Hybrid neural network modeling and particle swarm optimization for improved ethanol production from cashew apple juice
- PDF / 1,561,187 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 71 Downloads / 187 Views
RESEARCH PAPER
Hybrid neural network modeling and particle swarm optimization for improved ethanol production from cashew apple juice Andréa da Silva Pereira1 · Álvaro Daniel Teles Pinheiro2 · Maria Valderez Ponte Rocha1 · Luciana Rocha B. Gonçalves1 · Samuel Jorge Marques Cartaxo1 Received: 17 April 2020 / Accepted: 9 September 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract A hybrid neural model (HNM) and particle swarm optimization (PSO) was used to optimize ethanol production by a floc‑ culating yeast, grown on cashew apple juice. HNM was obtained by combining artificial neural network (ANN), which pre‑ dicted reaction specific rates, to mass balance equations for substrate (S), product and biomass (X) concentration, being an alternative method for predicting the behavior of complex systems. ANNs training was conducted using an experimental set of data of X and S, temperature and stirring speed. The HNM was statistically validated against a new dataset, being capable of representing the system behavior. The model was optimized based on a multiobjective function relating efficiency and productivity by applying the PSO. Optimal estimated conditions were: S 0 = 127 g L−1, X0 = 5.8 g L−1, 35 °C and 111 rpm. In −1 −1 this condition, an efficiency of 91.5% with a productivity of 8.0 g L h was obtained at approximately 7 h of fermentation. Keywords Ethanol production · Cashew apple juice · Artificial neural network (ANN) · Hybrid neural model (HNM) · Particle swarm optimization (PSO) Nomenclature A1 Initial value of horizontal asymptote A2 Final value of horizontal asymptote ANN Artificial neural network dx Model increment Ftab Tabled value for the Fisher Test F HNM Hybrid neural model n Number of samples nv Number of variables estimated N Stirring speed (rpm) p Number of model parameters P Product concentration (g L−1) Pf Final product concentration (g L−1) PSO Particle Swarm Optimization RSD Residual standard deviation (%) S Substrate concentration (g L−1) S0 Initial substrate concentration (g L−1)
Sf Final substrate concentration (g L−1) T Time (h) tf Final time (h) x Model variable x0 Average value between horizontal asymptotes X Cell concentration (g L−1) X0 Initial cell concentration (g L−1) ε Error (%) −1 μS Specific substrate consumption rate (gsubs.g−1 cell.h ) −1 −1 μP Specific product production rate (gproduct.gcell.h ) μX Specific growth rate of cells (h−1)
* Luciana Rocha B. Gonçalves [email protected]
Introduction
1
Chemical Engineering Department, Universidade Federal do Ceará, Campus do Pici, Bloco 709, Fortaleza, CE 60455‑76, Brazil
Engineering and Technology Department, Universidade Federal Rural do Semiárido, Mossoró, RN, Brazil
2
Superscript cal Calculated exp Experimental min Minimum max Maximum n Normalized
Currently, ethanol is the main substitute for gasoline and can be obtained from alcoholic fermentation [1–3]. The main raw materials used for the industrial production of ethanol are corn and sugarcane. Other source
Data Loading...
