Abstract:Taking the warm extrusion concave die for connecting cuplid of automotive steering screw as an example,the die wear and life prediction was studied.With four major factors influencing the wear and life of warm extrusion concave die,including concave die’s entrance angle,mould initial hardness,mold initial temperature and friction factor as the process parameters,and selecting four different levels respectively,32 groups of warm extrusion concave die wear test program with four factors and four levels were determined.Numerical simulation of forming process was carried out through the Deform3D finite element numerical simulation software.The BP neural network was trained with different influence factors and corresponding die wear as samples,and the mapping relationship between the four major factors and the die wear volumes was established.Taking warm extrusion concave die wear volume as objective function,the combinatorial optimization of four factors were carried out by genetic algorithm to minimize the wear of the die and realize the longest life of die.