A Review of Industrial Non-Intrusive Load Monitoring

Non-intrusive load monitoring (NILM) is essential for enhancing industrial energy efficiency to obtain appliance-specific energy consumption statistics just through analyzing total electricity meter data. However, most studies of the NILM are on residential loads, and NILM models that work well on resident may not necessarily be suitable for industrial settings because of industrial datasets scarcity as well as behavioral characteristics and usage patterns of different factories. Therefore, we present a summary of three main topics, aiming to provide valuable insights for researchers in the industrial NILM: (a) the publicly available industrial datasets, and analysis of effect of different sampling rates on industrial datasets; (b) the latest developments of industrial NILM approaches by using the hidden Markov model (HMM) and deep neural networks (DNNs); (c) open challenges and research opportunities for load characteristics and NILM algorithms within the industrial setting.