ANFIS在局部板形控制中的建模与仿真

在轧制生产过程中，局部板形动态特性往往表现出非线性、多变量、强耦合和大惯性等特点，这使得难以对其建立比较精确的模型，从而难于精确表达轧制过程及实施优化控制．针对局部板形控制建模难的现状，为达到建立精确非线性模型的目的，采用自适应神经模糊系统（ANFIS）模糊建模方法．该方法通过对模糊系统的结构辨识和参数辨识，使神经模糊网络能够自主、迅速有效地收敛到要求的输入和输出关系，从而达到精确建模的目的．Matlab仿真结果表明，通过anfis训练及检验的模型是有效的，具有较高的控制精度．

Modeling and Simulation of Local Flatness Control with ANFIS

In the process of rolling production, the dynamic behavior of local flatness shows a characteristic of non-linearity, multi-variable, strong coupling and big inertia, which makes the modeling very difficult. As a resuct, the whole optimal control for rolling processes is impossible. According to this actuality and the objective of building accurate nonlinear model for rolling process, an adaptive neural-fuzzy inference system （ANFIS） is used. Through the structure and parameter identification of fuzzy identification system, the system can converge to the required input-output relationships independently and rapidly. Matlab simulation results show that the trained and tested antis model is effective, and has higher precision control.