论文部分内容阅读
运用人工神经网络技术对２００ＭＷ 核供热堆的故障诊断系统进行了研究， 并用事故工况下反应堆参数的实际值和趋势变化值分别对两个ＢＰ网络进行训练和检验， 两个网络诊断结果的综合得出最终诊断结果。经检验证明， 将两个网络结合的综合系统与单网络系统相比， 可提高诊断的准确性和适应性。
The fault diagnosis system of 200 MW nuclear reactor is researched by using artificial neural network technology. The two BP networks are trained and tested respectively by the actual value and trend change value of reactor parameters under accident conditions. The results of two network diagnoses The final diagnosis results. The test proved that the combination of two networks with a single network system can improve the accuracy and adaptability.