Abstract:Rolling bearings are widely used in industrial equipment. It is of great significance to study the degradation trend of rolling bearings. In this paper, an Adaptive Multi-population Genetic Algorithm (AMGA) is proposed. Firstly, Kernel Principal Component Analysis (KPCA) method is used to fuse the vibration signal in both time domain and frequency domain, which uses kernel to map the sample space to higher dimensional space and uses the higher dimensional space for linear dimensionality reduction. It can effectively reduce the dimension of nonlinear correlation variables and obtain the trend signal representing the Remaining Useful Life (RUL) characteristics of rolling bearing. Moreover, AMGA is proposed to optimize the number of neurons, initial weight, and initial threshold of the Back Propagation (BP) neural network prediction model. AMGA applies chaos algorithm to Genetic Algorithm (GA) to improve the diversity of the initial population. Meanwhile, the communication frequency between different populations is controlled by judging the similarity of the optimal solution among different populations, so as to effectively jump out of the local optimum and obtain the global optimal solution. Finally, the whole life data of the spin-up process of the rolling bearing from University of Cincinnati is taken as an example to analyze the performance of the algorithm. Compared with the traditional BP, the R2-score performance and the MAPE performance of KPCA-AMGA-BP are improved by 0.297 and 2.46% respectively. Furthermore, compared with the optimized BP, this method obtains the improved R2-score performance and the MAPE performance by 0.218 and 0.46%.

Bearing Performance Degradation Evaluation Using Recurrence Quantification Analysis and Auto-Regression Model

Bearing Degradation Evaluation Using Recurrence Quantification Analysis and Kalman Filter

Bearing Degradation Evaluation Using Improved Cross Recurrence Quantification Analysis and Nonlinear Auto-Regressive Neural Network

An Integrated Method for Bearing State Change Identification and Prognostics Based on Improved Relevance Vector Machine and Degradation Model

Damage Assessment of Mechanical Systems Based on Recurrence Quantification Analysis

Evaluation of Rolling Bearing Performance Degradation Using Wavelet Packet Energy Entropy and RBF Neural Network

A Data-Driven Degradation Prognostics Approach for Rolling Element Bearings

A novel based-performance degradation indicator RUL prediction model and its application in rolling bearing

Optimising Bearing Degradation Prediction: A Bayesian Vector Autoregressive-Assisted DeepAR Model Approach

Degradation Assessment of Bearings with Trend-Reconstruct-based Features Selection and Gated Recurrent Unit Network

Performance Degradation Assessment by Kolmogorov-Smirnov Test and Prognosis Based on AR Model

Rolling Bearing Defect Severity Evaluation Using Recurrence Plot Entropy

Bearing Performance Degradation Assessment Using Long Short-Term Memory Recurrent Network

A Nonlinear Degradation Model Based Method for Remaining Useful Life Prediction of Rolling Element Bearings

Research on Degradation Prediction of Rolling Bearing Based on Adaptive Multi-Ga-bp

A Hybrid Prognostics Approach for Estimating Remaining Useful Life of Rolling Element Bearings.

A rolling bearing fault diagnosis technique based on recurrence quantification analysis and Bayesian optimization SVM

Vibration performance prediction and reliability analysis for rolling bearing

An Intelligent Performance Degradation Assessment Method for Bearings

Bearing Degradation Prediction Based on Deep Latent Variable State Space Model with Differential Transformation

Novel data-driven performance degradation state monitoring of rolling bearing