按照实际情况,将人工神经网络M-P模型应用于核电站机组故障诊断系统的部分诊断网络。根据核电站故障诊断信号和故障事件特点,提出在特定诊断网络中使用M-P模型的可能性和具体构建诊断网络的方法,对其优缺点进行了比较。使用M-P模型对核电站重要厂用水系统建立了简单诊断网络,进行了验证。该模型简单直接,可以避免若干使用其他人工神经网络模型时遇到的问题。该模型与其他模型结合使用于核电站故障诊断系统,尤其在处理逻辑判断性较强的诊断任务时,可以达到更好的效果。 According to the actual situation, the artificial neural network M-P model is applied to the part of diagnostic network of nuclear power plant fault diagnosis system. According to the fault diagnosis signals of nuclear power plant and the characteristics of fault events, the possibility of using M-P model in a specific diagnosis network and the method of constructing a diagnosis network are proposed. The advantages and disadvantages are compared. The M-P model is used to establish a simple diagnosis network for the important water system of nuclear power plants. The model is straightforward and avoids several problems encountered when using other artificial neural network models. This model, combined with other models, can be used in nuclear power plant fault diagnosis system to achieve better results, especially when dealing with logic diagnostic tasks.