This action might not be possible to undo. Are you sure you want to continue?

7, July 2012

**Modeling and Control of CSTR using Model based Neural Network Predictive Control
**

Piyush Shrivastava

Assistant Professor, Electrical& Electronics EngineeringDepartment, Takshshila Institute of Engineering & Technology, Jabalpur, Madhya Pradesh, India . consider plant behavior over a future horizon in time. Thus, the effects of both feedforward and feedback disturbances can be anticipated and eliminated, fact which permits the controller to drive the process output more closely to the reference trajectory. The classical MBPC algorithms use linear models of the process to predict the output of the process over a certain horizon, and to evaluate a future sequence of control signals in order to minimize a certain cost function that takes account of the future output prediction errors over a reference trajectory, as well as control efforts. Although industrial processes especially continuous and batch processes in chemical and petrochemical plants usually contain complex nonlinearities, most of the MPC algorithms are based on a linear model of the process and such predictive control algorithms may not give rise to satisfactory control performance [3, 4]. Linear models such as step response and impulse response models are preferred, because they can be identified in a straightforward manner from process test data. In addition, the goal for most of the applications is to maintain the system at a desired steady state, rather than moving rapidly between different operating points, so a precisely identified linear model is sufficiently accurate in the neighborhood of a single operating point. As linear models are reliable from this point of view, they will provide most of the benefits with MPC technology. Even so, if the process is highly nonlinear and subject to large frequent disturbances; a nonlinear model will be necessary to describe the behavior of the process. Also in servo control problems where the operating point is frequently changing, a nonlinear model of the plant is indispensable. In situations like the ones mentioned above, the task of obtaining a high-fidelity model is more difficult to build for nonlinear processes. In recent years, the use of neural networks for nonlinear system identification has proved to be extremely successful [59]. The aim of this paper is to develop a nonlinear control technique to provide high-quality control in the presence of nonlinearities, as well as a better understanding of the design process when using these emerging technologies, i.e., neural network control algorithm. The combination of neural networks and model-based predictive control seems to be a good choice to achieve good performance in the control. In this paper, we will use an optimization algorithm to minimize the

Abstract—this paper presents a predictive control strategy based on neural network model of the plant is applied to Continuous Stirred Tank Reactor (CSTR). This system is a highly nonlinear process; therefore, a nonlinear predictive method, e.g., neural network predictive control, can be a better match to govern the system dynamics. In the paper, the NN model and the way in which it can be used to predict the behavior of the CSTR process over a certain prediction horizon are described, and some comments about the optimization procedure are made. Predictive control algorithm is applied to control the concentration in a continuous stirred tank reactor (CSTR), whose parameters are optimally determined by solving quadratic performance index using the optimization algorithm. An efficient control of the product concentration in cstr can be achieved only through accurate model. Here an attempt is made to alleviate the modeling difficulties using Artificial Intelligent technique such as Neural Network. Simulation results demonstrate the feasibility and effectiveness of the NNMPC technique. Keywords-Continuous Stirred Tank Reactor; Neural Network based Predictive Control; Nonlinear Auto Regressive with eXogenous signal.

I.

INTRODUCTION

One of the main aims in industry is to reduce operating costs. This implies improvements in the final product quality, as well as making better use of the energy resources. Advanced control systems are in fact designed to cope with these requirements. Model based predictive control (MBPC) [1,2] is now widely used in industry and a large number of implementation algorithms due to its ability to handle difficult control problems which involve multivariable process interactions, constraints in the system variables, time delays, etc. The most important advantage of the MPC technology comes from the process model itself, which allows the controller to deal with an exact replica of the real process dynamics, implying a much better control quality. The inclusion of the constraints is the feature that most clearly distinguishes MPC from other process control techniques, leading to a tighter control and a more reliable controller. Another important characteristic, which contributes to the success of the MPC technique, is that the MPC algorithms

