Accuracy and Performance Evaluation of Low Density Internal and External Flow Predictions using CFD and DSMC

Area of Science:

• Fluid Dynamics
• Computational Physics
• Nanotechnology

Background:

• Direct Simulation Monte Carlo (DSMC) is standard for high Knudsen number flows.
• Classical Computational Fluid Dynamics (CFD) is preferred for low Knudsen numbers (< 0.1) due to accuracy and cost.
• Rarefied flows with a wide Knudsen number range are relevant in aerosolized nanoparticle imaging.

Purpose of the Study:

• Analyze DSMC and CFD simulation advantages and drawbacks for rarefied flows.
• Evaluate accuracy, computational effort, and energy consumption of both methods.
• Assess the feasibility of hybrid CFD-DSMC approaches.

Main Methods:

• Conducted DSMC simulations using SPARTA software.
• Performed CFD simulations utilizing OpenFOAM software.
• Simulated internal expanding flow in a nozzle and external shock-generating flow around a cone.

Main Results:

• Compared simulation results with experimental data.
• Evaluated time-to-solution and energy consumption for each method.
• Assessed the performance and potential of hybrid CFD-DSMC methods.

Conclusions:

• CFD offers better accuracy at lower Knudsen numbers, while DSMC is suitable for higher Knudsen numbers.
• Hybrid approaches present a promising direction for simulating complex rarefied flows.
• The study provides insights into selecting appropriate simulation methods based on flow conditions and research objectives.