论文部分内容阅读
针对实际生产中扰动的时变性,提出了一种扰动自适应的鲁棒预测控制(RAMPC)算法以提高扰动抑制性能。采用时间序列(ARMA)模型在线辨识系统的不可测扰动,通过基于多次迭代思想的递推辨识算法(multi-iteration pseudo-linear regression,MIPLR)来保证在线辨识的质量和收敛速度。考虑到数据与辨识模型的不确定性,改用min-max形式描述MPC算法的控制作用优化命题,并将在线辨识过程中的误差数据引入min-max命题,使在线辨识与控制作用鲁棒优化求解紧密结合起来,提高算法鲁棒性。进一步将此min-max问题转换为一个等效的非线性min问题,并采用多步线性化方法实现快速求解,解决了传统min-max方法在线计算负荷高的问题。仿真结果表明了该算法的有效性。
Aiming at the time-varying of perturbation in real production, a robust adaptive predictive control (RAMPC) algorithm is proposed to improve the disturbance rejection performance. The time series (ARMA) model is used to identify the unperturbed disturbance of the system online. The quality and convergence speed of on-line identification are ensured by multi-iteration pseudo-linear regression (MIPLR). Taking into account the uncertainty of the data and identification model, the min-max form is used to describe the MPC algorithm’s control action optimization proposition, and the error data in the online identification process is introduced into the min-max proposition so that the on-line identification and control effects are robustly optimized Solve closely together, improve the robustness of the algorithm. The min-max problem is further transformed into an equivalent nonlinear min-problem, and the multistep linearization method is used to solve the problem quickly. This solves the problem of high online computational load of the traditional min-max method. Simulation results show the effectiveness of the algorithm.