WIT Press

Application Of A Neural Mini-net To Identification Of Kinetic Parameters Of Chemical Processes

Price

Free (open access)

Paper DOI

10.2495/AI940051

Volume

6

Pages

8

Published

1994

Size

439 kb

Author(s)

D. Popovic & S. Zhou

Abstract

— Two methods for estimation of kinetic parameters in the Arrhenius reaction rate function, the logarithmic and the learning method using a neural mini— net, are proposed, and sucessfully applied to a certain reaction taking place in a CSTR. More implortantly, the application example shows the plausibility of introduction of the up datable activation function into neural networks, in the background of a chemical kinetics. I. INTRODUCTION Continuous stirred tank reactors (CSTRs) are widely applied in chemical and biochemical engineerings for the conversion of reactants into products [4], The Arrhenius reaction rate function k(c,T) = &<, exp(— E(c)/AT), with E(c) = E<> 4- Jc c, plays here an important role in the stability analysis of the reaction, the l

Keywords