Modelling of contacts in Adams flex bodies using Mamba

Abstract: Simulations are a powerful tool used to reduce development time and cost in the vehicle industry. However, the models used in simulations are simplifications of reality, and there are commonly contradicting requirements between accuracy and computational efficiency. Vehicles are constructed with a large number of parts joined together with bolts, welds and other joints. At these locations and other places where different surfaces come in contact, the contact can have a considerable effect on the dynamic behaviour of the part and capturing this effect in simulations can be difficult because of their nonlinear nature.  This thesis aims to evaluate different methods for simulating contacts and their effects under dynamic conditions. The thesis is performed at Scania, one of the top commercial vehicle manufacturer, which aims to increase the use of simulations in the design of their vehicles to have better designs with lower cost and time investments.  This thesis uses the multibody simulation software Adams Car to evaluate two different contact simulation methods, one developed by Adams and the other is the softwareMamba developed by Magna. Mamba defines special modes called Joint-interface modes, which capture the deformation in the vicinity of the contact. The contact simulations are compared with some of the existing simulation methods and physical test data by implementing it on cases where contact plays an important role in dynamic behaviour. Two such cases are identified as the leaf spring and the frame-subframe assembly. For the leaf spring example, the stiffness and energy dissipation were compared for Mamba simulations with the Adams built-in multi-beam model and physical test data. The stiffness with the Mamba contact model better matched the test data, but the energy dissipation was better modelled with the Adams leaf spring model. For the frame-subframe assembly, the effect of the Mamba contact modes was evaluated for random vibration tests by comparing acceleration Power Spectral Density (PSD), relative displacement and Operating Deflection Shape (ODS) analysis. The Adams contact model was also implemented for this case but could not converge to a solution. Improvement in accuracy was observed with Mamba contact simulations compared to simulations which ignored contact, but with the drawback of significantly higher simulation times.