RUS  ENG
Full version
JOURNALS // Vladikavkazskii Matematicheskii Zhurnal // Archive

Vladikavkaz. Mat. Zh., 2024 Volume 26, Number 3, Pages 33–46 (Mi vmj919)

Some analytical solutions in problems of optimization of variable thermal conductivity coefficient

A. O. Vatulyana, S. A. Nesterovb

a Southern Federal University, 8 a Milchakova St., Rostov-on-Don 344090, Russia
b Southern Mathematical Institute VSC RAS, 53 Vatutina St., Vladikavkaz 362025, Russia

Abstract: New formulations and solutions to problems of optimization of a variable thermal conductivity coefficient for an inhomogeneous pipe and a flat wall with mixed boundary conditions are presented. The quality functionals are either the average temperature or the maximum temperature, and as a limitation – either the condition of constancy of the integral thermal conductivity coefficient, or a priori information about the change in the thermal conductivity coefficient in a known range. To solve problems for a pipe, two optimization methods are used: 1) a variational approach based on the introduction of conjugate functions and the construction of an extended Lagrange functional; 2) Pontryagin’s maximum principle. To solve the optimization problem for a flat wall under the assumption of weak material inhomogeneity, the expansion method in terms of a small physical parameter is used. As the fourth problem, optimization of the variable thermal conductivity coefficient of a non-uniform flat wall with boundary conditions of the first kind is considered. The solution to a singular optimization problem is found among broken extremals. Using specific examples, a comparison was made of the values of minimized functionals for bodies with a constant thermal conductivity coefficient and an optimal variable coefficient. The gain from optimization is estimated.

Key words: optimization, thermal conductivity coefficient, functionally graded material, flat wall, pipe, Lagrange variational method, Pontryagin's maximum principle, small parameter expansion method, singular problem.

UDC: 517.929.4

MSC: 80M30, 80M50

Received: 19.04.2024

DOI: 10.46698/v9056-4395-2233-f



© Steklov Math. Inst. of RAS, 2024