Development of Inference Methodology for Supporting Understanding of Composting Processes

Abstract

Composting, one of the promising biotechnologies for solid waste management, is a process where organisms convert organic materials into a hygienic and bio-stable humus-like product. Food waste composting has gained increased attention in the past decade. Efficient operations of composting relies on insights of relationships between state variables (e.g. oxygen concentration, ash content, moisture content, and pH) and specific characteristics (e.g. microbial activities, maturity, and stability). Previously, many experimental approaches were developed in analyzing these relationships. However, experiment-based evaluations could hardly help quantify the interactions among multiple composting state variables. In comparison, a model-based analysis could help examine the inherent impacts of various factors on the biological and physiochemical processes and gain an in-depth insight into the related mechanisms. This study attempted to develop inference methodology based on multivariate analysis to describe the nonlinear relationships between the selected state variables and characteristics of interest in food waste composting. The experimental data from bench-scale composting reactors were used to demonstrate the applicability of proposed methods. These methods would help identify the most significant relationships, understand the interactive mechanisms, and infer the hard-to-obtain characteristics in an easier manner, during composting and many other environmental processes.