Research on Fault Diagnosis Method Based on Gate Recurrent Unit Network
Nuclear power plant has the characteristics of internal feedback and strong coupling,which makes the internal working environment complex,and once the abnormal situation is more likely to occur high-risk accidents.In order to improve the safety of nuclear power plant operation,a fault diagnosis method based on gated cycle network is proposed in this paper,in this method,the transient and time-varying features are extracted from the data generated during the operation of nuclear power plant.The experimental results show that the model can judge each fault more accurately than other models,and the accuracy of fault diag-nosis is over 99.9%.
nuclear power systemfault diagnosisgradient boosting decision treegated recurrent units neural network
国家科技期刊平台
NSTL
万方数据
