A fuzzy predictive control method with enhanced robustness based on disturbance observer

Abstract

The present invention provides a fuzzy predictive control method based on the enhanced robustness characteristic of a disturbance observer. The steps are as follows: 1. Establish a discrete fuzzy disturbance observer model, an auxiliary controller and a robust predictive controller; 2. Solve the feedback gain of the auxiliary controller and the gain of the disturbance observer, obtain the minimum robust invariant set, and calculate the tight constraint set of the control input and state variables of the robust predictive controller; 3. Initialize the system state variables and assign them to the state variables of the nominal model; 4. For the The current nominal model state variables, solve the optimization problem that minimizes the upper bound γ of the predictive control performance, and obtain the current disturbance estimation value; 5. Calculate the control input of the system and act on the controlled object; 6. Apply the control of the nominal system The input acts on the nominal model and calculates the output of the current state, and the output of the current state quantity of the sampling system; 7. Substitute with and set k=k+1, and then jump to step 4. The present invention improves the stability of the predictive control system.

Description

technical field [0001] The invention belongs to the technical field of thermal control, and in particular relates to a fuzzy prediction control method with enhanced robustness characteristics based on a disturbance observer. Background technique [0002] Power plants operate in complex environments, such as changes in coal products, changes in ambient temperature and humidity, and load disturbances in the power grid. At the same time, there are serious mismatches in the internal model of the power plant, including modeling errors, coking of heat exchanger pipes, and coking of furnaces. This requires that the design of the control system should not be limited to stable conditions, and the robustness and anti-disturbance capability of the system should also be considered. Because the machine-furnace power generation system is highly nonlinear in nature, it is difficult for the predictive controller designed based on the nominal model to realize the large-scale variable operat...

AI Technical Summary

Problems solved by technology

Due to the highly nonlinear nature of the machine furnace power generation system, it is difficult to realize the large-scale variable operating conditions of the system by the predictive controller designed based on the nominal model

Traditional pipeline predictive control can explicitly deal with disturbances, but its design is usually based on a linear model or a time-varying linear system with multiple cells. The design is unavoidably conservative. At the same time, it faces the problems of unknown strong disturbances and model mismatches. The performance of predictive control will decrease, and even unstable results will appear

Images