MPI-Based Computational Algorithm for Modeling a Cylindrical Catalyst Grain During Oxidative Regeneration Full article
Conference |
Parallel Computational Technologies : 17th International Conference 28-30 Mar 2023 , Saint Petersburg |
||||||
---|---|---|---|---|---|---|---|
Source | Parallel ComputationalTechnologies: 17th International Conference, PCT 2023, Saint Petersburg, Russia, March 28–30, 2023 Compilation, Springer Nature Switzerland AG. Cham, Switzerland.2023. 373 c. ISBN 978-3-031-38864-4. |
||||||
Output data | Year: 2023, Pages: 336-350 Pages count : 15 DOI: 10.1007/978-3-031-38864-4_24 | ||||||
Tags | Oxidative regeneration; Nonlinear model; Chemical kinetics; MPI technology; Numerical methods | ||||||
Authors |
|
||||||
Affiliations |
|
Funding (2)
1 | Ministry of Science and Higher Education of the Russian Federation | 0239-2021-0014 |
2 | Ministry of Science and Higher Education of the Russian Federation | FMRS-2022-0078 |
Abstract:
The article is devoted to the development of an efficient parallel algorithm for the numerical simulation of a cylindrical catalyst grain during the process of burning coke residua off the catalyst. The three-dimensional problem is reduced to a task in an axisymmetric formulation. The mathematical model of the process consists of a system of nonlinear parabolic partial differential equations and takes into account the mass transfer (diffusion and Stefan flow), the heat transfer, and chemical reactions. Reaction’s effect is included as a source term in the parabolic equations. The effective thermophysical characteristics are interpolated by time-dependent polynomials in conformity with reference data. The simulated processes occur with different characteristic times, hence the high degree of stiffness in the studied system. The model’s difference analog is given in a dimensionless form. We use an approach based on splitting by physical processes. Also, we isolate the solution of the chemical kinetics equations into a separate block. The chemical kinetics equations are stiff and require specialized solution methods. In the present research, we choose the Radau IIA method for this. For the equations of mass transfer and heat transfer, we use the integro-interpolation method. The numerical algorithm is implemented in C++ using MPI technology. The two-dimensional computational domain is divided into equal spatial cells. Each processor stores the data for computing the rates of chemical reactions and effective thermophysical characteristics. It should be noted that only boundary conditions are involved in the interprocessor exchange, which significantly reduces the program runtime. We investigate the efficiency of the developed parallel algorithm on the example of the process of oxidative regeneration of a cylindrical catalyst grain. We provide the corresponding results on the speedup of computations on several processors with different numbers of cells, as well as the graphs of substance concentrations and the temperature of the catalyst grain under various process conditions.
Cite:
Yazovtseva O.S.
, Gubaydullin I.M.
, Peskova E.E.
, Usmanova A.A.
, Zagoruiko A.N.
MPI-Based Computational Algorithm for Modeling a Cylindrical Catalyst Grain During Oxidative Regeneration
In compilation Parallel ComputationalTechnologies: 17th International Conference, PCT 2023, Saint Petersburg, Russia, March 28–30, 2023. – Springer Nature Switzerland AG., 2023. – C.336-350. – ISBN 978-3-031-38864-4. DOI: 10.1007/978-3-031-38864-4_24 Scopus OpenAlex
MPI-Based Computational Algorithm for Modeling a Cylindrical Catalyst Grain During Oxidative Regeneration
In compilation Parallel ComputationalTechnologies: 17th International Conference, PCT 2023, Saint Petersburg, Russia, March 28–30, 2023. – Springer Nature Switzerland AG., 2023. – C.336-350. – ISBN 978-3-031-38864-4. DOI: 10.1007/978-3-031-38864-4_24 Scopus OpenAlex
Dates:
Published online: | Jul 25, 2023 |
Identifiers:
Scopus: | 2-s2.0-85172680336 |
OpenAlex: | W4385211947 |