KNOWLEDGE BASED DESIGN OF EXPANSION TUBE-AXIAL SPLITTING IMPACT ABSORBING MODULE

• Main Author: Nugroho Aji, Hanif
• Format: Theses
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/52257
• Institution: Institut Teknologi Bandung

Vehicle crash are a common occurence in Indonesia. The consequence of such accident can be reduced by using crashworthiness. This concept can be applied on various kind of vehicles. In response to the rapid development of railway industry and planned regulation that will make crashworthiness compulsory, this research was aimed toward applying crashwothiness to railway vehicles. Impact energy absorber is one of the most important component for crashworthiness implementation. One of the best type of impact absorber module is expansion tube-axial splitting because of its high specific energy absorbed. Using finite element method to evaluate its performance in order to optimize the design of this component usually takes time, so an effective and efficient method is needed. In order to reduce design time and improve accuracy of the optimization process, this research used a simple metamodel created from a database that contains the characteristics of expansion tube-axial splitting module. This research was conducted by creating the database once and applying it to design impact energy absorber every time it is needed. Database was created using input from Latin Hypercube Sampling and output from finite element method. Then metamodel was created from the basis data using Kriging method and validated using LOO-CV. Finite element model used as the basis of the metamodel has been validated using field experiment with 13.15% peak force and 6.94% mean force error. From the metamodelling, 13.46% accuracy has been reached. Design application was conducted with single objective function to maximize energy absorption using Genetic Algorithm for 2 case studies, which are K1 train and Metro Kapsul. The optimum solution of these cases (inner diameter, thickness, and final diameter of expander) were: 204 mm, 6.5 mm, 224 mm for K1 train; and 92,04 mm, 2,9 mm, 101 mm for Metro Kapsul. Compared with the results of finite element method, optimization using metamodel has 16 % error for energy absorption. In conclusion, knowledge based design using metamodel from a database that contained the characteristics of expansion tube-axial splitting impact energy absorber can produce accurate optimum solution and can be used to design expansion-axial splitting impact energy absorber for various application.