Assessment Of Diverse Controllers For A Cylindrical Tank Level Process

IJIRST International Journal for Innovative Research in Science & Technology Volume 1 Issue 6 November 2014 ISSN (online): 2349-6010 Assessment Of Diverse Controllers For A Cylindrical Tank Level Process A.Thamemul Ansari K.Sujitha K.Tharani Raja H.Kala Assistant Professor Saranathan College of Engineering, Trichy, Tamilnadu,India S.Abirami Assistant Professor Saranathan College of Engineering, Trichy, Tamilnadu,India Abstract This manuscript deals with the proposal of the controller for a Level process. In this cram the level process is controlled by PID controller using diverse control methods such as IMC, MPC and the result of assorted control algorithms has been compared in terms of time domain specifications like settling time (ts), rise time (tr), maximum overshoot (Mp) and performance indices like ISE, IAE, ITAE, MSE. The conventional PID controller has a shortcoming that, it makes a counteractive exploit only after the error has developed but not in progress whereas MPC endow with a corrective exploit in advance. The intention of this cram is to scrutinize the MPC strategy, analyse and contrast the control upshot with conventional control stratagem in maintaining a water level system. The concert of the conventional controller, IMC and MPC has been compared in which MPC controller gives a enhanced system performance in terms of settling time (ts), rise time (tr), maximum overshoot (Mp) and performance indices akin to ISE, IAE, ITAE, MSE. Keywords: Level process, MATLAB, IMC, MPC, PID control, and FOPDT. I. INTRODUCTION In process industries, the control of liquid level is a basic problem especially in industries like petro-chemical industries, paper mixing process or mixing treatment in the tanks. It is essential for a control systems engineer to understand the problem and to control the level. Level control is very important for mixing reactant products. The quantity of the product of mixture depends on the level of the reactants in the mixing tank. Chemical process industries and food processing industries employs mixing reactant process. Hence, level control is very important in process industries. Nowadays, control system engineers in process industry are using computer aided control system design for modelling, system identification and estimation. In this paper, the controller is designed and simulated in MATLAB environment. In the liquid level control system, the manipulated variables of the liquid level are controlled. In the industrial production process, there are many places where the liquid levels have to be controlled and then manipulate the liquid level to maintain accurately for a given value. The traditional controller methods use a classical PID method and the advanced control strategies include IMC and MPC. In this paper, the tuning is done using Z-N method, modified Z-N method, T-L method, CHR, IMC and MPC and the result have been compared in terms of time domain specification. II. EXPERIMENT DESCRIPTION The process setup consists of a supply water tank fitted with pump for water circulation. The level sensor is fitted on a transparent process tank which is controlled by adjusting water flow to the tank by pneumatic control valve. These units along with necessary piping and fittings are mounted in support housing designed to stand on bench top. The control cubicle houses process indicator or microcontroller, output indicator, power supply for level transmitter, control switches etc., the process parameter is controlled through computer or microprocessor controller by manipulating water flow to the process. The controller used here is direct controller, since it increases in error when the controller output increases All rights reserved by www.ijirst.org 82

