Topology Design of Vehicle Structures for Crashworthiness Using Variable Design Time

Prasad Tapkir, Purdue University


The passenger safety is one of the most important factors in the automotive industries. At the same time, in order to improve the overall efficiency of passenger cars, lightweight structures are preferred while designing the vehicle structures. Among various structural optimization techniques, topology optimization techniques are usually preferred to address the issue of crashworthiness. The hybrid cellular automaton (HCA) is a truly nonlinear explicit topology design method developed for obtaining conceptual designs of crashworthy vehicle components. In comparison to linear implicit methods, such as equivalent static loads, and partially nonlinear implicit methods, the HCA method fully captures all the relevant aspect of a fully nonlinear, transient dynamic crash simulation. Traditionally, the focus of the HCA method has been on designing load paths in the crash component that increase the uniform internal energy absorption ability; thus far, other relevant crashworthiness indicators such as peak crushing force and displacement have been less studied. The objective of this research is to extend the HCA method to synthesize load paths to obtain the different acceleration-displacement profiles, which allow reduced peak crushing force as well as reduced penetration during a crash event. To achieve this goal, this work introduces the concept of achieving uniform energy distribution at variable design simulation times. In the proposed work, the design time is used as a new design parameter in topology optimization. The desired volume fraction of the final design and the design time provided two dimensional design space for topology optimization, which is followed by the formulation of design of experiments (DOEs). The nonlinear analyses of the corresponding DOEs are performed using nonlinear explicit code LS-DYNA, which is followed by topology synthesis in HCA. The performance of the resulting structures showed that the short design times lead to design obtained by linear optimizers, while long simulation times lead to designs obtained by the traditional HCA method. To achieve the target crucial crash responses such as maximum acceleration and maximum displacement of the structure under the dynamic load, the geological predictor has been implemented. The concept of design time is further developed to improve structural performance of a vehicle component under the multiple loads using the method of multi-design time. Finally, the design time is implemented to generated merged designs by performing binary operations on topology-optimized designs. Numerical example of the simplified front frame is utilized to demonstrate the capabilities of the proposed approach.




Tovar, Purdue University.

Subject Area

Mechanical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server