bn CENTRIFUGAL FANS THEIR COMMERCIAL APPLICATIONS THESIS FOR DEGREE oF M. E WILLIAM E. PIPER Boe aaa aman alae’ : : : THESIS > om. ow rom} Pad ey Pos ag a o BERKELEY. CAL é At i w CENTRIFUGAL FANS Their Commercial Applications. THESIS CENTRIFUGAL FANS Their Commercial Applications. I. Statements Applying to all Applications. le General form of the problem of fan application. Two essentials. Corollaries. Le Necessity for reliable data. Se Method of obtaining test data. 4. Principles underlying applicability of test data. Two essentials. Corollaries. 6. Forms in which data is presented for use in computations. Two forms. Comparative merits. 6. Formation of "Characteristic Curves". 7. Peouliar usefulness of “Characteristic Curves". Examples. Be Derived curves. Examples. II. Specific Applications. le Heating and Ventilating. Be How to figure amount of air required. Re ._ * the static pressure required. ce " £=©®* gelect the fan-kind and size. de " "™ figure the fan performance. (a, b, o, and @ illustrated by working a problem. Various aspects treated, such as heating and vent- ilation of:- Factories, Theatres, Churches, Schools, Paper Mills, Dye houses, etc. a, b, c and d@ disposed of in each case.) Ze Ventilation. Specific cases, ventilation only. Special methods of application, and special fan materials and con struction. Cooling. 3e Mechanical Draft. Forced. Induced. Uses of Mechanical Draft. How to figure a, b, c and @ for both cases. Forge and Cupola blowers. LOV3B TL 4. 5. Ge Te 8. 10. lle Moving of Materials. a, b, c and @ figured for various applications. Formulae given for various materials. Drying. Special methods of application. Temperatures, time for drying, etc. Air Washer applications. Effect of air washers on fan computations. Organ blowers. Single stage centrifugal type only. Gas blowers and boosters. Cooling towers. Transformer cooling. Disco Fans. Characteristics. When used rather than centrifugal fans. ADDENDA. Suggestions for thesis work with fans. CENTRIFUGAL FANS. Their Commercial Applications. I. Statements Applying to all Applications. (1) General Form of the Problem of Fan Application. Just as the capacity of a pump is expressed in Gal~ lons per minute against pounds pressure or feet head, so the delivery of a fan should be stated as cubic feet per minute against a static resistance of ounces per square inch, or a head of inches of water. The two essentials are:~ (a) Cubic Feet per Minute. (db) Static pressure. The formerly common practice of specifying cubic feet per minute and total pressure (usually in ounces per sq.in.) though still largly followed through habit, is erroneous and misleading. The total or dynamic pressure at the fan's outlet is the sum of the static pressure (resistance) and the fan's outlet velocity pressure. Therefore, specifying the total pressure without specifying fan outlet velocity makes the static pressure to depend on the size of the fan outlet, which fixes the velocity pressure for a given cubic feet per minute. This is erroneous, for the static pressure is the resistance of the distributing system and is constant for that system for any given cubic feet per minute, (CFM) regardless of the fan outlet size. On the other hand, specifying the fan outlet size limits or prohibits competition by giving undue preference to the make of fan having that outlet size, for few manu~ facturers have identical fan dimensions. The proper and impartial specification names the c.f.m. and static pressure, with certain subsidiary points. Aside from structural details affecting strength and rigidity, which may be covered by a general specifica~ tion demanding these qualities, there are several re~ quirements which vary for different cases, but they may be summed up in the following corollaries. (1) Horsepower. This involves the efficiency of the fan and is a fair ground for strictness. It is best to require each bidder to make his own guaranty of power consumption. It is the product of total pressure times c.f.m., divided by efficiency, which is proportional to horsepower, so for a given static pressure,and cfm a large fan of low efficiency may require less power than a smaller less efficient one, due to the latter's greater velocity pressure and consequently greater total Preagure. Le (2) Speed. To be impartial, the revolutions per min~ ute (rpm) of the fan should not be specified unless it — is directly connected to some fixed speed driver, such as a constant speed motor. This is because different makes of fans require different speeds for the same results. There are some cases wherein it is fair and necessary to limit the fan's peripheral velocity, i.e. the vel~ ocity of the tips of the fan wheel blades. (3) Peripheral Velocity. (P.V.) Most centrifugal fans become too noisy for church, school or theatre work when their P,.V. exceeds 4000 ft. per minute. Disc fans should be limited to 3750 ft. per minute. The newer types of many~bladed@ centrifugal fans do not need to attain this velocity with the static pressures usu~ ally met in such cases, but it is well to limit it to 4000 ft. per minute for such work. (4) Outlet or Inlet Velocity. More than 2750 ft. per minute has been found noisy in cases as mentioned in (3) above. This limit should be placed where noise is objectionable. (2) Necessity for Reliable Data. Fan engineering is rapidly becoming standardized. Formerly most fan applications were made without strict requirements as to c.f.m. or resistamce to be Overcome. Results were often guaranteed by the fan maker, who undertook to heat a building to 70 degrees in zero weather, to dry 1000 board