CHEMICAL ENGINEERING TRANSACTIONS VOL. 74, 2019 A publication of The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš, Laura Piazza Copyright © 2019, AIDIC Servizi S.r.l. I SBN 978-88-95608-71-6; I SSN 2283-9216 Development of Intelligent Models for the Prediction the Dynamics of Nonlinear Process Eric M. L. Luz, Brunno F. Santos* Department of Chemical and Material Engineering (DEQM). Pontifical Catholic University of Rio de Janeiro (PUC-Rio). Rua Marquês de São Vicente, 225 – Gávea, Rio de Janeiro – RJ, 22430-060, Brazil. bsantos@puc-rio.br The use of artificial intelligence (AI) techniques to address nonlinearity problems is gaining importance, mainly in context of industry 4.0. The present work was development of models based on artificial intelligence was performed to predict the concentration of cyclopentenol in a CSTR chemical reactor with the Van De Vusse kinetics. This reaction is known for its nonlinear behavior, which makes prediction difficult. Using simulations of this reactor in open and closed mesh combined with certain perturbations, a database was generated to train the models. The software MATLAB 2018a was used to implement the neural and the ANFIS systems. The input layer could be feed flow, feed temperature, heat provided by the thermal jacket, outlet temperature. The output were concentration of the component B and reactor temperature one step ahead. The models had good agreement, once presented remarkable performance reproducing the dynamics, obtaining R² values near to 1 and error index near to 0. 1. Introduction In recent years, the use of artificial intelligence techniques to deal with nonlinearity problems has been gaining importance, especially in the current competitive world, where cost reduction may be the differential. In the works Yang et al. (2018), Zhou et al. (2018) and Tsai (2011), artificial intelligence techniques were used to solve nonlinear problems. Predictive models involving process with chemical reaction can be difficult due to nonlinearity of chemical kinetics and the possibility of serial and parallel reactions. In this context, the CSTR chemical reactor with Van de Vusse kinetics is used as the standard nonlinear chemical process for the performance evaluation of controllers and predictive models. This reaction was studied in the Engell et al. (1993), Schäfer, et al. (2018) and Cassol et al. (2018). Figure 1: Van de Vusse CSTR reactor DOI: 10.3303/CET1974128 Paper Received: 17 July 2018; Revised: 28 December 2018; Accepted: 29 April 2019 Please cite this article as: Luz E., Santos B.F., 2019, Development of Intelligent Models for the Prediction the Dynamics of Nonlinear Process, Chemical Engineering Transactions, 74, 763-768 DOI:10.3303/CET1974128 763