38

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 10, No. 7, July 2012

cost function and obtain the control input. The paper analyses a neural network based nonlinear predictive controller for a Continuous Stirred Tank Reactor (CSTR), which is a highly nonlinear process. The procedure is based on construction of a neural model for the process and the proper use of that in the optimization process. This paper begins with an introduction about the predictive control and then the description of the nonlinear predictive control and the way in which it is implemented. The neural model and the way in which it can be used to predict the behavior of the CSTR process over a certain prediction horizon are described, and some comments about the optimization procedure are made. Afterwards, the control aims, the steps in the design of the control system, and some simulation results are discussed. II. PREDICTIVE CONTROL The predictive controller, in summary, is characterized by computing future control actions based on output values predicted by a model, with vast literature and academic and industrial interest (Clarke, 1987; Garcia et all, 1989; Arnaldo, 1998) [4]. This section presents the concepts of predictive control based on NPC, using the usual optimization functions and control laws, applied to the conventional predictive controllers. A. Optimization functions The optimization function, usually represented by the index J, represents the function that the control action tries to minimize. In an intuitive way, the error between the plant output and the desired value is the simplest example of an optimization function, and it is expressed by: (1) Where: y(k) represent the plant output y k ref ( )represent the desired response e(k) represent the estimation error k is the sample time One of the most usual optimization functions is based on the square error and it is represented as: (2) But the optimization index can take forms of more complex functions. For predictive controllers, whose models are capable to predict N steps ahead, the simple application of the square error approach can present satisfactory results. This case admits that the optimization function is not limited to an only point, but an entire vector of N predicted errors. It seeks to optimize the whole trajectory of the future control actions in a horizon of N steps ahead. (3)

More complex optimization functions can consider the control effort. It is the specific case of GPC (Generalized Predictive Control), where the optimization index J can be expressed as:

(4) where: y(k ) - is the output plant estimation at instant = k Δu - is the control action increment. N1 - is the minimum horizon of prediction. NU - is the control horizon. NY - is the maximum horizon of prediction. The objective of the control problem is to minimize the index J, with respect to the control actions, looking for the points where the first order differential is null. III. NEURAL NETWORK PREDICTIVE CONTROL

By the knowledge of the identified neural model of the nonlinear plant which is capable of doing multi step ahead predictions, Predictive control algorithm is applied to control nonlinear process. The idea of predictive control is to minimize cost function, J at each sampling point:

ˆ J(t,U(k)) = ∑[ r(k +i) − y (k+i)] + ∑ρ[ Δu(k +i −1)]

t=N1 i=1

N2

2

Nu

2

(5)

With respect to the Nu future controls,

U ( k ) = [u ( k ).....u ( k + N u − 1)]T (6)

and subject to constraints:

Nu ≤ i ≤ ( N2 − nk )

(7)

Using the predictive control strategy with identified NARX model (NNMPC) it is possible to calculate the optimal control sequence for nonlinear plant. Here, term r(k+i) is the ˆ required reference plant output, y (k+i) is predicted NN model output, Δ u ( k + i − 1) is the control increment, N1 and N2 are the minimum and maximum prediction (or cost) horizons, Nu is the control horizon, and ρ is the control penalty factor[4]. The predictive control approach is also termed as a receding horizon strategy, as it solves the above-defined optimization problem [5] for a finite future, at a current time and implements the first optimal control input as the current control input. The vector u = [Δu(k),Δu(k+1),…Δu(k + Nu-1)] is calculated by minimizing cost function, J at each sample k for selected values of the control parameters {N1, N2, Nu, ρ}.

39

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 10, No. 7, July 2012

The ese control p parameters de efines the pr redictive control per rformance. N1 i usually set to a value 1 that is equal to t is t the tim delay, and N2is set to defin the predictio horizon i.e. t me ne on the num mber of time-st teps in the futu for which th plant respon ure he nse is r recursively pred dicted.

