A stepwise-cluster microbial biomass inference model in food waste composting.

A stepwise-cluster microbial biomass inference (SMI) model was developed through introducing stepwise-cluster analysis (SCA) into composting process modeling to tackle the nonlinear relationships among state variables and microbial activities. The essence of SCA is to form a classification tree based on a series of cutting or mergence processes according to given statistical criteria. Eight runs of designed experiments in bench-scale reactors in a laboratory were constructed to demonstrate the feasibility of the proposed method. The results indicated that SMI could help establish a statistical relationship between state variables and composting microbial characteristics, where discrete and nonlinear complexities exist. Significance levels of cutting/merging were provided such that the accuracies of the developed forecasting trees were controllable. Through an attempted definition of input effects on the output in SMI, the effects of the state variables on thermophilic bacteria were ranged in a descending order as: Time (day)>moisture content (%)>ash content (%, dry)>Lower Temperature ( degrees C)>pH>NH(4)(+)-N (mg/Kg, dry)>Total N (%, dry)>Total C (%, dry); the effects on mesophilic bacteria were ordered as: Time>Upper Temperature ( degrees C)>Total N>moisture content>NH(4)(+)-N>Total C>pH. This study made the first attempt in applying SCA to mapping the nonlinear and discrete relationships in composting processes.