Sensitivity Analysis and Parameter Optimization of a Heat Loss Model for a Composting System

Abstract

ABSTRACT.A combined multi-parameter sensitivity analysis/frequency array analysis technique was employed to assess the impact of design conditions and matrix-specific properties on the rate of heat loss from a self-heating composting process. This method was specifically used to identify the range of parameters over which a model predicting the overall heat-transfer coefficients (U-values) for the heat loss process are particularly sensitive. The model was found to be most sensitive for the following range of input parameters: the internal diameter of reactor varying from 0.8 m to 1.2 m, the combined reactor wall and insulation thickness ranging from 4 cm to 6 cm, the free airspace of the matrix varying from 50% to 60%, the reactor length varying from 1.0m to 1.5m, the thermal conductivity of the organic substrates ranging from 0.1 W/m.K to 0.2 W/m.K and the preferred thermal conductivity of the insulation material being less than 0.2 W/m.K. A second sensitivity analysis was performed to identify which input parameters actually influenced the model's response the most. This analysis compared the acceptable and unacceptable frequency distributions of each input parameter, with model outputs below a U-value of 4.5 W/m 2 .K being an acceptable composting performance from a thermodynamics consideration. This analysis disclosed the matrix free airspace, the internal diameter of the reactor and the combined thickness of reactor wall and insulation as the most important parameters which should be given high priority when designing invessel compost reactors.

Greenhouse Gas Emissions Reductions from In-Situ Aeration in a Landfill: A Multi-Parameter Sensitivity Analysis Approach

“…A visual assessment of the model sensitivity, plotted as a threedimensional surface which is obtained through the results of the matrix, helps to identify input parameters which influence the model's response the most. The mesh plot generated provides an indication of the potential for obtaining a significant response for a given combination of the primary parameters (Mudhoo and Mohee, 2006). An elevated value in the plot would indicate a combination that will result in a potential "hot-spot" (Parker, 1997).…”

“…In this respect, only MCCH4,v,k,q, MCCH4,s,i,q, SGs,i,q, L0,i and fdg,i have been used for the MPSA. The discretisation of the range of each parameter into individual values must span the endpoints and also reflect the relative distribution of values that may be encountered in practice (Mudhoo and Mohee, 2006). In this respect, and in order to eliminate any bias during the model assessment, the logarithmic approach developed by Parker (1997) was used to determine the discretized values.…”