These slides were presented at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

The growing shift towards electric vehicles and the imperative to reduce CO2 emissions are driving the automotive industry towards lightweight vehicles. For both gasoline and hybrid vehicles, decreasing weight significantly enhances fuel efficiency and lowers CO2 emissions. In electric vehicles, weight reduction is especially important for extending driving range. The vehicle body contributes around 40% of the total weight, and the use of aluminum profiles and castings is increasing to help reduce this weight. However, bonding these components effectively is challenging because traditional adhesives are inadequate for such applications. As a solution, innovative adhesives have been formulated specifically for these needs. These adhesives are injected into joint gaps, flowing along designated channels that cannot be sealed laterally due to necessary joining tolerances. They are engineered to flow within these unsealed channels without intruding into adjacent gaps. A key property of these adhesives is their special rheological behavior, which is essential for achieving the desired bonding performance. In this context, simulation is crucial for predicting adhesive behavior, including its spreading time and minimizing unwanted intrusion into lateral gaps. However, simulating adhesive injection presents several challenges, such as accurately defining non-Newtonian and temperature-dependent viscosity, managing transient free surface flow in complex geometries, and addressing heat transfer between adhesive and surrounding walls. To tackle these challenges, a Moving Particle Simulation model has been developed and validated. Moving Particle Simulation (MPS) is a meshless CFD technique designed to solve the Navier-Stokes equations. It was initially created for transient free surface flows and employs a fully implicit solver for simulating highly viscous non-Newtonian materials. MPS can handle heat transfer and accounts for viscosity changes with temperature, making it well-suited for simulating adhesive behaviors. The MPS model has been developed using the same geometry and conditions as a test channel located at the SIKA AUTOMOTIVE AG facilities. This model simulates the flow within the main channel, and two key outputs are compared with experimental measurements: the time taken to reach a specified distance from the injection point and the extent of intrusion into lateral gaps. Simulation and experimental tests are conducted for various channel geometries and different heights of the lateral gaps. The analysis of simulation outcomes and measurements shows that MPS can reliably forecast the adhesive injection process and how it spreads, as long as the characteristics of the fluid are correctly specified and the impact of fluid temperature on viscosity is precisely calculated throughout the adhesive route. The research pointed out that the simulation of real car components could require extensive computational time. However, this issue can be mitigated by the MPS approach's high efficiency in utilizing multi-GPU computing. The paper presents the Moving Particle Simulation technique, the model based on the SIKA AUTOMOTIVE AG geometry and conditions, along with a qualitative and quantitative analysis of the simulation outcomes compared to experimental data for three test scenarios.