摘要
为获得铝合金半球形件的优化压边力,提出了成形结果的综合评价标准用于不同变压边力曲线模拟实验的评价,以正交实验法设计模拟试验获得样本,建立了预测不同压边力曲线下成形结果的神经网络模型。在精度可靠的神经网络基础上,运用遗传算法优化压边力,获得了先增后减的"最优"压边力曲线。
To get an optimized VBHF(Variable Blank Holder Force) curve for aluminum alloy hemispheric part, the paper proposed a standard to evaluate the forming in the VBHF simulation experiments; in the experiments, the samples were taken by the orthogonal plan method to set up the neural network model for the forming prediction. The paper pointed out the BHF optimization is available with the genetic algorithm on the basis of the neural network.
出处
《有色金属加工》
CAS
2013年第6期22-25,58,共5页
Nonferrous Metals Processing
基金
国家自然科学基金资助项目(10702018)
国家自然科学基金重点项目(51135004)
关键词
铝合金
半球形件
数值模拟
神经网络
遗传算法
aluminum alloy
hemispherical part
numerical simulation
neural network
genetic algorithm