Four Water Advanced Control Software Development

Quadruple tank is a typical multivariable nonlinear process in Beijing Key Laboratory of Field-bus Technology and Automation in North China University of Technology. It is a benchmark platform that can be used to verify many control theory and easily achieve the experiment of nonlinear, multivariable, phase adjustable process. To study the control of this plant in industrial environments and software realization is of great significance in actual industrial process control area. Especially with nonlinear multivariable control difficulty, advanced control algorithms such as multivariable predictive control and hybrid model identification and control on quadruple tank is studied, which achieved real-time advanced control software independent of the MATLAB simulation environment.A multi-input multi-output hybrid system model identification method based on data-driven and a real-time nonlinear predictive control method based on multi-parametric programming are proposed for a class of nonlinear and multivariable industrial processes. A clustering classification method of high-dimensional data space is adopted to solve a MIMO piecewise affine autoregressive exogenous (PWARX) model identification problem. With the help of HYSDEL modeling language of hybrid systems and its compiler tools, MIMO-PWARX model is transformed into nonlinear piecewise affine (PWA) model. Using multi-parametric programming method, a predictive controller based on PWA model is designed to achieve on-line control of nonlinear process. Through this method, the hybrid model information of systems is obtained from a set of process data and PWA model of complex process is automatically generated. Based on the PWA Process model, real-time nonlinear predictive controller is implemented. Finally, the hybrid identification and control method we proposed is applied to the typical MIMO nonlinear quadruple tank system. Higher identification accuracy and more satisfied control performance is obtained compared with linear model based control method.Finally, based on powerful capability of algorithm design and numerical calculation in MATLAB, the advanced control step-by-step algorithm is coded and compiled into.NET components by means of Matlab deploy tool, and advanced control software was truly realized in.NET environment separated from the Matlab environment. Through OPC communication technology, it is ideal and easy to use the software to control quadruple tank in real industrial environments, achieving predictive control and other advanced control methods with satisfactory control effects. The success of this scheme provides important ideas and references to real-time advanced control method and software design of complex industrial object.