The purpose of our neura network mo e al odel is to do eries prediction of the plant output. Given a series of n t n time se control signals l

% d u and past data yt it is desired t predict the to

plant o output series yN.The network is trained to do one step k ahead p prediction[9], i.e. to predict the plant outp yt+1 given put the cur rrent control si ignal ut and p plant output yt . The neural network will impleme the function k ent n

ˆ yt +1 = f (ut , yt )

(12)

Figure 1: NNMP principle app F PC plied to CSTR ch hemical process

e n s ion The minimization of criterion, J in NNMPCis an optimizati pro oblem minimi ized iterativel ly. Similar t NN traini to ing stra ategies, iterativ search meth ve hods are appli to determi ied ine the minimum.

θ (ii+1) =θ (i ) +μ (i) .d(i)

(8) where θ ( i ) specif e, fies

the current iterate (number ‘i’), d (i) is the sea e arch direction a and is the step size. Variou types of a us algorithms exi ist, cha aracterized by the way in which search di w irection and st tep size are selecte e ed. In the present work Newton bas p sed Lev venberg–Marqu uardt (LM) al lgorithm is im mplemented. T The sea arch direction a applied in LM algorithm is: a

μ (ii)

As it is discussed above, yt h has to contai sufficient in ation for this p prediction to be possible.It is assumed that e informa yt is mu ultivariable. One problem is that this metho will cause od a rapidly increasing d divergence due to accumulati of errors. e ion efore puts high demands on a h accuracy of the model. The e It there better t model mat the tches the actua plant the les significant al ss the acc cumulated error A sampling time as large a possible is r. as an effe ective method to reduce the error accum e mulation as it effectiv vely reduces t number of steps needed for a given the f d time ho orizon. The neu network tr ural rained to do on step ahead ne predicti will model the plant. Th acquisition o this model ion l he of is also r referred to as S System Identifi fication. IV V. MODELIN OF NEURAL NETWORK PRE NG EDICTIVE CONTROL (NN NPC)

ˆ d (H[U i (t)] +λ i I)d i = -G[U i (t)]

The thr steps involv in the ANN model develo ree ved N opment are (9) A. Gen neration of Inp put-Output data a The da generated to train the netw ata o work should co ontain all the relevan information about the dyn nt namics of the CSTR. The e input w given to th conventiona model of th CSTR and was he al he from t the convention model, th input and output were nal he sampled for 0.02 sam mpling instants and the requ s uired sampled e rain rk. data are obtained to tr the networ

with Gradien vector and H nt Hessian matrix as:

G[U i (t)] =

∂ J(t,U(t)) | i ∂ U(t) U ( t ) =U ( t ) % ∂ U(t) % % = − 2ϕ T [U i (t)]E(t)+2 ρ U (t ) |U ( t ) =U i ( t ) ∂ U(t) (1 10)

% U(t)) ∂ 2 J(t,U | i t) ∂U(t 2 U ( t ) =U ( t ) T ˆ % % ⎞ ∂U (t) ∂U (t ) ∂ ⎛ ∂Y(t) E(t) ⎟ +2ρ = | i ⎜ ∂U(t) ⎝ ∂U(t) ∂U(t) ∂U (t ) U ( t ) =U ( t ) ⎠

H[U i (t)] = U

(1 11)

where B(i) specifi the approxi ies imation of the inverse Hessi e ian d h the and G[U(i)(t)] is the gradient of the J with respect to t con ntrol inputs. Th most popula formula kno he ar own as Broyde enFle etcher-Goldfarb b-Shanno (BFG algorithm to approxima GS) m ate the inverse Hessi is used her ian re[8]. The prop posed scheme of plementing the NNMPC is sh e hown in Figure 2. e imp me iction with Neu Networks ural Tim Series Predi

Figure 2: Input-Outp Sequence e put B. Neu Network A ural Architecture The fee forward net ed twork with sig gmoidal activa ation function was ch hosen based o the trials w on with different structures of multila ayer perceptron n.

