An Evaluation of Elastic-Plastic Methods Applied to Crack Growth Resistance Measurements

Abstract

Information from tests on blunt notched specimens for J-integral calibration by conventional J = -1/B ∂U/∂a analysis is used to evaluate the significance of plastic zone adjustment to physical crack length in crack growth resistance, KR, calculations. Secants are drawn to load versus displacement test records to determine plastic zone adjusted crack lengths. Tests on three specimen geometries [compact specimen (CS), single-edge notched bend (SENB), and center-notched tension (CNT)] and on two materials (HY130 steel and 2024 aluminum) have shown that this procedure develops KR values that are equivalent to J. This demonstration opens possibilities that J can now be applied to cases where there is subcritical crack growth such as in R-curve work, KIscc, and possibly in creep cracking studies. Also, this provides a simplified method for computing J experimentally on complex geometries for which elastic KI solutions are available. Alternative J computational procedures are compared. These include J by a Ramberg-Osgood work-hardening law fit to load-displacement records, and J by area approximation methods. The Ramberg-Osgood modeling appeared to work reasonably well in tests on compact specimens but was found to be unreliable on SENB and CNT specimens and therefore is not recommended. With no stable crack propagation, the area approximation procedures for J determination produce reasonably accurate estimates of J as might have been anticipated from past experience. Tests on the compact specimen geometry required a Merkle-Corten correction procedure which worked well on large crack aspect ratios, where a/w ⩾ 0.5, but tended to overcorrect for short cracks, giving nonconservative results.