Investigation of the effectiveness of edible oils as solvent in reactive extraction of some hydroxycarboxylic acids and modeling with multiple artificial intelligence models

Sevindik Y. E., GÖK A., LALİKOĞLU M., Gülgün S., GÜVEN E. Y., Gürkaş-Aydın Z., ...Daha Fazla

Biomass Conversion and Biorefinery, cilt.13, sa.14, ss.13253-13265, 2023 (SCI-Expanded) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 13 Sayı: 14
  • Basım Tarihi: 2023
  • Doi Numarası: 10.1007/s13399-023-03853-2
  • Dergi Adı: Biomass Conversion and Biorefinery
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC
  • Sayfa Sayıları: ss.13253-13265
  • Anahtar Kelimeler: Carboxylic acid, Chemical experiment prediction model, Edible oil, Machine learning, Reactive extraction
  • İstanbul Üniversitesi Adresli: Evet

Özet

© 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.This study investigated the usability of different vegetable oils as solvents for separating citric, malic, and glycolic acids from aqueous solutions by reactive extraction method. A machine learning model was developed to predict intermediate values from the dataset created using the experimental results using multiple linear regression (MLR) and extreme gradient boosting (XGB). We used sunflower oil, corn oil, linseed oil, sweet almond oil, sesame oil, and castor oil in six types of vegetable oil. Trioctylamine (TOA) was used as an extractant in reactive extraction studies. The results obtained showed that approximately 99% of acids can be separated from their aqueous solutions when suitable mixtures of organic phases are used. Based on the results, we discovered that the XGB method outperforms the MLR method for each dataset. Thanks to the high-performance prediction model developed, it was possible to reach higher separation efficiencies by determining the optimum experimental conditions. In addition, the costs and wastes associated with experiments decreased due to the developed high-performance estimation model. The reactive extraction estimation model was publicly available on GitHub and open to other researchers.