| name | pump-performance-curves |
| description | Generate and interpret H-Q curves, apply affinity laws, and predict off-design performance |
| category | thinking |
| domain | mechanical |
| complexity | intermediate |
| dependencies | ["numpy","matplotlib","scipy"] |
Pump Performance Curves
Overview
This skill provides comprehensive methods for generating, interpreting, and manipulating pump performance curves. Performance curves are fundamental to pump selection, system design, and operational analysis. The skill covers the four primary curve types (H-Q, η-Q, P-Q, and NPSH-Q), affinity laws for scaling performance, curve fitting techniques, and operating point determination.
Performance Curve Components
1. Head-Flow (H-Q) Curve
The H-Q curve shows the relationship between total dynamic head and volumetric flow rate.
Characteristics:
- Shutoff Head (H₀): Head at zero flow (Q = 0)
- Design Point: Nominal operating condition (Q_design, H_design)
- Runout Flow: Maximum flow at zero head
- Curve Shape: Typically descending, influenced by impeller design
Mathematical Representation:
H(Q) = H₀ - a·Q - b·Q²
Or polynomial form:
H(Q) = c₀ + c₁·Q + c₂·Q²
Physical Interpretation:
- Steep curve: Radial impeller, low specific speed
- Flat curve: Axial impeller, high specific speed
- Rising curve (unstable): Poor design, avoid
Stability Criteria:
dH/dQ < 0 (stable operation)
dH/dQ > 0 (unstable - avoid)
2. Efficiency-Flow (η-Q) Curve
The efficiency curve shows how effectively the pump converts shaft power to hydraulic power.
Key Points:
- Best Efficiency Point (BEP): Peak of efficiency curve
- Preferred Operating Range: 70-120% of BEP flow
- Acceptable Range: 50-150% of BEP flow
- Avoid: Operation below 40% or above 150% of BEP
Efficiency Definition:
η = (ρ·g·Q·H) / P_shaft
Typical Efficiency Curve Shape:
η(Q) = η_max · (1 - k₁·(Q/Q_BEP - 1)² - k₂·(Q/Q_BEP - 1)⁴)
Efficiency by Pump Size:
- Small pumps (P < 20 kW): η = 50-75%
- Medium pumps (20-200 kW): η = 70-85%
- Large pumps (P > 200 kW): η = 80-92%
3. Power-Flow (P-Q) Curve
The power curve shows shaft power consumption vs. flow rate.
Power Calculation:
P_shaft = (ρ·g·Q·H) / η
Curve Shapes by Impeller Type:
Backward-Curved Blades (most common):
- Power increases with flow
- Non-overloading characteristic
- Motor sized for maximum flow
Radial Blades:
- Steep power increase
- Can overload motor
- Requires careful motor selection
Forward-Curved Blades (rare):
- Very steep power curve
- Highly overloading
- Generally avoided
Motor Selection:
P_motor = P_max / η_motor · SF
Where SF = safety factor (typically 1.10-1.25)
4. NPSH-Flow (NPSH_req-Q) Curve
Net Positive Suction Head Required curve shows minimum inlet pressure needed to avoid cavitation.
NPSH Definition:
NPSH_req = (P_inlet - P_vapor)/(ρ·g) + v²/(2g)
Typical Behavior:
NPSH_req(Q) = NPSH₀ + k·(Q/Q_BEP)^n
Where n ≈ 1.5-2.5
Safety Margin:
NPSH_available > NPSH_required · 1.3 (minimum)
NPSH_available > NPSH_required · 2.0 (recommended)
Thoma Cavitation Parameter:
σ = NPSH_req / H
Typical values:
- Radial pumps: σ = 0.06-0.12
- Mixed flow: σ = 0.10-0.20
- Axial pumps: σ = 0.15-0.35
Curve Characteristics and Interpretation
Shutoff Head
Definition: Maximum head at zero flow.
Typical Ratios:
H₀/H_BEP = 1.10-1.35 (centrifugal pumps)
Significance:
- Protection against deadhead operation
- Check valve selection
- System pressure relief sizing
Best Efficiency Point (BEP)
Definition: Operating point with maximum hydraulic efficiency.
Significance:
- Minimum wear and vibration
- Longest bearing life
- Lowest energy consumption
- Reference for operating range
Design Guidelines:
- Size pump so normal duty is 70-110% of BEP
- Avoid continuous operation far from BEP
- Maximum deviation: ±30% of Q_BEP
BEP Identification:
dη/dQ = 0 (at BEP)
d²η/dQ² < 0 (maximum, not minimum)
Operating Range
Preferred Operating Range (POR):
0.7·Q_BEP ≤ Q ≤ 1.2·Q_BEP
Acceptable Operating Range (AOR):
0.5·Q_BEP ≤ Q ≤ 1.5·Q_BEP
Continuous Operation Limits:
- Minimum: 40% Q_BEP (avoid recirculation)
- Maximum: 150% Q_BEP (avoid cavitation and vibration)
Stable vs. Unstable Regions
Stable Curve:
dH/dQ < 0 for all Q
- Single intersection with system curve
- Predictable operation
- Desirable characteristic
Unstable Curve:
dH/dQ > 0 for some Q range
- Multiple intersections possible
- Hunting and oscillation
- Parallel pump problems
- Avoid in design
Dip in H-Q Curve:
- Common at low specific speeds
- Caused by recirculation
- Creates unstable region
- May require minimum flow bypass
Affinity Laws
The affinity laws allow scaling pump performance for speed or diameter changes.
