Biogas Production Prediction From Anaerobic Co-Digestion Of Food Waste

In recent years,the production of food waste has increased rapidly in China,which will cause serious environmental pollution and resource waste if food waste is not properly treated.The anaerobic co-digestion of food waste and excess sludge can not only reduce the waste,but also recover biogas and produce economic benefits,which has become an important research direction in the treatment of food waste.The key to anaerobic digester is to maximize the production of biogas,and it is of great significance and value to establish a model for anaerobic digestion to predict biogas production(especially methane production)for the optimization of fermentation parameters and improvement of gas production efficiency.Based on the principle of artificial neural network,the anaerobic digestion process of food waste with rich oil and fat in Chengdu area was studied in this paper,and the methane yield per day was predicted and analyzed.Based on a large amount of experimental data,salt,VFA,pH,ammonia nitrogen and alkalinity are selected as the control parameters of the models.The prediction model of methane production of multivariable linear regression model,BP neural network and BP neural network optimized by genetic algorithm are established,and the prediction effect of models is compared.The results show that the average prediction accuracy of the multivariable linear regression model is 78.80%,and the network predictive value can largely reflect the change trend of methane production,but the prediction accuracy of the peaks and troughs is poor.The average prediction accuracy of BP neural network model is 84.76%,the prediction value can reflect the change trend of methane production,and most of the value of peaks and troughs is fitting better.The average prediction accuracy of BP neural network model optimized by genetic algorithm is 93.35%,and the value of peaks and troughs is fitting much better,which has better prediction effect on methane yield.In conclusion,it is shown that the artificial neural network models including BP neural network and BP neural network model optimized by genetic algorithm are more suitable for nonlinear anaerobic digestion than multivariable linear regression model,and better results can be obtained.