3) Procedure

(1) Flow: The volume (cfm) of air delivered by the fan will be a product of the outlet velocity as measured by the anemometer times the area of discharge. We will arbitrarily divide the installation types into two classes, - one where the fan is used with no stack or a very short stack on the outlet, and one where a relatively long stack is used.

Where a fan is used with no stack or with an outlet stack less than 1/2 fan diameter high above the plane of the blades, use the area of the torus between the fan hub and cylinder or stack as the discharge area. Divide the area of this torus to the center of area of each. Then choose two stack diameters at right angles to each other and take ten readings of outlet velocity of 1/2 minute duration on each diameter, using the intersections of the diameters with the circles formed by the radii of equal area segments to determine the anemometer position.

• Stack Height Less Than One-Half Fan Diameter
  

  

• Each Increment of Area = (A_t - A_h) / 5
  At: Total area of discharge cylinder, ft²
  Ah: Area of hub, ft²
• Each radius for measurement then becomes:
  r₁ is the radius of a circle whose area is A_h + (A_t - A_h)/10
  r₂ is the radius of a circle whose area is A_h + 3(A_t - A_h)/10
  r₃ is the radius of a circle whose area is A_h + 5(A_t - A_h)/10
  r₄ is the radius of a circle whose area is A_h + 7(A_t - A_h)/10
  r₅ is the radius of a circle whose area is A_h + 9(A_t - A_h)/10
• Stack Height More Than One-Half Fan Diameter
  r₁ = 0.316 R
  r₂ = 0.548 R
  r₃ = 0.707 R
  r₄ = 0.837 R
  r₅ = 0.948 R

Where a fan is used with a discharge stack greater than 1/2 fan diameter high above the plane of blades, the entire area of the inside of the fan stack or cylinder at the plane of measurement becomes the discharge area. Again, divide the stack area into five equal areas and determine the radii to the centers of area of each. Choose two stack diameters at right angles to each other and proceed as above. Again, the sketch in above figure will help to clarify this procedure.

There are two methods of using the anemometer. The discharge from the fan is far from axial and contains varying amounts of rotation. The angle of discharge immediately above the fan varies from hub to tip and also varies with the pressure the fan is working against. As the discharge stack becomes higher, the an stream tends to lose some of this rotational component, but usually there is some angle of yaw even with a stack one fan diameter high and a low static head. By placing a simple protractor on the extension rod of the anemometer and using a light thread attached to the anemometer housing, the angle of yaw can be determined just before the 1/2 minute readings are taken.

The anemometer can be held either normal to the plane of fan rotation or normal to the flow. The correction is less tedious if the first method is used and also depends less on the operator? judgment so this is the method we suggest. If an accuracy of 2% is acceptable and the yaw angles do not exceed 20^o as is quite common, no correction for yaw need be made. If higher accuracy is required the readings can be corrected by using below table.

After the anemometer readings are taken for one-half minute duration they must be doubled to obtain ft/min then corrected for anemometer calibration, and corrected again for yaw angle. Since these readings were all taken at the centers of area of five equal areas, the average of the twenty corrected readings multiplied by the fan discharge area in square feet will give the volume of air (cfm) delivered by the fan.

The effect of air densities other than calibration density (0.075 lb/ft³) is rather low, especially at the velocities found in a cooling fan discharge. Therefore, no correction usually is made. The toleration in density ratio for a one percent variation in accuracy is shown in below table.

If desired, the average velocities at each of the five radii can be determined from the data. These can then be plotted against the radius to graphically portray the fan velocity profile. This has no bearing on the test results, but will be of interest to the fan designer and the tower manufacturer in case remedial action is necessary.

(2) Static Pressure: Static pressure measurements should be taken in the plenum beneath the fan deck in air that is a quite as possible, and in a plane that is 6" to 12" below the fan stack entrance. It is suggested that the 1/4" tube be inserted through the fan deck at four points approximately midway between the base of the fan cylinder and the corner of the cell on the two cell diagonals as shown in below figure.

Care must be taken that the tube end is not near an opening in a cell wall or near a break in partition wall so that it will not be affected by a jet or stream of high velocity air. Care must also be taken to shield the manometer from high velocity air currents. Remember that the magnitude of the readings can be quite small, but will have a great effect on the fan efficiency as will be shown later.

Readings should be taken at each of the four locations at the beginning and the end of the flow tests. The average of these eight readings will be the fan static pressure. This is only true if there are no obstructions on the fan discharge. A guard over the fan outlet or on top of the fan stack, for example, adds to the static head against which the fan is working. It is extremely difficult to measure this pressure and it is recommended that it be calculated. Various authorities have given estimates for the resistance of guards in terms of the average velocity pressure in the stack. A reasonable estimate for various guards is shown in below table.

Example: If a guard made of 1?x 14 gage commercial expanded metal is placed over a stack with an average velocity of 1800 fpm, the approximate resistance pressure is: SP = (27.2/100) x (1800/4005)² = 0.055"

Although the static pressure on the fan discharge is positive and that on the inlet is negative, they must be added arithmetically disregarding the sign since they represent the entire static pressure against which the fan is working.

