Worst-Case Topology Optimization

This paper presents a worst-case design approach for the multidisciplinary topology optimization of an automotive hood design. The study considers the impact of a pedestrian’s head against the hood, static loads, and the minimum weight of the hood – all required to meet general design code requirements in automobile industry. Among the design code requirements of the hood design, the biggest challenge is to handle hundreds of head impact locations specified in the Euro NCAP pedestrian testing protocol, due to the high computational expense of hundreds and thousands of structural analyses demanded in the structural optimization. To overcome this challenge, we accordingly introduce a general framework for the worst-case topology optimization which investigates the worst impact locations on the hood by evaluating the maximum head injury criterion and the maximum deflection of the hood separately, reducing the burden to consider multiple disciplines simultaneously at hundreds of impact locations all at once. At the end, these worst impact locations are combined with a static load case and formulated into a single multidisciplinary design optimization problem that needs only tens of structural analyses per iteration for numerical gradients computation, enabling the proposed design framework suitable for large scale topology optimization problems.