Comparative Study of Response Surface Methodology (RSM) and Artificial Intelligence (AI) Integrated Modelling for Empty Fruit Bunch (EFB) Derived Biochar for Dye Adsorption

Empty fruit bunch (EFB) is an agricultural biomass that was utilised to synthesise high-performance biochar for the removal of methylene blue via batch adsorption. This study presented a comparative evaluation between traditional Response Surface Methodology (RSM) and an Artificial Intelligence (AI)-integrated framework to model and optimise the biochar production with adsorption as the response. Six supervised machine learning algorithms, including Linear, Ridge, Rasso, Random Forest (RF), Gradient Boosting, and Support Vector Regression, were employed to decode the relationships between pyrolysis temperature and residence time.  Methylene blue dye adsorption exhibited complex nonlinear behaviour, which was best mapped by Random Forest Regressor (R۲=۰.۹۱۲), aligning with a significant RSM quadratic model (q<۰.۰۰۰۱, R۲=۰.۹۸۶۷). Both frameworks had established temperature as the dominant control level, accounting for ۹۱.۸۸%of the predictive weight.  The response surface analysis identified an asymptotic adsorption plateau beyond ۶۰۰ °C, indicating that surface activation reaches thermodynamic equilibrium early in the pyrolysis process. This dual-modelling approach supports the statistical significance of RSM with predictive generalisation of AI, providing a high-precision template to transform agricultural biomass into engineered adsorbents for wastewater treatment.

کلیدواژه ها:

Biomass valorisation ، Dye Adsorption ، Response surface methodology ، artificial intelligence ، Pyrolysis