Skip navigation
Please use this identifier to cite or link to this item:
Title: Statistical modeling of supercritical extraction of hemp (Cannabis sativa) and papaya (Carica papaya) seed oils through artificial neural network and central composite design
Authors: Devi V.
Khanam, Shabina
Published in: Soft Computing
Abstract: Supercritical fluid extraction (SFE) is an effective and ecofriendly alternative for the oil extraction from natural products. However, its large-scale operation requires lots of resources like money, time and efforts. In this regard, it is always advantageous to get a prior idea about the process scale up and its optimization, which can be achieved through process modeling. In the present study, two statistical models; artificial neural network (ANN) and central composite design (CCD) are studied for SFE of hemp and papaya seeds. Hemp seed oil (HSO) and papaya seed oil (PSO) are rich in nutrition and have various health benefits, but not much studied. Studies on papaya seed extraction are almost negligible. Therefore, CCD and ANN modeling are investigated to predict and optimize the response of the SFE process for HSO and PSO. In CCD, face-centered and rotatable design is adopted. For ANN, feed-forward back propagation (FFBP) with multilayer perceptron networks is applied using Levenberg–Marquardt algorithm. The accuracy of the developed models is investigated through statistical parameters like average absolute relative deviation, mean square error, and correlation coefficient, R2. The most optimized FFBP-ANN model is obtained for [5 8 1] neurons configuration for HSO and PSO. Quadratic models, developed through CCD, exhibit good agreement with experimental data. In comparison, CCD modeling is observed to be slightly better at the prediction of experimental data as compared to ANN modeling. Effect of operating variables is investigated on the extraction yield of HSO and PSO, obtained through SFE and predicted through ANN and CCD. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.
Citation: Soft Computing
Issue Date: 2021
Publisher: Springer Science and Business Media Deutschland GmbH
Keywords: Artificial neural network
Central composite design
Supercritical fluid extraction
ISSN: 14327643
Author Scopus IDs: 55533845200
Author Affiliations: Devi, V., Department of Chemical Engineering, Indian Institute of Technology Roorkee, Roorkee, 247667, India
Khanam, S., Department of Chemical Engineering, Indian Institute of Technology Roorkee, Roorkee, 247667, India
Corresponding Author: Devi, V.; Department of Chemical Engineering, India; email:
Appears in Collections:Journal Publications [CH]

Files in This Item:
There are no files associated with this item.
Show full item record

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.