论文部分内容阅读
将有限元模拟技术和神经网络理论相结合,建立了钛合金等温锻造工艺设计的优化系统,实现了CAD-CAE-优化-CAD的循环,在循环过程中,CAE对CAD产生了数据反馈。将FEM模拟结果作为初始数据导入优化系统,作为ANN的训练参数,利用训练好的神经网络对设计做出预测,循环系统不断运行直至得到良好的结果。利用该优化系统能获得像顶级专家主持一样的优良效果,使得设计过程在时间和效果上都得到提高。
The finite element simulation technology and neural network theory are combined to establish the optimization system of titanium alloy isothermal forging process design, and realize the cycle of CAD-CAE-optimization-CAD. CAE produces data feedback to CAD during the cycle. The FEM simulation results are introduced into the optimization system as initial data, which is used as training parameters of ANN. The trained neural network is used to predict the design, and the circulatory system is continuously operated until the good result is obtained. Using this optimization system gives the same excellent results as a top-level expert, making the design process both time-and-effective.