digilib@itb.ac.id +62 812 2508 8800

ABSTRAK Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

COVER Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

BAB1 Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

BAB2 Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

BAB4 Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

BAB5 Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

PUSTAKA Prayoga
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

Green technology in transportation has driven the electric vehicles (EVs) growth in last 5 years. 10 million EVs will have been grounded by the end of 2020 in 2020. In Indonesia, 1900 EVs sold in the first semester of 2021. But the utilization of Lithium-Ion batteries in EVs would cause fire due to low voltage and post-crash accidents such as side pole impact. Therefore, safe and lightweight battery protection is needed in EVs. It could be achieved by optimizing the cross-car beam and chassis system geometry and material selection. Cross car beam maintains the underbody car structure due to side impact. The 254-diameter rigid pole is modelled based on FMVSS 214. The stationary rigid pole receives kinetic energy of the vehicle at 32 km/h. The chassis system model is based on the electric vehicle chassis. Various configurations of the structure are modelled based on the data sampling from Latin Hypercube Sampling (LHS) method. Then, the result will be simulated by using Finite Element Method (FEM) Specific Energy Absorption (SEA) and battery stress of the model. Machine Learning will predict the optimum design with maximum SEA and minimum battery stress such as the utilization of Artificial Neural Network (ANN) for model approximation and Nondominated Sorting Genetic Algorithm-II (NSGA-II) for variable optimization. Then, the TOPSIS algorithm offers a decision-making procedure for determining the optimal configuration. This research will be varied the cross-car beam geometry (Top Hat, Closed Top Hat, Rectangular) and material (Carbon Steel, Aluminum 6061-T6, and Aluminum 5012) as the input for optimization using machine learning. The chassis also optimized by variate material selection and thickness. The optimization process resulted in the maximum Specific Energy Absorption and battery stress are 4.06 kJ/kg and 39 MPa with Top Hat configuration and aluminum 6061-T6 as material selection. Then, numerical simulation was conducted to validate the result. As a result, the error between numerical simulation and approximation results are 8.77% for SEA and 3.88% for battery stress that indicates the approximation model can predict the optimal result well. The optimal result increases SEA 180.35% and decreases battery stress 20.39%.