Author(s)

H. Martikka

Abstract

Methods of optimum design and experimentation were used to ensure sufficient fatigue life in a series of locomotive steel frames having variable cross sections. These were tested using resonance fatigue loading to establish the optimal fab- rication methods for improving fatigue life. Dynamics simulation was used to adjust the system's natural frequency to exciter frequency by optimizing the mass distribution. Shape optimization was done by two methods. The shape was optimized using a nonlinear learning random search algorithm with both weighted and fuzzy goals and with constraints on fatigue life and with a mix of continuous, integer and discrete design variables. The goal was formulated to express customer satisfaction on the cost and technical performance. The shape was

Keywords