粒子群神经网络的水轮发电机故障诊断研究

Hydraulic Generator Fault Diagnosis Based on PSO and Neural Network Model

研究水轮发电机故障准确诊断问题,水轮发电机一种复杂、非线性系统,故障产生原因具有多样性和不确性,传统神经网络存在收敛速度慢、易陷入局极小等缺陷,导致故障诊断精度低。为提高水轮发电机故障诊断精度率,提出一种粒子群(PSO)算法和BPNN(BPNN)相结合的水轮发电机故障诊断模型(PSO-BPNN)。首先将水轮发电机故障信息输入BPNN进行学习,并利用PSO对BPNN参数进行选择,最后对建立故障诊断模型进行验证性测试,结果表明,PSO-BPNN克服了传统方法不足,提高了水电机故障诊断准确率,具有很好的应用价值。

In order to improve the accuracy of hydraulic generator fault diagnosis, this paper put forward a fault diagnosis method of hydraulic generator based on particle swarm optimization algorithm and BPNN （ PSO- BPNN）. The hydraulic generator fault information was input to a BPNN for learning, and the parameters of BPNN were selected by PSO which has global search ability. Lastly, the model was tested, and the results show that the proposed model overcomes the shortcomings of the traditional models and improves fault diagnosis accuracy.