CHEMICAL ENGINEERING TRANSACTIONS VOL. 70, 2018 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest Editors: Timothy G. Walmsley, Petar S. Varbanov, Rongxin Su, Jiří J. Klemeš Copyright © 2018, AIDIC Servizi S.r.l. ISBN 978-88-95608-67-9; ISSN 2283-9216 Adjustment of the Minimum Spouting Velocity in a Conical Spouted Bed from Artificial Neural Networks Juan F. Saldarriaga a, *, Fabio Freire b , Jose Freire b a Dept.Civil and Environmental Engineering, Universida de los Andes, Carrera 1Este#19A-40, 111711 Bogotá, Colombia b Dept. Chemical Engineering, Universidade Federal de São Carlos, Washington Luís km 235, São Carlos, Brasil jf.saldarriaga@uniandes.edu.co The minimum spouting velocity is one of the fundamental parameters in the application of the conical spouted bed pyrolysis, gasification and combustion, therefore it is necessary the adequate calculation and the use of a correlation with good adjustment, for this an artificial neural network has been applied, in order to improve the correlations of the literature. Six variables that involve different geometrical parameters and operation of the bed have been used. With the purpose to compare the results of the model with the experimental data and those predicted by the empirical equations, the quadratic error has been used. Although there is a complex relationship between the input variables and the output variables, and despite numerous of available data, the training and test steps of the network show a good adjustment with respect to the experimental values. This shows that an artificial neural network is an agile method to predict the minimum velocity of the bed at the pump, especially when the relationship between geometric parameters, operating parameters and minimum speed is complex and difficult to define. 1. Introduction The spouted bed reactor can achieve an intensive gas-solid contact, an important factor for the valorisation of biomass from thermochemical processes. Likewise, it operates stably in a wide range of gas flow due to the properties of mass and heat transfer. In this, you can process materials of different types of particle diameter and geometry. The most recent applications of the spouted bed reactor are many and varied, standing out in the industrial physical processes: drying of seeds, pharmaceutical powders and biomaterials (Konopka et al., 2008), granulated (Borini et al., 2009), materials coating (Rocha et al., 1995), and solid mixture (Hao et al., 2008). Applications in industrial chemical processes include gasification (Spiegl et al., 2010), combustion (Rasul, 2001) and pyrolysis (Olazar et al., 2001). Different modifications of spouted bed reactor in the original have been proposed in the literature with the aim of improving its performance. These are mainly the geometry of the contactor and the gas input to the bed, this in order to increase the handling capacity of the coarse particles, decrease the pressure drop, improve the cyclic movement of the particles and keep the operation stable in a wide range of gas flow velocities (Olazar et al., 1992). In the same way, the operation can be carried out in a spouted bed reactor with very short gas residence times (as low as milliseconds) (Olazar et al., 1992). According to Altzibar et al. (2013a), a crucial parameter that limits the scaling of the spouted bed is the relationship between the inlet diameter and the particle diameter, likewise, they argue that the inlet diameter should not be 20 or 30 times larger than the average particles diameter in order to reach the condition of spouted bed. The usual solution for this drawback is the use of a suction tube. Different configurations of the aspiration tube have been developed: solid, porous and open (Altzibar et al., 2013b) and are especially suitable for vigorous contact. The minimum spouting velocity is an essential variable in the operation of a conical spouted bed and is one of the parameters required for design and scaling. Saldarriaga et al. (2016) calculated the minimum spouting velocity for different types of biomasses and configurations of both the contactor and internal devices, generating new correlations for irregular particles (biomass) because those provided in the literature did not have DOI: 10.3303/CET1870208 Please cite this article as: Saldarriaga J.F., Freire F., Freire J., 2018, Adjustment of the minimum spouting velocity in a conical spouted bed from artificial neural networks , Chemical Engineering Transactions, 70, 1243-1248 DOI:10.3303/CET1870208 1243