Sampling Control Design And Batch Operation Optimization For Temperature Regulation Device Of Crystallization Process

In this thesis,a sampling control system is designed in the discrete-time domain for the temperature regulation device of industrial crystallization process.The batch optimization method is studied,and a centralized temperature monitoring platform is designed for continuous crystallization process.Firstly,for an inverse response process with time delay(also known as non-minimum phase process),a two-degree-of-freedom control method based on a simplified generalized predictor(SGP)structure is proposed.The delay-free output is predicted based on the process dynamic model and a dead-time compensator(DTC),which can eliminate the negative influence from time delay and measurement noise on the process output.The disturbance rejection controller is designed based on the desired complementary sensitivity function,which can eliminate the affect from the process transfer function zeros in the right half-plane(RHP),and thus improve the disturbance rejection performance.The set-point tracking controller is designed based on internal model control theory(IMC)to achieve fast tracking the target output.The main advantage of the two controllers is that either has an adjustable parameter which can be monotonically adjusted to realize a compromise between the control performance and robust stability.The robust stability constraints in case of model uncertainty are given based on the small gain theorem.The feasibility and superiority of the proposed scheme are verified by simulation comparison with the recent literature.Secondly,the proportional(P)type and proportional differential(PD)type indirect iterative learning control(ILC)design schemes with dual-loop control structures are proposed for batch processes with time delay.The predictor based active disturbance rejection control(PDRC)structure is adopted in the inner control loop,which treats the external load disturbance and system uncertainty as "total interference" to perform disturbance rejection,and therefore can guarantee good dynamic control performance and robust stability along the time direction for open-loop stable,integral and unstable processes.The delay-free output can be predicted by a filter-based smith predictor(FSP),such that the effect of time delay to batch process control can be eliminated.The tracking performance can be gradually improved along the batch direction by the P-or PD-type ILC controller in the outer loop,which can realize the learning and updating of setpoint command based on the historical batch data.The sufficient conditions of robust stability along the time and batch directions are analyzed by establishing a two-dimensional(2D)system description for the ILC system,such that the controller parameter tuning method is given.Simulation on an injection molding batch process is conducted to verify the effectiveness and advantages of the proposed method.Then the proposed method is applied to batch experiments of temperature control for a 4-liter jacketed crystallization reactor,to verify the effectiveness and superiority.A centralized temperature monitoring platform design scheme is proposed for the DN15 tubular continuous crystallization reactor.The monitoring program,device communication,signal processing,data storage,and human-computer interaction interface of the host computer are designed based on the LabVIEW software to realize graphical operation.Multiple JULABO CF41 temperature control devices are regard as the slaver computers,and the real-time full-duplex data communication is designed based on the RS232 communication interface to realize fast communication with the host computer.The designed monitoring platform can ensure real-time control of the temperature in different sections of the continuous crystallizer,thus facilitating the subsequent development of advanced control methods and centralized monitoring systems.