feet of lumber in a given time, etc. Success depended on the good judgment, ex~ perience, or luck of the fan maker,who was thus by costly failures driven to be an expert engineer in many ‘lines which used fans. His knowledge was however large~ ly a "rule of thumb", not scientific, nor reduced to a technical basis. Certain sizes gave certain results and were expected to give the same results in similar cases. A certain size of engine was required for this fan before, so it must always be used for this size of fan. One manufacturer at least, for years @talogued his - engine driven fans according to the size of the engine, not of the fan, and in fact his separate fans were listed by the size of the engine which should be used to drive them. Such crude analysis ydelded to attempts at system= ising and classifying fan knowledge. Rules and formulae appeared, all more or less useful in a restricted sense, in that they were based on observed results, but they were not of universal application. For instance, one well worn rule for factory heating advised that the sum of the areas of the outlets in the distributing system should be a certain multiple of the fan outlet area, that the area of the branches should be another multiple of the fan outlet area, and that the area of the main trunk should be another multiple. This was irrespective of the number or length of branches or length of trunk, etc. Se Of course such methods were blind to the fundamental re- quirements of capacity and resistance. Custom still binds us to many 01d tricks of the trade. In exhausting the refuse from wood working machinery we still assign to each machine a certain size of pipe without knowing how much material is to be withdrawn there- by. Until very recently we had no real knowledge of the proper air velocities, but ran the fan at "three or four ounces pressure™, and the fan size was fonnd by adding branch areas progressively, making the main line equal the sum of the branches, seleoting a fan whoscinlet came nearest to fitting the main line size. This shows the need of a change. Such methods if sucessful are yet open to the suspicion of waste. We must quit guessing and get down to fundamentals. In some branch~ es of the business, such a change has taken place. Architects and enrineers are being employed on problems involving the uses of fans with the result that standards are being established and specifications written for fan performance. The manufacturer guarantees his goods to meet specifica tions while the customer relies upon the engineer for re~ sults. This requires that the manufacturer must know intimately and accurately the capacities of his fans. While it is still necessary for fan makers to be familiar with all lines of fan application, this necessity is becoming less, while the need for accurate data regarding their own fan performances is greatere Such information tends to better the product, as the superiority of sdme design and the weak points of andther are clearly shown in the analysis. The maker without reliable data will sustain financial loss from his failure to meet guarantees, or will be under~ bid by a competitor who can present a smaller,more efficient, or otherwise more desirable fan to meet specifications. While ££ is not necessary for the enfinee r to possess such data, since in any case a test will determine whether the requirements have been met, yet an accurate knowledge of various fan performances would enable him to save time and worry by eliminating ignorant, reckless, or irresponsible bidders whosework might otherwise be rejected only after installation and test. Such information is hard to obtain from those of the manufacturers who have it, since it is costly to prepare and it is guarded carefully from the sight of competitors, in the main. If an engineer has the chance he should make his own tests and this thesis is written to indicate a method of making such tests, how to focus the data into usable shape and how to apply tho tools thus formed, to various uses to Which fans are put in actual practice. The capacity tables published in fan catalogs are un- avoidably rigid and limited in usability since each app- lies to but one set of conditions. As will be seen later they are useful merely as a puide. The tabular speed and horsepower cannot obtain, unless the corresponding tabular capacity and pressure happen to be identical with those of the case in hand. . foo great reliance has been placed in such tables in the past, with the result that fans end their results have been condemned, and confidence lost, whereas the fault was in the fact that the fan was expected to maintain the same results under varying conditions. Considerations of the ahove nature show the necessity that the manufacturer and the engineer shall possess better knowledge of fan performance. (3) Method of Obtaining Test Date. The purpose is to obtain data which can be used to show the performance of any sized fan of a given type under any possible conditions of installation, i.e. for any capa- city and pressure. Therefore arrangements must be made to reproduce in the test all such conditions, or a nwmber sufficient for thse construction of smooth and practically precise performance curves. Beside these condition=producing arrangements, certain measurements must be arranged for, of the principal quantities involved, for each condition. These cardinal quantities are:~ a. Peripheral Velocity of the fan wheel. This is the velocity of the tips of the fan wheel blades. be Air Velocity, through the fan orifice. From this the quantity delivered can be derived, but