Life Prediction Method of Rolling Bearings Integrated Expectation Maximization and Particle Filter

Abstract

The life prediction model of rolling bearings integrating expectation maximization and particle filter, first uses the kurtosis index to monitor the health status of the bearing in real time, and determines the starting time of life prediction; when the starting condition of prediction is met, the effective value is used to predict the remaining life of the bearing ; In the prediction stage, the model parameters are evaluated by using the expectation maximization method, and the bearing state is evaluated by the particle filter method at the same time, by accurately evaluating the model parameters and the bearing state, the remaining life prediction accuracy is improved. The present invention can realize the model The parameters and state of rolling bearings are accurately evaluated, and it shows a better prediction effect than the traditional exponential model in the life prediction of rolling bearings.

Description

technical field [0001] The invention relates to the technical field of prediction of remaining life of rolling bearings, in particular to a life prediction model of rolling bearings integrating expectation maximization and particle filtering. Background technique [0002] Rolling bearings are widely used in mechanical equipment. Due to complex and changeable working conditions, rolling bearing failures occur from time to time. In order to ensure the safe operation of the equipment, the traditional regular maintenance strategy needs to invest a lot of manpower and material resources in the regular maintenance of the rolling bearings, and replace the rolling bearings with faults or potential safety hazards. Rolling bearings often go through a long period of decline from failure to complete failure. Replacing rolling bearings with early failures will inevitably shorten the effective service period of rolling bearings, resulting in a waste of resources. Using preventive mainten...

AI Technical Summary

Problems solved by technology

However, there is a disadvantage in the above research work, that is, the observation value of the rolling bearing is directly used as the state value for life prediction, and the real health status of the rolling bearing is not evaluated.

In fact, there is a certain difference between the observed value and the real state of the rolling bearing, and there is a large amount of random noise in the observed value

Therefore, the traditional index prediction model only evaluates the model parameters and ignores the evaluation of the state of the rolling bearing, resulting in a decrease in the prediction accuracy of the model

Images