Lecture 32
Lecture 32
https://www.grc.nasa.gov/www/k12/airplane/nseqs.html
Computational Fluid Dynamics
• Types of simulations
• 2D, quasi-3D, 3D
• 2D
• Conceptual design phase
• Long blades/vanes (LP turbines)
• Reasonable results
• Quasi-3D
• Area of flow path changes
• Extra source terms for acceleration/deceleration or
boundary layer growth
Computational Fluid Dynamics
• Types of simulations
• 3D
• True geometry required
• Simulate secondary flows, shock locations
• End wall boundary layers
• Transient or stationery
• Stationery simulations more common
• Transient: flow unsteadiness, vortex
shedding, wake interaction with rotors
Computational Fluid Dynamics
• Grid/mesh
• Structured, unstructured and hybrid grids
• Structured grid
• More suited for well-defined geometries
• More difficult to generate
• Easier to control near-wall clustering of cells
• Unstructured grid
• Primarily intended for complex geometries
• Easier to generate
• Not much control over the near-wall clustering of cells
• Easily automated
Grid Generation
Blocks
Unstructured grid
Grid Generation
• Topology
• Is a structure off blocks that acts as a framework
for placing mesh elements.
• Blocks are laid out without gaps with shared edges
and corners.
• Blocks contain same number of elements along
each side.
• Is usually invariant from hub to tip.
• Can be edited on 2-D layers from hub to tip
sections.
• Number of blocks will dictate the skewness of the
grid elements.
Grid Generation
O-grid topology
Grid Generation
• J-grid:
• Usually used near leading and trailing
edges
• Wraps up in opposite directions at the
leading and trailing edges
• H-grid:
• Tends to complete the meshing by
adding some blocks in an unstructured
manner
• The structured blocks extend from
upstream of the LE, downstream of
the TE and between the blades and
the periodic surface.
Grid Generation
Grid Generation
• 3D compute requirements
• Steady computations
• 1 blade 0.5 – 1 M cells (1-2 CPU hours)
• 1 stage 1-2 M cells (3 CPU hours)
• Unsteady computations
• 1 stage 50 – 100 M cells (20000 CPU hours)
• 1 component (5 stages) 500 M cells (0.1 M CPU hours)!
Errors and uncertainties
• Sources
– Type of simulation Systematic
procedure for
– Geometry errors
estimating these
– Modeling errors errors like say the
– Boundary ASME / AIAA
conditions standards for
– Numerical errors experimental
• Discretization uncertainty
• Round-off analysis
• Convergence