Multi-Scale ICME to Optimize the Bondline Performance of Adhesive Joints

An Integrated Computational Materials Engineering framework to optimize the performance of an adhesively bonded single leg lap shear joint was developed and demonstrated. Integrated imaging, experimental testing, sensitivity analysis, and uncertainty quantification informed model and methodology development saving time and effort by identifying the most influential parameters for full characterization, identifying essential experimental testing needed for validation, and informing the development of advanced methodology when complex physical behavior was observed. While developed for a relatively simplistic example, this framework can be readily enhanced for more complex scenarios, because it is based on separate computational methods at each scale. This approach allowed for the best computational solution at each scale with non-deterministic parameter sets passed between scales. A final reduced dimension surrogate model is produced for one or more design objectives for a fast-running surrogate model to rapidly explore the design space that is limited to the most influential parameters. The reduced dimension surrogate model is developed from high-fidelity, validated computational models. These models are available for advanced use such as identifying processing flaws (e.g. void size and shape) and locations that should be considered during inspection and NDE as well as how fast such a flaw will propagate to determine inspection schedule.