The motivation to improve the lattice structure core in the battery protection
system is due to some drawbacks in the previous design, which are considered
not optimum. This structure is used as the protection system of the Li-
ion battery, which is known as the primary power source in the electric
vehicle. Li-ion battery has a high risk of fire if deformed too much due
to high energy density; therefore, protection is needed. Previous research
has been done to protect the battery from ground impact, which may cause
battery deformation, using a lattice structure. However, it is still too heavy
and can be improved to achieve a lighter structure without compromising
the battery's safety.
In this research, the optimization of lattice structure parameters is done
using 2 machine learning algorithms: artificial neural network and non-
dominated sorting genetic algorithm II. Furthermore, the optimum parame-
ters are selected using TOPSIS. The ANN is used to model and predict the
mass and battery deformation, while NSGA-II is used to find the optimum
lattice shape parameter. Combining the machine learning algorithms and
TOPSIS, it is found that the optimum structure shaped is BCC-Z, with the
porosity of 21.2%, A/B and H/C ratio of 0.7 and 1.17, built using Ti-6Al-4V
material and stacked 2 times. This configuration results in a mass reduction
of 17% compared to the existing design while maintaining the battery safety
performance in excessive deformation, which may lead to short circuits and
potential fire.