A Stable Control of a Cstr with Input Multiplicities Using Artificial Neural Network Based Narma-l2

Page 1

www.as-se.org/ccse

Communications in Control Science and Engineering (CCSE) Volume 2, 2014

A Stable Control of a Cstr with Input Multiplicities Using Artificial Neural Network Based Narma-l2 Prabhaker Reddy Ginuga*, Radha Krishna Poondla Dept. of Chemical Engineering, University College of Technology, Osmania University, Hyderabad, INDIA *gpreddy_ouct@yahoo.com; rkpoondla@yahoo.com Abstract In this paper, the Neural network based NARMA-L2 controller is analyzed to an isothermal continuous stirred tank reactor (CSTR) which exhibits input multiplicities in space velocity on product of B. i.e., two values of space velocity will give the same value of product B. The Performance of Neural network based NARMA-L2 controller and conventional PI controller have been evaluated through simulation studies. As the NARMA-L2 controller provides always the two values of space velocity for control action and by selecting the value nearer to the operating point, it is found to give stable and better responses than conventional PI controller. The PI controller results in unstable condition or switch over from initia l lower input space velocity to higher input space velocity vice versa. Thus, NARMA-L2 controller is found to overcome the control problems of PI controller due to the input multiplicities. Keywords NARMA-L2 Controller; Isothermal CSTR; Input Multiplicities; Unstable

Introduction The t erm In put multi pli cit y mean s, more t han on e val ue of in put variable producing the same val ue of output. It is a kind of nonlinearity in the process. Input multiplicity occurs due to the competing effect s in t he processes. Dyn amic an d steady st ate behaviour of t he process wit h input multiplicity will remain di stinct at different input values for the same output. Processes wit h multiple reacti ons, multi react ors or recycle struct ures are shown t o [1, 2] . Conventional linear PI controller will have control problems like in stability, oscill atory exhi bit in put multiplicities [3-5] due to in put multipli cities in t he process. an d l ess economi cal In t he l ast t wo decades, a n ew direction to control has gained consi derable attention. This new approach to control is called ‘Intelligent control’. The term ‘intelligent control’ addresses to more general control probl ems. It may refer to systems, which cannot be adequatel y described by a differential equations framework. There are three basic approaches to intelligent control: knowledge- based expert s systems, fuzzy l ogi c an d n eural n et works. The t erm ‘conventional control’ refers to theories and methods that are employed to control dynamic systems whose beh aviour is pri marily described by differential equations. Among these intelligent controllers, neural networks control has [6, 7] for control of dynamic process, demonstrating the ability of handling non linearity. become popular tool In this work, the design and eval uation of neural n etwork based NARMA-L2 controller are present ed to overcome the control problems associated with conventional PI controller due to input multipliciti es. Description of CSTR with Input Mu ltiplicities We consider here a continuous stirred tank reactor (CSTR) with the following isothermal series an d parallel reactions; A

k1

2A

k2

B k3

C D

(1) (2)

The product B is the desired one. The mass balance equations for A an d B are given by: 2

Where,

72

dX 1/dt = - k1*X 1 – k3*X 1 + (CA,0 – X 1)*u

(3)

dX 2/dt = k1*X 1 – k1*X 2+ X 2*u

(4)


Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.
A Stable Control of a Cstr with Input Multiplicities Using Artificial Neural Network Based Narma-l2 by Shirley Wang - Issuu