40

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 10, No. 7, July 2012

Figu 3: ANN mo of the CSTR ure odel e o yer. The lowest error corresponds to 7 neurons in the hidden lay Hen it is selecte as optimal architecture of ANN. The AN nce ed a NN sele ected here co onsists of 4 neurons in the input layer, 7 n e neu urons in the hid dden layer and one neuron in the output lay n yer. The ANN archite e ecture used in the present w work is shown in Fig gure 3. The trai ining algorithm used in the C m CSTR modeling is g bac propagation algorithm. Before traini ck n ing the proce ess wei ights are initia alized to small random numb bers. The weigh hts are adjusted till error gets mi inimized for a training se all ets. hen or t ing Wh the error fo the entire set is acceptably low, the traini is s stopped. Tab 2 shows th parameters used in developing the AN ble he NN model for the CST TR

Fi igure 4: (a) One step ahead pr rediction of mo odel, (b) Prediction error betw ween model ou utput and predi icted output Vali idation tests o test set: on

Parame eters Input neu urons Output Ne eurons Hidden l layer Neuro ons No. of hidd layer den Activation f function Training alg gorithm Iterati ion Architec cture Initial we eights

Va alues 4 1 7 7 Sigm moidal Levenberg g-Marquardt 10 0000 Feedf forward 1

Figur 5 :(a) one ste ahead predic re ep ction of model (validation set), (b Prediction e b) error between m model output an predicted nd output (validati set) o ion V. CONT TINUOUS STIRRE TANK REAC ED CTOR

Table 2: ANN Parameters fo CSTR model N or ling C. Model Validat tion The final step in developing the model is v e n t validation of t the model [11]. Valid dation is perfo ormed by evalu uating the mod del per rformance usin trained data and test data The input a ng a a. and targ were prese get ented to the network and t network w n the was trai ined using Lev venberg-Marqu uardt algorithm.

The Continuous Stirred Tank Reactor [6] is shown in e Figure 6.This CSTR m model in used a the nonlinea system. as ar

The im mage part with relationship ID rId50 was not found in the file.

Va alidation tests on training se et:

Figure 6: C Continuous Stir rred Tank Reac ctor The equations w e which shows th dynamic m he model of the system is

41

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(14)

(15) d uct where h (t) is the liquid level, Cb(t) is the produ ncentration at th output of th process, w1(t is the flow ra he he t) ate con of t concentrate feed Cb1 an w2(t) is the flow rate of t the ed nd the dilu uted feed Cb2 .The input con ncentration are set to Cb1=24 e 4.9 and Cb2= 0.1.Th constants associated w d he with the rate of con nsumption are k1=k2=1. e oller is to main ntain the produ uct The objective of the contro con ncentration by a adjusting the fl w1 (t), w2 ( =0.1.The lev low (t) vel of t tank h is n controlled. The designed controller uses a the not s neu ural network m model to pred future CST responses to dict TR pot tential control s signals. The tra aining data we obtained fro ere om the nonlinear mod of CSTR. del VI. SIM MULATION RES SULTS AND CON NCLUSION

