文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 自然科学 > 数学 > Research on Nonlinear Model Predictive Control
侵权投诉

Research on Nonlinear Model Predictive Control

Proceeding of the IEEE

International Conference on Automation and Logistics

Zhengzhou, China, August 2012

Research on Nonlinear Model Predictive Control Technology for Ship Dynamic Positioning System

Yuanhui Wang and Yufeng Sui Department of Automation Harbin Engineering University

Harbin, Heilongjiang Province, 150001, China wangyuanhui@hrbeu.edu.cn, suiyufeng@hrbeu.edu.cn

Jing Wuand Jianfang Jiao Department of Automation Harbin Engineering University

Harbin, Heilongjiang Province, 150001, China wujing0428@hrbeu.edu.cn, jiao_jianfang@163.com algorithm avoids the system shocking when the controller switches between the different states[3].

The main purpose of this paper is to design a DP controller based on the theory and method of Non-Switch Analytical Model Predictive Control Algorithm, which can control the ship quickly move to the appointed position and maintain the position, and verify the efficiency of the controller through the simulation experiments.

II. MATHEMATIC MODEL OF DYNAMIC POSITIONING SHIP As known, the movement of the ship has six degrees of freedom. For Dynamic positioning surface ship, we generally only consider three degrees of freedom, Surge Sway and Yaw. And we get the equations of dynamic positioning ship as follows[4-6].

d1 = 11u+τX u

mm x=ucosψ vsinψ y1=x 1111

, y=y.(1) d23d221 =usinψ+vcosψ, y = vv r+Y 2 y=ψ ψm22m22m22

=r 3 dd1 = 32v 33r+N r

mmm333333

III. DESIGN OF CONTROLLER

Nonlinear Model Predictive Control is an algorithm based

on optimization control strategy, and has the basic theories of predictive model rolling optimization and feedback correction. NMPC is widely used, because of its suitability for the process control system of strongly nonlinear, multi-constrains[7-8]. Control objective is to design a controller(τX,τY,τN) to make the ship move to the set state quickly and accurately and keep the state. Assumption: System output (x,y,ψ) and the

are continuous and can do a desired output (xd,yd,ψd)

sufficient number of differential operators. We rewrite the

system into canonical form as follow

=f(x)+g(x)u (2) x

y=h(x)

are state vector control input Where, x∈Rn,u∈Rl,y∈Rm

and system output, and n=6,l=3,m=3.

x=[x1x2x3x4x5x6]T=[xyψuvr]T (3)

Abstract - As an excellent technology, model predictive control has become a more pleasing and popular research topic. In order to solve the higher precision control problem for the nolinear, uncertained and strong-coupled ship dynamic positioning system, this paper introduces the nonlinear model predictive control (NMPC) technology into this system based on lots of foregoing work about its linear application. Firstly, a 3 degree of freedom nonlinear mathematic model of dynamic position ship is established based on ship dynamics. Secondly, the model is simplified according to the demand of design of the controller. Then, in terms of the concept of the lie notation, a nonlinear model predictive controller is designed based on NMPC theory. The controller is judged to be a Positive definite one. Finally, a simulation experiment is carried out and the simulation results show that the designed NMPC controller is effective and satisfactorily for ship dynamic positioning system. Index Terms - Nonlinear system, dynamic positioning, nonlinear model predictive control, constraints.

I. INTRODUCTION

Ship Dynamic Positioning (DP) system can make the ship keep its position with the help of thrusters under to overcome the disturbances of wind, wave and current. DP system has the characteristics of high precision and good flexibility, and its cost does not increase as the increasing depth[1]. Model predictive control was born in the 1970s, and it is been widely developed in the industry as the advanced control technology. Model predictive control can deal with the multivariable problems with the different types of constraints and control law cannot calculate offline. It has the function of predict, so it can correct the system before the deviation produced, and it has better robust to the change of system and environmental parameters. With the development of model predictive control, a lot of new research results has been made, such as, some predictive control algorithms based on nonlinear model (Wiener Model, Hammerstein Model, Volterra Model and so on) and some algorithms combining with intelligent control algorithm, Fuzzy predictive control, neural network predictive control, model predictive based on support vector machine and so on. Based on the achievement of the research before, Chen puts forward the Switch Analytic Nonlinear Model Predictive Control Algorithm[2]. This algorithm has become an important algorithm, because it has the general characteristics of Model Predictive Control and has advantages of the feedback linearization. Zhang Guoyin puts forward an improved algorithm based on the relevance called Non-Switch Analytical Model Predictive Control Algorithm, and the

978-1-4673-0364-4/12/$31.00 ©2012 IEEE

348

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email