TY - JOUR AU - Bohaienko, Vsevolod AU - Gladky, Anatolij PY - 2020/10/22 Y2 - 2024/03/28 TI - On the selection of fractional-differential model of convective diffusion with mass exchange JF - Modeling, Control and Information Technologies: Proceedings of International scientific and practical conference JA - MCIT VL - 0 IS - 4 SE - Mathematical modelling and computational methods DO - 10.31713/MCIT.2020.02 UR - https://itconfdoc.nuwm.edu.ua/index.php/ITConf/article/view/108 SP - 7-10 AB - <p><span style="font-size: small;">The paper considers two fractional-differential models of convective diffusion with mass exchange and proposes a decision-making algorithm for determining the optimal model at the time of concentration field observation. As for soils of fractal structure, direct experimental determination of model parameters’ values and type of mass exchange process is in many cases impossible, calibration and determination of the most adequate models is performed mainly solving inverse problems by, in particular, meta-heuristic algorithms that are computationally complex. In order to reduce the computational complexity, we study the qualitative differences between diffusion processes described by fractional-differential models with non-local mass exchange on the base of the Caputo derivative and local non-linear mass exchange based on the non-equilibrium sorption equation that corresponds to the description by the Caputo-Fabrizio derivative. We determine under which conditions both models within a given accuracy describe the same set of measurements at a certain moment of time. When the solutions are close at a certain initial stage of process development, the model with the Caputo derivative describes its faster approach to a steady state. Based on the obtained estimates of differences in solutions, a decision-making algorithm is proposed to determine the most accurate model and the values of its parameters concurrently with the acquisition of measurements. This algorithm’s usage reduces the time spent on solving inverse calibration problems.</span></p> ER -