Optimizing multi-variables of microbial fuel cell for electricity generation with an integrated modeling and experimental approach

Microbial fuel cell (MFC) is a device that transforms chemical energy in wastewater into electricity, and its performance is influenced by multi-variables. Mathematic modeling approach could be a useful alternative to design and optimize such a complex system for power generation and wastewater treatment. Here we develop a novel integrated modeling approach with uniform design (UD), a machine learning approach of relevance vector machine (RVM) and a global searching algorithm of accelerating genetic algorithm (AGA) to optimize the operation of multi-variable MFCs after they are constructed. With the integrated UD–RVM–AGA approach, a maximum Coulombic efficiency of 73.0% and power density of 1097mW/m3 of MFC are estimated under the optimal conditions of ionic concentration of 102mM, initial pH of 7.75, medium nitrogen concentration of 48.4mg/L, and temperature of 30.6°C. The Coulombic efficiency and power density in the verification experiments, 70.9% and 1156mW/m3, are close to those calculated by the modeling approach. The results demonstrate that the integrated UD–RVM–AGA approach is effective and reliable to optimize the complex MFC and improve its performance.