基于麻雀搜索算法优化PNN的核动力装置运行工况异常识别研究

Research on abnormal identification of operating conditions of nuclear power plants based on sparrow search algorithm to optimize PNN

为减少操作人员的判断失误,有效提高核动力装置运行工况异常识别准确率,通过分析核动力装置的强相关参数,一种新兴智能优化的麻雀搜索算法与概率神经网络相结合的异常识别模型。由于概率神经网络的平滑因子对输出结果影响较大,因此采用麻雀搜索算法对平滑因子进行参数寻优,将优化结果赋值给概率神经网络模型进行参数训练,得到用于异常识别的最优网络模型。选取17种核动力装置异常运行工况,依托核动力装置事故分析平台进行模拟计算并提取特征参数。Matlab仿真结果表明,该优化网络模型比原始网络模型具有更高的识别精度。

The normal operation of nuclear power plant reactors,and the judgment of faults,accident disposal measures and other operations all need the operator’s manual intervention,so the safe operation of the guarantee system depends on the operator very much.In order to reduce the error of operators’judgment and effectively improve the accuracy of abnormal identification of nuclear power plant operating conditions,a novel anomaly identification model based on the combination of Sparrow Search Algorithm(SSA)and Probability Neural Network(PNN)is proposed by analyzing strong correlation parameters of the nuclear power plant.Since the smoothing factor of PNN has a great influence on the output results,the SSA is used to optimize the parameters of the smoothing factor,and the optimized results are assigned to the PNN network model for parameter training,so as to obtain the optimal network model for anomaly identification.Based on the accident analysis platform of nuclear power plants,17 abnormal operating conditions of nuclear power plants are selected for simulation calculation and characteristic parameters are extracted.MATLAB simulation results show that the optimized network model has higher recognition accuracy than the original network model.