Resolution of an Inverse Problem for identification of thermophysical properties of substances typical in the biofuels industry.
NETO, P. P. S.1; CAVALCANTI, F. C.1
1 Universidade Federal da Paraíba
doi:10.20906/CPS/COB-2015-0779
Resumo
Within the biofuels, industry there is a constant concern with regard to the correct sizing of the equipment used during the extraction processes, transport and processing of the raw material for biofuels. Engineers has worked in the development and resolution of complex mathematical models in order to predict the thermal and dynamic behavior of substances in those industries, which are usually vegetable oils and sugarcane juice. However, the reliability of these models is somewhat limited by the lack of accurate data about the thermophysical properties of such substances. The demand for reliable values is urgent and requires a high theoretical and experimental effort. This work consisted in solving an Inverse Problem for the estimation of the thermal conductivity and the specific heat of vegetable oils. A transient direct and one-dimensional model of the heat conduction (thermophysical properties depending on the temperature) was solved using the finite difference method in order to determine the temperature field within the sample and simulate the functioning of the experimental device. The solution of the direct model was compared with simulations in ANSYS software for the model validation. Next, a detailed analysis of the sensitivity coefficients was made in order to establish the best identification strategy. Finally, an identification algorithm based on the Levenberg-Marquardt method was developed and the thermophysical properties were estimated simultaneously.
Palavras-chave: FINITE DIFFERENCE METHOD; PARAMETER ESTIMATION; INVERSE PROBLEM; BIOFUEL; LEVENBERG-MARQUARDT'S METHOD