In this talk, we will focus on the high-fidelity computations of turbulence from the perspective of modeling and numeric. Firstly, the interaction of an incident shock wave with a Mach-6 laminar boundary layer is addressed using DNS and equilibrium wall-modeled LES (WMLES). The consequences of the interaction are that the boundary layer transitions to turbulence, and that transition causes a localized significant increase in the Stanton number and skin-friction coefficient. WMLES provides predictions of DNS peak loads within 10% at 150 times lower computational cost. In the fully-turbulent boundary layer, WMLES agrees well with DNS for the Reynolds-analogy factor, the mean velocity and temperature profiles, and the temperature/velocity correlations. Secondly, we will introduce the recently proposed high-order shock-capturing TENO schemes. For TENO, arbitrarily high-order, i.e. both even and odd order, scheme can be constructed in a unified framework. The novel ENO-type stencil selection strategy guarantees that the spectral properties of the counterpart linear scheme can be preserved up to intermediate wavenumbers. The core idea is that one candidate stencil is either abandoned when crossed by discontinuities or applied for the final reconstruction with the optimal weight. TENO schemes are demonstrated to be robust for shock capturing and perform better than classical WENO schemes in resolving small-scale structures.
11
September 2020
10am - 11am
Where
https://hkust.zoom.us/j/93559765024 (Passcode: 518056)
Organizer(S)
Department of Mathematics
Contact/Enquiries
Payment Details
Audience
Alumni, Faculty and Staff, PG Students, UG Students
Language
English
Other Events
Seminar, Lecture, Talk
PHYS Seminar - Expert Sharing Seminar
02
Nov 2020
Seminar, Lecture, Talk
MATH - Seminar on Data Science - Compression and Acceleration of Pre-trained Language Models
Recently, pre-trained language models based on the Transformer structure like BERT and RoBERTa have achieved remarkable results on various natural language processing tasks and even som...
28
Oct 2020