CPCE CFD Post-Processing, Validation & Verification 2 — Questions and Answers
Question 1: What does a Poincaré section (or phase portrait) reveal during post-processing of time-dependent CFD results?
- The spatial distribution of pressure at a fixed time step
- The periodic or chaotic structure of unsteady flow dynamics (Correct answer)
- The location of shockwaves in supersonic flows
- The cumulative mass flux through a control surface
Correct answer: The periodic or chaotic structure of unsteady flow dynamics
A Poincaré section samples the phase-space trajectory at regular intervals, revealing whether flow dynamics are periodic, quasi-periodic, or chaotic — essential for unsteady flow analysis.
Question 2: Richardson extrapolation in CFD is used to:
- Accelerate solver convergence by extrapolating residuals
- Estimate the exact solution by extrapolating solutions from two or more grid levels (Correct answer)
- Reduce interpolation error when mapping results between non-matching grids
- Extrapolate boundary condition profiles from coarse to fine meshes
Correct answer: Estimate the exact solution by extrapolating solutions from two or more grid levels
Richardson extrapolation combines solutions from systematically refined grids to estimate the continuum (grid-independent) solution and quantify the order of spatial discretization error.
Question 3: In validation of a CFD simulation against experimental data, 'validation uncertainty' accounts for:
- Only the numerical discretization error in the CFD solution
- The combined uncertainties from the simulation and the experimental measurement (Correct answer)
- The error introduced by incorrect boundary condition specification
- The difference in turbulence models between two CFD codes
Correct answer: The combined uncertainties from the simulation and the experimental measurement
Validation uncertainty includes both the simulation numerical uncertainty and the experimental measurement uncertainty, since comparison is only meaningful when both sources of error are quantified.
Question 4: A velocity vector plot in CFD post-processing shows arrows proportional to speed. What does a region of densely packed, curling arrows typically indicate?
- A zone of laminar flow with uniform velocity profiles
- A recirculation zone or vortex in the flow field (Correct answer)
- A region where the mesh is too coarse to resolve gradients
- The location of an inlet boundary condition
Correct answer: A recirculation zone or vortex in the flow field
Densely packed, curling velocity vectors indicate a recirculation or vortex region where fluid revolves around a local center, a common feature downstream of bluff bodies or in separated flows.
Question 5: Which post-processing quantity is most useful for identifying regions of potential cavitation in liquid flow simulations?
- Turbulent kinetic energy (TKE)
- Local static pressure compared to the fluid's vapor pressure (Correct answer)
- Wall shear stress distribution
- Temperature gradient magnitude
Correct answer: Local static pressure compared to the fluid's vapor pressure
Cavitation occurs when local static pressure drops below the vapor pressure of the liquid; post-processing pressure contours against the vapor pressure threshold identifies cavitation-prone regions.
Question 6: The concept of 'observed order of accuracy' in a grid refinement study is determined by:
- The theoretical order of the numerical scheme specified in the solver manual
- Computing the ratio of errors on successive grid levels using a known formula involving the refinement ratio (Correct answer)
- The number of iterations required for convergence on each grid
- Comparing the CFL number across different mesh resolutions
Correct answer: Computing the ratio of errors on successive grid levels using a known formula involving the refinement ratio
The observed order of accuracy is computed empirically from the ratio of solution differences on successive grids using Richardson extrapolation, and should match the theoretical order of the scheme to confirm proper implementation.
Question 7: When comparing CFD results with Particle Image Velocimetry (PIV) experimental data, which consideration is most important for a fair comparison?
- Ensuring both datasets use the same color map for visualization
- Matching the spatial resolution and interrogation window size of PIV with the CFD data extraction plane (Correct answer)
- Running the CFD simulation and the experiment at the same computer time
- Using the same turbulence model as the one assumed in the PIV algorithm
Correct answer: Matching the spatial resolution and interrogation window size of PIV with the CFD data extraction plane
PIV data is spatially averaged over interrogation windows; comparing with CFD requires spatially averaging or filtering the CFD velocity field to the same resolution to avoid artificial discrepancies.
CPCE CFD Post-Processing, Validation & Verification 2
What does a Poincaré section (or phase portrait) reveal during post-processing of time-dependent CFD results?