Failure strength prediction for adhesively bonded single lap joints

Niat Mahmud Rahman, Purdue University

Abstract

For adhesively bonded joint, failure strength depends on many factors such as material properties (both adhesive and adherend), specimen geometries, test environments, surface preparation procedures, etc. Failure occurs inside constitutive materials or along joint interfaces. Based on location, adhesively bonded failure mode can be classified as adhesive failure mode, cohesive failure mode and adherend failure mode. Failure mode directly affects the failure strength of joint. For last eight decades, researchers have developed analytical, empirical or semi-empirical methods capable of predicting failure strength for adhesively bonded joints generating either cohesive failure or adherend failure. Applicability of most of the methods is limited to particular cases. In this research, different failure modes for single lap joints (SLJs) were generated experimentally using epoxy based paste adhesive. Based on experimental data and analytical study, simplified failure prediction methods were developed for each failure mode. For adhesive failure mode, it is observed that peel stress distributions concur along interface near crack initiation points. All SLJs for this test endured consistent surface treatments. Geometric parameters of the joints were varied to study their effect on failure strength. Peel stress distributions were calculated using finite analysis (FEA). Based on peel stress distribution near crack initiation point, a failure model is proposed. Numerous analytical, empirical and semi-empirical models are available for predicting failure strengths of SLJs generating cohesive failures. However, most of the methods in the literature failed to capture failure behavior of SLJs having thickness of adhesive layer as variable. Cohesive failure mode was generated experimentally using aluminum as adherend and epoxy adhesive considering thickness of adhesive layers as variable within SLJs. Comparative study was performed among various methods. It was observed that experimentally measured crack tip opening angles (CTOA) using DCB tests and CTOAs extracted from FEA using SLJs at failure loads follow similar trend when adhesive thickness was considered as variable. A factor was proposed to accommodate differences. Proposed method was also adopted for SLJs with overlap length as variable. Reasons for discrepancies were identified for such case. For generating adherend failure, graphite fiber epoxy composite was used as adherend. To avoid singularity, an area based approach was adopted. It is considered that failure within composite adherend occurs when size of the damaged zone reaches a critical value. This value was determined using test data and then FEA of SLJ. Using this hypothesis, various strength of material based failure models were tested for SLJs with different geometries. According to the study, Azzi-Tsai model works well for SLJs having different geometric parameters. Only composite failure parallel to interface and within the first ply was considered here.

Degree

Ph.D.

Advisors

Sun, Purdue University.

Subject Area

Aerospace engineering|Mechanical engineering|Materials science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS