Modelling and Optimization of Methane from Anaerobic Digestion of Animal Waste (Pig Dung) with Adaptive Neuro Fuzzy Inference System (ANFIS)

Abstract

Agricultural wastes have increased in recent times due to an increased need for agricultural products by humans. These wastes causes environmental pollution and hazard to the society. Hence, the need to embark on anaerobic digestion to convert these waste to biogas. Biogas has been proven to be a good alternative to gases produced from fossil fuel due to its environmental friendliness. Production of biogas gases ensures cleaner environment and enhances clean growth. In this study, hydraulic retention time, temperature, moisture content, pH and carbon to nitrogen ratio are the factors used for generating the biogas from anaerobic digestion of pig dung. Independent variables range values obtained for each factor was subjected to design expert software application which gave the factor levels and number of runs under Central Composite Design (CCD) in Design of Experiment (DOE). The data generated from design expert was utilized in the laboratory for biogas yield via bio-digester. Pig dung was made to meet various process conditions after which 2kg of pig dung was feed into a 10liters digester and kept for various days. The yield obtained was subjected to four RSM models (linear model, interaction model, pure-quadratic model and quadratic model), ANFIS model and three kinetic models (gompert, modified gompertz and logistic model). Quadratic model was selected as the best RSM model having the best prediction accuracy performance with R-square value of 73.03%. When compared with ANFIS model, ANFIS model had R-square of 95.9983% there by making it a better model which can be utilized for further analysis. ANFIS was optimized and a three dimensional plot of surface and contour plot were used to determine the effect of the factors on the interaction and the optimal response obtained was 92.36% at hydraulic retention time of 11days, temperature of 50oC, moisture content of 93%, pH value of 7.01 and carbon to nitrogen ratio of 1 which was used as the first guess for kinetic studies. Furthermore, logistic model had the best prediction accuracy with R-square of 68.09% when compared with the other two kinetic models, hence, it was concluded that ANFIS model should be utilized in determination of the yield using the factors and logistic model, this could be used in the study of the anaerobic bio-digesters for the production of biogas from pig dung.