Font Size: a A A

Application Of Multi-modal Prediction In Three-tank Water Level Control System

Posted on:2019-01-04Degree:MasterType:Thesis
Country:ChinaCandidate:Y P LiFull Text:PDF
GTID:2348330545992102Subject:Control engineering
Abstract/Summary:PDF Full Text Request
Three-tank Water is an ideal experimental platform with non-linear and big lag feature.The whole or the parts of many controlled objects in industry can be abstracted into a mathematical model of the Three-tank Water.This thesis takes A3000 Three-tank Water as the controlled object,applies the Multi-model Generalized Predictive Controller(MGPC)algorithm,which is developed under the EDPF-NT Plus system and configures SAMA diagram.All these processes realize the remote monitoring of the liquid level control of the Three-tank Water.In terms of modeling,mechanism analysis method and two-point method are adopted to model the Three-tank Water in this thesis.The MATLAB software is used to simulate and compare the mathematical models established by the two methods.In addition,the curve fitting method is adopted to optimize the model parameters,which makes the established mathematical model close to the actual liquid level system.In terms of fusion algorithm,first of all,the Stair-like Generalized Predictive Controller(SGPC)algorithm is adopted.This algorithm converts the solution of traditional generalized predictive control(GPC)algorithm inverse matrix into a vector operation,which greatly reduces the amount of calculation.The influence of the controller parameter changes on the system performance is analyzed.Secondly,a multi-model strategy is introduced to establish a linear model with the same structure according to different operating conditions.The slip model weighted switching strategy is adopted to switch the system model according to operating conditions.Finally,this paper uses the MGPC algorithm to simulate the Three-tank water liquid system model.The results show that compared with the traditional GPC algorithm,the MGPC algorithm has the following advantages:(1)Calculations of the algorithm are reduced.(2)The modeling problem of nonlinear systems is solved.(3)On the premise of ensuring the stability of the system,the response speed of the system is improved.In this thesis,the Communication between MATLAB and the EDPF-NT Plus system of Beijing Guodian Zhishen Control Technology Co.Ltd.has been realized by OPC Sever.The remote monitoring of the liquid level system of the Three-tank Water is achieved by MGPC algorithm.The original algorithm library of the EDPF-NT Plus system is updated.The successfully compiled MGPC algorithm block implements SAMA diagram configuration.A liquid level rapid lift disturbance test is performed at the scene.The results show that the MGPC algorithm has a shorter adjusting time,a smaller overshoot and a more stable liquid level comparing to the cascaded PID control after parameter tuning.At present,the results of thisresearch will be applied to other complex conditions.
Keywords/Search Tags:Predictive Control, A3000 System, OPC, EDPF-NT Plus
PDF Full Text Request
Related items