Fan performance Standards

Until very recently there were more than 12 national Codes for fan testing, incorporating over 70 specific duct arrangements. However, three international Standards, ISO 5801, ISO 5802 and ISO 13347 for specifying the aerodynamic and noise performance of fans have received con­siderable attention. As they alone embody the latest agreements within ISO, their virtues have been extolled in many quarters. Nevertheless, misunderstandings as to their intent and accuracy are apparent.

This Chapter outlines the reasoning behind the various decisions made, how fan performance Standards may be compared and corrects current misunderstandings. ISO Standards are discussed and the differences with previous Standards explained. Shortcomings in the latter have been identified and are rectified.

Until the early 1920s, the methods for testing the aerodynamic performance of industrial fans were legion. It is no exaggeration to say that these were determined by the various manufactur­ers according to their own beliefs, prejudices or downright com­mercial considerations. At about that time, ASHVE (the Ameri­can Society of Heating & Ventilating Engineers, a forerunner of ASHRAE — the American society of Heating, Refrigerating and Air-conditioning Engineers) in the USAand IHVE, (a forerunner of CIBSE — the Chartered Institution of Building Services Engi­neers) in the United Kingdom both set up fan standardisation committees, which produced recommendations for the conduct of such tests and the calculation methods to be used. Subse­quently, these recommendations were incorporated into the appropriate national Standards.

The situation worsened however, as other organisations be­lieved that they had to issue documents if they were not to be left out of the “race”. It seemed that we had simply exchanged one set of problems for another, as ever more organisations felt impelled and qualified to issue their own versions of a fan Stan­dard. Not only did ASME issue its own Standard in the USA but the FMA (Fan Manufacturers Association) in the UK, recognis­ing the deficiencies of the then British Standard, also issued its own code in 1952. Many other National Standards bodies had by then joined the game so that by the 1960s proliferation had made the matter worse than ever.

Into the chaotic situation which existed, ISO stepped with great confidence. It set up Technical Committee TC117 in 1963 to discuss the formulation of an International Standard which could be agreed to be all the major industrial fan manufacturing nations. It started off with considerable optimism and after vari­ous excursions along the way (see Section 4.1.3) eventually settled into a dull routine where each nation sought to protect its own Code at the expense of all the others. Eventually, it dawned that compromise was essential if work was to be completed this side of the grave!

You may well ask “Why all the fuss?” Does it come as a surprise to know that not all those national Codes were of the same tech­nical merit, and serious discrepancies could result? Afew years ago the company for whom the author was then working, car­ried out a series of tests on one particular fan to various stan­dards. The supposed differences in performance (see Figures

4.1 and 4.3) were alarming. In fact, of course, nothing should have changed. If efficiencies had been plotted, then, with an un­changed fan absorbed power, these should have been propor­tional to the fan pressure. This latter is very much a convention (see Section 4.1.3). It should be noted that one of the fans was a tube axial type with appreciable outlet swirl. How this swirl en­ergy is treated can have an appreciable effect on the results. The diffusion at the fan outlet can also be important and how much velocity pressure is converted into useful static pressure may be dependent on the length of such ducting.

This is a world which endeavours to preach the value of free trade. Increasingly, it has had to accept the fact of globalisation. As a contribution to harmony between nations, it is essential that valid comparisons can be made between different compa­nies (and nation’s) products. Only if they are all tested to the same standard test code is this possible.

The fan engineer works under the disadvantage of handling a fluid, which cannot be seen or directly weighed. If necessary a pump flow could be determined by catching the water in a bucket. The engineer does not have the possibility of determin­ing airflows in that way. Furthermore, in the “real” world, air travels in three dimensions and is turbulent. If one is making measurements under actual installation conditions, it is there­fore desirable to take a great many measurements of velocity and direction.

This is the thinking behind ANSI/ASME PTC 11-1984, which is a Code, developed in the USA, for determining performance under operating conditions of large fan units such as those re­quired for mechanical draught in central power stations.

It normally requires the use of a calibrated 5 hole pitot tube combined with a temperature sensor, as shown in Figure 4.1. A traverse is taken directly on the fan discharge and the many measurements of pressure (total and static), direction (pitch and yaw) and temperature (wet and dry-bulb) are then inte­grated to obtain the total flow and pressure. This normally re­quires the aid of a computer to reduce the otherwise tedious hand calculations.

Figure 4.1 View of 5 hole Pitot tube

In Australia, the Standard AS 2936-1987 adopted a similar phi­losophy, but permits the adoption of a simplified 3 or 2 hole yaw meter. These methods have no devices for straightening the swirling airflow, but determine the velocity in the actual direction of flow. Vectoring is then applied to obtain the mean axial flow velocity and hence the volumetric flowrate.

It should be noted that these methods do not necessarily give an accurate result for the fan static pressure. Due to diffusion at the fan outlet, there will be an exchange of kinetic and static en­ergy such that the maximum pressure may be developed at least 3 duct diameters from the fan outlet.

Whilst design programmes exist which can closely predict the performance of a fan, it is nevertheless essential to conduct tests to confirm them. Even the most advanced design tech­niques such as CFD (Computational Fluid Dynamics) require the input of empirically determined correction factors.

Fan tests will be conducted for one or more of the following reasons:

I) Tests carried out during the development of a product range to confirm the design programme

Ii) Tests carried out to provide selection data for a catalogue (either paper copy or electronic)

Iii) Acceptance tests at the manufacturers’ works to confirm that a unit meets the customer’s specification

Iv) Acceptance tests on site to confirm that a unit meets the customer’s specification and/or to confirm that the system resistance is correct or needs modification.

Laboratory tests are essential if the full characteristics of the fan are to be determined from zero flow (shut-off) to full flow (free delivery). Field tests are invariably limited to a particular duty point unless artificial resistance can be inserted into the circuit.

Until very recently there were more than 12 national Codes for fan testing, incorporating over 70 specific duct arrangements. However, three international Standards, ISO 5801, ISO 5802 and ISO 13347 for specifying the aerodynamic and noise per­formance of fans have received considerable attention. As they alone embody the latest agreements within ISO, their vir­tues have been extolled in many quarters. Nevertheless, mis­understandings as to their intent and accuracy are apparent.

This Chapter outlines the reasoning behind the various deci­sions made, how performance to other Standards may be com­pared and corrects current misunderstandings. These ISO Standards are discussed and the differences with previous Standards explained. Shortcomings in the latter have been identified and are rectified.

The aim is to collect, steady and generally organize the flow in a suitable test airway, and this is achieved in the various labora­tory test methods. The major national Codes for fan perfor­mance permitted a fan to be tested in a number of ways. It has been calculated that there were over 70 distinct test methods in use. Many of the methods incorporated in ISO 5801 were taken from the American, British and French Standards. Not all of these are of the same technical merit, and it will come as no sur­prise that some discrepancies can still result. In a world com­mitted to free trade as its contribution to harmony between na­tions, this is a little strange. Some may think that the differences in measured fan performance are not serious. This is not the case. It is a cause for joy that ISO 5801 is currently under re­view resulting, hopefully, in these differences being minimized and in a reduction of its present 232 pages.

Fan performance

The performance of a fan is affected by the connections made to its inlet and outlet. Ducting, where fitted, not only has a pres­sure loss, but can act as an impedance, modifying the flow into or out of the fan casing. In extreme cases it can prevent the de­velopment of a full velocity profile. Ideally the flow velocity vec­tors should be symmetrical and axially aligned (free from yaw) and without swirl or spin (pre or contra) if the fan is to develop its design duty.

The outlet duct

In many former test Codes, the outlet duct was simulated in ei­ther of two ways:

I) Using a parallel duct, usually of a similar area to the fan outlet, for those fan types where the outlet flow permits ac­ceptable measurements of flowand pressure on the outlet side, e. g. centrifugal fans.

Ii) For other fan types-and, in particular, axial flow fans with­out guide vanes — a short length of parallel ducting of the same size and shape as the fan outlet. This means that all measurements of flow and pressure were made on the in­let side.

No doubt the majority of tests carried out in accordance with, these Codes yielded comparable results, but discrepancies could arise (see Figures 4.2 and 4.3).

In each of the figures the same fan was tested to various Stan­dards. For the tube axial fan i. e., without discharge guide vanes, not only is the peak pressure different according to the Code used, but there were considerable differences in the mea­sured Fan Static Pressure over the working range of flowrates. For the centrifugal fan, the conditions at the fan outlet are criti­cal, especially where a tongue piece is fitted and, as with a backward-bladed fan, the impeller is towards the back of the fan casing. The “increase” in performance and efficiency by adding a straight duct of the same cross-section as the outlet and only

Nominal Imp. speed 1425 (rpm) BS 848 :1960 CATEGORY D ———— AMCA 210 : 74 8. BS 848 : 1963 ———— DIN 24163 : 85 ……….. UNI 7179-73

Figure 4.2 Performance of 610 mm tube axial fan to different national Codes

Intake volume flowrate qv m3/s

USA (23)

— Amca 210-74

Britain (24)………….. BS 848 :1880

(?)—————— BS 848 plus 20 straight————————————————————- France © AFNOR NF

X" X10-200:1971

Germany (22)——————————- DIN 24163 :1978

Figure 4.3 Performance of 630 mm backward inclined centrifugal fan to differ­ent national Codes

Two equivalent diameters long before the circular outlet duct, will be noted.

1 (=0 T«

Installation type:

C=0

The conventions for velocity pressure also differed. In Scandi­navia and the Low Countries some companies used to present their data with a discharge loss, which was assessed for the dif­ferent outlet duct configurations. This loss was referred back to the velocity in the annulus between the outside and hub/stator diameters. Essentially the efficiency calculated was an impel­ler /stator efficiency and did not encompass the overall fan. In such cases the Fan Total Pressure would be apparently higher.

The discharge loss was calculated for a constant diffuser or cone efficiency and was the same no matter what the impeller pitch angle. It comprised a conventional duct velocity pressure and an impact loss. Apparently, the effects of residual swirl had been discounted. Where such swirl was present, the discharge losses would be greater than calculated, and the available pressure from the complete fan would be reduced.

ISO conventions

The International Standards Organization Committee TC117 was charged over thirty years ago with the production of a mu­tually acceptable performance test Code. Many of the Commit­tee arguments were fierce and some members adopted en­trenched positions, which they were reluctant to abandon.

With the approach of the true European Common Market on 1 st January 1993, a new sense of urgency developed, for it was in­tended that any resulting ISO Standard would be adopted as a CEN Standard by the so-called accelerated PQ procedure. Great Britain tried to anticipate the outcome of the deliberations by revising BS 848:Part 1 in a new edition published in 1980. There were some scares along the way, for at one stage the French and Belgian delegations were proposing that fan perfor­mance be reported as mass flowrate kg/s against specific en­ergy J/kg! Of course, Great Britain was not completely correct and changes were taking place even at a final meeting in Flor­ence on 5th May 1993. Nevertheless, agreement was reached and ISO 5801 was finally published in 1997. It has subse­quently been adopted in its entirety by Britain, France and Italy with dual numbering e. g., in the UK it is also BS848 Part 1 : 1997.

A number of the concepts included are new to those not familiar with BS 848:1980. Attention should be drawn to the following, which are of major importance:

I) It recognizes that a fan will perform differently according to how it is installed.

Type A with free inlet and outlet

Type В with free inlet and ducted outlet

Type C with ducted inlet and free outlet

Type D with ducted inlet and outlet

It will be seen that the two alternative connections previously mentioned have been combined to give the four possible instal­lation categories (Figure 4.4). In installations of type A, a parti­tion in which the fan is mounted may support a pressure differ­ence between the inlet and outlet sides.

Ii)

-Ma.

It allows considerable flexibility in the methods of measur­ing flowrate. Where these are based on the devices and coefficients described in ISO 5167 for orifice plates, noz­zles and venturi tubes they will have equal validity. In effect all the devices given in BS 848:Part 1:1980 have been in­cluded with the addition of other test assemblies such as the French “caisson-reduit” used with an inlet or outlet ori­fice and the American (AMCA) multi-venturi chambers.

The coefficients of discharge for the British conical inlets have again changed. The 1963 version of BS 848 gave values up to 0.975 at high Reynolds numbers. In the 1980 edition this was

Figure 4.4 Fan installation categories Reynolds number Red x 10’3

Figure 4.5 Compound flow coefficient of conical inlets

Reduced to 0.96. A review of all the data collected by NEL and others suggested that the value should be diameter related and that boundary layer effects are present (Figure 4.5). Calibration of the inlet was always allowed and will continue. For those who practised this, no changes were therefore necessary.

However it is noteworthy that a number of companies have not changed the data in their catalogues for many years. With the various changes in Standards, in some cases dating back to 1952, these cannot be correct.

Pitot-static tube traverses are permitted without calibration al­though these are now restricted to cylindrical ducts to minimize the uncertainty. Thefourmajortypes-NPLmodified ellipsoidal, CETIAT, AMCA and AVA may all be used at the low angles of pitch and yaw without correction.

Iii) Fan pressure is defined as the difference in stagnation

Pressures at fan inlet and outlet, see equation 4.1. Below about 2.0 kPa this is virtually the same as the previously defined Fan Total Pressure. For the ventilation and air conditioning industry, therefore, no problems arise, al­though it will be noted that there will be less emphasis on fan static pressure.

Pf — Psg2 — Psg1

Equ 4.1

Iv) It introduces the concept of “common parts” of the ducting adjacent to the fan inlet and/or outlet sufficient to ensure an accurate and consistent determination of fan pressure no matter what method of flow measurement or control is used. The dimensions of these parts have been specified such that the duct area must be closely matched to the fan

V Cо> 20, , .|.o..7

©

©

Compared with the use of the simulation, or common part, it can be stated that, in the presence of non-uniform and swirling flow from the fan outlet: A) a short length of duct benefits the fan B) a multi-cell straightener, as used in outlet side testing in other national Standards, tends to penalize the fan unduly. Common parts of ducting Satisfactory measurements of pressure cannot be taken imme­diately adjacent to the fan inlet or outlet and it is necessary to establish test stations some distance away, where the flow can be normalized. The quantity measured at these stations is the static pressure, to which is added some conventional velocity pressure to ob­tain the effective total pressure. Oversized ducts can enhance fan performance whilst insufficient length can also result in in­accurate measurements of fan pressures. The common parts include a duct on the outlet side of the fan, having a length of five equivalent diameters to the pressure measuring point and incorporating a standardised flow straightener. Without such parts, different values of pressure can result according to the character of the airflow at the fan outlet. The velocity distribu­tion at this point often contains considerable swirl. Even when free from swirl it is far from uniform. This results in an excess of kinetic energy or velocity pressure over the conventional allow­ance of 1/4pv2 caused by the proportionality of kinetic energy to the local value of rv3 (mass flow x velocity pressure) so that the excess where v is high exceeds the deficit where v is low. Now the non-uniformity of the axial velocity components dimin­ishes as the flow proceeds down the duct and the excess en­ergy reaches a minimum of a few percent of M>pv2 within a length equal to two or three duct diameters, but full uniformity is not reached until about 4.5 diameters, (Figure 4.7). Part of the original excess is lost, but part is converted into additional static pressure, the conventional velocity pressure remaining con­stant. This addition is available for overcoming external resis­tance, and in order to credit it to the fan, as it should be for type B and type D installations, it has been determined that the test station for outlet side pressure measurement should be more than five duct diameters from the outlet (Figure 4.8).

I) Ai • circular fan m •twi 0, • 0,

— ~ ® ! ‘ °’ C Ii fl,,» o ‘1 $ V4- G 0, I.. •Wifi 1 V O <>*

CЈ>

To) Ai • cwcuiar fan mIm •tiinO, #0, Ot* >0980,’ Oi < 1 50, NOT! THewwwwwcn

Te) Ai a xctanfi* ‘** U ■ A): Tha MCI ton ad»aca»t to ma »an min ha« tha urn ractanyutar croaa taction 6 ■ A. a* tna fan miai to afMtfi it N atucNd and its lanfth >i ptax bHaw ± o»:

O®

Il At a orcMim tan ou flat •nan 0, • 01

Lb) At a ctruia» fan outlet «tan 0« * 0, 0H< (0« /0, I* < 1 D7 In *o* HOT« TtaOMNMKMiK

5 S 04

TT^T-TЦ ^ -1‘MM

©

©

4 5 0t

10, 0J#<isr<,Д7 Lji • 1 00, «nan«{a

>L. j==g: ;3o

<r

Lt i • O.7SUAI0« «non а > >A

©

Figure 4.6 Common parts for ducting on fan inlet and outlet Inlet/outlet area as relevant, whilst their length is generally longer than those previously used. (Figure 4.6). V) It specifies the use of a “conditioner” on the outlet of instal­lation type B or D fans. This is designed to dissipate any swirl energy, which is not normally available for overcom­ing the system resistance. Vi) It defines the inlet and outlet areas of the fan as the gross areas inside the casing at the appropriate plane. Vii) Site testing is considered of sufficient importance to be transferred to a separate document (ISO 5802). The tra­versing techniques for a variety of duct cross-sections are detailed. ISO considered it illogical and unacceptable for different fan types of the same installation category to need different test methods because of the differing outlet flow. Thus the neces­sity to devise an outlet simulation, which had the combined re­quirements of conditioning the flow to permit worthwhile mea­surements without severely hampering the fan by excessive pressure losses. These losses were likely to be an important part of the fan pressure determination and would be calculated on the basis of straight, fully developed flow. Then arose another requirement for the common part — to match its actual increase in pressure loss in the presence of non-uniform and swirling flow to that corresponding to a long, straight uniform duct. This was considered a fair requirement, which would neither unduly penalize nor benefit a fan with such an outlet flow. Unfortunately politics intervened and some ex­ceptions to this desirable situation continue to be permitted.

Excess Velocity Pressure Conventional Velocity Pressure

Fan Static Pressure.

Figure 4.7 Velocity diffusion downstream of a fan

Country of Origin	Test Coded	Date	Ducted Outlet Simulation	Straightener	Figure No.	Comments
United Kingdom	BS 848	1980 1963	ISO common parts Duct	Etoile (8 radial vanes)	12-15 16	Equates with ISO 5801 within limited of uncertainty Fan will “benefit” compared with ISO 5801 if appreciable swirl is present.
Untied States	AMCA 210	1985	2D or 3D if test on inlet. Duct + straightener if tested on outlet	Multi-cell	17	Fan will benefit if inlet test method chosen. May be penalized if outlet method chosen — especially if velocity profile is poor and swirl is present.
France	AFNOR NFX10200	1971 1986	Straightener + diverging duct Outlet common part including straightener + diverging duct	Croisillon (vanes) Etoile (18 vanes)	18 19	Pressure may be overstated due to reduced number of straightener vanes and also because pressure is not measured at fan outlet area. Provided pressure is measured in common part, will equate with ISO 5801 within limits of uncertainty.
Germany	DIN 24163	1985	Duct		20	Fan benefits when there is swirl. Regretfully ISO recommendations have not been incorporated despite its recent date.
Italy	UNI 7179-73P	1973	Duct + straightener	Multi-cell	21	Outlet tests may be optimistic, due to increased duct size allowed where pressure is measured. This is partially offset by increased resistance of straightener

Table 4.1 A comparison of national Standards

Type C Type D Short outlet duct Outlet diffuser

BS 848 1980 100JG MK3-U70rpm

*• [I

EC

S3

N

VOLUME FLOW (m3/s)

Figure 4.8 Fan characteristic with outlet swirl

A transition section may be used to accommodate a difference of area and/or shape but to minimize the effects of any change in aerodynamic impedance, it is specified that the duct area shall be within the limits of 5% less and 7% more than the fan discharge area. The dimensions of the transition are also spec­ified to give a small valley angle. National Standard comparisons Figures 4.9 and 4.10 show the requirements in BS 848:1980 and ISO 5801 for the outlet duct simulation. Bearing in mind the difficulties concerning fans with non-uniform and swirling flow at the outlet, the effect of using various national Codes for testing such fans for installation categories B and D compared with ISO 5801 is shown in Table 4.1. Common parts on the fan inlet are shorter and the pressure measurement station need be only three equivalent duct diam-

C ♦ DIFF NON GV 20 CODE 0 NON G. V 20

Figure 4.9 Effect of outlet connections on low pitch angle performance Eters from the fan inlet. This reflects the more regularized con­ditions, which apply on this side. For the same reasons, in an accelerating flow, a greater deviation in the upper limit of duct diameter is permitted. The lower limit is set at 5% less area of

Straightener or conditioner will do this. If it removed just the swirl energy and no more, the minimum energy convention would be satisfied. However, the energy actually removed is very dependent on the combination of swirl pattern and straightener. Again, the need for an agreed standard outlet duct will be appreciated.

In practice, a fan with a lot of outlet swirl ought not to be selected for use with a long straight outlet side duct, because the friction loss in the latter will be substantially increased. Guide vanes should be fitted which will remove and recover (instead of re­moving and destroying) the swirl energy. The flow straightener will then just ensure that test conditions are satisfactory in the downstream duct: the relatively small outlet swirl components from centrifugal, guide-vane axial or contra-rotating fans will be removed without measurable disturbance to the performance.

-X — CODE C NON G. V 32 C »DIFF NON G-V 32 — Y — C *2D NON G V 32 -0- CODE D NON GV 32

The actual design of straighteners to be used in the standard­ised test ducts is therefore of great importance. It is appropriate to review the two types which were considered, and which are also used in ISO 7194.

A) The AMCA straightener is used only to prevent the growth of swirl in a normally axial flow, and does not improve asymmetric velocity distributions. It consists of a nest of equal cells of square cross-section and has a very low-pressure loss. Typical use is either side of an auxiliary booster fan where this is necessary to overcome the resis­tance of the airway when a complete fan characteristic is required. It is especially preferred adjacent to a flow-mea­suring device. This type of straightener is illustrated in Fig­ure 4.12

B)

Figure 4.10 Effect of outlet connections on high pitch angle performance

H Common part {-

Flow measurement by

3a

» M—— r

The Etoile straightener is again designed to eliminate swirl and is of little use in the equalization of asymmetric veloc­ity distributions. The eight radial vanes should be of suffi­cient thickness to provide adequate strength but should not exceed 0.007 D4 for pressure loss considerations. This straightener has a similar pressure drop to the AMCA straightener, i. e., approximately 0.25 times the approach velocity pressure, but is also easier to manufacture. More importantly, it allows the static pressure to equalize radi­ally as the air flows through it. This is not the case with the AMCA straightener, which can produce variations in the

Venturi-nozzle

EC

Immersed orifice

Outlet orifice

Pitot-static traverse

B

Various methods on inlet side of inlet chamber

Figure 4.12 AMCA multi-cell straightener

Figure 4.11 Principle of the common parts applied to type B test airways Duct to fan inlet. Again the transition angles are specified to minimize the effects of flow separation. The principle of the common parts applied to type B test airways is shown in Figure 4.11.

BS/ISO Etoile (Star) 8 Radial blades

Flow conditioners The swirl energy at the fan outlet is only recovered in a straight uniform duct if more than about 100 diameters long. In the presence of swirl, simple measurements of effective pressure or volume flow are impossible, and it must, therefore, be re­moved when tests are to be taken in a duct on the outlet side of the fan, to give information on performance. An effective flow

AFNOR Croisillon (Cross) 4 Radial blades

Static pressure across the duct downstream. The Etoile straightener is therefore preferred in the common duct on the fan outlet and is shown in Figure 4.13.

It should be noted that well designed centrifugal fans or axial and mixed flow fans with efficient outlet guide vanes will not be penalized at design duty by the incorporation of flow condition­ers in the proposed test ducting. However, an axial flow fan without outlet guide vanes will be penalized by the 1980 Stan­dard up to as much as 13 points on peak efficiency and over 20% on pressure. Centrifugal fans with poor outlet velocity pro­files may also suffer. When operating away from the best effi­ciency point i. e. “off-design”, residual swirl may be present in all types of axially ducted fans, such that the straightener will reduce the pressure developed.

Posted in Fans Ventilation A Practical Guide