Variable structural nonlinear model predictor controller

Abstract

The invention relates to changeable structure non linear model predictive controller. It is made up of online real time configuration machine, non linear model predictive control calculator, control action output processor, controlled variable constrained optimizer, and operational variable constrained optimizer. The online real time configuration machine can self organize current proper controller structure by which the non linear time lag calculating and on line real time predictive calculating can be processed. It has the features of multi predictive time domain online correction, single value predictive control, state feedback, non linear weighting, multi-cycle control etc, forms multi variable changeable structure non linear model predictive controller and control system on model and structure to adapt controlled process, make it always be in optimize running state.

Description

technical field [0001] The present invention relates to a strategy, method and variable structure control method for performing advanced control (Advanced Control) on a continuously changing process or equipment, especially on the continuous production process of the existing distributed control system (DCS) and the PID controller contained therein. Linear Model Predictive Controller. Background technique [0002] Since the PID (proportional, integral, differential) controller is relatively simple and practical, it has been widely used in the production process for automatic control since the 1940s. With the development of production and science and technology, especially the continuous expansion of production scale and frequent changes in market demand, it is difficult for PID control to meet the requirements of stable control and multivariable production process with large time lag, multivariable, constrained and variable structure. Constraint coordination and variable st...

AI Technical Summary

Problems solved by technology

Since the superposition principle cannot be applied to the nonlinear model, generally there is no analytical solution, the control strategy adopts multi-value estimation and state constraints, and the controlled process is m

Images