How adhesives should flow during manufacturing

Abstract

This thesis proposes a methodology to optimize the application pattern of adhesively bonded joints, so that the liquid adhesive automatically flows into the desired bond shape during manufacturing — leaving no spots over- or underfilled, thus the adhesive is used in the most efficient way. Besides preventing adhesive excess and waste that eliminates the need for cleaning and reworking steps after joining, this optimization approach ultimately leads to a more ecological and economic adhesive manufacturing processes.

To achieve this goal, an optimization algorithm is developed that iteratively modifies the shape of adhesive application pattern until it flows into the desired bond shape. Established methods from topology optimisation for solids are used, re-interpreted, and applied in the context of fluid mechanics in a new way. At each iteration, the resulting distribution obtained after squeezing an initial adhesive application pattern is calculated. The resulting distribution is then compared to the desired adhesive distribution, and overflown adhesive is traced back to its origin in the initial pattern (using the calculated velocity and direction). These overflown areas are then removed from the initial pattern and the iteration loop continues until the resulting distribution matches the desired bond geometry within an acceptable tolerance.

This optimization strategy relies on the accurate simulation of squeeze flow processes for which a special numerical model was chosen. This model was extensively validated in a comprehensive experimental study in advance, to prove its suitability. By combining the Hele-Shaw approach, a method to investigate squeeze flow processes, with Particle Image Velocimetry, a method to measure fluid flows, it is demonstrated that the simulation model is sufficiently accurate and can be implemented to calculate the quantities that are required by the optimization algorithm. Further, the surface structure has proven to be of secondary importance and doesn't need to be considered in the optimization.

The algorithm was applied to optimize a variety of different application patterns. It was found that the optimized shapes have recurring shapes such as pointy fingers which flow into corners and bays that flow into straight edges. Furthermore, quantities like the degree of compression, number of cells and considered flow-laws were varied to investigate their influence on the algorithms results. However, it was found that the algorithm is mesh and flow-law independent, which highly improves its applicability and simplicity for practical use. It is capable to identify optimal patterns for frequently used adhesive layer thicknesses and degrees of compression.

Finally, the optimized patterns obtained by theoretical analysis were validated through real world experiments, that confirmed their ability to flow into the desired shapes. Overall, this thesis presents an innovative approach to optimize adhesive application patterns. Due to the performed optimization, the environmentally questionable adhesive waste and costly post-cleaning work can be eliminated — thus, a more ecological and economical manufacturing process is achieved.