Estimation of the Thermal Process in the Honeycomb Panel by a Monte Carlo Method

2017 International Conference on Aerospace Technology, Communications and Energy Systems

航空航天工程;无线电电子学;能源学

Gusev, S.A.^1,2 ; Nikolaev, V.N.^3

ICM and MG Sib. Br. RAS, pr. Academika Lavrenjeva 6, Novosibirsk

630090, Russia^1

NSTU, pr. K. Marksa 20, Novosibirsk

630073, Russia^2

SibNIA Named after S. A. Chaplygin, Polzunov Street 21, Novosibirsk

630051, Russia^3

关键词: Approximate solution;    Boundary conditions of the third kinds;    Computation costs;    Computational effort;    Diffusion process;    Parabolic Equations;    Parabolic problems;    Statistical modelling;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/302/1/012045/pdf
DOI  :  10.1088/1757-899X/302/1/012045

学科分类：航空航天科学

来源: IOP

【 摘 要 】

A new Monte Carlo method for estimating the thermal state of the heat insulation containing honeycomb panels is proposed in the paper. The heat transfer in the honeycomb panel is described by a boundary value problem for a parabolic equation with discontinuous diffusion coefficient and boundary conditions of the third kind. To obtain an approximate solution, it is proposed to use the smoothing of the diffusion coefficient. After that, the obtained problem is solved on the basis of the probability representation. The probability representation is the expectation of the functional of the diffusion process corresponding to the boundary value problem. The process of solving the problem is reduced to numerical statistical modelling of a large number of trajectories of the diffusion process corresponding to the parabolic problem. It was used earlier the Euler method for this object, but that requires a large computational effort. In this paper the method is modified by using combination of the Euler and the random walk on moving spheres methods. The new approach allows us to significantly reduce the computation costs.