Thermogenesis: Comparative study of several deconvolution methods in time-invariant calorimetry

Abstract

A comparative analysis of the performance of different deconvolution methods is presented. This includes harmonic analysis (HA), inverse filtering (IF), dynamic optimization (DO), optimal control (OC) and Z-transform (ZT). The methods are applied on simulated thermograms corresponding to time-invariant representative models of heat conduction calorimeters. The results suggest that: (1) IF and ZT are the simplest methods and give the best reconstructions of the heat power; (2) HA spends more computing time and memory space to achieve a similar quality in the thermogenesis; and (3) DO and OC, with the same expense in memory and time as HA, give much more inaccurate results when the power dissipation lasts a long time. On the other hand, it should be noted that to the present, IF and, conceptually, ZT and OC may be generalized to time-varying calorimetry.