Research On Fretting Fatigue Life Reliability Of Dovetail Joint

Fretting fatigue is one of the main failures in the aeroengine dovetail joint. It is important to do the fretting fatigue life assessment during dovetail joint design. The contact area is in a complex state of stress under fretting conditions. It is more reasonable and accurate to predict fretting fatigue life using multiaxial fatigue models. In addition, there is certain randomness in the material properties, geometry etc of the dovetail joint. The fretting fatigue life is also scattered and its dispersion is much larger than static strength's. It is significant in theory and enginerring to develop a fretting fatigue life reliability analysis method for dovetail joint.In this paper, a brief introduction on the current fretting fatigue life prediction and reliability analysis method was represented. A fretting fatigue life reliability analysis method was devopled for dovetail joint.Fisrtly, based on multiaxial fatigue theory, three multiaxial fatigue life models were established and analyzed. The better parameters, SWT and CCB, in predicting the fretting fatigue life under LCF and HCF load were chosen respectively to setup a mixed fretting fatigue life model on the base of linear cumulative damage theory. In this model, CCB parameter was modified considering stress ratio effect. And the model can be used to predict the fretting fatigue life under HCF-LCF load. Compared to the experiment, it was shown that the model is better, and the maxminum predicting error is 22.2%.Secondly, elastic modulus, friction coefficient and two coefficients in the life prediction model were regarded as random variables. Combined with the response surface method, fretting fatigue life samples were obtained by Monte-Carlo simulation. The probability-weighted moment method was used to get the life reliability model. The reliability life model was established under HCF-LCF load on the base of on probability Miner rule. Compared to the experiment, the result showed that fretting fatigue life under LCF or HCF load subjects to three-parameter Weibull distribution within the confidence level of 99%, and fretting fatigue life under HCF-LCF load subjects to lognormal distribution.In further, on the basis of sensitivity analysis on above four random variables, material constant bL , bH and elastic modulus, which have a greater impact on fretting fatigue life, were selected to establish simplified probability life prediction models.Finally, as for material constant bL , bH and friction coefficient, different coefficient of variation and distribution were considered to study their impact on the reliable life.