Improvement And Simulation Research On Predictive Functional Control Algorithm

Modern control theory must be based on precise object parametric model, but some features like nonlinearity, time-variance, closed coupling and uncertainty exist in every actual system. And precise parametric models may be difficultly attained. So it’s difficult to achieve the expected control effect by modern control theory. Model Predictive Control (MPC) was derived in such situations, and it’s a control strategy based on practice operation, and combines control theory. Great achievement has been acquired in the field of petrol, electrical and aviation system where precise parametric models are difficult to get, especially in petrochemical industrial. However, MPC has a more complex algorithm, more difficult design procedure, and larger calculation amount which cannot meet the requirement of real-time performance in comparison to conventional control strategy, so that MPC is usually used in the systems where good real-time performance is not imperative. Its application needs professional companies, so average engineers may have difficulties in mastering it. Predictive Functional Control (PFC) is a new kind of model predictive control strategy based on predictive control theory and it meets the need of fast system. PFC has such portraits as lower degree of precision to parametric models, easier algorithm, smaller amount of calculation, higher accuracy, and it’s relatively easy to design, convenient for engineers to master.PFC has made successful achievements in many fields abroad, like high-precision tracking in industrial robot, target tracking in military fields, chemical batch reaction process and steel rolling. Systems for both high and not high speed requirements exist in the procedures. In China, similar to conventional predictive algorithms, PFC is mainly used in systems where high control speed is not required, such as polyvinyl chloride polymerization producing process, control of PH value in penicillin fermentation process, and chlorinated polyethylene producing process. Plenty of economic value has obtained, but some drawbacks still exist like inconvenience in online adjustment, high requirement for models and bad robustness. To solve these problems, two enhanced PFC models are studied in this essay and they are researched in typical fast system (excitation control system) and slow system (liquid level control system). Two enhanced MPC algorithms have been acquired to meet the need of increasingly higher standards of modern industry.The first part combined PID algorithm with predictive functional control algorithm to derive the new predictive functional control based on PID algorithm (PIDPFC).As we all know that the actual controlled system is non-linear and the state of the system changes with the time going. It’s clearly inappropriate to calculate a fixed set to parameters before the system begins to run. The traditional PID control has the advantage of simple principle, strong robustness and easier adjustment with three adjustable parameters. By using the characteristics of the PID algorithm to improve the objective function of PFC, which derive a generalized improved predictive functional control with generalized PID parameters. Simulate the improved PID-based predictive functional control in excitation control system and process level control system.The second part combined fractional-order PID algorithm with predictive functional control algorithm to deduce the predictive functional control based on fractional-order PID algorithm (FOPIDPFC).In traditional PID control, the index of integral and differential link is a fixed integer, while the exponent of integral and differential link in actual system have greater flexibility that it can be fraction. It’s more meeting the actual system extending the PID control algorithm to fractional PID algorithm. The fractional order PID has five adjustable parameters and the design is more flexible. Therefore improve the objective function of predictive functional control algorithm through fractional PID algorithm to get a form of fractional order PID quadratic objective function. Thus derive improved predictive function based on fractional order PID algorithm and simulate in excitation control system and liquid level control system.The above research shows that the excitation control system is a typical fast demanding and difficult to model identification system, but the simulation studies show that the predictive functional control based on PID control algorithm and based on fractional order PID function algorithm can control the excitation system very well. It achieved more excellent effect than the conventional excitation control method.In the case of excitation control system having a large model mismatch, two types of algorithms can also play better control effect.In the process control industry, liquid level control system is a typical controlled plant. In the study we select a more complex dual-tank water level control system as a controlled plant. Simulation studies show that predictive functional control based on PID control algorithm and based on fractional order PID function algorithm has better control effects than traditional predictive control algorithm. Its flexibility is greater, the computer system overhead is small and the controller is more accurate than the PID controller.The research results show that compared to traditional control algorithm, the improved predictive functional control algorithm has the advantages of fast rising velocity, small steady-state errors, better robustness and strong anti-jamming performance etc. It provides two better control strategies for industrial control.