Numerical Simulation of Aircraft Engine Related Two-Phase Flows

Principal Investigator:
Benjamin Sauer

TU Darmstadt

Local Project ID:

HPC Platform used:
SuperMUC of LRZ

Date published:

The DNS calculations are part of a numerical method denoted as embedded DNS (eDNS). The development and application of the eDNS concept to primary breakup of aircraft engine related two-phase flows is the focus of this project. With the embedded DNS concept only the area of interest (the breakup region) is simulated using DNS, while the DNS domain is embedded in a larger and wider domain which is simulated using Large Eddy Simulations (LES). Both simulations are coupled by the boundaries they share. The LES computations were executed locally, converted and copied to the corresponding two-phase flow DNS on SuperMUC.

DNS provides the possibility to study the breakup without any models as all relevant length scales are resolved. In addition the resolution of liquid structures and droplets (< 10 µm) requires an extremely refined mesh. Thus, the computational costs rises significantly, but can be handled with the eDNS concept. A single run is computed with 720 CPU cores on SuperMUC. As primary breakup can only be described by the analysis of 3D field data, the state of the system has to be written with a high output frequency leading to a required disk space of up to 10 TB for each single run. Figure 1 shows snapshots of two liquid sheets during atomization with different operating conditions.

In order to understand the primary breakup of airblasted liquid sheets and to prove the applicability of the eDNS concept, computations of several operating conditions are necessary. The results up to now provide a detailed view into the influence of operating conditions and numerical parameters. Characteristic breakup mechanisms can be shown and detailed information about the breakup products are kept. These are valuable information for further design computations.

[1] B. Sauer. Direct numerical simulation of the primary breakup of aircraft engine related two-phase flows. PhD thesis, Technische Universität Darmstadt, 2014.

German Research Foundation (DFG), financial support under contract no. JA 544/39-1 and GRK 1344
Gauss Centre for Supercomputing e.V., computing time on GCS Supercomputer SuperMUC at Leibniz Supercomputing Centre.

Research Team:
Benjamin Sauer, Amsini Sadiki, Johannes Janicka
Department of Energy and Power Plant Technology
Technische Universität Darmstadt

Tags: Technische Universität Darmstadt