论文部分内容阅读
针对1450HC轧机,利用大型非线性有限元软件MSC.Marc建立仿真模型,对多种轧制工况进行了模拟,得到了板凸度值。研究了不同板带参数、工艺参数、板形调控参数对轧后板凸度的影响规律。以有限元计算值为训练样本,利用BP神经网络强大的非线性映射功能,建立了板凸度预报模型,在训练过程中采用了改进的快速BP训练算法,从而提高了训练速度,加快了网络收敛速度,增加了算法的可行性。该网络模型解决了有限元计算时间长,难以在线应用的问题。
For the 1450HC rolling mill, the simulation model was established by using the large-scale nonlinear finite element software MSC.Marc, and the rolling conditions were simulated and the crown value was obtained. The influence of different strip parameters, process parameters and shape control parameters on the crown of rolled plate was studied. Taking the finite element calculation value as the training sample, the plate convexity prediction model is established by using the strong nonlinear mapping function of BP neural network, and the improved fast BP training algorithm is adopted in the training process so as to improve the training speed and speed up the network The convergence speed increases the feasibility of the algorithm. The network model solves the problem of finite element calculation time-consuming and difficult to apply online.