Research On Power Grid Fault Early Warning Method Based On Meteorological Factors

As the most important energy source for human beings,electric power provides a continuous source of power for the development of today’s society.The safe and stable operation of the power grid is not only the foundation for the development of the power industry,but also the premise for the power industry to serve social development.Most power grid facilities are exposed to the natural environment,and the failure of power grid equipment caused by external meteorological factors is one of the main reasons for power grid failures.Analyzing the role of meteorological factors in power systems and conducting early warning research on meteorological faults in power grids based on meteorological factors is of great significance for reducing economic losses caused by grid faults and strengthening the level of disaster prevention and mitigation in power grids.This paper aims to build a power grid fault early warning system based on refined meteorological data,conducts research from two aspects:meteorological data feature extraction and power grid meteorological fault prediction model construction,and proposes a power grid meteorological fault early warning method that combines Stacked Denoising Autoencoders(SDAE)and Improved Sparrow Search Algorithm(ISSA)optimized XGBoost model.Firstly,starting from the mechanism of meteorological disasters in power grids,analyze the impact of meteorological factors on power grids,summarize the main types of meteorological disasters in power grids,and construct a Stacked Denoising Autoencoders(SDAE)models to extract features from meteorological data from the perspective of power grid faults,making meteorological data play a greater role in power grid fault prediction and early warning research.Secondly,the shortcomings of Sparrow Search Algorithm(SSA),such as insufficient optimization accuracy and easy to fall into local optimal solutions are improved.Convert the PWLCM chaotic sequence into a sparrow position to complete population initialization and improve the quality of the initial population;Introducing an adaptive weight that takes a smaller value at the later stage of the iteration helps enhance the local search ability of the discoverer;An optimized iterative local search method and a random differential mutation strategy are used to improve the process of updating the enrollee’s position and improve the individual’s ability to jump out of the local optimal solution.Then,the XGBoost model is constructed to achieve grid meteorological fault classification and prediction.In order to solve the problem of model parameter adjustment,the Improved Sparrow Search Algorithm(ISSA)is used to find the optimal hyperparametric combination of the model.After the XGBoost model is optimized by the ISSA algorithm,the fault prediction ability is significantly improved.Finally,considering the severity of meteorological factors and the influence degree of meteorological factors on the power grid under different fault conditions,a hierarchical meteorological fault early warning method for power grids is proposed.Experimental results show that compared to other methods,the proposed power grid fault early warning method based on SDAE-ISSA-XGBoost model has higher accurate warning rate,lower error warning rate,missed warning rate,and false warning rate.