Data-Driven Circular Economy of Biowaste to Bioenergy with Conventional Prediction Modelling and Machine Learning

Abstract

Rapid population growth has not only increased energy demand, but waste generation that has increased and introduced emerging pollutants into waste streams, posing sanitary and environmental risks. The purpose of this research was to investigate waste to energy (anaerobic digestion-bioenergy process) on integrated waste management and digitalization of biomethane production as a data-driven circular economy model across the value chain of the carbon cycle (transition from take-make-dispose to use-make-return) through waste quantification, characterization, biomethane potential test (BMP) to prediction of the biomethane production. Anaerobic digestion experiment was conducted at a laboratory scale to analyze biomethane production from diverse substrates such as food waste, cow manure, sewage sludge, and chicken manure, with an average pH of 7.58. The experimental results obtained were then modelled and simulated with Modified Gompertz, Logistic, and Richards models and compared to machine learning simulation using Python with Gompertz, Logistics, and Richards models. Modified Logistic model was shown to be the best-fit curve, with a coefficient of determination (R2) > 0.9 validating the conventional mathematical modeling and simulation performance. During simulation with machine learning (Python), the experimental results obtained from the cow manure substrate provided the best fitting curve to the training curve compared to other substrates with the highest average of R2 as 1.0 for training, validation, and test data. Cow manure had the best validation performance at MSE (mean squared error) of 25.36 at epoch 1. Graphical Abstract: (Figure presented.)