基于神经网络的变压器故障智能检测与分析

利用神经网络BP算法提出变压器常见故障诊断方法，实现了对变压器故障的智能检测与分析。在不同的训练样本、传递函数和训练函数的情况下使用Matlab对建立的神经网络诊断模型进行仿真。通过仿真找到了合适的传递函数、训练函数和隐含层节点数，构建合理的网络。并通过仿真实验验明了该方法比传统方法有更好的效果。

Intelligent Detection and Analysis for Transformer Faults Based on Neural Network

The common fault diagnosis method of transformer is put forward by use of neural network BP algorithm, and the transformer＇s fault intelligence detection and analysis is achieved. In this paper,simulations of Matlab are carried out for different neural network structures, different activation transfer functions, different quantities of training samples and different training functions so that the best diagnosis scheme of neural network can be determined. Through simulation, the proper activation transfer function and training function are found. The experimental results show that the proposed method is able to obtain higher quality solutions efficiently than the conventional approaches.