In order to solve the problems of oil‐paper insulation deterioration of oil‐immersed power transformers, including the instability of performance degradation quantity at the early stage, non‐linear deterioration process and uncertainty of life prediction, an improved Wiener stochastic process based on strong tracking filter (STF) is proposed to predict the remaining life of transformer oil‐paper insulation in this paper. Firstly, in order to grasp the dynamic process from healthy state to current deterioration state, a deterioration model is established based on Wiener random process. Secondly, in order to improve the accuracy, the initial values of model parameters are estimated based on expectation maximization (EM) algorithm, then with the increase of monitoring data, model parameters are updated by using STF algorithm, and the probability density function of remaining life is derived to predict the remaining life of oil‐paper insulation. Finally, the accuracy of the proposed method is verified by using furfural content as performance degradation index and accelerated thermal aging experimental data. Compared with the Bayesian–Wiener algorithm, the MSE of the proposed algorithm is reduced to 0.2138 and its MAPE is reduced more than 4%. Besides, the proposed method has the advantages of strong robustness and low uncertainty of residual life prediction.

To solve the problem of low accuracy of the previous electricity theft detection methods, the authors propose a multi-domain feature (MDF) fusion electricity theft detection method based on improved tensor fusion (ITF). Firstly, the original electricity consumption series is transformed by gram angle field (GAF) to obtain the time-domain matrix. The original electricity consumption series is converted into frequency-domain by Maximal Overlap Discrete Wavelet Transform (MODWT) to obtain the frequency-domain matrix. Then, the convolutional neural networks (CNN) are used to extract features of the time-domain matrix and frequency-domain matrix, respectively. Next, in order to fuse single-domain feature information and MDF interaction information while reducing redundant information, the authors propose an ITF method to obtain a multi-domain fusion tensor. Finally, the multi-domain fusion tensor is input into the electricity theft inference module to judge whether the user implements electricity theft behaviour. The authors simulate six electricity theft types and evaluate the method's performance separately for each electricity theft type. The results show that the proposed method outperforms other methods.

Traditional offline cable diagnosis methods need power outages during detection, affecting power supply reliability. Here, a hierarchical diagnosis method of cable aged segment based on transfer function is proposed. Firstly, the calculation model of cable transfer function with the aged segment is established; on this basis, the correlation between transfer function and cable aging is analysed. Then, a structure with combined sparse autoencoder and convolutional neural network is trained to estimate the aging location, and a hierarchical diagnosis model of distribution cable based on transfer function is proposed. The sensitivity and accuracy of aged segment detection are improved after hierarchical diagnosis. Finally, the simulation results show that the method proposed in this paper can effectively realize the online identification and location of the cable aged segment. The proposed method makes use of the advantage that the cable transfer function can be obtained online. Compared with the existing methods, this method does not need power outages in the diagnosis process, and the aged segment can be located without a lot of additional equipment.