Using data-driven learning methodology for a solid waste-to-energy scheme and developed regression analyses for performance prediction

Abstract

Adopting innovative technologies like machine learning is crucial for achieving our sustainability goals. It has great potential for improving waste management and energy generation. The development of energy systems with machine learning algorithms is advancing rapidly. By conducting an in-depth comparison between the linear and non-linear regression models, this study makes a significant contribution to the field of waste-to-energy systems. The focus of this research is to forecast the performance of a newly designed solid waste-to-multi-generation energy system. This multi-generational energy system is designed to provide multiple outputs simultaneously, including power, heat, hydrogen, oxygen, and distilled water. In order to estimate the outputs of this system, both linear and non-linear algorithms are utilized and their respective performances are thoroughly analyzed and compared. The linear algorithms demonstrate notable precision through the creation of models that exhibit R-square values exceeding 96 %. In contrast, the non-linear algorithms demonstrate increased precision with R-square values surpassing 97 %, and even suggesting R-square values as impressive as 99 %, thereby attesting to the superior performance of these algorithms. Linear regression models are capable of providing predictions and identifying trends. However, non-linear regression models exhibit enhanced accuracy in predicting outcomes and are more efficient in capturing dynamic trends.