M-P人工神经网络用于核电站故障诊断系统

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。

M-P artificial neural network model for nuclear power plant fault diagnostic system

An M-P artificial neural network (ANN) was used to model the diagnostic network of a nuclear power plant unit fault diagnostic system. The M-P model was developed based on the characteristics of the signals and fault events in nuclear power plant faults. The advantages and disadvantages of the M-P model were analyzed. A diagnostic network was built using the M-P model for the safety-related water service system. The M-P model avoids some problems related to the building and training of other neural networks. A combination of the M-P model and other models provides an improved nuclear power plant diagnostic system, especially for processing diagnostic tasks which depend heavily on logic judgments.