In fluid mechanics, known as computational fluid dynamics (CFD), problems in fluid flow and heat transfer behavior are investigated and solved using numerical techniques and algorithms. It can be used in a wide variety of scientific and industrial domains. Various academic and industrial domains use computational fluid dynamics (CFD). It is applied to the design of efficient wind turbines and power plants in the energy sector, mixing and chemical processes in the manufacturing sector, oceanography and climate forecasting in the environmental sciences, structural analysis and flood modeling in civil engineering. Design of energy-efficient buildings in the building industry. It is also applied in aerospace and automotive engineering to improve aerodynamics and engine performance.
These capabilities are made possible by great advances in computing algorithms, physical model building, and data analytics. In addition, high-performance computing (HPC) systems have dramatically improved availability, speed, and efficiency, enabling high-fidelity flow simulations with increasing resolution and considering complex physical processes.
To better understand these phenomena, the study of turbulence is ubiquitous in environmental and engineering fluid flow. Direct Numerical Simulation (DNS), which accurately depicts the unsteady three-dimensional flow field without any approximations or simplifications, is useful for understanding turbulent flows. While appealing, such simulations require a lot of processing power to accurately depict fluid-flow patterns at various geographic scales.
Therefore, to simplify this problem, researchers have developed a simulation formulation that can enable the calculation of fluid flow with TPU. The researchers built this TPU to utilize the latest advances in hardware design and TensorFlow software. They emphasize that this framework exhibits efficient scalability to adapt to different problem sizes, resulting in improved runtime performance.
It uses the graph-based TensorFlow programming paradigm. The accuracy and performance of this framework are studied numerically and analytically, focusing specifically on the effects of TPU-native single-precision floating point arithmetic. The algorithm and implementation are validated with canonical 2D and 3D Taylor–Green vortex simulations.
Throughout the development of CFD solvers, idealized benchmark problems have been used repeatedly, many of which have been incorporated into this research effort. An essential benchmark for turbulence analysis is homogeneous isotropic turbulence (a standardized and well-studied flow in which statistical properties, such as kinetic energy, are invariant under translation and rotation of coordinate axes). The researchers implemented a fine-resolution grid with eight billion points.
The researchers tested its ability to simulate turbulent flows. To achieve this, simulations were performed for two specific configurations: a homogeneous isotropic decay and a turbulent planar jet. The researchers found that both simulations exhibited strong statistical agreement with benchmark answers.
The researchers also used four different test conditions involving 2D and 3D Taylor-Green vortex flow, decaying homogeneous isotropic turbulence, and a turbulent planar jet. The simulation results showed that round-off errors do not affect the solutions, indicating a second-order accuracy level.
check Papers and Google AI articles. All credit for this research goes to the researchers in this project. Also, don’t forget to participate Our 30k+ ML SubReddit, 40k+ Facebook community, Discord channelAnd Email newsletterWhere we share the latest AI research news, cool AI projects and more.
If you like our work, you will like our newsletter.
Rachit Ranjan is a Consulting Intern at MarkTechPost. Currently he is pursuing his B.Tech from Indian Institute of Technology (IIT) Patna. He is actively shaping his career in the field of Artificial Intelligence and Data Science and is passionate and dedicated to explore these fields.