COMPOSITE BRAKE MATERIAL DESIGN USING KNOWLEDGE BASED DESIGN METHODOLOGY

• Main Author: Yulianto
• Format: Theses
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/38891
• Institution: Institut Teknologi Bandung

One important part of braking system is the friction material. The development of friction material for braking system nowadays leads to a composite brake technology which has the major advantage such as wear resistance, light weight, friction stability, heat resistance, and cheaper production costs. Each composite brake application require different characteristics from each other. Today, the optimum design and production of composite brake for various applications based on “trial and error” research and sometimes take a long time to develop new brakes application. In this research, a more systematic and efficient methodology of composite brakes design is developed based on knowledge based design methodology. The applications of knowledge based design methodology for composite brake started with the preparation of composite brake materials which were tested empirically to obtain the three main characteristics of composite brake, i.e. crush strength, cross-breaking strength, and coefficient of friction, metamodel building with Radial Basis Function method and optimization process using Memetic Algorithm method. Metamodeling produces equations of characteristics approach, which is used to compose material based on the database model and then used to process optimization numerically. The methodology was tested by simulated cases systematically before tested with actual cases, which is the design of composite brake for trains and motorcycles. The results produces the maximum error between optimization results and targets result is 5.8% and produces optimum composition of composed material for each characteristics with variations value of each is 16.3%, 12.8%, and 11.5%, and the most of the design included in that range. The range of variation is obtained based on data base error that is the error obtained from analysis of data correlation and the confidence level of empirical data base. The amount of error show that the need for the addition of a data base to reduce the error and improvement quality of experimental data.