基于人工智能的钛合金热变形工艺参数优化  

Optimization of hot deformation process for titanium based on artificial intelligence

作　　者：李萍[1] 薛克敏[1] 

机构地区：[1]合肥工业大学材料科学与工程学院,合肥230009

出　　处：《中国有色金属学报》2006年第7期1202-1206,共5页The Chinese Journal of Nonferrous Metals

基　　金：国家自然科学基金资助项目(50405020);安徽省优秀青年科技基金资助项目(04044058)

摘　　要：在深入分析热变形工艺参数对Ti-15-3合金显微组织及成形载荷的影响的基础上,以变形温度、变形程度和变形速率等热变形工艺参数作为设计变量,以显微组织和成形力的最佳综合为目标,建立了该合金热塑性成形工艺参数的多目标优化数学模型。以显微组织参数和成形力的人工神经网络预测模型作为优化算法的知识源,将人工神经网络与修正的遗传算法相结合,对Ti-15-3合金的热塑性成形工艺参数进行优化。结果表明,提出的修正的遗传算法是有效的,采用将其与人工神经网络相结合的方法对钛合金的热塑性成形工艺参数进行优化是可行的。The systematic analyses of the effects of hot deformation process parameters on microstructure and load of Ti-15-3 alloy were accomplished. Based on the results, a multi-objection optimization model was established for hot deformation process of Ti-15-3 alloy. In the model, temperature, strain and strain rate are treated as design variables and the objective is to obtain uniform fine-grain microstructures under the smaller load. Optimization of hot deformation process parameters for Ti-15-3 alloy was conducted by introducing artificial neural network prediction models of microstructures and forming load into a modified genetic algorithm. The results indicate that the modified genetic algorithm is effective and the optimization method based on artificial neural network and the modified genetic algorithm is feasible.

关 键 词：Ti一15—3合金 优化 修正的遗传算法 人工神经网络 热变形参数 

分 类 号：TG166.5[金属学及工艺—热处理]