Non-intrusive Load Decomposition Method based on the Factor Hidden Markov Model

Aiming at the problems of demanding high frequency sampling devices and the factor hidden markov model algorithm without considering the correlation between the power equipments' running states, a novel non-intrusive load monitoring and decomposition (NILMD) method which based on factor hidden markov mode is proposed in this paper. This method considers appliances' current as load feature, and establishes a mathematical model between the total current and the currents of each appliance that considers the correlation between the running states of the equipment. Then, factor hidden markov algorithm is applied to search the running states of each appliance that realizes the decomposition of power load. The experiment shows that the resolution precision of each electric equipment is improved, and the principle of the method is simple. Meanwhile, the load data required in this method can be obtained directly by the universal smart meters on the market which reduces the cost input of the hardware.