A model-based study of deconvolution in time-varying or injection calorimeters

Abstract

A deconvolution algorithm based on time-varying inverse filtering is applied to a calorimetric model. The model is designed to simulate a device in which continuous liquid injection takes place. By means of a proper separation of different contributions to the thermogenesis, classical time-invariant deconvolution routines can be applied even if the model presents zeros in the time-invariant transfer function. The corrections achieved on the thermogenesis are evaluated for two different kinds of thermal power dissipation.