Servo Mechanism Technique based Anti-Reset Windup PI Controller for Pressure Process Station

Transcription

1 Indian Journal of Science and Technology, Vol 9(11), DOI: /ijst/2016/v9i11/89298, March 2016 ISSN (Print) : ISSN (Online) : Servo Mechanism Technique based Anti-Reset Windu PI Controller for Pressure Process Station Eadalasarath Yadav and Thirunavuarasu Indiran * Deartment of ICE, MIT, Manial , Karnataa, India; it.arasu@manial.edu, sarath.eadala@gmail.com Abstract Bacground/Objectives: In this aer a new architecture is followed which is based on maniulating feedbac structure for ressure rocess. Methods/Statistical Analysis: This wor coes with the research based on anti-reset windu mechanism. Mathematical Modeling of ressure station is carried out using two-oint method. Windu action is being comlication in FOPDT rocess, thereforedesigning a controller for such roblem is always a big challenge for control engineers. As a art of controller design direct synthesis method has been used.aart from controller design, architecture of system will also influence the system erformance. Tending towards architectural maniulationa new structure has been followed. Findings: Result shows the significance and ositive aroach towards the roosed model for ressure rocessin both simulation and real time. Controller action can be otimized in a reliable way and energy utilization of actuator gradually reduced. Alications/Imrovements: In some real time nonlinear rocesses windu action maniulates lant rocess more than required limit, where rocess leads to saturation. In such scenarios the roosed aroach rovides lant to regain its osition and act within limits of its boundary. Keywords: Anti-Reset Windu, Architecture Maniulation, FOPDT, PID Tuning, Pressure Station Servo Mechanism 1. Introduction Classical PID control is being used in industries from last few decades and continuing the usage is because of its simlicity in designing and easily understandable architecture, however for most comlex industrial roblems the classical controller alone cannot be recommended for better erformance and stability requirements. Hence there comes the need to develo advanced controller for nonlinearity, robustness and otimality etc. In real time or industrial rocesses almost all rocess will be limited to their control action, so the controllers used will be within the setu. Hence user won t be having the access to maniulate the control arameters. The roblem arises for such ind of rocess is when it needs any maniulation or the external arameters influence the rocess. In such scenarios the rocess reaches saturation or there will be mismatch in rocess arameters. Considering this constraint, researchers following different aroaches to overcome and imrove the erformance in dynamic rocess. In this aer a different aroach is followed to overcome resetwindu issue by considering ressure rocess which is FOPDT model 1. The bloc diagram of control system including saturation nonlinearity is shown in Figure 1 where r is reference inut, y is outut, e is error (difference of between outut (y) and reference inut (r), C and P are Controller and Process lant resectively. u max and u min are higher and lower limits of saturation. u is inut and u * is outut of the saturation. * Author for corresondence

2 Servo Mechanism Technique based Anti-Reset Windu PI Controller for Pressure Process Station ste resonse. The selected oints for modelling are 23.8 and 63.2 from the ste resonse curve. The standard FOPDT model and resonse curve as follows 5 Figure 1. General bloc diagram with saturation. Conditions for Saturation is given by u* = + u sat u > +u sat (1) tds Qs () = e τ s + 1 (4) u* = u - u sat u + u sat (2) u* = u - u sat u < + u sat (3) 2. Overview of Integral Windu Windu roblem in the system occurs when the controller outut saturates or the controller outut is more than the limit of final control element (actuator) when subjected to transient resonse. Usually the reset windu occurs when the reference signal is alied rather than load disturbance. Furthermore, themost imortant effect of integral windu is considered when the order of the lant transfer function is low, which means that the value of dead time is lower than the time constant (τ). In order to run the lant in controlled region, its imortant reduces or eliminate the actuator saturation. The control action of reset windu can be considered in two ways. Firstly considering nonlinearity from the initial stage and designing the controller accordingly or designing nominal PID controller and noting the effect of nonlinearity on lant and alying comensation to it. In cases where nonlinearity exists in/out of the lant avoiding or limiting saturation maes the system to gain its stability. This can be done by smoothening set-oint changes or detuning the controller (by selecting more sluggish controller), which is not advisable in ractical cases.reset windu is one of the issue to be considered in such scenarios. There are different advancedand controller design techniques being used when windu roblemsarise 2-4. Figure 2. Selection of variables from ste resonse. The selection of different rocess variables from Figure 2 is given by Y = U F F YI U I τ = 1.5 (t 2 - t 1 ) (6) (5) 1 td = 1.5[( t1 t0) ( t2 t0)] (7) 3 Figure 3 deicts the real time model of ressure lant and rocess lant equation followed in the Equation (8) through substituting Equations (5), (6) and (7) in (4). 3. Modelling Pressure Station In this aer the lant is modelled by using two oint methods in which oerating conditions were based on linear characteristics and the FCE saturation limit of the flow rocess. The lant can be modelled by considering Figure 3. Real-time ressure rocess station. 2 Vol 9 (11) March Indian Journal of Science and Technology

