Development of a multi-criteria selection method of kinetic model for substrate inhibition in ceramic separator microbial fuel cells used for wastewater treatment and bioelectricity production

Ceramic-separator microbial fuel cells (MFCs) using earthen pots (100 mL) with stainless steel mesh anodes (42.6 cm 2 ) and graphite plate cathodes were operated to investigate kinetic relationships between substrate concentration (COD: 500–5000 mg/L; NO 3 − –N: 50–1850 mg/L at 5:1 C: N) and degradation rate/power generation. Steady-state datasets from four relationships: chemical oxygen demand (COD) versus COD degradation, COD versus power density, nitrate vs. nitrate degradation, and nitrate versus power density-were analysed using non-linear regression of 9 kinetic models in MATLAB (R2020a). A novel multi-criteria selection framework (R 2 + Root mean square error (RMSE) + mean absolute percentage error (MAPE) + corrected Akaike Information Criterion (AIC C ) + Bayesian Information Criterion (BIC)) was designed, which helped identify Edward model as optimal for COD degradation, Luong model for COD-power generation, and Edward model for both nitrate degradation and nitrate-power generation relationships. This study also helped establish precise operating windows, i.e., COD < 5.25 kg/m 3 and nitrate < 0.78 kg N/m 3 , to avoid inhibition. Substrate affinities (COD: 0.96–1.56 kg/m 3 ; Nitrate: 0.028–0.672 kg N/m 3 ) highlight ceramic MFCs’ suitability for low-to-moderate strength wastewaters. Compared to the literature, current systems show comparable affinity but lower inhibition thresholds, confirming concentration control as critical for optimization of performance. This statistically efficient multi-criteria approach challenges traditional R 2 -only model selection for engineering ceramic MFCs.