The Correct Formula for Using The Affinity Laws When

There Is a Minimum Pressure Requirement

Richard R. Vaillencourt, PE
Canterbury Engineering Associates

Abstract
When a pump or fan system, or its operation, is not simply overcoming friction, the
Affinity Laws appear to overstate the savings. However, by carefully using the Affinity
Laws, the correct adjustment to the speed ratio for a system with a minimum pressure
requirement can be mathematically defined and applied. With the correct speed ratio
formula, applying the various Affinity Law equations without modification will deliver
the correct results.

Introduction
The application of Variable Frequency Drives (VFD) has proven to save energy for
pumps and fans with variable flow requirements. However, there has been considerable
debate about how to apply the Affinity Laws to calculate the savings. Many individuals
have argued that the Affinity Laws are theoretical and do not apply to actual situations. A
common response has been to modify the Affinity Laws by arbitrarily “adjusting” the
exponent in the horsepower equation to some number between 2 and 3, based upon
personal experience.
Certainly, when there is a minimum system pressure requirement, the direct application
of the equations results in an overstated amount of savings. The problem, however, lies
in the choice of the formula for the speed ratio, not the exponent.

Discussion
The Affinity Laws define the relationship between RPM, flow, pressure, and horsepower
by equating the percent change in RPM with the percent change in these three
parameters. The three common formulas are:

Equation 1
 N2  F2
 
 N1  F1
Equation 2
2
 N2 H2
  
 N1  H1
Equation 3
3
 N2  HP 2
  
 N 1  HP1

N = RPM; F = Flow (GPM, CFM); HP = Horsepower

It will be useful to refer to the various ratios as the (%N), (%F), (%H), and (%HP).
(These are not true percentages unless multiplied by 100, but it is easier than continually
repeating something like “the ratio of (…) expressed as a decimal”, etc.)
These equations can be directly applied when the piping distribution system and end
loads represent only friction losses.
Eq. 1 is the equation of a straight line: y = x. The general “slope-intercept” equation for a
straight line is:
Equation 4
y = mx + b
Where m = the slope of the line and b = the y-intercept.
When there are only friction losses: m = 1 and b = 0. Under these conditions y = x, or
 N2  F2
  , and the curve is a straight line with a 1:1 slope passing through the origin
 N1  F1
on a graph and the point (1,1).

Figure 1
A system curve relates pressure to flow in the basic equation: y = ax 2. A system curve
with only friction losses will take the shape in Fig. 2. When x = 0, y = 0 and the curve
starts at (0,0).
Figure 2

However, if there is a static pressure head to overcome, i.e., an open cooling tower, or
any other minimum pressure requirement, the system curve is moved up the Y-axis by the
minimum pressure. Fig. 3 represents the system curve for a system with a 35’ minimum
pressure requirement (Hmin). This could be could be the actual height difference between
the suction level of a sump and the discharge level of the system. Or it could be a control
setpoint for a variable frequency drive to provide enough pressure to meet the P
requirements of an evaporator bundle and chilled water coils of a chilled water
distribution system, etc.
Figure 3

Hmin
If the pump or fan is operating at a speed that will not produce sufficient pressure to
overcome the minimum pressure requirement, there will be zero flow in the system. The
correct application of the second Affinity Law will tell you what that minimum pump
speed is that will be necessary to produce that minimum pressure even at zero flow.
Under these conditions (a 35’ minimum pressure requirement), Eq. 1 is no longer simply
y = x. The equation now has all the elements of Equation 4: y = mx + b. The graph of
Eq. 1 will now look like Fig. 4. We now need to determine the values of the slope (m)
and the y-intercept (b) to develop the correct speed vs. flow equation.
Figure 4

To find (m) and (b) we must look at the operating curve of the pump (often called the
“pump curve”). The following is the typical shape of a pump curve:
Figure 5

At a specified design flow rate (F D), there is a corresponding design pressure (H D). This
information is readily available on the nameplate of most pumps. These are the
parameters when the pump is operating at the nameplate speed of the motor (N1). We can
now refer to N1 as ND.

