A Hybrid XGBoost-Based Model Approach To Evaluate the Effect of Biomass Fly Ash in Vegetable and Fruit Waste Composting

Abstract

The global energy crisis has increased interest in renewable energy sources, particularly biomass power plants. Fly ash is a significant waste product generated by biomass power plants. Sustainable waste management necessitates the exploration of alternative options for integrating ash into the circular economy. This study aims to determine the effect of biomass fly ash as an additive material in vegetable and fruit waste composting. The composting process was modeled with a two-stage hybrid model that combines the advantages of statisticalbased Response Surface Methodology and Machine Learning-based Extreme Gradient Boosting. The hybrid model was compared with Response Surface Methodology and Artificial Neural Network-based models. The results showed that the hybrid modeling tool had the best modeling performance with mean absolute percentage error values of less than 1 %. The prediction model was optimized with a Genetic Algorithm. Optimization results showed that biomass fly ash rates of 9.833 %, 9.776 % and 5.783 % were effective with desirability levels above 99 % for moisture content, total nitrogen and total organic carbon losses, respectively. The results presented a new strategy for the recycling of biomass fly ash by vegetable and fruit waste composting and showed that the proposed hybrid model was effective for obtaining a high-value-added product.