Figure 8: c control signal b the controller by In this pa aper modelin of CSTR has been ng R implem mented using artificial neur networks. The neural ral model has been train using data set obtained fr ned rom dynamic ons. Feed forw ward neural n network has b been used to equatio model the CSTR. Th neural mod has been d he del designed as a black b model. Th simulation results from conventional box he model a the neural model were co and ompared for th given input he variatio and the re ons esults have been found satis sfactory. The simulat tion shows tha implementat at tion of the Neu Network ural based a advanced contr rollers for the s set-point tracki case were ing able to force process output varia o ables to their t target values smooth and within r hly reasonable rise and settling ti e imes. VII. REFERE ENCES [1] Garcia C. E and Morari, M. 1982. In E. , nternal model l-I. g some new resu ults,Industrial control “A unifying review and s Engine eering Chemica Process ”. De 21, 308--32 al ev. 23. [2] L.G. Lightbody and G. W Irwin, “Neu networks W. ural onlinear adapt tive control, “in Proc. IF FAC Symp. for no Algorit thms Architect tures Real-Tim Control, B me Bangor, U.K. pp. 1–1 1992. 13, [3] arke, C. Mo ohtadi and P S. Tuffs, P. D. W. Cla ralized Predic ctive Control Basic “Gener Algorithm”, Automa atica, Vol.23, n no.2, pp: 137- 148, 1987. [4] Morari, M. and Lee, J. H 1999. Mod predictive H. del l: nt nd control past, presen and future, Computers an Chemical Engine eering, 23, 667-682. [5] X. Zhu, D.E Seborg, “N E. Nonlinear mod predictive del l mmerstein mod dels”, in. Proc. International control based on Ham Symposium on Proce System Eng ess gineering, Seou Korea, pp. ul, 000, 995–10 1994. [6] ova ural network Vasičkanino and M. Bakošova, “Neu predicti control of a chemical reactor” Proce ive f eedings 23rd Europe Conference on Modellin and Simulat ean e ng tion ©ECMS Javier O Otamendi, And drzej Bargiela, 2006. [7] J. D. Mornin ngred, B. E. Pa aden, D. E. Se eborg, and D. llichamp, “An adaptive nonli inear predictive controller,” e A. Mel in Pro Amer. Co oc. ontrol Confer rence., vol. 2 2,,pp. 1614– 1619,19 990. [8] N. Kishor, “Nonlinear p predictive cont trol to track ed NNARX model of a hydro deviate power of an identified N plant”. Expert Syste ems with App plications 35, 1741–1751, 2008. Tan, Y. and Cauwenbergh A. “Non-lin he, near one step [9] orks: control strategy and ahead control using neural netwo

e f s govern theCST TR The objective of the control strategy is to g dyn namics to force the system concentration t track a certa c to ain set-points. In this system, the in nput is the cool flow rate a lant and the output is th concentratio of the pro he on ocess [12]. T The ntifier is train and initialized before th control acti ned he ion iden star The input v rts. vector of the identifier inclu i udes coolant flo ow rate at different t es time steps (the sampling time is 20sec). e The performance of the propose controller is shown in Figu e ed s ure 7. E Evidently, the concentration values of the p plant could tra ack the set-point valu excellent. It is to be noted that to impro ues t d ove ponse, one ma consider a larger predicti ay ion the transient resp me. markable to note that bec n cause of high hly tim It is rem non nlinearity natur of CSTR process, using the convention re p nal con ntrol technique could not rea the contro task. It can be e ach ol see in figure 7 th controller output is track en hat o king the referen nce sign nal.

Fig gure 7: Respons graph with and without co se a ontroller

42

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

stability design”, Automatica, vol 32, no. 12, 1701-1706, 1996. [10] Dan, W.P., 1996. Artificial Neural Networks- Theory and Applications. Prentice Hall, Upper Saddle River, New Jersey, USA. [11] S.A.Billings, and W.S.F. Voon, “Correlation based model validity tests for nonlinear models. International Journal of Control, 44, 235–244.1986.

AUTHORS PROFILE

Author is presently working as Assistant Professor in Electrical and Electronics Department of Takshshila Institute of Engineering and Technology. He received the Masters degree in Electrical Engineering with specialization in Control Systems Engineering from Jabalpur Engineering College. His area of specialization is in Neural Networks, Control Systems, Fuzzy Logic.

