Investigation of comparative machine learning models in effluent dephenolization process onto H3PO4-anchored corn cob

Abstract

The objective of the present study is to investigate the application of artificial intelligence tools in dephenolization of simulated wastewater. The adsorbent was waste corncob activated and impregnated with tetraoxophosphate V acid (H 3 PO 4 ). The adsorbent was characterized via SEM and FTIR analysis. Artificial intelligence models such as adaptive neuro-fuzzy inference system (ANFIS), artificial neural network (ANN), and response surface methodology (RSM) were used in modeling the dephenolization process. Genetic algorithm (GA) was used to optimize the best models. The results indicated that H 3 PO 4 -assisted impregnation enhanced the adsorptive dephenolization capability of the adsorbent. The BET surface area analysis of the modified adsorbent showed a surface area of 903.7 m 2 /g, average pore width of 5.55nm and micropore volume of 0.389 cm 2 /g. SEM micro-graph showed that the modified activated carbon has asymmetrical and interspatial pores which indicated desirable sorption properties of the adsorbent. The correlation coefficient (R 2 ) of the ANFIS, ANN and RSM models were 0.9998, 0.9998, and 0.9430 respectively while the Adjusted R 2 were 0.9339, 0.9998, and 0.9998 respectively. This implied that ANN and ANFIS models were more significant than RSM. Further statistical analysis suggested that ANFIS and ANN have almost equal capability in modeling the dephenolization process. The ANN-GA and ANFIS-GA optimization gave maximum percentage of phenol removal as 92.44% and 92.34% respectively under optimum process conditions. The maximum adsorption capacity of 121mg/g obtained from the GA optimization was comparably higher than most reported works. The point of zero charge was 5.65 while the regeneration with 0.2M NaOH showed best adsorbent reusable capacity. These results suggest that these waste corn cob activated carbon can be utilized as a high performance and efficient adsorbent in dephenolization of waste water. This is in addition to the reduction of the huge environmental waste associated with corn cobs. Furthermore there is no harmful environmental impact and no generation of toxic by-products during the adsorption process. • Synthesis and characterization of H 3 PO 4 activated corn cob. • A comparative assessment of modelling techniques using error indices. • Adsorbent dosage (F-value = 96.18) was the most influential process parameter in dephenolization of the simulated wastewater. • At optimum conditions, ANN-GA and ANFIS-GA optimization gave maximum percentage of phenol removal as 92.44% and 92.34% respectively.