Definition of correlations between automotive test environments through mechanical fatigue damage approaches

Abstract

Abstract

The aim of validating the global endurance of a mechanical system is to make sure that it shows no signs of failure during series of specific tests. The latter are carried out in more severe environments than those to which the system is subjected during normal use. The application described here concerns ground vehicles, and the study proposes methods and tools for optimizing and organizing tests. This optimization is based on correlations between different existing test environments. Two main methods of correlation are proposed and developed: correlation based on mechanical loadings and correlation based on mechanical damage. These methods respectively permit analysis of the mechanical loadings that a system undergoes during tests and prediction of the consequences of these loadings. Both analyses help to reduce the duration and number of tests. The analysis of measured loadings appearing in a mechanical system consists in knowing the statistical contents of load signals according to various forms (global statistical distribution, local statistical distribution, distribution of extreme values, dynamic signal behaviour, Markov matrix). Analysis of the consequences of loadings consists in building a spectrum of robustness, a spectrum of mechanical damage by fatigue, and a matrix of mechanical damage. Tools stemming from these analyses permit better evaluation of fatigue life expectancy and better definition of trial environments for endurance testing with an overall severity equivalent to that applied during normal use of the system.