Non-Invasive Power Load Identification Technology Based on MFCC and SVM
Yu Liang He, Yingchun Wang, Jun Li, Wei Wei, Yaojun Xu · 2023
Non-intrusive load identification (NILI) technology can provide important data information for users, power companies, power design and operation units, which is simple, economic and effective. With the popularization of high-speed data acquisition terminals, the technology of power load transient feature recognition with high accuracy has a good application prospect. Aiming at the difference of power load transient waveform, proposes a method of power load identification based on MFCC feature extraction and support vector machine (SVM) feature classification. After slicing and wavelet filtering normalization of power load transient waveform data, MFCC is used to extract the slice feature information, and then support vector machine is used to classify the features representing the load waveform information, Simulation and experimental data are used to verify the effectiveness of the proposed algorithm.