Time-accurate pressure-based algorithms for flow over stationary and moving bodies
Date
2022-08
Authors
Yin, Jiangli
Major Professor
Advisor
Sharma, Anupam
Sarkar, Partha
Ward, Thomas
Leiffson, Leifur
Yan, Jue
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Numerical simulation of flow over multiple bodies in relative motion puts additional demands on computational fluid dynamics (CFD) algorithms. Examples of such flows include interaction of a helicopter rotor with the fuselage, wind turbine rotor blades moving past the turbine tower, etc. In CFD parlance, these are called moving body problems. Despite recent advances in CFD and computer hardware, this class of problems remains challenging and computationally expensive.
This thesis focuses on developing high-performance CFD algorithms to simulate moving body problems. The contributions of the current research are categorized in four areas. First, a new implicit pressure-based algorithm, named SPARK (Segregated Pressure-based All-speed Runge Kutta), is developed by extending the IRK-SIMPLER algorithm to the compressible flow regime. The algorithm is validated for several cases, ranging from the incompressible flow limit to supersonic flow. Comparison against the all-speed version of SIMPLE/SIMPLER shows that the new algorithm offers higher computational efficiency to reach the same level of accuracy in all cases considered.
Second, two unstructured finite volume algorithms, IRK-SIMPLER and Nav2D, are compared for accuracy and efficiency in simulating incompressible flow over stationary bodies. The results show that the pressure-based IRK-SIMPLER algorithm with FCM (Flux Correction Method) is at least an order of magnitude (10 x) faster than the coupled, density-based Nav2D algorithm with low-speed preconditioning.
Third, a series of pressure-based algorithms for moving bodies on unstructured grids have been tested and validated. A new FCM formula for deforming mesh is derived and applied to improve the accuracy. Remeshing and edge flipping techniques are employed to handle large displacement of moving bodies. Current results show that the IRK-SIMPLER with three-stage diagonally implicit Runge-Kutta method and FCM outperforms other algorithms in terms of efficiency and accuracy.
Finally, turbulence modeling and momentum source methods are integrated in the IRK-SIMPLER algorithm for real-world moving body problems. The implementation is validated against experimental data and prior simulations using established methods.
Series Number
Journal Issue
Is Version Of
Versions
Series
Academic or Administrative Unit
Type
dissertation