SPECIFICATION: Product Product code Type of control Control unit Communication Level transmitter I/P converter Control valve Rotameter Pump Process tank Supply tank Air filter regulator Pressure gauge Overall dimensions Level control trainer 313 313A DDC SCADA Interfacing unit with control module with digital ADC/DAC conversion indicating controller RS232 Type capacitance,two wire,range 0-300 mm,output 4-20Ma Input 4-20Ma,output 3-15 psig Type pneumatic;size:1/4,input:3-15 psig,air to close,characteristics:linear 10-100 LPH Fractional horse power,type centrifugal Transparent,Acrylic,with 0-100% graduated scale SS304 Range 0-2.5 kg\cm2 Range 0-2.5 g\cm2(1no),range0-7 kg\cm2(1no) 440Wx445Dx750H mm Fig 1: Level process setup III. ZEIGLER-NICHOLS METHOD Zeigler-Nichols technique is possibly the largely recognized and the generally used process for tuning of PID controller. It is a trial and error technique. This technique is based on sustained oscillations. Z-N is also known as online (or) continuous cycling (or) ultimate gain tuning method. The design criterion for this method is ¼ decay ratio. Z-N method does not necessitate process model but it forces the process into a condition of marginal stability that may lead to unstable operation. P 0.5K u - PI 0.45k u P u /1.2 - PID 0.6k u P u /2 P u /8 IV. MODIFIED Z-N METHOD Modified Z-N settings is preferred in control loops, where the assess of oscillation afford by ¼ decay ratio and the corresponding large overshoots for set point changes are undesirable. Some overshoot 0.33K u P u /2 P u /3 No overshoot 0.2k u P u /2 P u /3 V. TYREUS-LUYBEN METHOD The Tyreus-Luyben method is a method, which is related to Z-N method but the final controller settings are diverse. Also, this technique only proposes settings for PI and PID. This method is based on ultimate gain and period. PID K u /3.2 2.2P u P u /6.3 All rights reserved by www.ijirst.org 83

VI. C-H-R METHOD The open-loop Z-N method is modified and proposed by Chein - Hrones and Reswich, which is recognized as CHR method. They use a design criterion that quickest response without overshoot or quickest response with 20% overshoots. The tuning procedure is similar to Z-N method. PID(0 % overshoot) 2.4d 0.42d PID(20% overshoot) 2d 0.42d PID(0 % overshoot) m 0.5d PID(20% overshoot) 1.4 m 0.47d VII. MODEL PREDICTIVE CONTROL Many process industries such as chemical plants, refining/petrochemical industries and oil refineries use an advanced method of process control known as Model Predictive Control (MPC). MPC criterion relies on dynamic models of the process and most of the linear empirical model obtained by system identification. MPC predict the future response of a plant. It optimizes the future plant behaviour by computing a sequence of future manipulated variable adjustments. Model Predictive Control focuses on constructing a controller that the control action before a change in the output set point actually occurs. When MPC combined with a traditional feedback operations, it enables a controller to make adjustments that are smoother and closer to optimal control action values. The advantages of MPC are that it ensures reduced maintenance and longer plant life and alarms the possible future problem in the plant. Fig 2: Block Diagram of MPC controller VIII. INTERNAL MODEL CONTROL IMC is another model based controller used in many industries. IMC makes use of a process model to infer the effects of immeasurable disturbance on the process output and eliminate that effect. The IMC controller consists of an inverse of the process model. IMC is a model based control technique. We develop a model-based procedure, where a process model is embedded in the controller by explicitly using process knowledge; by virtue of the process model improved performance can be obtained. Process model mismatch is common. The process model cannot simply be inverted to form the controller. It must be factored so that the resulting controller is stable and realizable. The open loop control strategy will not be able to maintain the output at set point. The closed loop oscillation technique developed by Ziegler and Nichols did not require a model of process. Direct synthesis was based the use of desired closed loop response and a process model to synthesize a control law. IMC + All rights reserved by www.ijirst.org 84

