Remaining Useful Life Prediction of Crane Rolling Bearings Based on RMS Trend Consistency
As a key component of crane equipment rotating mechanism,the health status of rolling bearings will directly affect the safety of crane equipment.Due to the crane operation in harsh environment,often lead to sensor collected vibration signal contains a lot of noise,which seriously interfere with the accuracy of feature extraction.In order to improve the prediction accuracy of rolling bearing remaining useful life(RUL),a prediction model combining variational mode decomposition(VMD)and stack-gated recurrent neural network(SGRU),namely VMD-SGRU model,was proposed in this thesis.First,the VMD is optimized by rime optimization algorithm(RIME)to decompose the original vibration signal effectively.Then,the optimal intrinsic mode function(IMF)component is selected by the maximum mutual information coefficient(MIC)to accomplish accurate signal reconstruction.On this basis,the key degradation features are extracted from the reconstructed signals,and the feature dimension is reduced by principal component analysis(PCA).Finally,the dimensionally reduced features are fed into the stack-gated cycle unit(SGRU)network to predict the RUL of the bearing.In this study,the VMD-SGRU model was validated using PRONOSTIA standard dataset.The experimental results show that the model can effectively extract the degraded features from the denoised signals and take them as the input of SGRU network,which significantly improves the prediction accuracy.
cranerolling bearingsrime optimization algorithmvariational mode decompositionremaining useful life prediction
万方数据
维普
