Research On The Relationship Between Heat Treatment And Mechanical Properties Of Ti₂ AlC/TiAl Composites By Network

With the development of aerial high-tech, the performance of the materials became more and more highly demanded. In high temperature structured material field, Ti₂AlC/TiAl composite materials are promising high temperature structured material due to their low density, high specific strength, high specific stiffness and high temperature strength ratio. Ti-Al alloy as new high temperature structured material which used in aerial engine and rocket driving system was candidate material. As China proposed the targets and requirements of saving energy and reducing emissions, energy saving and emission reduction in the scientific experiments are also very important. Artificial neural network is a new research method, which is considered as one of the most promising methods to solve material design or performance prediction. Artificial neural networks can save costs, manpower and material resources, avoiding excessive test.In this theme, the heat burst method was used to compose TiAl and TiAl(Nb, B)based alloy powder. Spark plasma sintering (SPS) technology was used in-situ to fabricate the Ti₂AlC/TiAl (Nb, B) composites. The two method of heat treatment Ti₂AlC/TiA(lNb, B)composites were multistep heat treatment and rapid temperature rise heat treatment. And the effects of two heat treatments on mechanical properties (microhardness and bending strength) of Ti₂AlC/TiAl (Nb, B) composites were studied.In this paper, MATLAB was used to predict and optimize the Ti₂AlC/TiAl composites. Unlike other high-level language, MATLAB language was known as fourth computer language which is simple and quick. People extricated from complex codes by MATLAB language. BP neural network optimization are not only an effective way to optimize forging process of Ti₂AlC/TiAl composites,but also save the cost and can improve research efficiency. Based on the preparation and heat treatment of Ti₂AlC/TiAl (Nb, B) composites process analysis, simulation and theoretical analysis of the mechanism were used. The method of combining model and BP neural network was used to predict of properties of Ti₂AlCTiAl compound materials. And the results show that the error between predictive value and experimental value is small. It was applied to analyze the effecting factors by orthonormal planing method. So this is a really newly and accurate method. So this method not only can predict the property but also analyze the property of Ti₂AlC/TiAl (Nb, B) composites. When samples which not used input to the BP neural network model, it can output the correct results, and its generalization accuracy within the limits of the process. Therefore, the prediction model instead of the real test, and predictive value instead of the model results was feasible. So the BP neural network model can guide the experiments.This prediction method used the test data, so the BP neural network's prediction results and accuracy are not dependent on the mathematic model of Ti₂AlC/TiAl compound material. Due to the less experiment, it can save lots of research costs and period. By MATLAB, we can btain a model of prediction of processing-properties of Ti₂AlCTiAl compound materials, which can guide the production practice.