Integrating Life Cycle Assessment and Machine Learning to Enhance Black Soldier Fly Larvae-Based Composting of Kitchen Waste

Abstract

Around 40% to 60% of municipal solid waste originates from kitchens, offering a valuable resource for compost production. Traditional composting methods such as windrow, vermi-, and bin composting are space-intensive and time-consuming. Black soldier fly larvae (BSFL) present a promising alternative, requiring less space and offering ease of handling. This research encompasses experimental data collection, life cycle assessment, and machine learning, and employs the Levenberg–Marquardt algorithm in an Artificial Neural Network, to optimize kitchen waste treatment using BSFL. Factors such as time, larval population, aeration frequency, waste composition, and container surface area were considered. Results showed that BSFL achieved significant waste reduction, ranging from 70% to 93% by weight and 65% to 85% by volume under optimal conditions. Key findings included a 15-day treatment duration, four times per day aeration frequency, 600 larvae per kilogram of waste, layering during feeding, and kitchen waste as the preferred feed. The larvae exhibited a weight gain of 2.2% to 6.5% during composting. Comparing the quality of BSFL compost to that obtained with conventional methods revealed its superiority in terms of waste reduction (50% to 73% more) and compost quality. Life cycle assessment confirmed the sustainability advantages of BSFL. Machine learning achieved high accuracy of prediction reaching 99.5%.