Evaluation of kerf as a crack arresting method

The study presents research work using kerf as a crack arresting method in Fracture Mechanics analysis. Kerf is a slit notch cut by machining process and it can emulate the crack interaction and able to lower the Stress Intensity Factor (SIF) by disturbing the stress flow near the crack tip. The main objective of this thesis is to evaluate the kerf as a crack arresting method using Finite Element Analysis (FEA) software. An experiment using fatigue test machine was conducted to verify the finding. Parametric studies were performed using FE model with various geometries consist of kerf to crack ratio (b/a), crack to width ratio (a/W) and distance between kerf-crack to width ratio (d/W). The FE model involves a specimen that contains an edge crack without any kerf and a specimen which contains an edge crack together with four different configurations of kerf. Four types of configuration were used which are single edge kerf, double edge kerf, single central kerf and double central kerf. The configuration was chosen based on the estimation of best kerf’s position to disturb the stress distribution near the crack tip. The results of FEA are based on percentage reduction of the SIF compared to the FE model without kerf. Results show that the best configuration of kerf is double edge kerf which produces maximum SIF reduction of 34.3%. Additional FE models which focused on double edge kerfs were further performed to obtain accurate results. It can be concluded that maximum SIF reduction occurs when the d/W=0.15, b/a=0.9 and a/W=0.15. Mathematical equation to represent non-linear graph of maximum SIF reduction with respect to crack length to width ratio is then developed using curve fit method. Mathematical equation based on Hoerl regression model is best fit for this non-linear model. Experimental results based on the best geometries showed an improvement compared to specimen without kerfs (10800 cycles). The results show that the double edge kerf specimen with the longest kerf (b/a= 0.9) yield the best result (34900 cycles). Slight modification to the specimen was performed by introducing a stop hole at the kerf tip due to new crack initiation occurring at the kerf tip. The best result for stop-hole is better than the no stop-hole design with 88000 cycles when b/a= 0.9.