基于动态递归模糊神经网络的轧机辊缝动态辨识

Dynamic recognize roll gap of rolling mill based on dynamic recursive fuzzy neural network

模糊系统和神经网络,由于具有逼近任意连续非线性映射的特性,而广泛应用于系统的辨识与控制。但是传统的模糊神经网络是一种静态映射,不适用于动态系统的辨识,而轧制过程中影响轧机辊缝的因素复杂,外界干扰严重,过程参数难以确定,为提高轧机辊缝动态的辨识精度,提出了一种基于动态递归模糊神经网络的辨识模型。轧制仿真结果表明,该模型具有很高的辨识精度。

Because of the specialities of approaching any continuous nonlinear mapping, the fuzzy system and the neural network are generally applied to the system identification and control. But, the traditional fuzzy neural network is an static mapping, and not acceptable for the dynamic system identification. For the complex factors that influencing the roll gap in the rolling process, and the serious external disturbances, the procedure parameters are difficult to determine. In order to improve the dynamic identification precision of the gap, a identification model based on the dynamic recursion fuzzy neural network is put forward. The results of the simulation indicated that the model has higher identification precision.