Operating Reliability Assessment Method Based On FVMD Multi-scale Permutation Entropy And KFCM Clustering

The reliable operation of mechanical equipment is not only related to the product quality of mechanical equipment,but also related to the actual production of mechanical equipment,so the operational reliability of mechanical equipment directly determines its core competitiveness.An operating reliability assessment method based on fast variational mode decomposition(FVMD),parameter optimized multi-scale permutation entropy and kernel fuzzy C-means(KFCM)clustering is proposed.Starting from the signal processing method,the vibration signal is processed,and then the feature of the processed signal is extracted.Finally,the operational reliability of the extracted feature data is calculated,and the operation reliability evaluation of the mechanical equipment is completed.Firstly,the fast variational mode decomposition method is used to process the vibration signal.The fast variational mode decomposition method is an improvement of the variational mode decomposition method.By introducing an iterative operator to update the original Lagrangian multiplication operator twice,the iterative number and running time can be reduced under the same iterative termination conditions.The fast variational mode decomposition method is verified by the decomposition of simulation signals,then,it is applied to the decomposition of measured signals.Secondly,the influence of the parameter scale factor s,the time series length N,the delay time τ and the embedding dimension m on the entropy value of the multi-scale permutation entropy method is analyzed.Therefore,the optimization algorithm is used to optimize the parameter selection of the multi-scale permutation entropy method,and the directed self-organising dynamic topology hybrid swarm intelligence algorithm is proposed to optimize the multi-scale permutation entropy method.Combined with the fast variational mode decomposition method,the signal is processed firstly,and then the state information of the signal is quantized by using multi-scale permutation entropy,which is reserved for the following method.Then,the feasibility of the feature extraction method is verified by the bearing vibration signal.Thirdly,in order to solve the problem that the damage type can not be judged by the operating reliability assessment,the operating reliability calculation method based on kernel fuzzy C-means clustering is proposed.By calculating the distance from each group of multi-scale permutation entropy to the cluster center of normal state,the reliability of mechanical equipment is determined,then the damage type of the sample in 10 states to be evaluated is determined according to the proximity.The operational reliability of vibration signals of rolling bearings under different damage levels and different damage types is evaluated,and the feasibility of the proposed evaluation method is verified.Finally,the operational reliability experiment is carry out,the vibration signals of rolling bearings under different rotating speeds and different damage types are collected by experiments.First,the vibration signals collected in the experiment are decomposed and reconstructed by the fast variational modal decomposition method,and then the state information of the reconstructed signals is extracted by the parameter optimization multi-scale permutation entropy method.Then,the kernel fuzzy C-means clustering method is applied to evaluate the operational reliability of the system.The validity and feasibility of the proposed method are verified.