Experimental validation of optimized performance of microbial fuel cell-based horned lizard algorithm and artificial intelligence

Microbial fuel cell (MFC) has shown promise for simultaneous wastewater treatment and electrical power production. Improving the performance of MFC is the objective of this study. Firstly, three input parameters including Ni (mg/m²), COD (mg/L) and aeration (mL/min) are investigated experimentally to measure the performance index of the MFC. The performance index of MFC includes the power density (PD), COD removal (CODr) and coulombic efficiency (CE). Secondly, using the experimental data, an adaptive neuro-fuzzy inference system (ANFIS) model was created to simulate the MFC in terms of Ni, COD and aeration. To assess the modelling stage, the results are compared with ANOVA. For the PD model, the predicted R² increased from 0.902 to 0.93 by around 3.1% compared to ANOVA, whereas for the ANFIS model for the CODr, the predicted R² increased from 0.58 to 0.81 by around 39.6% compared to ANOVA. For the ANFIS model of the CE, the predicted R² value increased from 0.81 to 0.91 by around 12.3% compared to ANOVA. This demonstrated the robustness of ANFIS model of the MFC. Thirdly, the optimal values of Ni, COD and aeration are identified based on integration between ANFIS model of the MFC and the horned lizard algorithm.