Research Of Fault Detection Based On Optimized Neural Network In Power Network

Nowadays,electric energy has become an indispensable and important resource in human life.Once a fault occurs in the operation of power grid,it will bring huge losses to social production and people's lives.Therefore,when the power grid fails,the necessary measures should be taken to locate the fault area quickly and accurately,find out the specific fault lines,and improve the speed and detection accuracy of power supply recovery after the power grid failures.BP neural network has good self-learning ability and self-adaptability,and it is one of the most widely used fault detection methods in power grid.However,BP neural network has many shortcomings,such as its own gradient descent algorithm has some shortcomings,such as low training accuracy and slow convergence speed.Artificial Tree algorithm is a new swarm intelligence bionic algorithm.The concept of inertia weight and random probability is introduced to further improve the Artificial Tree algorithm.The improved Artificial Tree algorithm optimizes the weights and thresholds of BP neural network,and designs the number of nodes in the hidden layer of the neural network to make the network more efficient.With higher training accuracy and faster convergence speed,the accuracy of network fault detection can be effectively improved.When the power grid fault occurs,the signal is vulnerable to noise interference affecting the accuracy of signal acquisition.Therefore,the improved wavelet threshold denoising method is used to remove the noise source in the signal and improve the accuracy of signal acquisition.A power system fault model based on switching quantity is established to obtain the decision table of power system fault information,train the neural network,and detect the fault occurrence interval.The fault model based on electrical quantity is carried in the MATLAB/Simulink simulation software.The fault signals are collected to form training samples of the neural network.After training the neural network,some data are selected to detect the accuracy of the network fault detection.Experiments show that the improved Artificial Tree algorithm optimizes the training speed of the neural network faster and has a higher detection accuracy of 96%.Figure19;Table6;Reference 50...