基于MATLAB和VB神经网络在轧制力预报中的应用

Application of Neural Netuork Based on MATLAB and VB in Rolling Force Prediction

为方便实际应用,结合MATLAB神经网络强大的计算能力和VB友好界面的优势,将二者混合编程,并应用于轧制规程中,规程中采用改进的BP网络LM优化算法对轧制力进行快速预报,通过仿真证明该优化算法能够提高轧制力预报精度,混合编程能降低程序复杂性,提高效率。

With the full use of the calculation ability of MATLAB NN and VB advantages, this paper presents a method of hybrid programming between MATLAB and VB, and applies to rolling schedule in which the BPNN adopted Levenberg-Marquardt optimized algorithm has been used to predict rolling force. The simulation shows that the algorithm has significant prediction performance, and hybrid programming can improve efficiency.