Figure 6

HD

FD

Adding the system curve with minimum pressure requirement to the pump curve looks
like this:
Figure 7

HD

Hmin

FD

By applying the second Affinity Law we can determine the pump speed required to
develop Hmin.
Equation 5

2 2
 N2  H2  N min  H min
   Therefore:   
 N1  H1  ND  HD

Solving for the % N:

Equation 6

 N min  H min
 
 ND  HD

The ratio (Nmin / ND) can be called the “Minimum % Speed”. Operating the pump at an
RPM below Nmin will result in a pressure output below the minimum pressure
requirement for flow (Hmin).
The question is: what is the formula for the % N that will allow us to insert it into the
Affinity Law equations and get the right answer when there is a minimum pressure
requirement? In other words: what is the equation of the line for Fig. 4?
Eq. 7 is the equation for the line in Fig. 4 after substituting %N for y and %F for x.
Once we have this equation, we need to determine (m) and (b) in Eq. 7.

Equation 7
  N2   F2 
   m b
 N1   F1 

Since %F (x) is zero when the %N is at the minimum speed which we established was

equal to:
H min , then:
HD
Equation 8

H min
 m 0   b , therefore
H min  b .
HD
HD
Having determined the y-intercept, the slope (m) can be defined in terms of (b) by using
another standard equation for a straight line: the Point-Slope Form.

Equation 9

 y 2  y1 
m  
 x 2  x1 

We know the values of %N and %F at two points: the design point and the H min point. At
the design point, N2 = ND and F2 = FD. Therefore; equation 1 becomes:

Equation 10

 ND  FD
  1
 ND  FD

Therefore, the first point we know is the point (1,1).

The other point that we know is the y-intercept: (0,b). That is: %F = 0 when the %N

drops to
H min . Therefore the Point-Slope formula becomes:
HD
Equation 11

 1 b 
m  
 1 0 

Substituting and simplifying, the slope can be defined in terms of the Minimum % Speed.
Equation 12

 
 1 H min 
 H D 
m 
 1 
Therefore: m  1 H min
  HD
 

Finally, the question after Eq. 6; “What is the formula for the % N that will allow us to
insert it into the Affinity Law equations and get the right answer when there is a
minimum pressure requirement?” can be answered.

Substituting our expressions for (m) and (b) into Eq. 7 gives us the general equation for
the % N that will provide the correct answer for all cases when applying the Affinity
Laws.

Equation 13

 N 2   H min   F2 
 H min
   1    F 1  
 N1   HD  HD
 N2 
This entire formula is substituted for   when applying the Affinity Laws. Therefore
 N1 
Eq. 2 becomes:

Equation 14

2
 H min   F2  H min   H2
1   
  FD     
 HD  HD   H1 
Eq. 3 becomes:

Equation 15

3
 H min   F2  H min   HP 2 
1  
  FD     
 HD  HD   HP1 

Example 1
Suppose you wanted to put a VFD on a boiler feedwater pump. The pump operates
continuously against a throttling valve controlled by the boiler water level. Clearly the
valve is wasting energy whenever it is partially closed, which should be most of the time.
An alternative to throttling is to control the flow into the boiler by controlling the speed
of the pump. The pump will speed up and slow down as the water level falls and rises.
Assumptions:
Pump design = 60 gpm at 175 feet TDH and 80% mechanical efficiency; boiler operating
pressure is 50 psig; boiler capacity is 20,000 PPH .

Calculate the horsepower at 30 gpm (50% flow) using a VFD.

The pump design brake horsepower is:

60  175
bhp   3.3
3960  0.80

Using the Affinity Laws directly:

3
 F2 
HP 2  HP1  
 F1 
Therefore:
3
 30 
HP 2  3.3    0.413
 60 

However, if we look at the second Affinity Law we can see what the output pressure from
the pump would be at the 50% speed.
2
 F2 
H 2  H1 
 F1 

Therefore:
2
 30 
H 2  175    43.75 feet
 60 

Unfortunately 43.75 feet is equal to approximately 19 psig. Since the boiler is operating
at 50 psig (equal to approximately 115 feet), the output from the pump is well below the
pressure of the boiler and no water will flow into the boiler at 50% pump speed.

Under actual operating conditions, the water level in the boiler will continue to drop
sending an ever increasing signal to speed up the pump. As the pump speeds up, the
output pressure will increase until water successfully enters the boiler. However, the
0.413 bhp calculation is well below what the actual bhp will be.

Using the proposed method:

Assumptions: Hmin = 50 psig (115’); HD = 175’ (75.8 psig)

The minimum % speed is given by Equation 6:

 N min  H min
 
 ND  HD
 H min 115
  0.66  0.81
HD 175

Therefore, the minimum % speed that the motor must operate at to be able to deliver
water to the boiler is 81%.

