Modeling of Residual Stress Fields and Their Effects on Fatigue Crack Growth

Abstract

Fatigue Crack Growth (FCG) is a deleterious physical phenomenon in engineering materials, which is intensified by the presence of tensile Residual Stress Fields (RSF), while compressive RSF has been shown to delay the FCG phenomenon. However, several challenges make it difficult to fully incorporate the beneficial effects of compressive RSF into the design process in aerospace and other engineering industries. As such, this study is designed to understand and quantify the effects of RSF on the FCG phenomenon in thick aluminum alloy specimens. Experimental studies were conducted on specimens made of 7050-T7451 aluminum alloy in order to obtain the material properties required for a FCG model. In addition, FCG tests on Single-Edge Notched Tension (SENT) specimens with well-defined RSF were conducted for the verification of the FCG model. Finite Element Analysis (FEA) software (ABAQUS™) was used to simulate the FCG in RSF, and to analyze the redistribution of RSF due to FCG. ABAQUS was used first to introduce a RSF through a well-controlled four-point bending simulation, which was set as an initial condition to the FCG simulation. Several FCG test simulations were conducted to evaluate the crack closure and plastic wake effects on FCG. As part of these simulations, three test cases were considered: a large stress ratio (R = 0.7), a low stress ratio (R = 0.05) and a negative stress ratio (R = -1). For test case 1 (R = 0.7), the calculation of the FCG rate shows no indication of the crack growth retardation due to the presence of the compressive RSF. However, the FCG rate was retarded in the test cases with a low and a negative stress ratios (R = 0.05 and -1). The plastic deformation induced at the crack tip due to FCG decreased under a positive effect of the compressive RSF. In addition, the crack closure and plastic wake effects had a greater influence in the test case with a low stress ratio (R = 0.05), while their possible effects were not significant at R = 0.7 and -1, as demonstrated by the experimentally verified numerical model presented in this manuscript.