摘要
进行了核动力装置神经网络故障诊断方法的研究，用装置参数的变化量来训练神经网络，并采用了多项式拟合差分滤波技术来确保得到足够准确的参数变化量和滤波阀值的方法，从而保证了神经网络快速。
Studies of fault diagnosing in nuclear power plant using neural network technology was made. The neural network is trained using the value of variation of plant parameters. A polynomial fitting-deviation filter technology is used to attain enough accuracy of the values. Filter technology and filter threshold are adopted to ensure quick and accurate diagnosis.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
1999年第2期21-25,共5页
Journal of Harbin Engineering University