Using Equation 15:

3
 H min   F2  H min 
HP 2  HP1 1  
  FD   
 HD  HD 

3
 115   30  115 
HP 2  3.3 1      
 175   60  175 


HP 2  3.3 1  
0.66   0.5  0.66  3


HP 2  3.3  1  0.81   0.5  0.81  3.3  0.90 

3 3

In other words, in order to deliver 50% flow at sufficient pressure to overcome the boiler
pressure, the pump speed must be 90%.

Therefore:

HP 2  3.3  0.90   2.4bhp

3

Clearly the 0.413 bhp calculation would result in a gross overstatement of the savings. It
may even be likely that the 0.9 bhp savings is not sufficient to economically justify the
use of a VFD.

Example 2
Suppose you wanted to put a VFD on a hot water circulation pump. The pump operates
continuously against throttling control valves at each hot water coil controlled by the
space temperature. The pump operates at a relatively constant horsepower because as the
flow goes down the pressure will go up. An alternative to letting the pressure rise is to
control the total flow from the pump by controlling the speed of the pump to maintain a
constant differential pressure in the system. The pump will speed up and slow down as
the differential pressure falls and rises.

Assumptions:
Pump design = 140 gpm at 75 feet TDH and 80% mechanical efficiency; pressure drop
through the boiler is 15 feet; pressure drop through the hot water coils is 10 feet .

Calculate the horsepower at 70 gpm (50% flow) using a VFD.

The pump design brake horsepower is:

140  75
bhp   3.3
3960  0.80

Using the Affinity Laws directly:

3
 F2 
HP 2  HP1  
 F1 
Therefore:
3
 70 
HP 2  3.3    0.413
 140 

However, if we look at the second Affinity Law we can see what the output pressure from
the pump would be at the 50% speed.
2
 F2 
H 2  H1 
 F1 

Therefore:
2
 70 
H 2  75    18.75 feet
 140 

Unfortunately the pressure drop through the boiler and the coils is equal to 25 feet. And
this does not include the friction losses from the piping system.

Under actual operating conditions, two things will happen. First, with low flow the space
temperature will not be met and the control valves will open further to attempt to get
additional flow. Second, the differential pressure sensor will continue to send an ever
increasing signal to speed up the pump. As the pump speeds up, the output pressure will
increase until sufficient water flows in the system to satisfy space conditions. However,
the 0.413 bhp calculation is well below what the actual bhp will be.

Using the proposed method:

Assumptions: Hmin = 25 feet; HD = 75 feet

The minimum % speed is given by Equation 6:

 N min  H min
 
 ND  HD

H min 25
  0.33  0.57
HD 75

Therefore, the minimum % speed that the motor must operate at to be able to deliver the
correct flow to the system is 57%.
Using Equation 15:
3
 H min   F2  H min 
HP 2  HP1 1  
  FD   
 HD  HD 

3
 25   70  25 
HP 2  3.3 1      
 75   140  75 


HP 2  3.3 1  
0.33   0.5  0.33  3


HP 2  3.3  1  0.57    0.5  0.57   3.3  0.785

3 3

In other words, in order to deliver 50% flow at sufficient pressure to overcome the
pressure drops through the boiler and hot water coils, the pump speed must be 78.5%.

Therefore:
HP 2  3.3  0.785   1.6bhp
3

While there is a significant amount of savings from the application of a VFD in this
example, clearly the 0.413 bhp calculation would result in a gross overstatement of the
savings.

Conclusion
The Affinity Laws do not require modifications to make the theory work for real-world
applications. The only requirement is to address the fact that there may be a minimum
speed requirement to meet a minimum pressure requirement. Once that is included in the
formula, no changes to the Affinity Law equations are required.

Note that as Hmin approaches 0, the equations become the original Affinity Laws. Also, as
Hmin approaches HD the % speed approaches 100% indicating that the flatter the system
curve, i.e., the closer the minimum pressure requirement is to the design pressure, the
lower the savings potential from a VFD. Savings are only available for pressures
between the minimum pressure and the design pressure.

Biography

Richard R. Vaillencourt, PE, is the sole proprietor of Canterbury Engineering Associates,

a consulting firm specializing in supporting other energy engineering companies with
energy audits and assisting commercial, industrial, and institutional customers in the
intelligent use of energy.

Canterbury Engineering Associates

PO Box 459
Canterbury, CT 06331
(860) 546-1124
rrvaillencourt@aol.com