ADM1-based simulation of thermophilic biohydrogen production from enzymatically pretreated palm oil mill effluent

Abstract

Palm oil mill effluent (POME) is an abundant agro-industrial wastewater with chemical oxygen demand (COD) exceeding 30,000 mg/L. While it holds promise as a substrate for sustainable biohydrogen production, its complex composition restricts microbial accessibility, resulting in low hydrogen yields. Moreover, empirical models such as the Gompertz equation provide good curve fitting but lack mechanistic depth, whereas default Anaerobic Digestion Model No. 1 (ADM1) formulations often underpredict volatile fatty acid (VFA) accumulation and hydrogen dynamics. This study aimed to calibrate and validate ADM1 for thermophilic biohydrogen production from enzymatically pretreated POME. Batch experiments were conducted under optimised conditions (pH 6.5, 4.3 % w/v enzyme loading, 55 °C), and the model was evaluated against experimental hydrogen evolution, reducing sugar release, and VFA profiles. Sensitivity analysis was performed to identify the key kinetic parameters that influence hydrogen production. Enzymatic pretreatment increased reducing sugar availability by approximately 182 %, resulting in a fivefold increase in cumulative hydrogen production (444 mL) compared with untreated POME. ADM1 simulations achieved high predictive accuracy (R² > 0.90; RMSE ≤ 0.2), successfully reproducing the trends in hydrogen evolution and VFA accumulation. The substrate uptake rate (2.47 kg COD kg⁻¹ COD d⁻¹) and biomass decay constant (0.62 d⁻¹) emerged as critical drivers of hydrogen flux and microbial stability. This study demonstrates a validated application of ADM1 to thermophilic fermentation of enzymatically pretreated POME, establishing a robust mechanistic framework for process optimisation. Future work should extend this approach to continuous systems and hybrid ADM1–data-driven integration for scale-up and real-time monitoring.