OPTIMIZATION OF CO-DIGESTION PROCESS OF FOOD WASTE AND SEWAGE SLUDGE USING ARTIFICIAL NEURAL NETWORK AND GENETIC ALGORITHM

Abstract

Food waste is a major global issue especially in developed countries. This is because of the abundance of food waste in landfills has contributed to the emission of greenhouse gas (GHG). Therefore, by using anaerobic co-digestion technology, food waste (FW) can be used as a substrate with sewage sludge (SS) to produce a valuable product such as methane gas. In order to find the optimal ratio of FW to SS as well as substrate-to-inoculum (SI) ratio for the highest methane production, the present study utilizes the Artificial Neural Network (ANN) and Genetic Algorithm (GA) model. This study is based on the secondary data sources from various research papers and articles. The digester operational parameters such as mixed substrate ratio and SI ratio were considered. The optimal feedstock ratio was evaluated based on its biochemical methane potential (BMP). The performance of the ANN model was verified to be effective in predicting the methane production accurately with a desirable R²-value of 0.9838 and 0.9976. The trained ANN model was coupled with GA to optimize the methane production and to find the optimal feedstock ratio. The result of optimal mixed substrates ratio of FW:SS and SI ratio are similar which is 50:50 with the highest methane production of 454.4 mL CH⁴/kg volatile solids (VS).  However, the comparison of BMP from different substrates ratio shows inconsistency on the optimal ratio. Hence, other parameters such as particle size and mixing rate should be considered.