43

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

- Journal of Computer Science IJCSIS March 2016 Part II
- Journal of Computer Science IJCSIS March 2016 Part I
- Journal of Computer Science IJCSIS April 2016 Part II
- Journal of Computer Science IJCSIS April 2016 Part I
- Journal of Computer Science IJCSIS February 2016
- Journal of Computer Science IJCSIS Special Issue February 2016
- Journal of Computer Science IJCSIS January 2016
- Journal of Computer Science IJCSIS December 2015
- Journal of Computer Science IJCSIS November 2015
- Journal of Computer Science IJCSIS October 2015
- Journal of Computer Science IJCSIS June 2015
- Journal of Computer Science IJCSIS July 2015
- International Journal of Computer Science IJCSIS September 2015
- Journal of Computer Science IJCSIS August 2015
- Journal of Computer Science IJCSIS April 2015
- Journal of Computer Science IJCSIS March 2015
- Fraudulent Electronic Transaction Detection Using Dynamic KDA Model
- Embedded Mobile Agent (EMA) for Distributed Information Retrieval
- A Survey
- Security Architecture with NAC using Crescent University as Case study
- An Analysis of Various Algorithms For Text Spam Classification and Clustering Using RapidMiner and Weka
- Unweighted Class Specific Soft Voting based ensemble of Extreme Learning Machine and its variant
- An Efficient Model to Automatically Find Index in Databases
- Base Station Radiation’s Optimization using Two Phase Shifting Dipoles
- Low Footprint Hybrid Finite Field Multiplier for Embedded Cryptography

This paper presents a predictive control strategy based on neural network model of the plant is applied to Continuous Stirred Tank Reactor (CSTR). This system is a highly nonlinear process; therefo...

This paper presents a predictive control strategy based on neural network model of the plant is applied to Continuous Stirred Tank Reactor (CSTR). This system is a highly nonlinear process; therefore, a nonlinear predictive method, e.g., neural network predictive control, can be a better match to govern the system dynamics. In the paper, the NN model and the way in which it can be used to predict the behavior of the CSTR process over a certain prediction horizon are described, and some comments about the optimization procedure are made. Predictive

control algorithm is applied to control the concentration in a continuous stirred tank reactor (CSTR), whose parameters are optimally determined by solving quadratic performance index using the optimization algorithm. An efficient control of the product concentration in CSTR can be achieved only through accurate model. Here an attempt is made to alleviate the modeling difficulties using Artificial Intelligent technique such as Neural Network. Simulation results demonstrate the feasibility and effectiveness of the NNMPC technique.

control algorithm is applied to control the concentration in a continuous stirred tank reactor (CSTR), whose parameters are optimally determined by solving quadratic performance index using the optimization algorithm. An efficient control of the product concentration in CSTR can be achieved only through accurate model. Here an attempt is made to alleviate the modeling difficulties using Artificial Intelligent technique such as Neural Network. Simulation results demonstrate the feasibility and effectiveness of the NNMPC technique.

- A REVIEW OF OPTIMIZATION TECHNIQUES IN ARTIFICIAL NETWORKS.
- Artificial Neural Networks an Overview
- 05973664
- DCDIS-2007
- HazaNuzlyAbdullHamedMFSKSM2006TTT
- Jordan Neural Network MPC Maciej
- matlab
- Geodezija_2012_04_Chapter6
- Computational Intelligence Approaches for Pattern Discovery in Biological Systems
- Power System Control by Embedded Neural Network in Hybrid System Modeling
- r07320505 Neural Networks
- ANN_SET2
- new paper
- 2
- Particle Swarm Optimization Using C#
- Optimization of injection molding process parameters using integrated artificial neural network model and expected improvement function method
- Swarm Optimized Modular Neural Network Based Diagnostic System for Breast Cancer Diagnosis
- A 03910109
- Net Elastic Integration Analysis on the Self-Organizing Map
- Gecco 2010 k Out Nik
- 1-s2.0-S0925231201005999-main
- Application of Artificial Intelligence in Gas Storage Management
- HAND WRITING RECOGNITION USING HYBRID PARTICLE SWARM OPTIMIZATION & BACK PROPAGATION ALGORITHM
- okt09_bengio_ldhr_01
- 9
- sdarticle
- 114863579-b-04090610
- International Journal of Engineering Research and Development (IJERD)
- Evolving an Artificial Neural Network Classifier for Condition
- h 04094751
- Modeling and Control of CSTR using Model based Neural Network Predictive Control

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd