Fans are generally classified as centrifugal, axial, mixed flow, or cross flow according to the direction of airflow through the impeller. Figure 1 shows the general configuration of a housed centrifugal fan. The components of an axial-flow fan are shown in Figure 2.

Unhoused centrifugal fan impellers are used as circulators in some industrial applications (e.g., heat-treating ovens) and are identified as plug fans. In this case, there is no duct connection to the fan because it simply circulates the air within the oven. In some HVAC installations, the unhoused fan impeller is located in a plenum chamber with the fan inlet connected to an inlet duct or cone. This fan arrangement is identified as a plenum fan. Table 1 compares typical characteristics of some of the most common fan types.

All fans produce pressure by altering the airflow’s velocity vector. Velocity change is in the tangential and radial velocity directions for centrifugal fans, and in the axial and tangential velocity directions for axial-flow fans.

Source: Adapted from AMCA Publication 201-02, Fans and Systems, with written permission from Air Movement and Control Association International, Inc.

Figure 1. Centrifugal Fan Components

Figure 2. Axial Fan Components

Centrifugal fan impellers increase pressure of air flowing through them by (1) centrifugal force created by rotating the air column between the blades and (2) kinetic energy imparted to the air by its velocity leaving the impeller. This velocity is a combination of rotational velocity of the impeller and airspeed relative to the impeller. When the blades are inclined forward, these two velocities are cumulative; when backward, oppositional. Fans with backward-curved blades are generally more efficient than ones with forward-curved blades.

Axial fan impellers increase pressure principally by the change in air velocity as it passes through the impeller blades, with none being produced by centrifugal force. These fans are divided into three types: propeller, tubeaxial, and vaneaxial. Propeller fans are nonducted fans customarily used at or near free air delivery, usually having a small-hub-to-tip-ratio impeller mounted in a fan panel, orifice plate, or inlet ring. Tubeaxial fans usually have reduced blade tip clearance and operate at higher tip speeds, giving them a higher pressure capability than the propeller fan. Vaneaxial fans are essentially tubeaxial fans with guide vanes and reduced blade tip clearance, which results in improved pressure, efficiency, and noise characteristics.

Table 1 Types of Fans

	Type		Impeller Design				Housing Design
Centrifugal Fans	Backward-Inclined (includes AF, BC, and BI)			Blades inclined away from direction of rotation and can be single-thickness flat (BI), single-thickness curved (BC), or airfoil (AF) contour. Deep blades allow efficient expansion in blade passages. Air leaves impeller at velocity less than tip speed. For given duty, has highest speed of centrifugal fan designs.				Single- or double-inlet scroll design for efficient conversion of tangential velocity pressure to static pressure. Maximum efficiency requires close clearance and alignment between impeller and inlet.
	Radial and Radial Tipped			Blades are either fully radial (R) or backward inclined with a radial curve at outer edge (RT). Fully radial blades can have back plate and shroud, back plate only, or neither (open). Radial tipped blades normally have back plate and shroud.				Single or double inlet scroll similar to other centrifugal fan designs. Fit between impeller and inlet not as critical as for backward inclined fans.
	Forward- Curved			Large number of thin curved blades with outer edge tipped toward direction of rotation. Air leaves impeller with a high tangential velocity. Relies on scroll housing to convert this velocity pressure to static pressure. For given duty, has lowest speed of centrifugal fan designs.				Single or double inlet scroll shaped design necessary for conversion of tangential velocity pressure to static pressure. Fit between impeller and inlet not as critical as for backward inclined fans.
	Plenum/ Plug			Single inlet centrifugal impellers can be airfoil, backward curved, or backward inclined. Mixed flow or radial impellers can also be used for specific applications.				No integral housing. The equivalent of a housing, or plenum chamber (dashed line), depends on the application. Due to the lack of a housing, all pressure development occurs in the impeller. The components of the drive system for the plug fan are located outside the airstream.
Axial Fans	Propeller			Small number of blades (two to six), either airfoil or single thickness curved, attached to a relatively small hub.				Simple circular ring, orifice plate, or venturi. Optimum design is close to blade tips and forms smooth airflow into impeller. Improved static efficiency can be obtained with an extension of the venturi into an expanding conical discharge.
	Tubeaxial			Usually has four to eight blades with airfoil or single thickness cross section. Blades may have fixed or adjustable pitch. Hub is usually less than half the fan tip diameter.				Cylindrical tube with close clearance to blade tips. Free inlet applications should use a bell mouth inlet to minimize entrance losses.
	Vaneaxial			Typically has more blades than tubeaxial fans. Blades may have fixed, adjustable, or controllable pitch. Hub is usually greater than half fan tip diameter. Most efficient designs have airfoil blades.				Cylindrical tube with close clearance to blade tips. Guide vanes upstream or downstream from impeller increase pressure capability and efficiency.
Mixed-Flow	Mixed- Flow			Combination of axial and centrifugal characteristics. Contoured back or hub; airfoil, curved, or straight blades, either shrouded or not. Airflow through impeller has both radial and axial components.				The majority of mixed-flow fans have a tubular housing for use in ducted applications and include outlet turning vanes.

Table 1 Types of Fans (Continued  )

Performance Curves*		Performance Characteristics			Applications
		Highest efficiency of all centrifugal fan designs, with peak efficiencies occurring at 50 to 60% of wide-open volume. Fan has a non-overloading characteristic, which means power reaches maximum near peak efficiency and becomes lower, or self-limiting, toward free delivery. Airfoil blades are most efficient, followed by curved and then flat blades.			General HVAC. Used in ducted applications covering a large range of pressures. Airfoil blades can be used for clean air industrial operations. For industrial applications where the environment may corrode or erode airfoil blades, consider using single thickness blades instead.
		Higher pressure characteristics than airfoil and backward curved fans. Power rises continually to free delivery, which is an overloading characteristic. Radial tipped blades are slightly more efficient than straight blades. Shrouded impellers are more efficient than nonshrouded.			Primarily for material handling in industrial plants where high duct velocities are required to keep materials airborne. Also for some high-pressure industrial requirements. Rugged impeller is simple to repair in the field. Choice of impeller type generally depends on materials being transported. Impeller sometimes coated with special material.
		Pressure curve less steep than that of backward inclined fans. Curve dips to left of peak pressure. Highest efficiency occurs at 40 to 50% of wide open volume. Operate fan to right of peak pressure. Power rises continually to free delivery which is an overloading characteristic.			Primarily for low- to medium-pressure HVAC applications, such as residential furnaces, central station units, and packaged air conditioners. Slower speed and lower tonal noise characteristics allow them to be used closer to occupied spaces. A possible exception is central station air handling units because of low frequency noise that is more difficult to attenuate.
		Similar to housed airfoil/backward curved/backward inclined fans but slightly less efficient due to the lack of conversion of kinetic energy in the discharge airstream. Because plenum and plug fans do not have usable outlet velocity pressure, they should always be selected based on fan static pressure. Both fans are susceptible to performance degradation caused by poor inlet installation and insufficient clearance to surrounding walls.			Plenum fans are used for HVAC equipment such as air handlers. Advantages include flexibility of equipment discharge and potential for smaller equipment footprint. Multiple plenum fans in parallel (fan arrays) can be used to further reduce the axial length of air handling equipment. Plug fans are commonly used for high-temperature process applications.
		High flow rate, but very low pressure capabilities. Discharge pattern circular or conical and airstream swirls.			For low-pressure, high-volume air-moving applications, such as ventilation through a wall without ductwork or air-cooled condenser fans.
		High flow rate, medium pressure capabilities. Pressure curve can dip to left of peak pressure at higher blade pitches. Avoid operating fan in this stall region. Discharge pattern circular and airstream rotates or swirls downstream of fan.			Low- and medium-pressure ducted HVAC applications where straight through flow and compact installation are required. Used in some industrial applications, such as drying ovens, paint spray booths, and fume exhausts.
		High pressure characteristics with medium volume flow capabilities. Pressure curve can dip to left of peak pressure. Avoid operating fan in this stall region. Guide vanes correct circular motion imparted by impeller, which improves pressure characteristics and fan efficiency.			General HVAC systems in medium- and high-pressure applications where straight-through flow and compact installation are required. Has swirl-free downstream airflow. Used in industrial applications in place of tubeaxial fans. More compact than centrifugal fans for same duty.
		Characteristic pressure curve between axial fans and centrifugal fans. Higher pressure than axial fans and higher-volume flow than centrifugal fans.			Similar HVAC applications to centrifugal fans or in applications where an axial fan cannot generate sufficient pressure rise.

Table 1 Types of Fans (Concluded)

	Type		Impeller Design				Housing Design
Cross-flow	Cross-flow (Tangential)			Impeller with forward-curved blades. Typically impeller width is much greater than diameter.				Specially designed housing for 90° or straight-through airflow. Housing causes air to enter the impeller at its periphery, flow through the impeller, and discharge at a different point of its periphery.
Other Designs	Inline Centrifugal			Single-inlet centrifugal impellers can be airfoil, backward curved, or backward inclined.				Cylindrical tube similar to vaneaxial fan, except clearance to impeller is not as close. Air discharges radially from impeller and turns 90° to flow through guide vanes. Variations include cylindrical and square housings with or without guide vanes. Square housings can include side discharge.
	Power Roof Ventilators	Centrifugal		Single-inlet centrifugal impellers can be airfoil, backward curved, or backward inclined.				Weather-protected housing with means of mounting to a building opening. Usually does not include configuration to recover velocity pressure component. Radial discharge from impeller can be directed either toward building or away from building.
		Axial		Small number of blades (two to six), similar to an axial propeller fan.				Similar housing to an axial propeller fan with added weather protection and means of mounting to a building opening. Can supply or exhaust air from a building, and housing can direct air either toward building or away from building.
		Induced-Flow Exhaust		Centrifugal backward inclined or mixed flow.				Can be scroll or tubular inline with converging cone or nozzle to increase nozzle velocity, which induces additional airflow through outlet.
	Jet Fan	Centrifugal		Single-inlet centrifugal impellers can be airfoil, backward curved, or backward inclined.				Rectangular, low profile. Outlet tapers to increase discharge velocity. May contain outlet straighteners and inlet protection guards.
		Axial		Could be bidirectional or reversible in operation; similar flow rates in either direction is common.				Tubular housing may or may not have guide vanes. Typically fitted with inlet and outlet silencers and protection screens.
		Circulating		Includes small propeller fans up to large-diameter ceiling mounted fans. Typically two to eight blades, either airfoil or single-thickness curved, attached to a comparatively small-diameter hub.				Often unhoused, but can have housing.

Table 1 includes typical performance curves for various types of fans. These curves (unless otherwise stated in the table) show the general total or static pressure Pt_f of P_sf, total or static efficiency η_f or η_s, and fan input power H_i characteristics of various fans at a single constant operating speed as they are normally used. They do not reflect fan characteristics reduced to common denominators such as constant speed or constant propeller diameter, because fans are not selected on the basis of these constants. A particular fan (size, speed) must be selected by evaluating actual characteristics.

Table 1 Types of Fans (Concluded)

Performance Curves*		Performance Characteristics			Applications
		Similar to forward-curved fans. Power rises continually to free delivery, which is an overloading characteristic. Lowest efficiency of any centrifugal fan type.			Low-pressure HVAC components (e.g., fan heaters, fireplace inserts, electronic cooling, air curtains) where a long, narrow discharge is needed.
		Performance similar to backward-inclined fan, except capacity and pressure are lower. Lower efficiency than backward-inclined fan because air turns 90°.			Ducted HVAC applications with air discharging in axial direction (e.g., low- to medium-pressure return air systems in HVAC applications).
		Less efficient than scroll-type housed fan. Centrifugal units are slightly quieter than axial units.			Low-pressure ducted exhaust systems (e.g., general factory, kitchen, warehouse, some commercial installations).
		Usually installed without ductwork; therefore, operates at very low pressure and high volume.			Low-pressure exhaust or supply systems (e.g., general factory, kitchen, warehouse, some commercial installations).
		Performance similar to BI and MF but with reduced efficiency due to high velocity needed for induction. Outlet airflow greater than inlet airflow.			Typically used for laboratory or hazardous chemical exhaust where dilution is required and tall stacks are not desired.
		Performance measured and rated in thrust Ft. Designed for high-velocity discharge to add momentum to larger volume of air. Note that performance curves shown are not at a single speed, but at variable speeds N.			Unducted for space-constrained parking garages and emergency fire and smoke evacuation. Comparable initial cost with ducted solutions but more flexibility for exhaust and inlet air locations.
		Performance measured and rated in thrust Ft. Designed for high-velocity discharge to add momentum to larger volume of air. Typically operates at high speed, which necessitates inlet and outlet silencers. Note that performance curves shown are not at a single speed, but at variable speeds N.			Road or rail tunnels, parking garages, and emergency fire and smoke evacuation. Comparable initial cost with ducted solutions but more flexibility for the exhaust and inlet air locations.
		Performance typically measured and rated in airflow rate versus input power and impeller rotational speed N. Usually operated at free air with high-volume flow rate. Note that performance curves shown are not at a single speed, but at variable speeds.			Used for generating elevated air speeds to provide cooling of people or animals and prevent stratification of temperature or humidity (e.g., inside warehouses, airports, gymnasiums, barns). Can also be used for drying processes.
* These performance curves reflect general characteristics of various fans as commonly applied. They are not intended to provide complete selection criteria, because other parameters, such as diameter and speed, are not defined.

ASHRAE Standard 51 (AMCA Standard 210) specifies procedures and test setups to be used in laboratory testing fans and other air-moving devices. The most common type of test uses multiple nozzle inlet or outlet chambers. Figure 3 shows a pitot-static tube traverse procedure for developing characteristics of a fan. Fan performance is determined from free delivery conditions to shutoff conditions. At free delivery, outlet resistance is reduced to zero; at shutoff, the fan is completely blocked off. Between these two conditions, an auxiliary fan and various airflow restrictions are used to simulate various operating conditions on the fan. Enough points are obtained to define the curve between free air delivery and shutoff conditions. For each case, the specific point on the curve must be defined by referring to the airflow rate and corresponding fan total or static pressure. Other test setups described in ASHRAE Standard 51 should produce a similar performance curve, within the limits of each setup.

Fans designed for use with duct systems are tested with a length of duct between the fan and measuring station. The length of duct evens out the air velocity profile discharged from the fan outlet to provide stable, uniform airflow conditions at the plane of measurement. Pressure losses of the ductwork and flow straightener between the fan outlet and the plane of measurement are added to the measured pressure at the plane of measurement to determine the actual fan performance. Fans designed for use without ducts, including almost all propeller fans and power roof ventilators, are tested without ductwork.

Figure 3. Method of Obtaining Fan Performance Curves

Not all fan sizes are tested for rating. Test information and the fan laws may be used to calculate performance of geometrically similar larger fans, but such information should not be extrapolated to smaller fans. Test information and the fan laws can also be used to calculate performance at other speeds. To determine performance of one fan using the known performance of another, the two fans must be geometrically, kinematically, and dynamically similar. Strict similarity requires that the important nondimensional parameters (those that affect aerodynamic characteristics, e.g., Mach number, Reynolds number, surface roughness, gap size) vary in only insignificant ways. For more specific information, consult the manufacturer’s application manual, engineering data, or Howden Buffalo (1999).

Figure 4. Example Application of Fan Laws

Aerodynamic performance of a fan as installed almost always differs from the performance determined by laboratory testing. Performance differences primarily derive from system configuration differences between the field and laboratory setup, such as added elbows, obstructions in the path of the airflow, and sudden changes of duct cross-sectional area in the field installation.

Because of the performance differences, it is sometimes necessary to determine a fan’s in situ performance. Typical reasons include the following:

In North America, most general fan system evaluations in the field can be accomplished using the guidelines outlined in AMCA Publication 203. When field testing is required and more accurate results are needed to meet a stringent contract, AMCA Standard 803 or ASME Code PTC 11 is often used. In Europe and other areas of the world, ISO Standard 5802 is sometimes used. Because finding a suitable flow measurement plane can often be difficult in the field, it is highly recommended to include provisions for a calibrated flow measuring station near the fan inlet and/or outlet or pressure ports on the fan as part of system design.

The classical form of the fan laws (see Table 2) relate performance variables for any geometrically, kinematically, and dynamically similar fans. They are a simplification of the generalized fan laws (Sardar 2001) in that they ignore Reynolds number effects. The variables are fan size D , rotational speed N , gas density ρ, volume airflow rate Q , pressure P_tf   or P_sf, impeller input power H, and total efficiency of the fan η_t . It is important to recognize that the fan laws only apply to fans and must not be used to calculate performance of other system components, such as drive belts, motors, or variable-frequency drives. More information about fan law deviations related to Reynolds number, Mach number, and bearing and drive losses can be found in Annex E of ASHRAE Standard 51-2016, Phelan et al. (1979), and Sardar (2001).

Fan Law 1 shows the effect of changing size, speed, or density on volume airflow rate, pressure, and power. Fan Law 2 shows the effect of changing size, pressure, or density on volume airflow rate, speed, and power. Fan Law 3 shows the effect of changing size, volume airflow rate, or density on speed, pressure, and power.

As indicated by the fan laws, fan performance is affected by gas density. Unless otherwise identified, fan performance data are based on dry air at standard conditions: 14.696 psi and 70°F (0.075 lb/ft³). In actual applications, the fan may be required to handle air or gas at some other density. Changes in density may be caused by temperature, humidity, composition of the gas, or altitude. With constant size and speed, power and pressure vary in accordance with the ratio of gas density to standard air density.

Figure 4 shows application of the fan laws for a change in fan speed N for the same fan (i.e., D₁ = D₂). The computed P_tf curve is derived from the base curve. For example, point E (N₁ = 650) is computed from point D (N₂ = 600) as follows:

At point D,

Using Fan Law 1a at point E,

Using Fan Law 1b (ρ₁ = ρ₂),

The total pressure curve P_tf at N = 650 rpm may be generated by computing additional points from data on the base curve, such as point G from point F.

If equivalent points of rating are joined, as shown by the dashed lines in Figure 4, they form parabolas, which are defined by the relationship expressed in Equation (2).

Each point on the base P_t f curve determines only one point on the computed curve. For example, point H cannot be calculated from either point D or point F. Point H is, however, related to some point between these two points on the base curve, and only that point can be used to locate point H. Furthermore, point D cannot be used to calculate point F on the base curve. The entire base curve must be defined by test.

Air traveling through a fan and air distribution system contains both static energy and kinetic energy. At any plane in the flow direction, these energies are represented by static pressure P_s and average velocity pressure P_v, respectively. Energy conversion occurs within the fan and distribution system, and is indicated by changes in static pressure to velocity pressure, and vice versa. Total pressure P_t at any plane, representing the total energy, is the sum of the static and velocity pressures. Although static and velocity pressures can increase of decrease with changes in cross-sectional area, total pressure always decreases in the direction of flow (except in the fan, where energy is added and total system pressure increases).

System pressure loss ΔP is the sum of all individual total pressure losses imposed by the air distribution system elements on both the inlet and outlet sides of the fan. Chapter 21 of the 2017 ASHRAE Handbook—Fundamentals provides further information about airflow versus pressure relationships in air distribution systems. System pressure loss is not simply the sum of the static pressure losses in individual distribution system elements (e.g., ducts, coils, filters), except in special cases where air velocities are the same throughout all of the distribution system elements.

ASHRAE Standard 51-2016 describes laboratory methods of test to determine fan aerodynamic performance data, such as fan pressure versus airflow characteristics. The terminology used to describe fan pressures P_tf, P_sf, and P_vf and the applicability of those terms must be clearly understood. Fan total pressure P_tf is the difference between the total pressures at the fan outlet and inlet (P_t2 – P_t1). In contrast, fan static pressure P_sf is the difference between the fan total pressure and the fan velocity pressure (P_tf – P_vf). Therefore, it is also the difference between the static pressure at the fan outlet and the total pressure at the fan inlet (P_s2 – P_t1). Except in special cases, fan static pressure P_sf is not the measured difference between static pressures at the fan inlet and outlet planes (P_s2 – P_s1).

Figure 5. Pressure Changes for Fan with Equal-Sized Ducted Inlet and Outlet Systems

Figures 5 to 7 depict pressure changes throughout various fan and system configurations. The datum line shown in these figures refers to zero pressure for a typical system. Figure 5 shows a fan with ducted inlet and outlet systems, and Figure 6 shows a fan with a ducted outlet system but no connected inlet system. An inlet bell may be required to achieve published fan performance in the absence of an inlet duct. In this case, published fan performance effectively extends from plane 0 to plane 2, and fan static pressure is equal to the static pressure rise across the fan and inlet bell. Finally, Figure 7 shows a fan with a ducted inlet system but no outlet system.

Figure 6. Pressure Changes for Fan with Outlet System Only

Figure 7. Pressure Changes for Fan with Ducted Inlet System Only

Fan selection (covered in detail in the section on Selection) involves, in part, matching the fan pressure capability with system pressure loss at a design condition (often at the maximum required airflow). For fan systems having ducted outlets (Figures 5 and 6), fan total pressure P_tf should be used for fan selection. For a fan without an outlet duct (Figure 7), however, fan static pressure P_sf should be used instead. In the latter case, the fan velocity pressure at the discharge dissipates immediately and cannot contribute to overcoming system resistance. Therefore, the fan static pressure P_sf, which is the static pressure at the fan outlet P_s2 minus the total pressure at the fan inlet P_t1, is matched to the system pressure loss.

Figure 8. Pressure Changes for Fan with Diverging Cone at Outlet

All systems shown in Figures 5 to 7 have inlet or outlet ducts that match the fan connections in size. Often, however, the duct size is not identical to the fan outlet or inlet, thus introducing a further complication. To illustrate pressure changes in this case, Figure 8 shows a diverging outlet cone, which is a common type of fan connection. Static pressure in the cone increases in the direction of airflow (static regain). The regain occurs because the cross-sectional area increases through the cone and the average velocity decreases correspondingly. The total pressure decreases slightly because the conversion from average velocity pressure to static pressure involves losses. As with the other two previous ducted-outlet systems, fan total pressure P_tf should be used for fan selection, and static pressure across the fan (P_s2 – P_s1) does not equal the fan static pressure P_sf.

In certain applications, it may be desirable to calculate the air temperature rise across the fan. For low pressure rises (<10 in. of water), estimate the air temperature rise by

(1)

where

If the motor is not in the airstream, efficiency is the fan total efficiency. If the motor and transmission (e.g., drive belt, if used) are in the airstream, the efficiency is instead the product of the motor, transmission, and fan efficiencies. It is important to recognize that fan, transmission, and motor efficiencies are not necessarily constant, and in many cases vary with the fraction of full-load operation.

Also, heat rejected by variable-frequency drives (if used, and usually not located in the airstream) and the fraction of motor and transmission heat not transferred to the airstream should be accounted for as building loads instead.

Figure 9 shows a simplified duct system with three 90° elbows. These elbows represent the resistance offered by the ductwork and account for duct friction losses and fitting losses. A given airflow through a system requires a definite total pressure in the system. If the rate of airflow changes, the resulting total pressure required will vary (Equation [2], which is true for systems with no coils or filters [e.g., return or exhaust air systems]). HVAC systems such as these generally follow this law very closely.

(2)

In some systems, particularly constant- or variable-volume air systems that include coils and filters and that may also control pressures at some points, the system resistance curves can deviate widely from Equation (2), even if many of the system elements are described by this equation. Sherman and Wray (2010) provide more details about these types of systems, including the effects of system leakage.

Figure 9. Simple Duct System with Resistance to Flow Represented by Three 90° Elbows

Equation (2) can plot this simple system’s pressure loss Δ P curve from one known operating condition (see Figure 4). The fixed system must operate at some point on this simple quadratic system curve as the volume flow rate changes. For example, in Figure 10, at point A of system curve A, when the flow rate through a duct system such as that shown in Figure 9 is 10,000 cfm, the total pressure loss is 3 in. of water. If these values are substituted in Equation (2) for Δ P₁ and Q₁, other points along the system curve (Figure 10) can be determined.

Figure 10. Example System Total Pressure Loss ΔP Curves

For 6000 cfm (Point D in Figure 10):

If the system is changed so total pressure at the design flow rate is increased, the system will no longer operate on the previous system curve, and a new system curve will be defined.

For example, in Figure 11, an elbow added to the duct system shown in Figure 9 increases the total pressure loss of the system. If the total pressure loss at 10,000 cfm is increased by 1.00 in. of water, the system total pressure loss at this point is now 4.00 in. of water, as shown by point B in Figure 10.

Figure 11. Resistance Added to Duct System of Figure 9

If the system in Figure 9 is changed by removing one of the elbows (Figure 12), the resulting system total pressure loss decreases, and the new system curve is curve C of Figure 10. For curve C, a total pressure loss of 1.00 in. of water has been assumed when 10,000 cfm flows through the system; thus, the point of operation is at 2.00 in. of water, as shown by point C.

Figure 12. Resistance Removed from Duct System of Figure 9

The three system curves in Figure 10 follow the relationship expressed in Equation (2). These curves result from changes in the system itself and do not change the fan performance. During duct design, such system total pressure changes may occur because of alternative duct routing, differences in duct sizes, allowance for future duct extensions, or the design safety factor being applied to the system.

In an actual operating constant-air-volume system without coils or air filters, these three system curves can represent three system characteristic lines caused by three different positions of a throttling control damper. Curve C is the most open position, and curve B is the most closed. A control damper forms a continuous series of these system curves as it moves from wide open to completely closed and covers a much wider range of operation than is shown here. Note that pressure-independent throttling of VAV box dampers does not result in different system curves (Sherman and Wray 2010).

A fan is normally tested under one of several standardized laboratory conditions (e.g., wide-open inlet and a long, straight duct attached to the outlet), which result in uniform flow into the fan and efficient static pressure recovery on the fan outlet. When a fan is installed in an application, laboratory conditions are usually not preserved. Typically, inlet and outlet conditions in installations are not the same as in the testing environment, because the fan discharges to a plenum, the discharge or intake side of the fan is too close to a transition (change in area and direction of flow) or wall, or any other interference in the flow field. This difference affects the actual fan performance.

The adverse influences of system connections on fan performance are commonly called system effects. To select and apply the fan properly, the system effects must be considered and the pressure requirements of the fan, as calculated by duct design procedures, must be increased accordingly. More importantly, because of the huge potential pressure (and thus energy) losses associated with system effects, great care should be taken in actual air system design and installation to eliminate or minimize system effects.

The magnitudes of system effects are typically characterized by system effect factors, which are values (usually in terms of pressure) suggested to compensate for the system effects. ASHRAE’s (2012) Duct Fitting Database provides information on converting system effect factors into loss coefficients so that they can be combined with other duct losses. AMCA Publication 201 provides further information on determining system effect factors for various conditions. These calculated system effect factors are only an approximation, however. Fans of different types, and even fans of the same type but supplied by different manufacturers, do not necessarily react to a system in the same way. Therefore, judgment based on experience must be applied to air system design.

To provide improved knowledge and additional test data on system effects, ASHRAE completed a series of research projects to study the inlet system effects on both air and sound for different types of fans. Conclusions from these research projects include the following:

More details are provided by Darvennes et al. (2008), Stevens and Schubert (2010), and Swim (1997). ASHRAE research projects RP-1216 and RP-1420 (Guedel et al. 2011, 2014) also provide further information on inlet and discharge installation effects on airfoil centrifugal fans.

After the system pressure loss curve of the air distribution system has been defined, a fan can be selected to meet the system requirements (Graham 1966, 1972). Fan manufacturers present performance data in either graphics (curves) (Figure 13), tabular form (multirating tables), or selection software. Multirating tables usually provide only performance data within the recommended operating range. The optimum selection range or peak efficiency point is identified in various ways by different manufacturers.

One of the most important selection criteria for fans is energy use. In recent years, fan energy consumption has been addressed by several energy codes and regulations. For example, a building code or regulation may require fans in most applications to meet a minimum fan energy index (FEI). The FEI is a metric used to characterize fan energy usage for fans inclusive of motors and drives. It is defined in AMCA Standard 208 and is being included in ISO Standard 12759-6. The FEI is the ratio of the electrical input power of a reference fan to the electrical input power of the actual fan for which the FEI is calculated, both at the same duty point (which is characterized by the value of airflow and fan pressure). The FEI can be calculated for each point on a fan curve. The FEI may be reported with the manufacturer’s performance data or electronic selection tools. If not, consult the fan manufacturer.

Figure 13. Fan Performance Curve

Other codes and regulation also apply to the extended fan system. For example, the European EU Directive 327/2011: Ecodesign—Fans requires the extended fan system to meet a minimum efficiency metric that is similar to fan motor efficiency grade (FMEG). The FMEG is a wire-to-air efficiency metric and is defined in ISO Standard 12759:2010 (soon to be ISO Standard 12759-4, under development). Determination of FMEG is based on overall fan system peak efficiency, electrical input power, efficiency category, and fan type.

In addition to codes and regulations as related to the extended fan system, some apply only to the stand-alone fan (without drive components) such as the fan efficiency grade (FEG). The FEG is a metric used to characterize fan energy efficiency and is defined in AMCA Standard 205-19 and ISO Standard 12759:2010 (soon to be ISO Standard 12759-3, under development). Fans that comply with a minimum FEG requirement must be selected so that the difference between peak total efficiency and total efficiency at the selection point is within a prescribed value (e.g., 10%, as shown in Figure 13).

Fans that do not meet the minimum requirements of a code or regulation cannot be considered for selection. The FEI, FMEG, and FEG are separate, distinct metrics for describing fan efficiency and energy usage. They can be used to establish compliance with applicable codes or regulations, but not as a means to make fan selections or to quantify actual fan performance in a given application.

Performance data as tabulated in typical manufacturers’ fan performance tables or as shown by fan performance maps are based on arbitrary increments of flow rate and pressure. In the tables, adjacent data, either horizontally or vertically, represent different points of operation (i.e., different points of rating) on the fan performance curve. These points of rating depend solely on the fan’s characteristics; they cannot be obtained from each other by the fan laws. However, points of operation listed in fan performance tables are usually close together, so intermediate points may be interpolated arithmetically with adequate accuracy for fan selection.

Selecting a fan for a particular air distribution system requires that the fan pressure capability matches with the system pressure loss at a design condition. For a fan system with a ducted outlet, the fan total pressure should be used for fan selection. On the other hand, for a fan system without an outlet duct, the fan static pressure should be used instead. Since the system pressure loss is the sum of all individual total pressure losses imposed by the air distribution system elements on both the inlet and outlet sides of the fan, the total system must be evaluated and airflow requirements, system resistances, system air leakage, and system effect factors at the fan inlet and outlet must be known (see Chapter 21 of the 2017 ASHRAE Handbook—Fundamentals and Chapter 19 of this volume). Fan speed and power requirements are then calculated, using multirating tables or single or multispeed performance curves or graphs.

Fan manufacturers provide catalogs or electronic tools for making fan selections. Fan performance data (e.g., airflow rate, pressure rise, input power) are typically presented over a recommended range of fan operating conditions. When the fan is selected within this recommended range, airflow over the aerodynamic surfaces, such as the inlet cone and impeller blades, smoothly follows the surfaces in a way that results in good aerodynamic efficiency. However, when the fan operates outside of this recommended range, airflow may be unable to follow these surfaces and it breaks away, or separates, resulting in a number of undesirable effects. This condition, generally known as stall, is characterized by extensive regions of separated flow, increased noise, and highly unsteady behavior in the key flow variables (airflow rate, static pressure rise, and input power). The latter effect is known as surge. Fans operating under these conditions are subject to mechanical damage because of the large unsteady forces involved.

Different fan types exhibit stall and surge to varying degrees. Care must be taken when selecting a fan to ensure that operation does not extend beyond the recommended boundaries provided by the fan manufacturer. Fan manufacturers provide a recommended selection range for each fan to avoid stall and flow pulsations. For further information about stall, refer to Eurovent (2007).

The selected point of operation (Figure 14) must represent a desirable point on the fan curve, to attain maximum efficiency and resistance to stall and pulsation. In systems where more than one point of operation is encountered, look at the range of performance and evaluate how the selected fan reacts within this complete range. This analysis is particularly necessary for variable-air-volume (VAV) systems, where not only the fan undergoes a performance change, but the entire system deviates from the relationships defined in Equation (2). In these cases, it is necessary to look at actual losses in the system at performance extremes and throughout the part-load operating range (where the system operates most of the time).

Special attention is needed to select the proper motor and any related drive components (e.g., belts, variable-speed drives). The motor and other component ratings must be sufficient to allow operation at all anticipated points of operation. Fans with overloading power characteristics and fans operating at elevated temperatures present areas of concern. Variable-air-volume systems may require special consideration if operated before system balancing. Techniques for assessing the effects of fan system component characteristics (i.e., fan, belt, motor, variable-frequency drive) on efficiency, power, and energy use throughout the operating range of a VAV system are described in DOE (2019).

Figure 14. Desirable Combination of P_tf and ΔP Curves

When two identical fans with stable performance characteristics throughout the range of their individual performance curves operate in parallel, ideally their combined performance curve can be created by doubling the airflow of a single fan at any pressure level. Examples of such fans include backward-inclined centrifugal fans and axial-flow fans operated with a low pitch blade angle. Figure 15A illustrates a combined performance curve of two fans with stable operating characteristics.

When two fans with unstable performance characteristics (i.e., a pressure reduction to the left of the peak pressure point) operate in parallel, an unstable flow condition may occur if the fans operate too close to the peak pressure point of the combined performance curve. Figure 15B illustrates the combined performance curve of two fans operating in parallel.

Figure 15. Two (A) Stable and (B) Unstable Fans in Parallel Operation

Curve A-A of Figure 15B represents the pressure characteristic of a single fan. Curve C-C is the combined performance of two fans operating in parallel. All operating points to the right of point CD, the peak pressure point, are the sum of two times the airflow values for a single fan at the same pressure. The points are along the stable operating range of each fan. All systems with a resistance curve intersecting the combined performance curve to the right of point CD, such as resistance curve D-D, have stable flow characteristics.

Points to the left of point CD are the sum of all airflow values possible at the same pressure level of the fans. The points are along the unstable operating range of each fan. The double-loop shape (∞) seen in Figure 15B to the left of the peak pressure of two-fan operation is formed by summing the multiple airflows possible at a single pressure level. Systems with a resistance curve intersecting the performance curve in this region, to the left of point CD (e.g., resistance curve E-E), have unstable flow characteristics. Unstable flow is shown by the intersection of the system resistance curve at multiple points on the combined fan performance curve, points CE and CE'. The fans oscillate between these two points with random changes in airflow, noise, and vibration levels.

Avoid operating at unstable conditions. Always select fans to operate in their stable range, well to the right of the peak pressure point subject to the guidance in the section on Selection. If possible, plot the combined fan performance curve, including the double loop, and select the fan operating point along the stable portion of the performance curve to the right of the double loop.

When three or more identical fans are installed in parallel, the resulting airflow is the individual fan performance multiplied by the number of fans in the system, for properly selected and installed fans. The designer must ensure that the operating point of each fan is well to the right of the peak to avoid oscillation of any one of the fans. Also, fan configuration (e.g., inlet and outlet duct orientation, distance between fans relative to impeller diameter) plays a major part in how well the system performs.

Two identical fans operating in series theoretically double the pressure rise without changing airflow (Figure 16). Actual performance of the two fans operating in series will be less than the theoretical, because losses occur in the transition between the two fans and the second fan operates less efficiently. Lack of sufficient spacing and/or straightening of flow between the fans further reduces performance. At fan pressures below 30 in. of water, air compressibility can be neglected. Above that limit, calculation of the operating point, fan material selection, and fan design should consider temperature rise caused by friction and compressible gases.

Fan noise is a function of the fan design, volume flow rate Q, total fan pressure P_tf, and efficiency η _t. After the proper type of fan for a given application has been determined (keeping in mind the system effects), the best size selection of that fan is commonly based on efficiency, because the most efficient operating range for a specific line of fans is normally the quietest. Low outlet velocity does not necessarily ensure quiet operation, so selections made on this basis alone are not appropriate. Also, noise comparisons of different types of fans, or fans offered by different manufacturers, made on the basis of rotational or tip speed are not valid. The only valid basis for comparison are the actual sound power levels generated by the different types of fans when they are all producing the required volume airflow rate and total pressure. Obtain octave-band sound power level data from the fan manufacturer for the specific fan being considered.

Source: Reprinted from AMCA Publication 201-02, Fans and Systems, with written permission from Air Movement and Control Association International, Inc.

Figure 16. Theoretical Characteristic Curve of Two Fans Operating in Series

Sound power levels L_w can be determined using several different laboratory methods, such as a reverberant room comparing the sound generated by the fan to the sound generated by a reference source of known sound power; using an anechoic room; or using sound intensity measurements. The reverberant room measuring technique is described in AMCA Standard 300; the enveloping surface method, which uses an anechoic room, is described in ISO Standard 13347-3; and the sound intensity method is described in AMCA Standard 320. These standards do not fully evaluate the pure tones generated by some fans; these tones can be quite objectionable when they radiate into occupied spaces. On critical installations, make special allowance by providing extra sound attenuation in the octave band containing the tone.

Discussions of sound and sound control may be found in Chapter 8 of the 2017 ASHRAE Handbook—Fundamentals and Chapter 49 of the 2019 ASHRAE Handbook—HVAC Applications.

Fan vibration is the structural response of a fan to excitations caused by impeller imbalance, unsteady aerodynamic forces, and drive torque pulsations. The magnitude and extent of the structural response, or vibration level, is determined by the stiffness of the fan components, drive alignment, and bearing properties, among other factors.

Excessive vibration levels can lead to premature failure of the fan, produce high noise levels, and transmit undesirable forces into the support structure. Although fan vibration can never be entirely eliminated, acceptable levels can be achieved through proper fan design, manufacture, and application.

Acceptable vibration levels have been established from a long history of practical fan experience. AMCA Standard 204 defines recommended fan balance and vibration levels based on the fan application and fan drive power. The fan balance and vibration (BV) category, shown in Table 3, determines acceptable levels for fan balance quality and vibration levels.

Fan balance quality, or balance grade (e.g., G 6.3), is used to establish the maximum residual imbalance specification for a given fan speed and size. Balance grades for each of the BV categories are shown in Table 4.

Acceptable vibration levels, as measured in in/s, are shown in Table 5. These levels are generally obtained using an accelerometer placed in the vicinity of the fan or motor. A newly commissioned fan would be expected to meet the start-up

values given in the table. If alarm levels are reached during operation, appropriate steps should be taken to identify, contain, and correct the source of the vibration. The fan should be taken out of service if vibration levels exceed the shutdown values. Operation at these levels can lead to premature failure of the fan.

Acceptable vibration performance at the fan design conditions can be achieved by following these recommendations. However, with the widespread use of variable-speed control, fans may still produce excessive vibration levels at certain critical speeds that correspond to natural frequencies of the fan and/or fan support structure. These critical speeds should be avoided, or eliminated through proper design of the fan components.

The vibration level of a fan changes with fan operating condition and time. Therefore, it is important to monitor, or periodically check, the fan vibration level to ensure safe and reliable operation.

Direction of rotation is determined from the drive side of the fan. On single-inlet centrifugal f ans, the drive side is usually considered the side opposite the fan inlet. The AMCA Standard 99 series defines standard nomenclature for fan arrangements.

For a duct connected to a fan outlet or inlet, a flexible connection should be used to minimize vibration transmission. Provide access to the fan impeller for periodic removal of any accumulations tending to unbalance the rotor. When operating against high resistance or when low noise levels are required, it is preferable to locate the fan in a room away from occupied areas or acoustically treated to prevent sound transmission. The lighter-mass building construction common today makes it desirable to mount fans and driving motors on resilient bases designed to prevent vibration transmission through floors to the building structure. Conduits, pipes, and other rigid members should not be attached to fans. Noise that results from obstructions, abrupt turns, grilles, and other items not connected with the fan may be present. Treatments for such problems, as well as the design of sound and vibration absorbers, are discussed in Chapter 49 of the 2019 ASHRAE Handbook—HVAC Applications.

Many fan applications require the air volume to vary in response to ventilation requirements. Control strategy selection is based on several considerations, such as frequency of airflow changes, effects on fan energy consumption, and first cost of control device(s).

Airflow control can be achieved by changing the system characteristic or the fan characteristic. Methods that change the system characteristic have low first cost, but generally consume more energy compared to methods that affect the fan characteristic.

System characteristic can be altered by installing dampers or orifice plates. This approach reduces airflow by increasing the system restriction. In general, the resulting input power is higher than that of a comparable fan correctly selected for the new airflow operating point. Dampers are usually the lowest-first-cost method of achieving airflow control and are sometimes used in cases where continuous control is needed, although other, more energy-efficient means are available.

Inlet vanes are a form of damper control that offer better energy efficiency. Inlet vanes are typically a series of triangular vanes located in the inlet of a fan. The vanes are controlled through a linkage mechanism, similar to a damper, with one important difference: the vanes are oriented such that incoming flow is turned to provide a preswirl into the fan impeller. In many cases, this can improve fan performance and offset the energy penalty of the vanes. Figure 17 illustrates the change in fan performance with inlet vane control. Curves A, B, C, D, and E are the pressure and power curves for various vane settings between wide open (A) and nearly closed (E).

Changing the fan characteristic (P_tf  curve) for airflow control can reduce power consumption. One option is to vary the fan’s rotational speed to produce the desired performance. If the change is infrequent, the speed of belt-driven fans may be adjusted by changing the drive pulley combination. More frequent changes using belt drive can be accomplished with adjustable sheaves that are manually, electrically, or hydraulically actuated. An alternative method for speed control, and the only option for direct-driven fans, is to vary the motor speed with a variable-speed control, such as a variable-frequency drive (VFD) or similar electronic device. Electronic motor speed control is often the best option from an energy consumption standpoint. When using variable-speed controls based on duct static pressure control, the fan characteristic can still be calculated with the fan laws.

Figure 17. Effect of Inlet Vane Control on Backward-Curved Centrifugal Fan Performance

An alternative to speed control is to change the fan geometry (especially blade setting angles) to optimize aerodynamic efficiency. This approach is more common in axial fans than with centrifugal fans.

Figure 18. Effect of Controlled Blade Pitch on Vaneaxial Fan Performance

Tubeaxial and vaneaxial fans are available with adjustable-pitch blades to allow balancing the airflow system, or to make infrequent adjustments. Vaneaxial fans can also be produced with controllable-pitch blades (i.e., pitch that can be varied while the fan is in operation) for frequent or continuous adjustment. Varying pitch angle retains high efficiencies over a wide range of conditions. The performance shown in Figure 18 is from a typical vaneaxial fan with variable-pitch blades. From the standpoint of noise, variable speed is somewhat better than variable blade pitch. However, both control methods give high operating efficiency and generate less noise than inlet vanes or dampers.

Table 6 summarizes control strategies and their relative energy consumption and first cost.

Further information on fan control as it relates to specific HVAC distribution systems may be found in Chapter 45 of this volume and in Chapter 48 of the 2019 ASHRAE Handbook—HVAC Applications. Additional information about estimating energy consumption during part-load operation is listed in the Fans, Pumps, and Distribution Systems section of Chapter 19 of the 2017 ASHRAE Handbook—Fundamentals. More detailed information about economic analyses is listed in Chapter 38 of the 2019 ASHRAE Handbook—HVAC Applications.

A fan inlet cone has a flared inlet and a narrow throat for efficiently receiving and guiding airflow into a rotating fan impeller. An instrumented fan inlet cone uses pressure taps to measure static pressures at both the inlet and near the throat diameter. The difference between these static pressures, adjusted for the empirically determined characteristics of the inlet cone, is used as an indication of the fan airflow mainly for control purposes (Figure 19).

The technique is based on the Bernoulli and continuity equations, which allow calculation of flow through a converging nozzle based on measurement of the static pressure drop across the nozzle.

Source: Reprinted from AMCA Publication 600-16, Application Manual for Airflow Measurement Stations, with written permission from Air Movement and Control Association International, Inc.

Figure 19. Illustration of Instrumented Fan Inlet Cone

Airflow is estimated by the formula

(3)

where

If the fan inlet is entirely unobstructed, then single pressure taps at the cone’s inlet and throat will suffice. Otherwise, two piezometer rings, each typically consisting of four conduit-connected pressure taps, should be used for pneumatic averaging. The inlet cone calibration factor should be established experimentally by test.