Research On Real-time Monitoring Method Of Nuclear Power System Fault

The nuclear system is a complex industrial system.During the operation,the failure of the equipment cannot be avoided,and the parameters of the nuclear system are numerous.It is difficult for the staff to take into account each parameter variable,which increases the difficulty of monitoring.Therefore,it is necessary to effectively monitor the state of the nuclear system during operation,detect possible faults in time,and improve the safety of the nuclear system operating environment.In this thesis,the fault monitoring model is established based on principal component analysis,and the research on improving the performance of fault monitoring is carried out.Firstly,the SWATSGA algorithm is used to select the principal components,and the advantages of Adam’ s adaptive learning rate and SGA fine tuning parameters are combined to retain the original data information to the greatest extent.The fault monitoring model is established for different principal components.The optimal principal component number is determined by comparing the monitoring accuracy of different models,and the optimal principal component model is established,and then the SWATSGA-PCA fault monitoring model is constructed.Secondly,for the random noise in multivariate time series data,the continuous point abnormal alarm is set by sliding window to improve the accuracy of fault alarm and reduce the false alarm rate of fault alarm.Finally,after monitoring the fault,the contribution rate diagram method is used to identify the fault variables and assist the engineer to take countermeasures.The SWATSGA-PCA fault monitoring model is used to carry out simulation experiments on the Tennessee Eastman data set and the simulated nuclear power plant thermal hydraulic system data set.The experimental results show that the fault monitoring model has good monitoring performance.After adding the sliding window,it can effectively reduce the false alarm rate of the fault.After monitoring the fault,the identification of the fault variable helps the engineer to take timely measures to make the operation process into a stable state.