Perpustakaan Digital ITB

In modern civilization, as it is today, transportation takes a significant role in human life. As modes of transportation become more diverse and more parties to use and develop them, people increasingly pay attention to the safety of transportation modes. Various kinds of efforts have been given by engineers and policyholders to achieve these goals, some of which regulate the use of seat belts, speed limits, driving attitudes, and road safety. Some of the rules and standards also control the specifications of the vehicle structure for some cases of accidents/collisions that aim to protect the driver/passenger that commonly referred to as crashworthiness. The energy absorbing system is one way to increase safety according to the crashworthiness principle, one of the energy absorbing systems that is commonly used is the crash box. One of the crash box appealing points is how it can be designed and utilized to meet the safety criteria, such as configuring its thickness, perimeter, and the number of crash box corners. In this paper, the design exploration and optimization of crash box design was performed under both axial and oblique loading via GPR model and Bayesian optimization. The optimization process considers thickness, perimeter, and the value of a, b, and d as the side length for the 20-corner crash box. Bayesian optimization is then performed to fine-tune the geometry of the 20-corner design with fixed perimeter and thickness of each crashworthiness characteristic from their best optimum geometry. It is found that geometry with thickness 3 mm and perimeter 400 mm is the most optimum geometry for the mean crushing force (Pm) with the value of a = 26.771 mm, and b = 13.319 mm is the most optimum mean crushing force (Pm) geometry dealing with the axial load. Hence, the geometry that has the value of a = 26.474 mm, and b = 19.474 is the most optimum Pm geometry for the oblique loading. On the other hand, geometry with a thickness of 1 mm and perimeter of 120 mm is the most optimum geometry for the specific energy absorption (SEA) with the value of a = 9.986 mm, and b = 3.351 mm is the most optimum SEA geometry dealing with the axial load. Hence, the geometry that has the value of a = 8.038 mm, and b = 2.781 is the most optimum SEA geometry for the oblique loading. Crushing force efficiency (CFE) cannot be analyzed during the oblique loading phenomena, because the value of peak force and mean crushing force will be the same for the angle above 10° on the oblique simulation. Whereas, geometry with thickness 3 mm and perimeter 120 mm is the most optimum geometry for the crushing force efficiency (CFE) with the value of a = 10.556 mm, and b = 3.556 mm is the most optimum CFE geometry dealing with the axial load.