Optimal Design Of Nano-Composite Cermet Die Materials With Immune Genetic Algorithm

At present, the design method of ceramic die materials still depend on the experiment, then the components and processing parameters were adjusted according to the experiment data. The materials with better performance can be obtained. But this method would expend a lot of manpower, material and time. If the computational intelligence technology was used in the optimal design, the design efficiency would be improved. The neural network and the genetic algorithm are used in the ceramic materials design extensively. In this paper the immune algorithm is used in the optimal design of ceramic materials, and the immune algorithm is combined with genetic algorithm and neural network. The components and processing parameters of ceramic materials are optimized with this method.In this paper, the immune algorithm is used to optimize the mechanical properties of Ti(C,N) based nano- composite cermet die materials. The mathematic relationships between fracture toughness, hardness, flexural strength and components are obtained with stepwise regression analysis method. The relationships are the objective function of immune algorithm. The optimal mechanical properties and components can be obtained after optimizing. The optimal results are as follow. The best fracture toughness is 7.7 MPa·m^1/2,and the components of ZrO₂ and WC is 10.67% and 20%, respectively; The best hardness is 15.31GPa, and the components of ZrO₂ and WC is 0% and 15.10%, respectively; The best flexural strength is 1057.39MPa, and the components of ZrO₂ and WC is 6.75% and 11.38%, respectively.The immune algorithm is combined with genetic algorithm successfully, and the immune genetic algorithm is used to optimizing the mechanical properties of Ti(C,N) based nano- composite metal ceramic die materials. The optimal results are the same as that of immune algorithm, but the iterations of immune genetic algorithm are fewer than the immune algorithm. The convergence rate is improved.The relationships between the mechanical properties of Ti(C,N) based nano- composite cermet die materials and the processing parameters are established with BP neural network. The forecast is done with BP neural network, and the forecast errors are less than 10%. The BP neural network models are the objective functions of immune genetic algorithm. The processing parameters when the mechanical properties are best are obtained. The optimal results are as follow, the best fracture toughness is 8.2 MPa·m^1/2, and the sintering temperature and holding time is 1455.5℃and 12.6 min, respectively; the best hardness is 17.3 GPa, and the sintering temperature and holding time is 1490.7℃and 24 min, respectively; the best flexural strength is 1097.6 MPa, and the sintering temperature and holding time is 1454.1℃and 10.9 min, respectively. The error between optimal results and experiment results is 4.65%, 0.23% and 4.05%, respectively. At last, the components of Ti(C,N) based nano-composite cermet die materials are optimized with this method. The best fracture toughness is 6.93 MPa·m^1/2,and the components of ZrO₂ and WC is 12.3% and 6.9%, respectively; The best hardness is 15.19GPa, and the components of ZrO₂ and WC is 13.75% and 19.74%, respectively; The best flexural strength is 924.68MPa, and the components of ZrO₂ and WC is 7.57% and 2.24%, respectively.