Optimisation of car crash box finite element model for crashworthiness analysis using model updating and factorial design method / Noor Am Zura Binti Abdullah

Faculty of Mechanical and Automotive Engineering Technology

Thesis (Doctor of Philosophy) -- Universiti Malaysia Pahang – 2023

Includes bibliographical references

Finite element modeling and analysis are widely used method for simulating the structural behaviour of a system in order to provide information of the structure under various loading conditions. However, the constructed model not always accurate due to several factors such as simplifications in modeling, material properties, or uncertainties in boundary conditions. The model updating method is an optimisation technique that aims to improve the accuracy of finite element models by incorporating dynamic response data obtained from experimental testing. In the field of crashworthiness, the complexity of modelling can lead to inaccuracies, particularly when incorporating structural enhancements. This study investigates the reliability of model updating methods in reducing inaccuracies in finite element models used for crashworthiness analysis. The primary objective of this research project is to assess the performance of updated finite element models in providing accurate crash data for a car crash box structure. The updated model was evaluated for crash analysis, and its accuracy was compared with experimental crash test data. For comparison, response optimisation using factorial design was also conducted to obtain an optimised model with minimal error in crashworthiness analysis. The results of this study show that the best initial correlation with the welded crash box structure was achieved using a one-dimensional bar element (CBAR) with an average error of 5.3%. For the bolted crash box, the best correlation was achieved using a joint model that combined the usage of CBAR and a rigid element (RBE), labeled as BOLTED2, with an average error of 4.9%. After applying model updating, the SOLID model showed the best correlation for the welded specimen with an average error of 4.5%, while the BOLTED2 model remained the best for bolted specimens with a reduced average error of 4.7%. However, the investigation of the updated model's performance did not show a significant improvement for crashworthiness analysis. The optimisation study yielded satisfactory results, with the crash output parameters showing closer values to the experimentally obtained data. The outcomes of this project will contribute to the field of automotive engineering and, in particular, to vehicle manufacturers' efforts to produce safer vehicles that meet consumers' demands. This study emphasizes the need to improve computational simulation and analysis methods to generate more accurate predictions and minimise the problem of model discrepancies that arise during computational analysis. In addition, the study on model updating methods for crashworthiness analysis of car crash box structures has several potential applications in the field of automotive engineering and safety.