Weighted Fusion Correlation Vector Machine Model for Remaining Life Prediction of Rolling Bearings

Abstract

The weighted fusion correlation vector machine model for the prediction of the remaining life of rolling bearings first uses the improved particle filter framework to reduce or eliminate the influence of outliers on the prediction effect of each kernel function model, and then based on the generalization ability of each single correlation vector machine model to the data , screen out the single correlation vector machine models with strong generalization ability, and perform weighted fusion on them to obtain the weighted fusion correlation vector machine model, realize the complementary advantages of the characteristics of each single correlation vector machine model, and improve the weighted fusion correlation vector machine to the rolling bearing For the prediction effect of the operating state and remaining life, the weighted fusion correlation vector machine prediction model obtained by the present invention has high prediction accuracy and strong robustness, and is more suitable for practical engineering applications.

Description

technical field [0001] The invention relates to the technical field of prediction of remaining life of rolling bearings, in particular to a weighted fusion correlation vector machine model for prediction of remaining life of rolling bearings. Background technique [0002] Rolling bearings are key components in rotating machinery and are also the most vulnerable elements. The structural characteristics of rolling bearings, manufacturing and assembly factors, and complex load-bearing conditions have buried hidden dangers for bearing failures or even failures. Once the rolling bearing fails, it will inevitably pose a serious threat to the safe service of the rotating machinery, ranging from production accidents of equipment downtime to serious disasters such as machine crashes and human deaths. Therefore, the monitoring and diagnosis of rolling bearings is of great significance. Although the traditional regular maintenance plan for rolling bearings can effectively reduce the a...

AI Technical Summary

Problems solved by technology

However, the current selection of correlation vector machines is mainly based on experience. Different types of correlation vector machines have different characteristics, which are reflected in the differences in the prediction effects of correlation vector machines. Specifically, the prediction accuracy of a single correlation vector machine model is low in stability and robustness

Images