In order to effectively utilize the large amount of high-dimensionality historical data generated by energy meters during operation, this paper proposes a DBN-MLP fusion neural network method for multi-dimensional analysis and fault-type diagnosis of smart energy meter fault data. In this paper, we first use DBN to strengthen the feature extraction ability of the network and solve the problem of many kinds of feature data and high dimensionality of historical data. After that, the processed feature information is input into the MLP neural network, and the strong processing ability of MLP for nonlinear numbers is used to solve the problem of weak correlation among data in the historical data set and improve the accuracy rate of faults diagnosis. The final results show that the DBN-MLP method used in this paper can effectively reduce the number of training iterations to reduce the training time and improve the accuracy of diagnosis.

The built-in relay in a meter is a key control component of a smart meter, and its reliability determines whether the user can use electricity safely and smoothly. In this paper, the degradation characteristics of the arc-burning energy are enhanced by the method of K-means clustering to replace degradation data, such as the overtravel time, release time, and other data. In existing methods, the meter needs to be disassembled to describe the degradation trend of the meter relay. The proposed method is combined with a bidirectional long short-term memory (Bi-LSTM) neural network to predict the degradation trend of the relay’s performance. In this paper, K-means clustering is used to enhance the extraction of arc energy data features, and then the arc energy data obtained from the reliability lifetime test is assessed to predict the degradation trend of the meter relay by means of a bidirectional LSTM.

Due to the high reliability of present meters, it is difficult to obtain the failure time of meters through accelerated life tests. Based on the failure data of the accelerated life test, this paper studies the mathematical model based on the Wiener process and establishes the degradation model of the instrument by the maximum likelihood to estimate the parameters of the Wiener model. With full consideration of the possible nonlinear effects in modeling, the time scale transformation method is used to study and obtain the reliability life prediction model of smart meters based on nonlinear data. Finally, the reliability life prediction model of meters is verified and evaluated through the example data of the accelerated life test of smart meters. Compared with the conventional method, this method has less error in calculating the reliability, greatly saves test time, and has a higher accuracy than estimating the lifetime model using the Wiener process directly.