The Algorithm Of Model Predictive Control And Its Application In The Cement Rotary Kiln

Cement rotary kiln is the core of the cement production equipment, we through the further study of its mechanism, select the relevant variables.This topic is to cement rotary kiln system as the research object, based on Model predictive Control(Model Predive Control, MPC) theory, by combining the interval linear MPC and nonlinear MPC, senior design integrates the MPC algorithm controller design.The controller to achieve good control complex multivariable system has important theoretical and practical significance.The specific research work are as follows:Firstly, From the preheating decomposition and kiln calcine, preheater waste gas processing and clinker cooling of four aspects to study the control of cement rotary kiln calcining process, and through the operator experience and theoretical research is analyzed and the control parameters of cement rotary kiln system, select the operation operation variable(the manipulated variable(MV) and the controlled variable(controlled variable, CV).Secondly, For linear MPC established a MPC algorithm based on interval soft features and variable constraint.Convenient concise by a mathematical formula using the position relationship between predicted values and the interval, imitate the value deviation of design principle, constructs the range tracking deviation and other variables of soft constraints is in classical MV of MPC algorithm on the basis of soft constraint and joined the CV of soft constraints, the two together constitute the new variable soft constraints.Thirdly, To ensure that the controller has feasible region under the constraint conditions, shall be exempt from the objective function at the intersection between the constraint solution domain is empty, and at the same time make the CV in the soft output constraint space as possible, to ensure the smooth operation of the optimal control.Aiming at this problem, the feasibility of the proposed solution first before every step of the rolling optimization solution, based on the input and output space soft constraint space constraint, the feasibility of the intersection of interval judgement, when the objective function constraints the intersection of slack variable was developed for the solution domain is empty, readjust the soft scope implementation constraints feasible solution set, and slack variable solving scheme is given.Finaly, Innovation is a kind of nonlinear process control strategy method.When the weak nonlinear MPC algorithm, using interval can be within the range of the nonlinear system with very good guarantee the stability of the control.When the controlled object in under the condition of strong nonlinear, switch to a nonlinear MPC algorithm based on LSSVM model, replaced by the method of multi-step iterative multi-step prediction, skip the process of setting up a nonlinear prediction model, only need to be trained in a single step prediction model of controlled object.This is a great progress on nonlinear prediction, has very important practical value.