IX. MINIMUM ERROR INTEGRAL CRITERIA Tuning for ¼ decay ratio often leads to oscillatory responses and also this criterion consider only two points of closed loop response. The controller can also be design based on performance index. Some of the performance indexes are: (1) Integral of the absolute value of the error(iae): IAE= (2) Integral of the square value of the error(ise): ISE= (3) Integral of the time weighted absolute value of the error(itae): ITAE= (4) Integral of the time weighted square value of the error(itse): ITSE= X. RESULT The simulated block diagrams were implemented in MATLAB environment using different conventional control algorithm and model based control technique. Different control methods have a different retort for a unit step input. For auxiliary scrutiny, the retort of diverse controller is taken into consideration. From the response curve, it has been verified that amongst diverse controllers, MPC controller has a better performance in terms of settling time, rise time, peak overshoot, peak time and MPC also has better performance indices. Table - 1 Performance of Various Controllers TUNING CONTROLLER PEAK OVERSHOOT PEAK TIME RISE TIME SETTLING TIME Zeigler-Nichols 1.15 77 42.51 360 Modified ZN(Some overshoot) 0.5 140 74 720 No overshoot 0.5 160 90 1240 TL 0.059 200 131 980 CHR(Load Rejection) 0% 0.479 80 55.5 384 20% 0.735 81 50 350 CHR(Set Point Tracking) 0% 0.042 215 150 1380 20% 0.07 65 63 560 Cohen-Coon 0.85 65 47.75 354 IMC - - - 190 MPC - - - 135 Table - 2 Error Performance Criteria for Various Controller CONTROLLER METHOD ISE IAE MSE Ziegler-Nichols 305.4085 257.3462 0.9757 Modified ZN(Some overshoot) 296.3378 237.5796 1.0451 No overshoot 272.0205 272.1714 1.0451 TL 280.9931 280.3197 0.7356 CHR(Load Rejection) 0% 285.9747 270.0000 0.8401 20% 249.3440 294.2629 0.9807 CHR(Set Point Tracking) 0% 280.5777 281.4813 0.7218 20% 275.1120 258.5871 0.8642 Cohen-Coon 302.7569 247.0703 1.0058 IMC 304.8537 301.7100 0.6168 All rights reserved by www.ijirst.org 85

RESPONSE OF DIFFERENT CONTROLLERS 2.5 2 1.5 1 0.5 0 0 200 400 600 800 1000 1200 1400 1600 ZN-PID Modified ZN -Some overshoot Modified ZN-No overshoot TL-PID CHR(Load Rejection)-0% CHR(Load Rejection)-20% IMC-PID CHR(Set Point Tracking)-0% CHR(Set Point Tracking)-20% Cohen-Coon MPC -0.5 Fig 3: Response of the level process XI. CONCLUSION We have designed diverse controllers which include traditional controller strategies and advanced controller strategies for cylindrical tank level process and the results are compared in terms of peak time, rise time, settling time and peak overshoot and also in terms of performance indices. From the result and discussions it has been proven that MPC which is an advanced control technique has minimum settling time, no peak overshoot and no rise time than traditional controller strategies. Further enhancement could be implemented in real time by which the level of the tank is controlled at a desired set point. REFERENCE [1] System Identification and Comparison of Ziegler-Nichols and Genetic Algorithm for Moisture Process. H. Kala, S. Abirami, S.M. Giriraj Kumar. [2] Wikipedia, PID controller, http://en.wikipedia.org/wiki/pid_controller. [3] J. B. Ziegler and N. B. Nichols, Optimum settings for automatic controllers, ASME Transactions, v64, pp. 759-768, 1942. [4] James B. Rawlings, Tutorial Overview of Model Predictive Control, IEEE Control System Magazine, June 2000. [5] Wikipedia, Model predictive control, http://en.wikipedia.org/wiki/model_predictive_control. [6] Ziegler, J.G., N. B. Nicholas, 1942. Optimum setting for automatic controller, Transaction ASME, 64:759-768. [7] Kala. H, Abirami. S, Muthumari. S, Venkatesh. S, 2012. Model Identification and Comparison Of Different Controllers for Humidity Process, International Conference on Electrical Sciences. [8] D. Mercy, S. M. Giriraj kumar Tuning of controllers for nonlinear process using Intelligent techniques, IJAREEIE Vol. 2, Issue 9, September 2013. [9] Morari, M. and J. Lee, Model Predictive Control: Past, Present and Future", Comp. Chem. Eng., 21, 667-682, 1999. [10] G. Stephanopoulos, Chemical Process Control: An Introduction to Theory and Practice, Prentice Hall, Upper Saddle River, NJ, USA, 1984. [11] Dr. V. Balaji, Study of Model Predictive Control using NI Lab VIEW, International Journal of Engineering Research & Technology, Volume 3,Issue 2, july- December (2012),pp.257-266. [12] Aidan o Dwyer, Hand book of PI and PID controller tuning rules, 3rd edition All rights reserved by www.ijirst.org 86