Development of a generic morphable motorcycle model for crash analysis

Abstract

There is currently a lack of openly available finite element motorcycle (MC) models suitable for crash analysis. This thesis has aimed to address this problem by creating a generic, morphable MC that can be adapted to geometrically represent a range of commonly encountered motorcycles on the road and reproduce their characteristic behaviour during crash simulations. Additionally, the goal is to make the model open source, allowing users to access, morph, modify, and share it freely. As part of this project, a base model was morphed into four different MC presets: the Harley-Davidson Sport Glide, BMW GS1300, Honda CB125F, and KTM Duke 390. These models were calibrated against physical test data obtained from frontal impacts with a rigid wall (RW), this was done by tuning the structural stiffness of the components and by tuning the spring-damper characteristics of the front suspension. In the physical tests, the RW was angled at 10° for the Harley-Davidson Sport Glide, BMW GS1300, and Honda CB125F, and at 0° for the KTM Duke 390. All MCs travelled at 40 km/h upon impact. After the calibration was completed, the KTM Duke 390 model was validated against a physical test of a KTM Duke 390 with a post-mortem human subject (PMHS) impacting the side of a Honda Accord. This was replicated in the simulation using a human body model (HBM), a Honda Accord car model (National Highway Traffic Safety Administration, n.d.), and the morphed KTM Duke 390 model. In the physical test, the car was angled at 30°, and the motorcycle had a velocity of 50 km/h upon impact (Carrol, Bolte, 2024). The calibration results show that, in many cases, the overall crash response of the bike is captured. Moreover, the validation shows good agreement with the physical test data, where the resultant behaviour of the MC matches the physical test. The results from the calibration and validation show that a simplified model, starting from a single generic model, can be morphed into specific MC models that replicate important behaviours crucial for crash outcomes and enable MC crash analysis using human surrogates.