Wearing random process test prediction method based on wearing mechanism

Abstract

The invention discloses a wearing random process test prediction method based on wearing mechanism, comprising steps of determining three affecting factors including a shifting force, rotation speed and hardness, processing test data, to wit, material abrasion loss in unit time, calculating related coefficients between variable groups, verifying multiple correlations between variable groups, restoring a model according to a normalized inversion process to obtain a regression coefficient value of an original variable, determining three phases of wearing which are a running-in period, a stabilized period, and an acute period, utilizing the data of the running-in period and the stabilized period in full-life test data to perform statistic and analysis and collect wearing data, solving a deviation coefficient of a particular parameter value u<-> and a process intensity (), and utilizing abrasion loss (W) at various times in the random wearing test data to perform variance estimation. The invention utilizes a orthogonality test design method, reduces the workload, fully considers the mutual relation between multiple factors in a horizontal direction, achieves high efficiency, fastness, and economy.

Description

technical field [0001] The invention relates to a wear mechanism-based test prediction method for a wear random process of a vehicle transmission. Background technique [0002] In engineering applications, the factors that affect the wear of vehicle transmissions are random and time-related, so it is necessary to accurately reflect the regularity and dynamics of vehicle transmission wear through random processes to achieve the purpose of test prediction . [0003] Based on the wear mechanism, the present invention uses the partial least squares regression method (Partial Least-Squares, PLS) to establish a prediction model of the wear rate of the slide block by using the wear orthogonal test data of the slide block of the vehicle transmission device. The life test determines the distribution type of the wear random process variable, establishes the basic model of the wear random process of the vehicle transmission, and clarifies the method of determining the model parameters...

AI Technical Summary

Problems solved by technology

In traditional multiple regression analysis, the multiple correlations of variables will seriously damage the parameter estimation, increase the error of the model, and reduce the stability, especially in the mechanical wear system of the transmission device, where the working conditions and environment are complex and have an impact on component wear There are many factors, and there are serious multiple correlations among the factor variables

[0010] The determination of the wear process lacks theoretical support, especially in the establishment of related stochastic process models, and there are very few related studies, and there is no relevant research and analysis on the wear problem of the slider of the vehicle transmission. The method does not fully consider the interaction between multiple factors and multiple levels, which is not enough to support the establishment of the model, and does not have engineering application value

Images