Яндекс.Метрика

D. Prokhorov, Y. Bazaikin, V. Lisitsa

Том: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Conference name: 23rd International Conference on Computational Science and Its Applications, ICCSA 2023 (Athens, Greece, July 3-6, 2023)
Том: 14104 LNCS , Год издания: 2023
Многотомное издание: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Conference name: 23rd International Conference on Computational Science and Its Applications, ICCSA 2023 (Athens, Greece, July 3-6, 2023)
Страницы: 693-707

Аннотация

The sintering process is widely used in modern industry because it allows for obtaining materials with predefined properties. Chemical or physical techniques can measure these properties. Besides the cost of such methods, it is worth noting that some techniques destroy samples, which causes difficulties in measuring the parameters evolution. Computer simulation of the sintering process allows for overcoming these difficulties. The sintering models based on the system of Cahn-Hilliard and Allen-Cahn equations require optimization if the number of grains is large. However, optimizations affect the solution; thus, a detailed quality assessment is required. The article presents such a study for our optimization: the Allen-Cahn equations are solved in small subdomains of the whole computational domain, which change over time. We provide comparative tests between solutions obtained by our algorithm and solutions obtained by solving the system in the whole domain. Besides common approaches, we use powerful tools of topology: Hausdorff distance and Betti numbers. The choice of the algorithm parameters is justified by obtained accuracy and efficiency.
индекс в базе ИАЦ: 027966