Thermal curve interpretation by spectral resolution into a basic set of rectangular pulse curves

Abstract

Listen

Translate article

There are several interesting solutions to the problem of the determination of quick thermal effects in calorimeters having large time constants [1, 2]. We have proposed a method for thermokinetic determination based on numerical functional analysis [3, 4] and the present paper describes a new improvement in this area. The problem of thermokinetic determination resolves itself into the numerical analysis of a thermal curve given as a set of experimental points and the reconstruction of the thermal effect from it, by using information about the calorimeter transition function. The transition function could cover everything known about the correlation between the thermal effects and their curves. As one of the forms of the transition function the Dirac curve can be taken, which is used in the optimization method [1] and the harmonic method [2] for thermokinetic determination. Of course, broadly speaking, the calorimeter transition is a functional which operates on a thermal pulse curve and produces its thermal curve. In our spectral resolution method the Dirac curve was used for thermal effect reconstruction. However, it is possible to use another thermal pulse curve for the purpose of calorimeter identification, as the information about the calorimeter, its f ingerprint, is present in any of its curves. The aim of the present paper is to generalize the spectral resolution method in the sense of using any kind of identification thermal pulse for determination of the thermokinetics. For several reasons this generalization seems to be interesting and useful. Firstly, we expect that using curves of thermal effects which are much closer to the effect we want to reconstruct for the purpose of calorimeter identification should give much better results. Secondly, the identification of the calorimeter by using only a Dirac curve is a serious limitation, because it disregards the