Mechanical Engineering MS Thesis Defense by Mr. Aakash Kardam
When: Monday,
August 13, 2018
1:00 PM
-
3:00 PM
Where: Textiles Building 101E
Description: Mechanical Engineering MS Thesis Defense
by Mr. Aakash Kardam
DATE:
August 13, 2018
TIME:
1:00 p.m. 3:00 p.m.
LOCATION:
Textile Building, Room 101E
TOPIC:
A Comparison of Nodal Discontinuous Galerkin Schemes for Hypersonic Flow Over
a Blunt Body
ABSTRACT:
This work focusses on the use of a family of compact discontinuous Galerkin (DG) methods in a systematic way to simulate unsteady hypersonic inviscid flow over a blunt body and compare the solution quality at a fixed resolution. These flow conditions gives rise to unsteady shocks. Mathematically speaking, these are discontinuities. The presence of discontinuities such as shocks/small-scale flow features make such simulations problematic for numerical simulation. In particular, we assess a high-order DG scheme's performance for standard benchmark tests while varying the shocking capturing limiter, numerical flux, and the scheme's order. In particular, we consider the classical MinMod limiter, its high-order generalization(s), and a recently introduced moment-based limiter by Moe, Rossmanith, and Seal. To couple our quadrilateral sub-domains we consider the popular Lax Friedrichs numerical flux, as well as the HLL, Roe, and Marquina numerical fluxes. The work will be summarizing the strengths and weaknesses of each flux/limiter combination for hypersonic inviscid flow over a blunt object. The comparisons between some of our best performing flux/limiter combinations with a second-order structured finite-volume solver. Our simulations are carried out using a newly developed partial differential equation solver, SpECTRE, that combines a discontinuous Galerkin method with a task-based parallelism model. SpECTRE has been developed with the philosophy of solving astrophysics problems and the present work uses this astrophysics code-base for problems in Mechanical and Aerospace Engineering which required to extend the code for certain additional functionality of putting solid objects in the domain and having set up for different solid wall boundary conditions which was not there before.Previous work shows the code's scalability on the Blue Waters supercomputer up to the machine's full capacity of 22,380 nodes. We will briefly touch on the SpECTRE's scalability for our problem and computational resources. A part of the thesis also presents the strong scaling analysis on MGHPCC machine that has been used extensively throughout the completion of this work.
ADVISOR:
Dr. Scott Field, Assistant Professor, Department of Mathematics at UMass Dartmouth
COMMITTEE MEMBERS:
Dr. Scott Field, Assistant Professor, Department of Mathematics at UMass Dartmouth
Dr. Alireza Mazaheri, Aerothermodynamics Branch, NASA Langley Research Center, Hampton, VA
Dr. Mehdi Raessi, Associate Professor, Department of Mechanical Engineering at UMass Dartmouth
Open to the public. All MNE students are encouraged to attend.
For more information, please contact Dr. Scott Field (sfield@umassd.edu, 508-999-8318).
Thank you,
Sue Cunha, Administrative Assistant
Mechanical Engineering Department
scunha@umassd.edu
508-999-8492
by Mr. Aakash Kardam
DATE:
August 13, 2018
TIME:
1:00 p.m. 3:00 p.m.
LOCATION:
Textile Building, Room 101E
TOPIC:
A Comparison of Nodal Discontinuous Galerkin Schemes for Hypersonic Flow Over
a Blunt Body
ABSTRACT:
This work focusses on the use of a family of compact discontinuous Galerkin (DG) methods in a systematic way to simulate unsteady hypersonic inviscid flow over a blunt body and compare the solution quality at a fixed resolution. These flow conditions gives rise to unsteady shocks. Mathematically speaking, these are discontinuities. The presence of discontinuities such as shocks/small-scale flow features make such simulations problematic for numerical simulation. In particular, we assess a high-order DG scheme's performance for standard benchmark tests while varying the shocking capturing limiter, numerical flux, and the scheme's order. In particular, we consider the classical MinMod limiter, its high-order generalization(s), and a recently introduced moment-based limiter by Moe, Rossmanith, and Seal. To couple our quadrilateral sub-domains we consider the popular Lax Friedrichs numerical flux, as well as the HLL, Roe, and Marquina numerical fluxes. The work will be summarizing the strengths and weaknesses of each flux/limiter combination for hypersonic inviscid flow over a blunt object. The comparisons between some of our best performing flux/limiter combinations with a second-order structured finite-volume solver. Our simulations are carried out using a newly developed partial differential equation solver, SpECTRE, that combines a discontinuous Galerkin method with a task-based parallelism model. SpECTRE has been developed with the philosophy of solving astrophysics problems and the present work uses this astrophysics code-base for problems in Mechanical and Aerospace Engineering which required to extend the code for certain additional functionality of putting solid objects in the domain and having set up for different solid wall boundary conditions which was not there before.Previous work shows the code's scalability on the Blue Waters supercomputer up to the machine's full capacity of 22,380 nodes. We will briefly touch on the SpECTRE's scalability for our problem and computational resources. A part of the thesis also presents the strong scaling analysis on MGHPCC machine that has been used extensively throughout the completion of this work.
ADVISOR:
Dr. Scott Field, Assistant Professor, Department of Mathematics at UMass Dartmouth
COMMITTEE MEMBERS:
Dr. Scott Field, Assistant Professor, Department of Mathematics at UMass Dartmouth
Dr. Alireza Mazaheri, Aerothermodynamics Branch, NASA Langley Research Center, Hampton, VA
Dr. Mehdi Raessi, Associate Professor, Department of Mechanical Engineering at UMass Dartmouth
Open to the public. All MNE students are encouraged to attend.
For more information, please contact Dr. Scott Field (sfield@umassd.edu, 508-999-8318).
Thank you,
Sue Cunha, Administrative Assistant
Mechanical Engineering Department
scunha@umassd.edu
508-999-8492
Topical Areas: Faculty, General Public, Staff and Administrators, Students, Students, Graduate, Students, Undergraduate, University Community, College of Engineering, Mechanical Engineering, Lectures and Seminars