Menu
Design centrifugal pumps using Euler equations, velocity triangles, and specific speed
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Query vapor pressures and NPSH requirements for cavitation assessment
Query thermodynamic properties for 100+ fluids from CoolProp database
Query loss coefficients for pipes, valves, fittings in pump systems
Query fluid viscosities, densities, and material properties vs temperature
Access atmospheric properties and aerospace fluid data from NASA Earthdata
Query high-accuracy thermodynamic properties from NIST REFPROP database (commercial)
| name | centrifugal-pump-design |
| description | Design centrifugal pumps using Euler equations, velocity triangles, and specific speed |
| category | thinking |
| domain | mechanical |
| complexity | advanced |
| dependencies | ["numpy","matplotlib"] |
This skill provides a comprehensive workflow for designing centrifugal pumps from first principles using Euler turbine equations, velocity triangles, and dimensionless parameters. The methodology follows classical pump design theory as presented in Stepanoff, Gülich, and Karassik.
Input Parameters:
Derived Parameters:
Specific Speed Definition:
The specific speed is a dimensionless parameter that characterizes pump geometry:
Ns = (N·Q^0.5) / H^0.75
Where:
Pump Type Selection Based on Ns:
| Specific Speed (Ns) | Pump Type | Impeller Shape | Typical Efficiency |
|---|---|---|---|
| 10-30 | Radial flow | Narrow, radial | 70-85% |
| 30-50 | Francis-vane | Medium width | 80-88% |
| 50-80 | Mixed flow | Wide, angled | 85-90% |
| 80-150 | Mixed flow | Very wide | 85-92% |
| 150-300 | Axial flow | Propeller | 82-88% |
Note: Specific speed units vary by convention. The European convention uses:
nq = N·√Q / H^0.75 [where Q in m³/s, H in m, N in rpm]
US convention multiplies by different constants. Always verify which convention is being used.
Stepanoff Correlation for Outlet Diameter:
D2 = 84.6 · (H / N)^0.5 [D2 in mm, H in m, N in rpm]
Or in SI units:
D2 = K · (H·g / u2²)^0.5 · u2 / (0.5·π·N/60)
Simplified form:
D2 = 60 / π · √(2·g·H / Ku²) / N [D2 in m]
Where Ku is the outlet velocity coefficient (typically 0.95-1.05)
Outlet Width Estimation:
b2 = K_b · D2 / Ns^0.65
Where K_b = 2.5-3.5 for radial pumps
Eye Diameter Estimation:
Based on suction specific speed:
D_eye = (4·Q / (π·c_m1))^0.5
Where c_m1 is the meridional velocity at inlet (typically 2-4 m/s)
Velocity Components:
At any point in the impeller:
Velocity Triangle Relations:
c = u + w (vector addition)
Component breakdown:
Inlet Triangle (Station 1 - Eye):
For axial inlet with no pre-rotation:
c_u1 = 0 (design condition)
c_m1 = Q / A1 = Q / (π·D_eye²/4)
c1 = c_m1
u1 = π·N·D_eye/60
w1 = √(u1² + c_m1²)
β1 = atan(c_m1 / u1) [blade angle at inlet]
Outlet Triangle (Station 2):
u2 = π·N·D2/60
c_m2 = Q / A2 = Q / (π·D2·b2)
From Euler equation and design choices:
c_u2 = g·H / u2 (for backward-curved blades)
c2 = √(c_m2² + c_u2²)
w2 = √((u2 - c_u2)² + c_m2²)
β2 = atan(c_m2 / (u2 - c_u2)) [blade angle at outlet]
Important: β2 is typically 15-35° for backward-curved blades
Fundamental Equation:
The theoretical head developed by a pump is:
H_th = (u2·c_u2 - u1·c_u1) / g
For radial entry (c_u1 = 0):
H_th = u2·c_u2 / g
In terms of blade angles:
H_th = u2² / g - u2·c_m2 / (g·tan(β2))
Slip Factor Correction:
Real pumps have finite number of blades, causing slip:
σ = 1 - (π·sin(β2)) / Z
Where Z is the number of blades (typically 5-9 for radial pumps)
Actual head:
H = σ·H_th
Inlet Blade Angle (β1):
β1 = atan(c_m1 / u1)
Typically: β1 = 15-30° for radial pumps
Outlet Blade Angle (β2):
β2 = atan(c_m2 / (u2 - c_u2))
Blade Type Selection:
Blade Number Selection:
Pfleiderer formula:
Z = 6.5 · (D2 + D1)/(D2 - D1) · sin((β1 + β2)/2)
Round to nearest integer (typically 5-9 blades)
Component Efficiencies:
Total efficiency:
η = η_h · η_vol · η_mech
Hydraulic Efficiency (η_h):
Accounts for friction and shock losses:
η_h = H / H_th ≈ 0.85-0.95
More detailed (Gülich):
η_h = 1 / (1 + k_friction + k_shock + k_recirculation)
Volumetric Efficiency (η_vol):
Accounts for leakage:
η_vol = Q / (Q + Q_leak) ≈ 0.96-0.99
Mechanical Efficiency (η_mech):
Accounts for bearing and seal friction:
η_mech = (P_hydraulic) / (P_shaft) ≈ 0.95-0.98
Overall Efficiency Estimation:
For preliminary design (Gülich correlation):
η_opt ≈ 0.94 - 0.0525·(Ns)^(-0.5) [for Ns in European units]
Or Anderson correlation:
η = 1 - 0.8 / (Ns/50)^0.25 [for well-designed pumps]
Head-Flow Characteristic:
The H-Q curve can be approximated as:
H = a·Q² + b·Q + c
Where coefficients are determined from:
Power Curve:
P = ρ·g·Q·H / η(Q)
NPSH Required:
NPSH_req = (c1²)/(2g) + k·(c1²)/(2g)
Where k accounts for acceleration and losses (k ≈ 1.5-2.5)
Major Loss Sources:
Total head loss coefficient:
k_total = k_friction + k_shock + k_incidence + k_diffusion
Specific Speed Range:
Blade Angles:
Velocities:
Dimensional Ratios:
Number of Blades:
Clearances:
Efficiency:
Operating Range:
NPSH Safety Margin:
Impeller Materials:
Surface Finish:
Incorrect Specific Speed Application:
Velocity Triangle Errors:
Excessive Blade Angles:
Ignoring Slip:
Poor Suction Design:
Unrealistic Efficiency:
Over-Designing:
Under-Designing:
Manufacturing Constraints:
System Mismatch:
Dimensional Analysis:
Affinity Laws:
Comparison with Similar Pumps:
For high heads:
Number of stages = H_total / H_per_stage
Select H_per_stage for optimal Ns (40-80 for best efficiency)
Using affinity laws:
Q ∝ N
H ∝ N²
P ∝ N³
Thoma cavitation parameter:
σ = NPSH_req / H
Typical values:
Modern design uses CFD for:
However, preliminary design still follows classical theory presented here.
Centrifugal pump design is a systematic process combining:
Success requires balancing theoretical calculations with practical experience and validation against existing designs.
See reference.md for detailed derivations, correlations, and literature references.