Speed Changes (Constant Diameter)
Flow Rate:
Q₂/Q₁ = N₂/N₁
Head:
H₂/H₁ = (N₂/N₁)²
Power:
P₂/P₁ = (N₂/N₁)³
Efficiency:
η₂ ≈ η₁ (approximately constant)
NPSH Required:
NPSH₂/NPSH₁ = (N₂/N₁)²
Validity:
- Accurate for ±20% speed change
- Assumes Reynolds number independence
- Efficiency slightly improves at higher speeds
Example Application:
N_ratio = 1450/1750
Q_new = 100 * 0.829 = 82.9 m³/h
H_new = 50 * 0.829² = 34.4 m
P_new = 20 * 0.829³ = 11.4 kW
Impeller Diameter Changes (Constant Speed)
Flow Rate:
Q₂/Q₁ = (D₂/D₁)
Head:
H₂/H₁ = (D₂/D₁)²
Power:
P₂/P₁ = (D₂/D₁)³
Efficiency:
η₂ ≈ η₁ - Δη
Where Δη accounts for increased clearance effects (typically 2-5% loss)
Practical Limits:
- Maximum trim: 75% of original diameter
- Below 75%: Efficiency penalty increases significantly
- Minimum recommended: 85% of original diameter
Trimming Guidelines:
D_min = 0.75·D_original (absolute minimum)
D_practical = 0.85·D_original (recommended minimum)
Example Application:
D_ratio = 270/300
Q_new = 200 * 0.90 = 180 m³/h
H_new = 80 * 0.90² = 64.8 m
P_new = 60 * 0.90³ = 43.7 kW
η_penalty = 0.02
Combined Changes
For simultaneous speed AND diameter changes:
Q₂/Q₁ = (N₂/N₁) · (D₂/D₁)
H₂/H₁ = (N₂/N₁)² · (D₂/D₁)²
P₂/P₁ = (ρ₂/ρ₁) · (N₂/N₁)³ · (D₂/D₁)³
Density Effects
For different fluids or temperatures:
H₂ = H₁ (head is independent of density)
P₂/P₁ = ρ₂/ρ₁ (power scales with density)
NPSH₂ = NPSH₁ (head-based, not pressure)
Important: Specific gravity affects power, not head!
Curve Fitting from Test Data
Polynomial Curve Fitting
H-Q Curve (2nd Order):
H = a₀ + a₁·Q + a₂·Q²
Fitting Methods:
- Least squares regression
- Three-point method (Q=0, Q=Q_BEP, Q=Q_max)
- Weighted least squares (emphasize BEP region)
Implementation:
import numpy as np
from scipy.optimize import curve_fit
def head_model(Q, a0, a1, a2):
return a0 + a1*Q + a2*Q**2
params, _ = curve_fit(head_model, Q_data, H_data)
Quality Metrics:
R² = 1 - SS_res/SS_tot (coefficient of determination)
RMSE = √(Σ(H_predicted - H_measured)²/n)
Target: R² > 0.99 for good fit
Efficiency Curve Fitting
Gaussian-like Model:
η(Q) = η_max · exp(-k·(Q - Q_BEP)²)
Polynomial Model:
η(Q) = b₀ + b₁·Q + b₂·Q² + b₃·Q³
Constraints:
- η(0) = 0
- η(Q_BEP) = η_max
- η → 0 as Q → ∞
Power Curve Fitting
Direct Calculation:
P(Q) = ρ·g·Q·H(Q) / η(Q)
Or Polynomial Fit:
P(Q) = c₀ + c₁·Q + c₂·Q² + c₃·Q³
NPSH Curve Fitting
Power Law Model:
NPSH_req(Q) = NPSH₀ + k·Q^n
Where n ≈ 1.5-2.0
Exponential Model:
NPSH_req(Q) = a·exp(b·Q) + c
Uncertainty Analysis
Measurement Uncertainties:
- Flow: ±2-5%
- Head: ±1-3%
- Power: ±1-2%
- Efficiency: ±2-5% (compound error)
Propagation:
σ_η = η·√((σ_Q/Q)² + (σ_H/H)² + (σ_P/P)²)
System Curve Intersection (Operating Point)
System Curve Definition
System Head Equation:
H_system = H_static + K·Q²
Where:
- H_static = elevation difference + pressure head
- K = friction coefficient = (8·f·L)/(π²·g·D⁵)
Components:
H_static = (z₂ - z₁) + (P₂ - P₁)/(ρ·g)
H_friction = (f·L·v²)/(D·2g) = K·Q²
H_minor = Σ(K_i·v²/(2g))
Operating Point Determination
Graphical Method:
- Plot H-Q curve (pump)
- Plot H_system curve
- Intersection = operating point
Analytical Method:
H_pump(Q) = H_system(Q)
a₀ + a₁·Q + a₂·Q² = H_s + K·Q²
Q_op = (-a₁ + √(a₁² - 4·(a₂-K)·(a₀-H_s)))/(2·(a₂-K))
Numerical Method (Newton-Raphson):
def find_operating_point(pump_curve, system_curve, Q_initial=1.0):
f = lambda Q: pump_curve(Q) - system_curve(Q)
Q_op = newton(f, Q_initial)
