基于自适应动态规划的矿渣微粉生产过程跟踪控制

矿渣微粉是一种新型绿色环保型建材，可以大大提高水泥混凝土的力学性能.本文以矿渣微粉生产过程为研究对象，针对该过程难以通过机理建模进行辨识和控制的特点，利用数据驱动的思想，建立矿渣微粉生产过程的递归神经网络模型.在此基础上，利用自适应动态规划，设计具有控制约束的跟踪控制器，并将其应用到矿渣微粉生产过程中.仿真分析表明，建立的数据驱动模型能够有效地辨识矿渣微粉生产过程，同时，本文提出的控制方法能够实现输入受限的微粉比表面积及磨内压差的最优跟踪控制.

Optimal Tracking Control for Slag Grinding Process Based on Adaptive Dynamic Programming

Super fine slag powder is a new kind of green environmental-friendly construction material, which can greatly improve the mechanical properties of cement concrete. However, the slag powder grinding process is hard to identify by a mechanism model. In this paper, a data-driven based recurrent neural network model is constructed utilizing the information measured from slag grinding system. Based on this model, an adaptive dynamic programming algorithm is proposed to realize the optimal tracking control with constrained control input. Further, this algorithm is applied to the slag grinding process. Simulation examples show that the data-based model can effectively identify the grinding process, and the control method can realize the optimal tracking control of specific surface area and mill differential pressure with control constraints.