Multi-domain Feature Extraction Combined with AdaBoost for Fault Diagnosis of Speed-up Turnout with Unknown Fault
In view of the problem that the unknown new faults of speed-increasing turnouts are misjudged,which will affect the operation safety of trains and the turnouts maintenance efficiency.A signal analysis and fault diagnosis model based on multi-domain feature extraction and adaptive boosting algorithm(AdaBoost)is proposed.Firstly,in order to deeply extract the fault features of turnouts,the original feature set is constructed by extracting fault features from time domain,frequency domain,and time-frequency domain respectively.Secondly,the classification models with different numbers of features are constructed based on the feature importance ranking obtained by the AdaBoost model,and the best feature subset is further obtained by using the classification accuracy of the model.Finally,the best feature subset is input into the AdaBoost fault diagnosis model with a decision mechanism to complete the diagnosis of unknown faults on the speed-increasing turnout.Meanwhile,through the retraining of the model,the adaptive update of the existing fault diagnosis model is realized.Research indicates:the algorithm in this paper can effectively extract the fault features and improve the diagnosis accuracy of the known faults of the turnout,at the same time,the method can effectively identify new faults that have not occurred ago.
国家科技期刊平台
NSTL
万方数据
