Design of Nano-Micro-Composite Ceramic Tool and Die Material with Back Propagation Neural Network and Genetic Algorithm

摘要

An algorithm combined with back propagation neural network (BPNN) and genetic algorithm (GA) was used in the optimum design of the compositions of an advanced ZrO₂/TiB₂/Al₂O₃ nano-micro-composite ceramic tool and die materials. GA was used to fully optimize the network topology, thresholds, and initial connection weights of BPNN. The input parameters are the contents of each compositions of ceramic tool and die materials and the output parameters are mechanical properties including hardness, flexural strength, and fracture toughness. The compositions with optimum mechanical properties can be chosen for materials preparation with less error and the result can be used to guide the experimental process. As a result, the nano-micro-composite ceramic tool and die material with good mechanical properties was then fabricated. It indicated that the algorithm can offer a robust and efficient way for the compositional design of ceramic tool and die materials.