(1. 中南大学 材料科学与工程学院，长沙 410083;
2. 华中科技大学 塑性成形模拟及模具技术国家重点实验室，
武汉 430074;
3. 云南省机械研究设计院， 昆明 650031)
摘 要: 基于MATLAB平台， 将BP神经网络、遗传算法和数值模拟技术应用于铝型材挤压模具参数优化设计。 采用三层BP神经网络建立型材挤压模具的数学模型， 由正交实验法安排模拟实验组合，采用有限元软件进行挤压过程的数值模拟， 并以具有不同工作带尺寸的挤压模具中金属流出模口平面上的Z向质点流速均方差作为模型目标值，将模拟结果作为神经网络的输入样本对训练网络并建立网络知识源， 通过遗传算法求得模型的全局优化解； 最后通过有限元数值模拟技术验证并比较优化所得工作带与经验法确定的工作带对金属流动均匀性的影响。数值模拟结果表明， 本研究对挤压模具工作带的优化是有效的。
关键字: 铝型材； BP人工神经网络； 遗传算法； 挤压模具；工作带； 有限元模拟
profile extrusion die
（ 1. School of Materials Science and Engineering,
Central South University, Changsha 410083, China;
2. State Key Laboratory of Plastic Forming Simulation and Die Technology,
Huazhong University of Science and Technology, Wuhan 430074, China;
3. Yunnan Mechanical Research and Design Institute, Kunming 650031, China）
Abstract: BP artificial neural network, genetic algorithm and FEM simulation were applied to optimize the design of profile extrusion die on MATLAB foundation. A three-layer neural network was used to set up mathematical model for profile extrusion dies with different bearing lengths. Orthogonal test was arranged for numerical simulation to get Z-velocity at the die land exit which was used as the target value of the model. The neural network is trained by the above Z-velocity values to form knowledge source, and the general optimized solution was attained through genetic algorithm. At last, the optimized bearing of the extrusion die was analyzed by FEM and compared to the design with experiential way. The simulation results show that the optimization of die bearing is effective.
Key words: aluminum profile; BP artificial neural network; genetic algorithm; extrusion die; bearing length; finite element simulation
