| name | diffuser-design |
| description | Diffuser design — static pressure recovery coefficient Cp, conical/planar/annular diffusers, area ratio, divergence angle, stall limits (Sovran-Klomp), performance charts, separation, turbomachinery diffusers, AMCA standards. |
| metadata | {"priority":7,"promptSignals":{"phrases":["diffuser design","diffuser pressure recovery","conical diffuser","diffuser stall","pressure recovery coefficient","Cp diffuser"],"minScore":3}} |
Diffuser Design — Complete Skill
Diffuser Fundamentals
Purpose: convert kinetic energy (velocity) to static pressure via area increase
Ideal (inviscid): Bernoulli → Cp_ideal = 1 - (A₁/A₂)² = 1 - (1/AR)² [AR = area ratio = A₂/A₁]
Actual: Cp_actual < Cp_ideal due to boundary layer separation, mixing losses
Pressure recovery coefficient Cp:
Cp = (P₂ - P₁) / (½ρV₁²) [dimensionless; = (static pressure rise) / (inlet dynamic pressure)]
Ideal Cp: Cp_ideal = 1 - 1/AR² [perfect recovery; no losses]
Efficiency: η_diff = Cp / Cp_ideal [or: Cp / (1-1/AR²)]
Total pressure loss:
ΔP₀ = P₀₁ - P₀₂ = ½ρV₁² × (1 - Cp - 1/AR²) [positive; represents losses]
Loss coefficient: ξ = ΔP₀ / (½ρV₁²) = 1 - Cp - Cp_ideal... [combines separation + friction]
Performance Parameters
Area ratio:
AR = A₂/A₁ = (D₂/D₁)² (conical); = W₂/W₁ (planar 2D)
Non-dimensional length:
L/D₁ or L/R₁ (conical); N/W₁ (planar 2D) [N = length; W₁ = inlet width]
Semi-divergence angle (conical):
α = arctan((D₂-D₁) / (2L)) = arctan((√AR - 1) × D₁ / (2L))
Blockage (boundary layer at inlet):
B = δ*/W₁ [displacement thickness / inlet width; B = 0.02–0.10 typical; reduces effective AR]
Effective AR = (W₁ - 2δ*) / W₁ × AR → use (1-B) × AR in performance
Stall Limits (Sovran-Klomp Design Chart)
Conical diffuser performance regimes:
At given AR and L/D₁ → Cp from Sovran-Klomp charts (1967; standard reference)
Stall boundaries:
Line a-a (first stall): oscillating stall at one wall; Cp begins to degrade
Line b-b (two-wall separation): full separation; Cp drops significantly
For best Cp at given L/D → operate near but below line a-a
Optimal 2D planar divergence angle:
α_opt ≈ 7–11° (half-angle) for L/W₁ = 5–10; steeper for shorter diffusers
Conical optimal: α_opt ≈ 5–7° (half-angle) for L/D₁ = 3–8
3C correlation (Kline, 1959):
Stall initiates when: 2α > arctan(1.8 × (N/W₁)^(-0.45)) [planar; N = length; W₁ = inlet height]
Conical Diffuser Design
Given: inlet conditions (V₁, D₁), target AR, space constraint L:
- Calculate ideal Cp_ideal = 1 - 1/AR²
- Calculate half-angle α = arctan((D₂-D₁)/(2L))
- Check stall: α < 7° → low stall risk; α > 12° → likely stall
- From Sovran-Klomp chart: actual Cp at (AR, L/D₁)
- Calculate actual pressure rise: ΔP = Cp × ½ρV₁²
Recommended ranges:
| L/D₁ | Optimal AR | α [°] | Cp/Cp_ideal |
|---|
| 2 | 1.5 | 7.2 | 0.70 |
| 4 | 2.0 | 6.7 | 0.80 |
| 8 | 2.5 | 5.7 | 0.85 |
| 12 | 3.0 | 5.6 | 0.87 |
Short diffuser (length constrained):
Add vaned diffuser (splitter vanes) → reduce effective angle per passage
Vaned conical: α up to 20° with vanes (comparable Cp to 7° without vanes)
Annular Diffuser (Turbomachinery)
Application: centrifugal compressor, turbine exit
Additional parameter: hub-to-tip ratio r_h/r_t
Effective area ratio (annular):
AR_eff = A₂/A₁ = (r_t2² - r_h2²) / (r_t1² - r_h1²)
Vaneless diffuser (centrifugal compressor):
Continuity: r × V_r = const → V_r decreases with r
Angular momentum: r × C_θ = const → C_θ decreases with r
Combined: radius ratio r₂/r₁ = 1.4–2.0 typical; Cp = 0.4–0.6 (lower due to swirl at inlet)
Vaned diffuser (wedge, channel):
Vane inlet angle: match absolute flow angle from impeller exit β₂ - 3° to 5° (avoidance shock)
Design: 15–30 vanes; passage AR = 2–4; divergence angle 5–8° per side
Cp = 0.5–0.7 for well-designed vaned diffuser
2D Planar Diffuser (Turning Vanes)
Without turning vanes:
Optimal α: 7–9° (half-angle); Cp = 0.6–0.75 at AR = 2, L/W₁ = 6
With turning vanes: effective angle per passage reduced; allows steeper walls
HVAC duct diffusers:
Target: Cp ≥ 0.6 at AR = 1.5–2; L/H₁ = 3–5
Use turning vanes in elbows to reduce separation + pressure loss
Pressure Recovery Enhancement
Boundary layer control:
Suction slots: remove BL; delay separation; Cp improvement +0.05–0.10
Vortex generators (VG): re-energize BL; tabs, triangular vanes h = 0.5δ; delay stall
Riblets: reduce skin friction in attached BL; secondary effect
Splitter vanes:
Divide passage → each passage sees smaller AR and shorter L/W → reduced stall risk
Optimal vane location: at r/R = 0.6 for single splitter
Diffuser Total Pressure Loss Correlation
Sovran-Klomp empirical correlation (conical, 2D):
ξ = 0.004 × (L/W₁) × (1 + 0.5 AR²) / AR [approximate; W₁ = inlet width or D₁]
Plus separation losses if operating near stall
1D continuity + loss model:
P₂ = P₁ + ½ρV₁² × (1 - 1/AR² - ξ)
Standards
| Standard | Scope |
|---|
| AMCA 210 | Laboratory methods for airflow testing |
| ISO 5801 | Fan performance testing |
| ASME PTC-11 | Fan testing |
| VDI 2044 | Turbocompressor diffuser design (German) |
Output
Provide: diffuser type (conical/planar/annular/vaneless/vaned), inlet velocity V₁ [m/s] and diameter D₁ (or width W₁) [mm], area ratio AR, diffuser length L [mm], half-angle α [°] vs. stall limit, ideal Cp_ideal and actual Cp (from Sovran-Klomp), pressure rise ΔP [Pa], total pressure loss ΔP₀ [Pa] and ξ, diffuser efficiency η_diff [%], stall margin (distance from line a-a), enhancement method if needed (vanes/VG/suction), and applicable standard (AMCA 210, ISO 5801).
