Health Management of Bearings Using Adaptive Parametric VMD and Flying Squirrel Search Algorithms to Optimize SVM
Open Access
Abstract
Bearing, as one of the core parts of rotating machinery, has a running state which is related to the overall operation of the system. Due to the bearing structure and its complex operating environment, running condition monitoring and fault diagnosis is always a key problem in the field of bearing health management, which is of great significance for bearing maintenance and equipment reliability and safety. In view of the difficulty in parameter selection and poor feature extraction ability of variational mode decomposition (VMD) in existing feature extraction, this paper uses the flying squirrel search algorithm (SSA) to optimize the parametric of decomposition layer k and penalty factor α in VMD, and forms an adaptive VMD signal decomposition method. To solve the problem of high dimensionality and long extraction time of multi-domain fault feature set, kernel principal component analysis (KPCA) is used to reduce feature dimensionality. Then, the processed features are input into the support vector machine (SVM) for fault diagnosis and classification, and the parameter optimization ability of SSA is used again to build the SSA-SVM fault diagnosis model. To evaluate the running state of bearings, an alarm threshold method based on the root mean square value calculated by cosine similarity and 3σ is proposed to divide samples of different health states. Finally, the method constructed in this paper is compared with other methods by using simulation and experimental data sets, and the running condition monitoring and fault diagnosis of rolling bearings are successfully realized, which shows the superiority and effectiveness of the method proposed in this paper.Keywords
Funding Information
- National Natural Science Foundation of China (52075088, LNYJG2022490, JG22DB480)
- Liaoning Province Graduate Education and Teaching Reform Research Funding Project (52075088, LNYJG2022490, JG22DB480)
- Liaoning Province education science “14th Five-Year plan” (52075088, LNYJG2022490, JG22DB480)
This publication has 26 references indexed in Scilit:
- Physics-informed multi-state temporal frequency network for RUL prediction of rolling bearingsReliability Engineering & System Safety, 2023
- Construction of bearing health indicator under time-varying operating conditions based on Isolation ForestEngineering Applications of Artificial Intelligence, 2023
- Research on the health status evaluation method of rolling bearing based on EMD‐GA‐BPQuality and Reliability Engineering International, 2023
- Performance Degradation Evaluation Model of Rolling Bearing Based on CAE-SVDD2023
- Novel morphological scale difference filter with application in localization diagnosis of outer raceway defect in rolling bearingsMechanism and Machine Theory, 2023
- Multi-fault diagnosis of rolling bearing using two-dimensional feature vector of WP-VMD and PSO-KELM algorithmSoft Computing, 2022
- A new intelligent fault diagnosis method for bearing in different speeds based on the FDAF-score algorithm, binary particle swarm optimization, and support vector machineSoft Computing, 2019
- Data-driven bearing fault identification using improved hidden Markov model and self-organizing mapComputers & Industrial Engineering, 2018
- A rolling bearing fault diagnosis strategy based on improved multiscale permutation entropy and least squares SVMJournal of Mechanical Science and Technology, 2017
- A fusion feature and its improvement based on locality preserving projections for rolling element bearing fault classificationJournal of Sound and Vibration, 2015