Optimization and prediction of chromium (iv) removal from synthetic acid mine drainage using green adsorbents: A Box–Behnken design and adaptive neuro‐fuzzy inference approach

Abstract Chromium (VI) is a highly toxic heavy metal ion linked to severe health issues, including kidney failure, gastrointestinal irritation, multi‐organ failure, and death, depending on exposure levels. Several chemical and traditional water purification methods have been developed in the past, but most are expensive, tedious, and ineffective. This current study employed cellulose nanocrystals (CNCs) prepared via hydrolysis of waste paper, followed by carboxylation and integration with sodium alginate to enhance Cr (VI) adsorptive properties. The adsorbents were solidified using calcium chloride and characterized for their chemical structure and surface energy optimization. MATLAB‐ANFIS was used to predict chromium (VI) adsorption at various optimization parameters such as pH, dosage, contact time, and initial concentration. The adaptive neuro‐fuzzy inference system (ANFIS) offered a practical approach to optimize adsorption processes, saving time and resources for real‐life applications. To enhance prediction accuracy, the study employed the Box‐Behnken design (BBD) to optimize membership functions (MFs) and their numbers. Eight widely used MFs, including triangular, trapezoidal, Gaussian, and generalized bell‐shaped, were evaluated. ANFIS models fitted with triangular and trapezoidal MFs proved statistically significant at the 95% confidence level, according to ANOVA analysis. Optimal MF numbers for inputs were identified as 5‐5‐2 for triangular and 9‐9‐3 for trapezoidal MFs. The optimized ANFIS model, employing triangular MFs, achieved a low root mean square error (RMSE) of 1.9084 and a high correlation coefficient ( R 2 ) of 0.9922, demonstrating its accuracy and reliability in predicting chromium (VI) adsorption capacity under the identified optimal conditions. This systematic approach demonstrates the effectiveness of ANFIS in accurately predicting and optimizing heavy metal adsorption processes.