AcademicArticleSCO_85021096682

abstract

• © 2017, Pontificia Universidad Javeriana. All rights reserved. This work considers the real-time prediction of physicochemical parameters for a sample heated in a uniform electromagnetic field. As a demonstrative model, we estimated the thermal conductivity and the volumetric heat capacity (the sample (with known geometry) was subjected to electromagnetic radiation, generating a uniform and time-independent volumetric heat flux within it. A real temperature profile was simulated by adding white Gaussian noise to the original data, which were obtained from the theoretical model. To solve the objective function, simulated annealing and genetic algorithms, along with the traditional Levenberg-Marquardt method, were used for comparative purposes. the results showed similar findings of all algorithms for three simulation scenarios as long as the signal-to-noise-ratio was at least 30 [dB]. For practical purposes, this means that the estimation procedure presented here requires both good experimental design and correctly specified electronic instrumentation. If both requirements are satisfied simultaneously, it is possible to estimate these types of parameters on-line without the need for an additional experimental setup.