Axial Compressor Stage
Axial Compressor Stage: Theoretical Foundations
Overview
Professor, axial compressors have many rows of blades lined up, right? How much does the pressure increase per stage?
Good question. For subsonic axial compressors, the pressure ratio per stage is roughly around 1.15 to 1.4. Achieving an overall pressure ratio of 10:1 or more by stacking multiple stages is typical for gas turbine compressors.
Only that much per stage? Why not compress it all at once?
Because the flow decelerates during compression, and rapid deceleration can cause boundary layer separation. Keeping the Diffusion Factor (DF) per blade row low is key to stable operation. According to Lieblein's criterion, DF < 0.6 is a guideline.
Governing Equations ― Euler Work Equation
What equation represents the energy transfer per stage?
It's Euler's turbomachinery equation, which is fundamental to turbomachinery.
Here, $U$ is the blade speed, and $C_{\theta}$ is the swirl component of the absolute flow. Subscript 1 denotes inlet, 2 denotes outlet. This equation is derived rigorously from energy conservation and holds true for both compressors and turbines.
So, a larger change in swirl velocity means more work is done, right?
Exactly. However, if the swirl increases too much, the blade loading becomes excessive. So, drawing velocity triangles to check the relative Mach number and flow angles is the ABCs of design. The degree of reaction $R$ is also used to manage the work distribution.
A 50% reaction stage with $R = 0.5$ results in symmetric velocity triangles for the rotor and stator, making it a widely used design that tends to minimize losses.
de Haller Criterion and Diffusion Limit
How much deceleration is allowed in a blade row?
The de Haller number is a simple, practical index often used.
$W$ is the relative velocity. If this ratio falls below 0.72, the risk of boundary layer separation increases sharply. For a more precise evaluation, Lieblein's Diffusion Factor is used.
$\sigma$ is the solidity (chord/pitch). DF < 0.6 is the guideline design limit.
Are these two criteria used in the preliminary 1D design stage, before CFD?
Exactly. The standard workflow is to determine the velocity triangles through 1D Mean-Line Analysis first, then proceed to CFD.
Commercial Tools and Blade Row Design
What software is used for axial compressor blade row design?
Let me list some representative ones.
| Tool | Purpose | Features |
|---|---|---|
| Ansys CFX / TurboGrid | 3D CFD + Dedicated Blade Row Mesher | High-quality structured meshing of inter-blade passages |
| NUMECA FINE/Turbo | 3D CFD (AutoGrid5) | Specialized for multi-stage turbomachinery, supports Non-Matching Interface |
| Concepts NREC (AXIAL) | 1D/2D Preliminary Design | Integrated Mean-Line + Throughflow analysis |
| AxSTREAM (SoftInWay) | 1D to 3D Integrated Design | Seamless linkage from preliminary design to CFD |
I've heard of TurboGrid. What's good about it?
Its ATM optimization (Automatic Topology and Meshing) automatically generates H/J/C/L type topologies suited to the blade shape. O-grids around the leading and trailing edges are also placed automatically, making boundary layer resolution significantly easier.
The Engineers of the Past Who Couldn't Draw Velocity Triangles
The "velocity triangle"—that vector diagram representing the relationship between absolute velocity, relative velocity, and blade speed, essential for axial compressor design—was solved by engineers in the late 19th century through hand calculations and trial and error. Although Euler's equation itself was established in the 1750s, systematic methods for applying it to blade row design (Mean-Line method) weren't fully developed until the 1940s. The rapid demand from jet engine development historically accelerated the "practical application of velocity triangles."
Computational Methods for Axial Compressor Stage
Rotating Frame Formulation
When solving for compressor rotor blades in CFD, how is the rotation handled?
The most common method is the MRF (Multiple Reference Frame) method, i.e., steady-state analysis in a rotating coordinate system. Coriolis and centrifugal body force terms are added to the momentum equation.
$\mathbf{v}_r$ is the relative velocity in the rotating frame, and $\boldsymbol{\omega}$ is the angular velocity vector.
The Coriolis force part is the addition. For steady-state calculation, the time derivative is zero, right?
Yes, for steady MRF, the first term on the left-hand side is zero. A Mixing Plane (circumferential averaging) is placed at the rotor-stator interface to connect blade rows with different pitches in a steady-state manner.
Turbulence Model Selection
Which turbulence model is standard for compressor CFD?
The practical standard is the SST (Shear Stress Transport) k-omega model. It can capture separation due to adverse pressure gradients on blade surfaces more accurately than k-epsilon models. In CFX, simply selecting "SST" enables automatic near-wall switching.
| Turbulence Model | Strengths | Weaknesses | Recommended Use |
|---|---|---|---|
| SST k-omega | Strong against adverse pressure gradients, good near-wall accuracy | Does not capture transition | Steady-state analysis at design point |
| Gamma-Theta Transition Model | Can predict laminar-turbulent transition | Requires calibration parameters | Performance evaluation of low-Re airfoils |
| SAS / DES | Resolves unsteady structures of large-scale separation | High computational cost | Analysis near stall / surge |
In what situations is a transition model necessary?
For low-speed fans or small compressors where the chord-based Reynolds number is below $5 \times 10^5$, the laminar region on the blade surface can be quite long. Since the transition location directly affects losses, the Gamma-Theta (Langtry-Menter) model is effective.
Boundary Condition Settings
What kind of conditions are set at the inlet and outlet?
Typical boundary conditions for a compressor stage are as follows.
- Inlet: Total temperature $T_0$, total pressure $p_0$, flow angle (swirl component), Turbulence Intensity (usually around 5%)
- Outlet: Static pressure or mass flow rate specification
- Blade Surface: No-slip, adiabatic wall is typical
- Hub/Shroud: No-slip rotating wall (rotor side) or stationary wall (stator side)
- Periodic Surface: Rotational periodic boundary condition for one pitch
So if the outlet is set to static pressure, the mass flow rate comes out as a result, right?
Yes. To draw a performance map, you repeatedly calculate by gradually increasing the back pressure at the outlet towards operating points near surge. The point where mass flow rate drops sharply is near the stall limit. However, steady-state calculations cannot fully capture true surge, so unsteady analysis becomes necessary near the limit.
How Boundary Conditions Change the Compressor Map
In axial compressor CFD, "which boundary condition to use" actually significantly affects the results. Whether you specify static pressure or mass flow rate at the outlet completely changes the behavior near the surge limit. In practice, it's often said that "back pressure specification converges more easily," but in return, the flow rate becomes a result of the calculation. The difficulty of turbomachinery CAE lies in these choices of "what to fix and what to solve for," which is where the engineer's skill comes into play.