Transformer Operating State Monitoring System Based on Wireless Sensor Networks
Xiaoyu Chen, Yong-Tao Hu, Dong Zuo-lin, Pengfei Zheng, Jiahong Wei
Abstract
Transformer operating state prediction is an important module of the transformer operating state maintenance system. Since the analysis of dissolved gas in oil is the prerequisite for realizing the analysis of transformer operating status, the key to predicting the transformer operating status is to predict the content of dissolved gas in oil. Before optimizing the parameters <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\alpha $ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\beta $ </tex-math></inline-formula> of the improved grey prediction model GM(1,1,B), this article determines that the average relative error of the model fitting is the objective function. It is stipulated that the search space for the optimal solutions of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\alpha $ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\beta $ </tex-math></inline-formula> are both [0~1], and the hybrid algorithm of genetic algorithm and particle swarm optimization is used to optimize the model parameters <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\alpha $ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\beta $ </tex-math></inline-formula> . By analyzing the characteristics of different types of abnormal values in transformer online monitoring data such as oil chromatogram and oil temperature, a fast analysis and detection method of online monitoring data stream based on multivariate time series and correlation analysis is proposed. For multi-dimensional monitoring data, from the perspective of data association and time series analysis, a sliding time window is used to record the occurrence time and type of abnormal points, establish a judgment model for candidate abnormal data sets, and use clustering algorithms to analyze candidate abnormalities. The data collection performs comprehensive abnormality judgment of multi-dimensional data. Experiments show that this method can detect abnormal operating states in online monitoring data streams in real time, and has high application value.