Tuesday  |  Conference Center B  |  10:20 AM–10:40 AM

#13477, Determination of Adhesion Strength through Computationally Informed Blister Tests

Adhesively bonded aluminum joints are widely used in multiple engineering sectors, especially the automotive industry. In long-term service, environmental conditions and stresses caused by factors such as internal deformation, residual stresses, or corrosion, cause the degradation of adhesion strength. Measuring the adhesion strength is a complex task. While many different adhesion tests are available, the BT is the only one that provides a measure of the normalized energy required to initiate and propagate a crack at the polymer/substrate interface. This test is performed by pressurizing a fluid through a hole in the substrate, which causes blistering on the polymer, ultimately leading to complete delamination. In this work, the adhesion strength is determined using a BT along with a Digital Image Correlation (DIC) system and inverse computational optimization analysis using the Hypercomplex-variable Finite element method (ZFEM). ZFEM enables the calculation of accurate arbitrary order sensitivities of the nodal displacements with respect to the cohesive fracture parameters. These sensitivities are provided as a byproduct of the analysis using the hypercomplex Taylor series expansion method. The sensitivity information is useful for inversely determining the cohesive fracture parameters from experimental or synthetic data using a finite element-based approach. Our results show that the performance of the gradient-based optimization scheme fed with ZFEM derivatives was shown to be superior to a gradient-free method, both in terms of runtime and relative error. Thus, ZFEM derivatives can be used to improve the efficiency of gradient-based algorithms used for optimizing the design of mechanical systems.

David Restrepo   The University of Texas at San Antonio
Drishya Dahal   The University of Texas at San Antonio
Brendy Rincon   The University of Texas at San Antonio