摘要
依据A356咖啡机顶盖高压铸造特点,采用FEM仿真软件对铸件成型工艺进行数值模拟,以L16(45)正交试验和6个补充试验作为BP神经网络的训练样本,建立模具热应力与浇注温度、模具预热温度、压射比压、压铸速度四个压铸工艺参数的非线性映射关系。在所定的压铸工艺参数范围内,随机选取6组工艺参数组合,结合FEM模拟软件和已经训练好的BP网络,预测在不同工艺条件下模具的热疲劳趋势,为同类压铸件工艺参数的选择提供了参考。
According to the feature of high pressure diecasting of dome of Model A356 Coffee Machine, the diecasting process of coffee machine dome has been simulated by finite element simulate software.The L16 （4^5） -orthogonal experiments and six complementary experiments have been chosen as the trained samples of back propagation neural network for building up a non-linear mapping between thermal stress of diecasting die and each major processing parameters of diecasting as pouring temperature, die pre-heat temperature,injection pressure and injection speed.In the range of determined diecasting process parameters,6 groups of process parameters have been randomly selected. The trend of die thermal fatigue has been predicted by using well trained BP neural network and FEM simulation software, which could provide certain extent guidance on producing similar diecasting parts.
出处
《中国铸造装备与技术》
CAS
2011年第3期47-51,共5页
China Foundry Machinery & Technology
基金
教育部
科技部
广东省产学研专项:基于多场耦合模拟轻质合金精密成型缺陷控制技术及产业化(2009B090600006)
关键词
压铸模具
热疲劳
BP神经网络
数值模拟
Die casting die
Thermal fatigue
BP Neural Network
Numerical simulation
