Abstract
Rolling bearings, as the main components of the large industrial rotating equipment, usually work under complex conditions and are prone to break down. It can provide a certain theoretical basis for identifying the sub-health state of the industrial equipment by the analysis from the incipient weak signals. Thus, a sub-health recognition offline algorithm based on Refined Composite Multiscale Dispersion Entropy (RCMDE) and Deep Belief Network-Extreme Learning Machine (DBN-ELM) optimized by Improved Firework Algorithm (IFWA) is proposed. First of all, in light of the drawbacks that it is easy to fall into local optima and cross the boundary for exploding fireworks in Firework Algorithm (FWA), Cauchy mutation and adaptive dynamic explosion radius factor coefficient is introduced into IFWA. Secondly, Maximum Correlation Kurtosis Deconvolution (MCKD) optimized by the improved parameters is used to process the incipient vibration signals with nonlinearity, nonstationary, and IFWA is used to adaptively adjust to the period T and the filter length L in MCKD(IFWA-MCKD). Then, each sequence of signals is further extracted the feature-RCMDE to rich sample diversity. Finally, combining the powerful unsupervised learning capability from DBN and the generalization capability from ELM, DBN-ELM can be established. What's more, in order to avoid the interference of human on the parameters, IFWA is used to optimize the number of hidden nodes in DBN-ELM, and the IFWA-DBN-ELM is established. It shows that the algorithm has the higher sub-health recognition accuracy, better robustness and generalization, which has a better industrial application prospect.
Highlights
In recent years, with the development of large-scale industrial equipment towards automation and intelligence, sub-health monitoring for the large-scale equipment has attracted much attention. ‘‘Sub-health’’ is a diseased state which is an incipient state with minor fault
Dispersion Entropy [6] (DE)which overcame the weaknesses of Sample Entropy [7] (SampEn) that had the high time complexity and was easy to mutate for similarity measure as well as Permutation Entropy [8] (PE) without considering the interaction of signal amplitude, had an excellent information extraction capability
It shows that the Improved Firework Algorithm (IFWA)-Deep Belief Network-Extreme Learning Machine (DBN-ELM) algorithm finds suitable hyperparameters through methods such as optimizing the number of hidden layer neurons and improving training ways
Summary
With the development of large-scale industrial equipment towards automation and intelligence, sub-health monitoring for the large-scale equipment has attracted much attention. ‘‘Sub-health’’ is a diseased state which is an incipient state with minor fault. Dispersion Entropy [6] (DE)which overcame the weaknesses of Sample Entropy [7] (SampEn) that had the high time complexity and was easy to mutate for similarity measure as well as Permutation Entropy [8] (PE) without considering the interaction of signal amplitude, had an excellent information extraction capability On this basis, Azami et al proposed Refined Composite Multiscale Dispersion Entropy (RCMDE). Shao et al proposed a method based on combining Particle Swarm Optimization (PSO) to optimize the number of the hidden layer neurons, learning rate and the momentum in DBN, which was applied to analyze experimental signal of the rolling bearings, and PSO-DBN could recognize fault state more intuitively and accurately [23]. Based on the related researches from domestic and foreign countries, this paper proposes a new improved fireworks algorithm to optimize DBN-ELM sub-health identification model. Train the DBN-ELM and classify the testing set, and verify the generalization capability under different loads
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
