An Approachable Anaerobic Bioreactor for Remote Biogas Production: Experimental Analysis and Neuroevolution Modeling

Abstract

This study introduces a straightforward batch-mode bioreactor for the production of biogas from two waste sources: animal slaughterhouse solid waste and fruit and vegetable solid waste, as well as poultry slaughter solid waste and fruit and vegetable solid waste. To measure the efficiency of methane and carbon dioxide (CO₂) production, the system was experimentally studied for 40 days, investigating different carbon-to-nitrogen (C/N) ratios: 20, 30, and 40. The highest biogas and methane contents were observed at a C/N ratio of 30. The poultry slaughterhouse waste (SHW) and fruit and vegetable waste (FVW) combination resulted in an impressive 201.7 L of biogas, with 149.2 L of pure methane. Similarly, the animal SHW and FVW mixture resulted in 241 L of biogas, containing 182.7 L of valuable methane. Recognizing the complex nature of factors that impact the anaerobic digestion (AD) process, this study employed kinetic models and artificial neural networks (ANNs) combined with three optimizers: Differential Evolution (DE), Bacterial Foraging Optimization (BFO), and the Dragonfly Algorithm (DA). The simulation data revealed that the BFO approach yielded the best models. Notably, the mean squared error (MSE) during the testing phase was remarkably low, measuring 0.000552 for cumulated CO₂ production and 0.001598 for cumulated methane production. Overall, the models introduced in this study exhibit excellent generalization capability and serve as reliable predictors for the systems’ output in various scenarios. The significance of these findings extends beyond the laboratory, as the proposed system and its model can effectively aid end-users in planning their consumption and correlating biogas utilization with peak production. This mainly benefits small consumers in remote areas, offering them sustainable energy solutions and paving the way for a greener future.