Modelling and optimization of crude oil removal from surface water via organic acid functionalized biomass using machine learning approach

Christian O. Asadu,Benjamin Nnamdi Ekwueme,Chijioke Elijah Onu,Thomas O. Onah,Innocent Sunday Ike,Chinonso Anthony Ezema
DOI: https://doi.org/10.1016/j.arabjc.2022.104025
IF: 6.212
2022-06-04
Arabian Journal of Chemistry
Abstract:Banana peel fiber adsorbent (BPF) with well-arranged substructure of pores was fabricated via esterification reaction betwixt organic acid and biomass. The emerged adsorbent (BPF) was employed in taking away crude oil from water surface. Three machine learning tools such as RSM, ANN and ANFIS was employed for the modelling and optimization of the process. From results, the optimal oil layer removal of 98.2% was achieved at oil water ratio of 0.2g /100cm 3 . For now, BPF displayed high adsorptive prospect at a very low pH of 4 with 96.8 % oil removal. On the other hand, the activation energy, enthalpy change and entropy change of the system are (18.56, 25.44, -0.751 KJ/mols) and (25.77, 29.16, -0.813 KJ/mols) designating a non-spontaneous system. The process of removal by BPF really matched the Langmuir isotherm model as proved by statistical error analysis with highest adsorption capacity of 49.33mg/g as shown through equilibrium modeling. RSM displayed the optimum conditions of the key variables such as temperature, oil concentration, adsorbent dosage, pH and time as 100°C, 0.2g/100cm 3 , 2.1g, 6 and 75 mins, respectively. Analysis of the three generic algorithm indicated significant oil removal prediction with quite remarkably similar coefficient of correlation of 0.999. Additional statistical analysis suggested that RSM was marginally better than ANN and ANFIS for the modelling of crude oil removal via esterified banana peels fiber.
chemistry, multidisciplinary
What problem does this paper attempt to address?