(3) Air Density: From readings of the dry bulb and wet bulb temperatures in the fan discharge and barometric pressure reading, the density of the air handled by the fan can be determined. Readings should be taken at four points and averaged. Standard tables and psychrometric chart are available, therefore the procedure will not be given here. Note that it is quite difficult to get extremely accurate dry bulb measurements in the fan discharge due to the presence of so much free water in the stream. However, the percentage error in air density for a 1^o error in dry bulb reading is extremely small (approximately 0.2%) and need cause no concern.

(4) Fan Speed: Most of the units we are considering have the fan mounted on a reduction gear driven by an electric motor. The motor rpm can be easily measured by a tachometer. Then by rotating the motor shaft by hand and counting the turns required to get 10 fan revolutions, the gear ratio can be determined. From this data and the motor speed, the fan rpm can easily be determined. The use of belt drive complicates this procedure slightly, since there will be some slippage at operating speed. The best measurement for such a case is either a revolution counter attached to the hub or a stroboscopic device calibrated in rpm.

(5) Power: On three phase electrical installations, the motor input horsepower can be calculated as follows:

HP_(input) = (Volts x Amperes x 3^1/2 x Power Factor) / 746 = Watts / 746

In order to determine the brake horsepower absorbed by the fan, it is necessary to know the motor and gear efficiency. This can be taken from the manufacturer? curves with very good accuracy. The fan horsepower then is: bhp fan = (hp_input) x Efficiency of Gear x (Efficiency of Motor).

Usually motor efficiency curves are not reliable when the operating voltage varies greater than +/- 10% of the motor nameplate.

4) Calculation of Fan Performance

We now have a set of values for flow (cfm), static pressure (sp) and fan brake horsepower (bhp) under the density conditions existing at the fan outlet. In order to relate these to the curve of the fan made from wind tunnel tests, we must convert these values to standard air with a density of 0.075 lb/ft³. This is very simple. The air flow (cfm) needs no conversion since the fan is a constant volume machine and the volume moved does not vary with density. The static pressure (sp), and brake horsepower (bhp), however, do vary directly as the density. The conversion then becomes:

• cfm (standard air) = cfm
• sp (standard air) = sp x (0.075 / test density)
• bhp (standard air) = bhp x (0.075 / test density)

In order to calculate the total efficiency of the fan, we must calculate the velocity head (vp) and add it tothe static head. We do this knowing the area of the fan stack. The formula is: vp = [cfm / (A x 4008.7)]², where A is area of inside of fan stack at the plane of the fan.

Please note again that although the velocity pressure is positive and the static pressure is negative, they are added arithmetically disregarding the sings. tp = sp + vp, tp = total pressure. We now have all the factors in standard air to calculate the fan total efficiency.

Total Efficiency = (cfm x tp)/(6356 x bhp)

We can now plot this one point on the fan curve from the tunnel test to check the relation. Generally, the field test points falls very near the test curve. On a combining factor of flow, pressure and brake horsepower, the field tests generally come within +/- 5% of the fan curve. Certainly this test method is a far cry from laboratory testing, but there are various factors that lead to this in addition to the obvious errors in the test method. Many times the fans are not tested with the proper AMCA type test. Experience indicates that if a fan is tested in accordance with AMCA Bulletin #210 using the chamber test set-up type B with outlet duct, then the field test and the tunnel test will line up very well.

Another factor that influences the performance of the fan is the effect of the drift eliminators in the towers. If these are less than 1/2 fan diameter away from the fan entrance, they can act as partial intake guide vanes either turning the air to effect an apparent increase or decrease in fan angle. This can very easily be the guide vane effect of the drift eliminators.

Another important factor in fan performance is the tip clearance of the blades. Excessive tip clearance becomes more and more of a factor as the resistance of the tower increases. This can completely throw out any correlation between field and tunnel testing. In one recent test tip clearance on a 6?fan was 1/2" too great. The cylinder was lined with fiber board, and the test was re-run, showing an increase of about 18% in fan efficiency. Lastly, the entrance used to the fan should be the same type as you do from a smooth bell entrance. The difference may amount to as much as 0.4vp added to the fan as effective static pressure.

A question that often arises is: ?ow much increase in air flow can be obtained by a given increase in fan horsepower??In other words, if more air is desired on a given tower with a constant fan speed, the fan blade angle can be changed and absorb more power. If the fan efficiency would remain constant, the change in flow would be directly proportional to the cube root of the horsepower ratio. Unfortunately, the fan total efficiency varies from free air to block-off, and also varies with blade angle. Strictly speaking, there is no hard and fast rule for this. However, it would be safe to apply this rule for variations in horsepower not to exceed 15% so long as the fan was originally operating well back from a stall point. The fan manufacturer should be consulted for his recommendations in such cases. There have been cases where the flow actually dropped when the angle was increased although the power absorbed by the fan increased. In this case, the increase in angle put the fan into the stall range.

If the change in horsepower can be effected by a change in fan speed only, the only limiting factor is the tip-speed limitation of the fan. You would not likely reach any critical section speeds in general applications. In this case, the flow would definitely vary as the cube root of the horsepower ratio.