Crashworthiness Optimisation of DP980/CFRP Hybrid Battery Pack Bottom Shell under Lateral Impact: A Data-Driven Approach Integrating Machine Learning Surrogates and Multi Criteria Decision Making

Electric Vehicles Are Redefining Modern Transportation. As the Global Fleet Surpasses 57 Million Units, Battery Safety Has Become a Critical Engineering Priority. A Single Lateral Impact Can Compress Battery Modules Beyond 20 MPa. Above That Threshold, Internal Short Circuits and Thermal Runaway Begin. This Thesis Attacks the Problem at Its Structural Root. It Presents the First Systematic Study of DP980 High-Strength Steel Paired with Unidirectional Carbon Fibre-Reinforced Polymer as a Hybrid Bottom Shell for EV Battery Packs under Lateral Bar Impact. Two Laminate Topologies Are Examined. In the SC-Type, DP980 Forms the Outer Layer. In the CS-Type, CFRP Takes the Outer Position and DP980 Serves as the Inner Liner.The Research Is Built on 258 Abaqus/Explicit Finite Element Simulations Across Five Impact Velocities and Two Impact Positions, Generated by Latin Hypercube Sampling. Eight Machine Learning Surrogate Models Are Trained on This Dataset. The MPA-SVR Architecture Emerges as the Top Performer. It Achieves an R² of .9941 for Energy Absorption and .9980 for Peak Deformation. Three Multi-Criteria Decision Methods Converge Unanimously on the Same Answer. TOPSIS, Grey Relational Analysis, and Hybrid GRA-TOPSIS All Select the CS-Type as the Primary Optimal Design. Its Specification Is a 1.5 mm CFRP Outer Layer Paired with a .6 mm DP980 Inner Liner. Against the Monolithic DP980 Baseline, This Delivers a 10.94% Gain in Specific Energy Absorption, a 3.69% Reduction in Deformation, and an 11.82% Reduction in Structural Mass. Two Independent Loading Scenarios Confirm the Result.