sector, significant consumers globally. building’s could be reduced by using feedback programs. Non-Intrusive Load Monitoring (NILM) approaches have emerged as viable options for disaggregation. This paper presents a deep learning algorithm Long Short-Term Memory (LSTM) models It employs low-frequency sampling power data collected private house. aggregated active and reactive powers are used inputs sliding window. obtained results show that proposed approach gives high performances term recognizing devices' operating states predicting consumed each device.