Nonintrusive identification and type recognition of household appliances based on the harmonic analysis of the steady-state current

This paper is devoted to the appliance load monitoring based on the harmonic analysis of the steady-state current. The use of the current harmonic content enables a high accuracy in the identification of nonlinear loads, but it is not suitable for linear loads. Namely, the current drawn by the linear loads reflects the supply voltage waveform, and accordingly, its higher harmonics are unpredictable and must be treated as noise. To overcome this problem, we introduce a two-stage classification algorithm which in the first step classifies appliances into two categories, linear and nonlinear. The proposed algorithm exploits steady-state changes in current harmonic vectors to estimate the total harmonic distortion of the current flowing through the target appliance, which is then used as a criterion for load-type classification. In the second classification stage, the device is identified based on the appropriate current harmonic components, linear devices by the fundamental harmonic component and nonlinear devices by the higher harmonic components. The proposed method is experimentally verified on a representative group of household devices. The experimental results show that the proposed method offers an accurate and computationally effective solution for the classification of appliances that are powered on/off into linear or nonlinear. Furthermore, the incorporation of the load classification step into the load monitoring procedure improves appliance identification accuracy.