Adaptive Sampling Distributed Predictive Control Strategy And Its Application

New Industrial Process(NIP)includes many aspects related to the development of people's livelihood,such as petrochemical industry,food and pharmaceutical industry,metallurgy and construction industry.With the development of social science,NIP presents the characteristics of strong integration and complicated division of labor.The industrial production capacity demand and economic benefits promote the development of NIP towards distributed control management.Distributed control is an effective way to save energy as well as,reduce emission.The traditional control strategy is unable to deal with the high-dimensional and multivariable coupling process of NIP in large industrial process,so the key problem to realize the whole process optimization of NIP is to effectively reduce dimension,and deal with various disturbances in the complex industrial productio n environment.Centralized Model Predictive Control(CMPC)cannot handle the increase of variable dimension.However,Distributed Model Predictive Control(DMPC)can better deal with NIP,which is a high-dimensional and multivariable coupling process.DMPC transforms a complex centralized optimization problem into an optimization problem which can be solved separately by multiple subsystems.The large number of subsystems brings new problems on how to choose the appropriate sampling strategy.In the traditional controller design,the sampling interval of each subsystem controller is the same.However,in the complex industrial manufacturing process,such subsystem is multi-input and multi-output.The production process of each subsystem is different,and the corresponding operation is also different.The dynamic behavior is necessarily different.It is unrealistic to adopt the same sampling interval everywhere in the system.The optimal sampling interval of each subsystem of complex industrial production is bound to change the dynamic behavior caused by the disturbance.Therefore,this paper proposes to adjust the sampling interval according to the dynamic behavior of each subsystem at different moments,so as to better capture the dynamic behavior of each subsystem.The performance of the control system can reach a higher level.This is the idea of Adaptive Sampling(AS),which takes the typical process of chemical process--reactor-storage tank-separator(RSS)as the controlled system,and the main research contents are as follows:1)Numerical modeling of control object RSS.On the basis of understanding the dynamic process of control object RSS of chemical process,th e corresponding dynamic process model is established according to its dy namic mechanism.The operation variables and working conditions of the control object RSS were analyzed,and the simulation was carried out to study the differences under the adaptive sampling CMPC strategy.2)DMPC strategy of RSS process.The distributed predictive performance of controlling object RSS is studied by means of theory,experimental settings and different simulations.On this basis,the distributed predictive control device is designed to master the main factors affecting the system coordination and s tability,so as to deeply understand the influence of subsystem dynamic behavior on the selection of optimal sampling interval.3)DMPC strategy under RSS process adaptive sampling method.In the adaptive sampling predictive control structure,the sampling time of the system can be reasonably selected to improve the working efficiency of the system.In RSS system,according to the output dynamic behavior data of each subsystem,the o ptimal sampling interval of each subsystem is calculated,so that th e dynamic behavior of the system can be collected more accurately.The results verify the correctness of the strategy.Under the DMPC control strategy based on adaptive sampling mechanism,the anti-jamming ability and dynamic adjustment ability of RSS system are better,and the overall performance of the system is improved.4)Distributed economic model predictive control based on adaptive sampling mechanism.Due to the realistic needs of enterprises,the economic benefits of NIP are the focus of attention.According to the main factors affecting the economic benefit,the cost function is established to replace the original objective function for the controlled system.The Adaptive Sampling Economic Model Predictive Control(AS-EMPC)of RSS was analyzed.Adaptive Sampling mechanism was applied to design model based on subsystem of Economic Model Predictive controller,which can greatly reduce the interference of external disturbances on the syst em normal working condition.Under AS-EMPC,the economic benefits of RSS system have been steadily improved,and the stability and security have been guaranteed.