طراحی کنترل پیش بین یادگیری تکراری مبتنی بر شبکه های عصبی برای سیستم های مکاترونیکی با تغییرات غیرخطی جدا
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|27362||2009||7 صفحه PDF||سفارش دهید||3601 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Process Control, Volume 19, Issue 1, January 2009, Pages 68–74
The paper presents a new nonlinear predictive control design for a kind of nonlinear mechatronic drive systems, which leads to the improvement of regulatory capacity for both reference input tracking and load disturbance rejection. The nonlinear system is first treated into an equal linear time-variant system plus a nonlinear part using a neural network, then an iterative learning linear predictive controller is developed with a similar structure of PI optimal regulator and with setpoint feed forward control. Because the overall control law is a linear one, this design gives a direct and also effective multi-step prediction method and avoids the complicated nonlinear optimization. The control law is also an accurate one compared with traditional linearized method. Besides, changes of the system state variables are considered in the objective function with control performance superior to conventional state space predictive control designs which only consider the predicted output errors. The proposed method is compared with conventional state space predictive control method and classical PI optimal control method. Tracking performance, robustness and disturbance rejection are enlightened.
Mechatronics is the synergetic integration of mechanical engineering with electronics and intelligent computer control in the design and manufacturing of industrial products and processes ,  and . In many cases, the mechanical part of the system is coupled with the electrical, thermodynamical, chemical or information processing part. Therefore, mechatronic systems are actually coupled processes exhibiting nonlinear characteristics. Different control methods, e.g., for the control of position, speed or force for various mechatronic systems are provided , ,  and  based on different control design theories . The methods are generally designed to compensate for the system’s nonlinearities. However, they are rather computationally demanding and require high processing capability of CPUs and the control performance greatly relies on the accuracy of the compensators . Typical examples focusing on compensating for nonlinearities are: (1) Compensation of nonlinear static characteristics , however, this method needs to design an inverse function compensator and suppose that the nonlinear function has an inverse function. (2) Friction compensation , however, the control performance greatly depends on the accuracy of the linearization and generally this linearization is not an accurate one. Therefore, the development of effective and reliable control methods with relatively simple structure is a challenge for the designers. The paper is to focus on the mechatronic drive systems , which are typically controlled by conventional state space predictive control (CSSPC)  and classical PI control (PI)  and recently studied by Rau et al. . The paper proposes a nonlinear iterative learning predictive control method (NILPC), which provides improvement of regulatory capacity for both reference input tracking and load disturbance rejection compared with CSSPC and PI. General description of model predictive control (MPC) can be found in . Generally, the procedure consists of two steps: (i) unlike traditional methods , the nonlinear part of the process is first transformed into a linear time-variant part and a nonlinear part, and the linear part of the system is used to design an overall convergent linear predictive control law (ii) secondly, different from CSSPC , NILPC is designed for the obtained equal linear time-variant system plus a nonlinear part.
نتیجه گیری انگلیسی
Neural network based model predictive control of mechatronic drive systems with isolated nonlinearity is considered in this study. Control performance is compared for three types of controllers: CSSPC controller , classical PI optimal controller  and NILPC controller. In CSSPC and classical PI, the nonlinear process is first linearized. In the case of NILPC control scheme, a new treating method has been performed instead of the linearization method. And in order to improve the control quality, NILPC controller is applied to the process. Results have shown that the proposed NILPC method yields improved performance.