POSTAL ADDRESS: GEETHA M, RESEARCH SCHOLAR, DEPARTMENT OF INSTRUMENTATION AND CONTROL ENGG, PSG COLLEGE OF TECHNOLOGY, COIMBATORE.
GUIDE NAME: Dr. JOVITHA JEROME
PROBLEM ADDRESSED: LIMITATIONS OF CLASSIC CONTROL SCHEME IN CSTR:
Continuous measurement of concentration is not possible.
So temperature is measured to infer about the concentration.
Cascade control scheme uses more number of transmitters and controllers.
Measurement inaccuracies are involved in conventional transmitter.
ABSTRACT:
The embedded based implementation of virtual feedback control of Continuous Stirred Tank Reactor (CSTR) using Compact RIO through LabVIEW Platform is attempted in this work. State measurements are done using hardware sensor devices in process industries. It has inaccuracies due to different disturbances and ageing effect. Hence soft sensor techniques which are specifically named as state estimation techniques are used. Estimation is the determination of constants or variables for any system according to a performance level and based on the measurements taken from the process. It is an important pre-requisite for the safe and economical process operations in plant analysis.
Estimation of states of a Continuous Stirred Tank Reactor (CSTR), which is highly non-linear is important for performance prediction, control application and simulation analysis. CSTR has wide applications in the field of petrochemical industries and biological processes. Estimation technique is used as the continuous measurement of reactant concentration and temperature in the reactor is difficult. Industrial reactor is usually controlled by a cascade structure consisting of a master controller for reactor temperature and an underlying slave controller for the cooling circuit. The cascade control provides a robust operation, but often lacks in control performance concerning the required strict temperance control. So we need state estimation techniques to overcome the difficulties of conventional control schemes.
EKF, a modified version of Kalman filter is widely used to estimate the states of non-linear systems. EKF is extension to non-linear domain through local linearization. It is known for its high convergence rate, which improves the transient performance significantly. Additionally, accurate estimation and convergence in steady state requires high-frequency signals, which are also inherently met by EKFs with the measurement noises included in the model. Also, it gives better performance under different process uncertainties. The proposed block diagram is shown in Figure 1.
Figure 1. Block Diagram of Virtual Feedback Control
The proposed methodology aims at real time implementation of soft sensor techniques in CSTR process, real time implementation of Virtual feedback control for CSTR Bench mark system using Compact RIO(Reconfigurable input/output system) in LabVIEW platform have been employed. It is used to control the state variables of CSTR process using soft sensor technique namely Extended Kalman Filter (EKF) in the feedback path to estimate concentration and temperature of reactants in the reactor. Simulation study has carried out to assess the performance of EKF and closed loop simulation was performed for conventional Temperature control and virtual feedback control for direct Concentration control as well as Temperature control under various conditions and the results were analysed. The P&I Diagram of CSTR bench mark system is shown in Figure 2.
Figure 2. P & I Diagram of CSTR Bench mark system
This proposed work tries to project the efficiency of the real time controller developed using compact RIO, the ease with which many such embedded controllers can be developed in a very short duration of time at a reduced cost of development and maintenance and at the ruggedness and flexibility offered by the controller developed. The independent embedded controller has the ability to function without the help of a computer, showing its tremendous capability of running processes in a plant even when the centralized computer is undergoing repair or maintenance.
