Work Description

Date: 30 March, 2020

Dataset Title: 2D Benchmark Reacting Flow Dataset for Reduced Order Modeling (ROM) Exploration

Dataset Creators: C. Huang

Dataset Contact: Cheng Huang huangche@umich.edu

Funding: Air Force Center of Excellence Grant FA9550-17-1-0195, titled “Multi-Fidelity Modeling of Rocket Combustor Dynamics.”

Research Overview:
A 2D planar representation of a generic laboratory-scale combustor is established to assess the capabilities of ROMs for representing realistic combustion flowfields. The purpose of this dataset is to provide a testbed to build reduced model for relevant challenging reacting flow problems using different methods. The dataset was generated under the Air Force Center of Excellence on Multi-Fidelity Modeling of Rocket Combustion Dynamics and the goal of the center is to advance the state-of-the-art in Reduced Order Models (ROMs) and enable efficient prediction of instabilities in liquid fueled rocket combustion systems.

Methodology:
The data are model output from an in-house CFD solver, General Equation and Mesh Solver (GEMS). Please refer to Ref 3 in related publications for detailed information on the solver.

Instrument and/or Software specifications:
- recommendation: Matlab and Tecplot

Files contained here:

- Data_150000to159999.tar: the unsteady flow field data from time step 150000 to 159999 (time increment, dt, between each time step is 1E-7 sec).

- Data_160000to169999.tar: the unsteady flow field data from time step 160000 to 169999 (time increment, dt, between each time step is 1E-7 sec).

- Data_170000to179999.tar: the unsteady flow field data from time step 170000 to 179999 (time increment, dt, between each time step is 1E-7 sec).

- Data_180000to189999.tar: the unsteady flow field data from time step 180000 to 189999 (time increment, dt, between each time step is 1E-7 sec).

- Data_190000to199999.tar: the unsteady flow field data from time step 190000 to 199999 (time increment, dt, between each time step is 1E-7 sec).

- Data_200000to209999.tar: the unsteady flow field data from time step 200000 to 209999 (time increment, dt, between each time step is 1E-7 sec).

- Data_210000to219999.tar: the unsteady flow field data from time step 210000 to 219999 (time increment, dt, between each time step is 1E-7 sec).

- Data_220000.tar: the unsteady flow field data at time step 220000.

- grid.dat: the topology of the CFD mesh used to generate this data (can be directly loaded in Tecplot).

- the folder "sample_code" contains the sample Matlab scripts to load and output the .dat files to help the researchers to get started. To run the script, the software Matlab is required and the researchers can simply run sampleIO.m script in Matlab to test the code.

Related publication(s):
1. McQuarrie, S., Huang, C., and Willcox, K., Data-driven reduced-order models via regularised Operator Inference for a single-injector combustion process, Journal of the Royal Society of New Zealand, 2021. (code available: https://github.com/Willcox-Research-Group/ROM-OpInf-Combustion-2D).

2. Swischuk, R., Kramer, B., Huang, C., and Willcox, K., Learning Physics-Based Reduced-Order Models for a Single-Injector Combustion Process , AIAA Journal, 2020.

3.Huang, C., Duraisamy, K., and Merkle, C.L., Investigations and Improvement of Robustness of Reduced-Order Models of Reacting Flow, AIAA Journal, 2019.

4. Harvazinski, M.E., Huang, C., Sankaran, V., Feldman, T.W., Anderson, W.E., Merkle, C.L., and Talley, D.G., Coupling between hydrodynamics, acoustics, and heat release in a self-excited unstable combustor, Physics of Fluids, 2015.

Use and Access:
- the files with .tar suffix are Tape Archive and the data inside the archive can be extracted using:
- for linux system: tar -xvf fileName.tar (e.g. tar -xvf Data_150000to159999.tar)
- for windows system: common software like 7-zip and wintar can be used to extract the dataset
- after extracting the archive, you will obtain the unsteady flow field data at each time step, for example: test_file_ncons_150000.dat contains the necessary flowfield information (pressure, velocities, temperature and species mass fractions) at all spatial coordinates described by grid.dat at the 150000th time step (i.e. at time = 150000 * 1E-7 = 0.015 sec)

- grid.dat contains the spatial coordinates of the CFD mesh used to generate the dataset and can be directly visualized in Tecplot (software)

- all the data is in ASCII format with an easy access for researchers who are not familiar with binary format. Each data file can be directly open with any software that can handle .dat files. To help the researchers to more easily get started on the data, we prepared a set of sample Matlab codes that describe how to load and output the data.

To Cite Data:
Huang, C. (2020). 2D Benchmark Reacting Flow Dataset for Reduced Order Modeling (ROM) Exploration [Data set]. University of Michigan Deep Blue Data Repository.
*note: please also cite the first two papers in the related publications