型材挤压模具寿命的高低 ,与模具的结构形状及几何尺寸有很大关系 .通过用BP神经网络建立型材挤压模具的数学模型 ,用弹性有限元法计算模具内部各点的等效应力值并作为BP神经网络训练的导师信号 ,再结合遗传算法优化模具结构 ,建立了铝型材挤压模具的力能参数实时计算系统和几何参数优化设计系统 .研究表明 :采用本方法设计的方管铝型材挤压模具 ,其危险点处的等效应力降低了 39.1 % ,且应力集中现象明显减小 ,模具结构趋于合理 ,提高了模具的承载能力和使用寿命 . The life of profile extrusion die has a great relationship with the structure and geometry of the die.Based on the mathematical model of extrusion die with BP neural network, the equivalent stress of each point in the die is calculated by elastic finite element method And as a guide signal of BP neural network training, combined with genetic algorithm to optimize the mold structure, the real-time calculation system of force and energy parameters and the optimization design system of geometric parameters of aluminum extrusion die are established.The results show that the square tube aluminum Profile extrusion die, the equivalent point of danger at the point of 39.1% reduction, and the stress concentration was significantly reduced, the mold structure tends to be reasonable and improve the bearing capacity and service life of the mold.