Power supply reliability evaluation of distribution network based on non-intrusive low-voltage power load identification and time series algorithm

In order to overcome the problem of low reliability evaluation accuracy existing in traditional power supply reliability evaluation methods, a new power supply reliability evaluation method based on non-invasive low-voltage power load identification and time series algorithm is proposed in this paper. Firstly, a non-invasive low-voltage power load acquisition device is designed, and the adaptive Gauss filtering method is used to denoise. Secondly, the characteristics of low-voltage power load are extracted, and the characteristic parameters are input into the limit learning machine model to complete the identification of low-voltage power load. Finally, the time series algorithm is used to calculate the power supply reliability evaluation index of the distribution network, and the power supply reliability evaluation of the distribution network is completed. The experimental results show that the proposed method has high accuracy of low-voltage power load identification and reliability evaluation, and the highest evaluation accuracy is 97%.