CAJ | 학술논문

针对模拟电路故障诊断的难点和传统诊断方法的不足之处，提出了一种基于PSO算法优化的RBF神经网络模拟电路故障诊断方法。为了约简网络结构从而提高诊断效率，采用主成分分析方法对故障特征进行有效提取。针对RBF网络传统训练算法中隐层节点中心及基函数宽度选取困难问题，提出采用PSO算法来优化训练RBF网络，以提高网络的训练速度和泛化性能。最后，通过电路仿真对所提方法的有效性进行了验证。
침대모의전로고장진단적난점화전통진단방법적불족지처，제출료일충기우PSO산법우화적RBF신경망락모의전로고장진단방법。위료약간망락결구종이제고진단효솔，채용주성분분석방법대고장특정진행유효제취。침대RBF망락전통훈련산법중은층절점중심급기함수관도선취곤난문제，제출채용PSO산법래우화훈련RBF망락，이제고망락적훈련속도화범화성능。최후，통과전로방진대소제방법적유효성진행료험증。
According to the deficiency of traditional fauh diagnosis methods, the method based on particle swarm optimization and radial basis function neural networks is proposed. In order to reduce the input dimension of neural networks and improve fault diagnosis accuracy, the principal component analysis is adopted to abstract the fault characteristics. Because the width and center position of radial basis function for the hidden layer of RBF neural networks are difficult to select, the particle swarm optimization algorithm is adopted to train the RBF neural networks. At last, the efficiency of the proposed method is verified by the simulation circuit.