H_op = pump_curve(Q_op)
return Q_op, H_op
Multiple Pump Configurations
Pumps in Series:
H_total = H₁ + H₂
Q_total = Q (same flow through both)
Use: High head applications
Pumps in Parallel:
Q_total = Q₁ + Q₂
H_total = H (same head for both)
Use: High flow applications
Parallel Operation Stability:
- Requires H-Q curves with dH/dQ < 0
- Unequal flow distribution if curves differ
- One pump may operate at shutoff
Variable Speed Operation
Operating Point Trajectory:
H/H₀ = (Q/Q₀)² (system curve)
H/H₀ = (N/N₀)² (affinity law)
Combined:
Q/Q₀ = N/N₀
Energy Savings:
P₂/P₁ = (Q₂/Q₁)³ (for system curve flow)
Reducing flow by 20% saves ~50% power!
Design Guidelines and Best Practices
Curve Selection
-
Select pump with BEP near duty point
- Target: Q_duty = 0.8-1.1·Q_BEP
- Margin for fouling and wear
-
Verify operating range
- All expected flows within 50-150% Q_BEP
- Consider seasonal variations
- Account for future expansion
-
Check curve stability
- Ensure dH/dQ < 0 throughout
- Avoid dips or humps
- Important for parallel operation
-
Validate NPSH margin
- NPSH_avail > 1.3·NPSH_req (minimum)
- Higher margin at elevated temperatures
- Consider transient conditions
Common Mistakes to Avoid
1. Ignoring Efficiency:
- ❌ Selecting pump based only on H-Q intersection
- ✅ Verify operating point near BEP
- ✅ Calculate lifecycle energy costs
2. Neglecting Operating Range:
- ❌ Designing for single duty point
- ✅ Consider full operating envelope
- ✅ Verify performance at extremes
3. Misapplying Affinity Laws:
- ❌ Extrapolating beyond ±20% speed change
- ❌ Ignoring efficiency changes with trimming
- ✅ Use manufacturer curves when available
- ✅ Account for Reynolds number effects
4. System Curve Errors:
- ❌ Using only friction losses (forgetting static head)
- ❌ Assuming constant K with control valves
- ✅ Include all system resistances
- ✅ Account for changing conditions
5. Parallel Pump Design:
- ❌ Assuming 2 pumps = 2× flow
- ❌ Ignoring individual pump operating points
- ✅ Plot combined H-Q curve
- ✅ Verify stable operation for all combinations
Verification Checklist
Before finalizing pump selection:
Advanced Topics
Variable Frequency Drive (VFD) Operation
Speed Control Advantage:
Power savings = 1 - (N₂/N₁)³
Considerations:
- Bearing minimum speed limits
- Cooling at reduced flow
- VFD efficiency (typically 95-97%)
Specific Speed Effects
Curve Shape vs. Specific Speed:
Ns = N·√Q / H^0.75
- Low Ns (10-30): Steep H-Q curve, flat efficiency
- Medium Ns (30-80): Moderate slope, peak efficiency
- High Ns (80-150): Flat H-Q curve, narrow efficiency peak
Temperature Effects
Viscosity Impact:
- Reynolds number: Re = ρ·v·D/μ
- Head reduction: H_viscous < H_water
- Efficiency penalty increases at low Re
- Use HI correction charts for Re < 10⁶
Vapor Pressure:
- NPSH_req increases with temperature
- Critical for hot liquids
- May require booster pumps
Summary
Pump performance curves are essential tools for:
- Pump selection and sizing
- Operating point prediction
- Energy consumption analysis
- System optimization
- Troubleshooting and diagnostics
Mastery requires understanding:
- Physical meaning of each curve type
- Proper application of affinity laws
- System-pump interaction
- Operating range considerations
- Impact of off-design operation
References
See reference.md for detailed standards, equations, and literature references.
See plotter.py for verified computational examples.