Configuration design of a mechanical product is performed in the early stage of conceptual design. Every design methodology and analytical tool can be used to configure the mechanical part in the design process. This paper describes the configuration design methodologies of using neural network and TRIZ theory, of which main purpose are giving some prediction result promptly, based on the existing data and giving more creative design with TRIZ’s knowledge base. Configuration design of trainsets of high-speed train have been studied using the neural network with backpropagation and these results are compared with the actual values. The configuration results estimate the basic trainset such as weight, power, the number of cars, the number of motorcars, and the number of traction motors of a high-speed train. And these values are used in the configuration design process of bogie to determine the basic specification. The configuration design of a train bogie is performed with a design expert system that uses function taxonomy and TRIZ theory. The design expert system can propose a functional model of a new part that is not in the existing parts list of a bogie.
