适用于任意应变比的应变寿命新模型

A New Strain-Life Model Accounting for Effects of Various Strain Ratios

关键词:  发动机轮盘  疲劳寿命  应变比  多元线性回归  体积效应

Key words:  Aero-engine disk  Fatigue life  Strain ratio  Multiple linear regression  Volume effect

摘要:

为了提高非对称应变寿命模型预估轮盘疲劳寿命的精度，通过对不同应变比下的试验数据进行多元线性拟合，建立了一种适用于试验数据应变比覆盖范围内任意应变比条件下的应变寿命新模型。结合多种材料的疲劳试验结果，对包含本模型在内的四种常用模型预测结果进行了对比分析。结果显示Morrow弹性修正模型和Morrow-Harford模型的预测结果不稳定；SWT模型预测精度一般，但稳定性较好。本文新模型对CC450不锈钢和SAE1045钢的预测结果基本在2倍分散带内；对GH4133光滑试棒的预估中位寿命比试验中位寿命仅高0.25%，对GH4133中心孔拉伸试件的预估中位寿命比试验中位寿命小3.67%，展示出很好的考虑应变比影响的疲劳寿命预测能力。

Abstract:

The strain-life fatigue models with mean strain effect are often employed to estimate fatigue lives of disks in aero-engine， however， the prediction accuracy of fatigue-life models commonly used are unsatisfactory under certain conditions. Therefore， a new strain-life fatigue model was proposed to accounting for the effect of any strain ratio among the strain ratios of fatigue test data sets， and the parameters in the new model can be determined by fatigue test data with various strain ratios through multiple linear regression. The estimated results of four commonly used fatigue life models， including the new model proposed in this study， were compared with the fatigue test data of CC450 stainless steel， SAE 1045 steel and standard bars and tension specimens with center-hole of GH4133 superalloy. It was found that the prediction accuracy of both the Morrow mean stress correction model and the Morrow-Harford model were unstable for different materials， while the SWT model showed acceptable accuracy with stable predictive capability. By contrast， the new model proposed in this paper showed superior accuracy to any other candidate mean strain relationships. The estimated lives of CC450 stainless steel and SAE1045 steel were basically within the double scatter band， and the results of the standard bars of GH4133 superalloy was merely 0.25% higher than fatigue test data; while the estimated life of tension specimens with center-hole was just 3.76% less than the experimental fatigue life. Thus， this new model showed satisfactory capability to estimate fatigue life with stain ratio effect for different materials.