Research And Application Of Distributed Predictive Functional Control

Predictive functional control is a kind of control method developed on the basis of themature model predictive control, which not only makes the control input have more regularity but also makes the computation small through making the control input structured, is anew direction in the development of predictive control. On the other hand, with the rapiddevelopment of the computer network, the predictive control is not limited to thecentralized control but more and more replaced by distributed control, which brings newrequirements and challenges to the traditional control problems. Based on the study ofpredictive functional control and distributed control, this paper presents the distributedpredictive function control and studies it from two aspects of its theory and application.The main contents include the following aspects：(1)The principles and algorithm of the single variable predictive functional control,which is based on the step response, are introduced in detail and its parameter selectionprinciples are expounded. The algorithm is simple, easy and suitable for the needs ofindustrial sites. The simulation results show that the control performance is good, whichcan track the set value accurately and quickly.(2) In view of the problems of large-scale systems control, this paper puts forward thedistributed predictive functional control method. This method disperses the controlproblem of a large-scale system into subsystems, which reduces the scale and thecomplexity of the control problem. At the same time, the subsystem can communicate witheach other and predictive functional control is adopted as the control strategy to make thecontrol input structured, effectively reducing the online computation.(3) It deduces in detail the distributed predictive functional control algorithm based onthe step response model and discusses the stability of the algorithm as well as proves theiterative convergence condition. And simulation research is done by using the multivariable coupling system as the object. The simulation results of the model mismatchare presented. This algorithm effectively completes the control task and improves thecontrol performance.(4) According to the characteristics of the four-tank system, the mutual couplingmultivariable model is established and to study the distributed predictive functional controlscheme. Distributed predictive functional control algorithm is written by SCL (StructuredControl Language) and is applied to the liquid level control based on (PLC)(Programmable Logic Controller), which realizes the combination of the hardware andsoftware. And interface design is done through the configuration software to set parametersand have real-time monitoring. The simulation results and the effects of the actual controlverify the feasibility and effectiveness of the proposed algorithm.