Bolt connection loosening prediction method and system based on Gaussian process regression

Abstract

The invention provides a bolt connection loosening prediction method and system based on Gaussian process regression, and the method comprises the steps: setting prediction parameters, and obtaining training set sample data and test set sample data related to the prediction parameters; training an initial bolt connection loosening prediction model based on Gaussian process regression according to the training set sample data; setting a test precision standard, and obtaining a bolt connection loosening prediction model conforming to the test precision standard according to the test set sample data, the initial bolt connection loosening prediction model and the test precision standard; and obtaining to-be-tested sample data, and inputting the to-be-tested sample data into the bolt connection loosening prediction model to obtain a prediction result. According to the method, the bolt connection loosening prediction model based on Gaussian process regression is trained, the dispersity and uncertainty of the pre-tightening force in the bolt loosening process are considered for accurate prediction, and the mean value and the variance of the target can be predicted at the same time, so that the uncertainty problem is more reasonably solved.

Description

technical field [0001] The present application relates to the technical field of bolt loosening prediction, and in particular, to a method and system for bolt loosening prediction based on Gaussian process regression. Background technique [0002] As the most common failure form in the connection structure, the failure of bolted connection loosening is common in engineering, and generally occurs before the failure of bolt fatigue fracture. There is a steady decline. In the early stage of bolt loosening, it may not have a significant impact on the normal operation of the equipment, but as the pre-tightening force continues to decrease, the structure will gradually appear abnormal noise, leakage and other failures, which will lead to other failure forms such as bolt fatigue fracture, so bolted connection Accurate prediction of the release process is of great significance. [0003] With the rapid development of artificial intelligence theory, data-based modeling methods have ...

AI Technical Summary

Problems solved by technology

[0004] This application provides a method and system for predicting bolt loosening based on Gaussian process regression, which can solve the problem of poor prediction accuracy of existing bolt loosening prediction models

Images