Research On Power Alarm System Based On Probabilistic Neural Network

With the rapid development of social science and technology,electric energy has become indispensable in people's daily life and national production,and the power grid plays a decisive role in the whole power system.Because of the distribution of power supply,geography and other factors,the operation mode and structure of power grid are becoming more and more complex,and the grid fault rate is rising.Therefore,the research to identify and diagnose faults accurately and quickly has became the focus of smart grid study.In this paper,a fault diagnosis method for power grid based on probabilistic neural network algorithm is proposed.Firstly,the effective information of each fault signal is extracted by feature extraction algorithms,and then the fault mode is classified by probabilistic neural network algorithm.In this paper,the probabilistic neural network algorithm based on the kernel principal component analysis is been studied,which extracts the principal component from the variables to replace the traditional probabilistic neural network input.At the same time,a large number of neuron samples will be replaced by a limited number of pattern combination neurons.The algorithm greatly optimizes the network structure.In addition,the probabilistic neural network algorithms based on partial least squares and T-distributed stochastic neighbor embeddingalgorithm are also studied in this paper.The three algorithms are simulated and tested,and the experimental results are analyzed.The simulation results show that the proposed method is feasible in power system fault diagnosis and it has more advantages than traditional artificial intelligence technology.It will provide some advice for the development of power system diagnosis and alarm technology.