3 Eadalasarath Yadav and Thirunavuarasu Indiran tds Qs () = e = e Ts s s 4. Architectural Model Usually in real time rocess based on requirement of outut the rocess the dynamics vary from time to time. As the dynamics of lant changes the tuned controller values may not show good effect on the lant. As the closed loo built-in controller will be fixed in its gain the resonse will be oor. Here comes the need of dynamic change in tuning with resect to rocess lant 6. In this Proosed architecture there exists two loos as shown in the Figure 4, in which one loo feuded to controller G c i and G c1. Here Gc = and Gc1 = K s The Bloc diagram as follows. (8) as saturation limits. Therefore the controller value K i is selected as a function of feedbac loo error ē and hence the integral gain will be always varying with resect to error ē. 5. Controller Design Even though there are many tuning methods exists for designing PID arameters, due to simlicity and flexible in design, direct synthesis is the method of tuning has been used in this aer. The same values are used both in conventional structure and roosed model also, the roortional gain K c and integral gain K i is given by 7 K c τ = (10) ( τθ) c T i = τ (11) Figure 4. Servo maniulation bloc diagram. The dynamic controller in this aer is conditions for selecting G c is given by r y < e <σ e = if 0 otherwise U max U min e = y r, σ = andδ = Where G c. The Figure 5. Model incororated with anti-reset windu. (9) Figure 5 shows the lant model incororated with anti-reset windu mechanism. U max and U min are selected Where K is rocess gain, τ is time constant and θ is time delay and τ c is usually selected as settling time of closed loo resonse. Table 5.1 reresents roortional gain (K c ) and integral gain (K i ). Therefore by calculating (10) and (11) Table 1. Controller arameters K c 5.1 K i Results and Comarison Below mentioned results clearly distinguish the comarison of conventional control effect and anti-reset windu controller effect on rocess station. Figure 6 shows simulation resonse of direct synthesis method with anti-reset windu. As the integral action is comletely deendent on rocess resonse, even the changes occurred in the rocess as disturbance (load) and reference, the controller action will be erformed with resect to those changes. Figure 7 shows the erformance of lant rocess with resect to setoint change and comarison is made to enable the difference between convention control and the controller with antireset windu. It can be clearly shown that the erformance with anti-reset windu has better resonse than conventional rocess. Figure 8 shows the controller action comarison in which controller with anti-reset windu has very smooth controller action comared with conventional rocess. Vol 9 (11) March Indian Journal of Science and Technology 3

4 Servo Mechanism Technique based Anti-Reset Windu PI Controller for Pressure Process Station and settling time of resonse with and without anti-reset windu. Figure 6. Simulation resonse of direct synthesis method with anti-reset windu. 6.1 Results Comarison Table System charateristics of first ste change of (20%). The Table 2 rovides the comarison of system characteristics lie ea time, rise time, ea over-shoot Figure 7. Comarison of rocess variable: Direct synthesis method with and without anti-reset windu. Table 2. Comarison of system charateristics (with and without anti-reset windu) Methods Pea time Rise time Pea Over-shoot (%) Settling time (samling time) (samling time) (samling time) Conventational Servo Mechanism Technique rocess error. Direct synthesis method is used for tuning and exerimental tests have been erformed in simulation as well as real time. The results clearly show that the rocess with anti-reset windu give better erformance comared with conventional method. Energy utilized by the controller is drastically reduced with resect to conventional method. 8. Acnowledgement Figure 8. Comarison of maniulated variable: Direct synthesis method with and without anti-reset windu. 7. Conclusion In this aer, Pressure station is considered as rocess lant, modelling of lant has been carried out by using two oint methods. A new aroach of architecture designing in rocess flow is followed in which the controller arameters acts as function of error and the integral gain (Ki) will be auto tuned with resect to the The authors would lie to than DST, Government of India for the sanction of this roject under the wide circular #SB/FTP/ETA-0308/2012 Dated 25th May, The authors also would lie to than MIT, Manial University for roviding the infrastructure facility for carrying out this Project. 9. References 1. Burns RS. Advanced control engineering. 1st ed. USA: Butterworth-Heinemann; Visioli A. Practical PID Control. London: Sringer-Verlag; Vol 9 (11) March Indian Journal of Science and Technology

5 Eadalasarath Yadav and Thirunavuarasu Indiran 3. Charavarthi MK, Vinay PK, Venatesan N. Design and simulation of internal model controller for a real time nonlinear rocess. Indian Journal of Science and Technology. 2015; 8(19): Tarbouriech S, Turner M. Anti-windu design: An overview of some recent advances and oen roblems. IET Control Theory and Alications. 2009; 3(1): Satheeshbabu R, Thirunavuarasu I. Certain discussions on 2DoF PID control with anti-windu schemes for the FOPDT. Proceedings of Control Instrumentation System Conference Proceedings; India Kiong TK, Htet KKK. Set-oint maniulation aroach toward overcoming limitation of closed architectural framewor. Proceedings of the International Multi Conference of Engineers and Comuter Scientists, IMECS; Hong Kong PID Controllers Design, Tuning and troubleshooting Available from: htt://teacher.buet.ac.bd/shouat/ ChE6303_handout5.df Vol 9 (11) March Indian Journal of Science and Technology 5