Abstract: A composite sleeper’s design load is 1 000 kN, it must meet deformation requirements of sleeper’s bending load. So the test tooling must be designed to meet the requirements of bending load. This paper established the response surface of target parameters to section sizes based on the theory of material mechanics and Kirging model. Based on this, the tooling was obtained by multi-objective genetic algorithm, and finally, was obtained the optimal combination of tooling’s section size. Then bending resistance of the tooling was calculated according to optimized section sizes by FEM. Thereafter, the tooling was manufactured based on the optimized section sizes. The result show that this tooling did not create any plastic deformation with a 1 000 kN load.

Key words: composite rail sleeper tooling, bending load, Kriging model, response surface, multi-objective genetic algorithm, structure optimization