it is the velocity, not quantity, which is recordeé@, for the data is to be work~ ed up in terms of unit outlet area. Strictly it is the~ Ge Air Velocity Pressure at fan orifice which is observed and recorded, since the measurement is usually made by Pitot tubes, not anemometers. The pressure and velocity are readily convertible terms. ad. Static Pressure of Fan orifice. If the fen is being tested as a blower, ey . oe / Lat? nel owe C, Mr. Treat designed this tube and it rightfully should bear his name, as the Taylor tube bears that of Capt. Tay- lo®¥, but as it was used in the testing department of the American Blower Co. it has been called the A.B.C. tube to some extent. For further instructions regarding the tak- ing of readings etc. see Mr. Treat's paper. For convenience, a brief statement of the main facts and essential operations will be given here for though not original with this thesis they are a necessary prelude to what will follow and stating them here will be better than if some reference such as the above, had to be consulted. The inner tube at the tip of the Pitot tube, which connects with the back tube on the stem or shank, reg- isters the combined velocity and static pressures, This combination of pressures is variously called "Dynamic Pres- sure", “Impact Pressure", and "Total Pressure®, The first term is apt enough whan applied to the blowing side of the fan, but not when the reading is on the suction side. The second term is erroneous except when there is zero static pressure, in which case it is all velocity pressure. The third term "total pressure® suits very well if it_be. borne in mind that on the suction side of a fan, the total head will be numerically less than the static head, since both are suctions, and that their algebraic sum but numerical difference will be a positive velocity head, just as on the blowing side. So the pressure indicated by the rear tube will be called "Total Pressure" (T.P.) or perhaps, if it is a blowing test, “Dynamic Pressure" (D.P.). The other is "Static Pressure" (S.P.). Strictly, these should be called "heads" not "pressures", since they are expressed in inches of water. However, the habit is established and the slight lack of technicality need cause no confusion, In a blowing test the static pressure represents the frictional and other resistances in the distributing system between the Pitot tube and the atmosphere. In a suction test it represents similar resistances in the collecting system, plus the suction necessary to produce the velocity existing in the test pipe when the Pitot tube is inserted. These facts are illustrated by the following diagrams. ) - Ft = <> > | | | | ee oi X i) i} 1] P 44 11] sel W-lieds Fb SU VP bp Hl Ty. =e % % ; Exhawsting. Blowing. The location of the tip in taking readings is important It should be held parallel to the axis of the pipe, which is presumably parallel to the air flow. Some observers so place the tip as to take velocity readings at various points and then plot upon a cross-sectional drawing of the pipe, to scale,these points with their respective pressures. Then as if drawing a contour map, they mark points of intermedi- ate pressures on the section, by interpolation. Points of equal pressure are then joined by contour lines. The areas between these lines are each assumed to have a uniform pres- sure which is the mean of the pressures represented by its boundary lines, The areas are found by planimeter and each area is multiplied by its assumed average pressure, The sum of these products, divided by the total area, is the average pressure. It is better to find the velocities for each area and use them instead of thepressures, the result being the average velocity. A more accurate method is to divide the cross section of the pipe into a number of concentric (imaginary ) rings of equal area, and take readings on alternate circles which form the boundaries of these rings, assuming that the ob- served velocity applies to the ring within and the ring without the circle upon which it was read, Since all such areas are equal, the averaging is simplified into a mere averaging of observed velocities. Twenty such rings, with ten readings is customary. It has been found sufficient and necessary to take readings in a horizontal and ina vertical line though the axis of the horizontal test pipe. If the corresponding horizontal and vertical readings dife- fer, the velocities should be averaged for that ring. The error of averaging pressures instead of averaging velocities may be expressed as follows. Let a m= area of each unit ring. Let p,= pressure for each area, if pressures are av- eraged, Let p= pressure for each area if velocities are av- eraged., Let R-=s average pressure if pressures are averaged. Let v= average velocity if pressures are averaged. Let V2’ average velocity if velocities are averaged. Let K =a velocity constant which times the square root of a pressure will give the corresponding | velocity. Let N = the number of unit rings i.e. the number of readings, usually ten. lf pressures are averaged, 2B p and KRY, (/) or '¥ velocities are averaged, 2A = Flea. neta) fa) fom}, U = K fB4 - (VZP (3} W- VY = ACA! _ KY ZB (4) The ratio of difference fo yrs a Mi _ /- Aly (5) JD. ince rhe [PCE SSUES (2 A000 2 are The SAME (1) COV es- ondirg areas, 7he ltrms 17) she summanor are égua/ €ach le each, Therefore, Za), UZIN 1g) fal) the pn OB or on lf all i€ p-Terms are eguaal, the above becomes [TOE pte P "PO and there is no difference between the first and the second method, as stands to reason, for the pressure would then be 9. bobo uniform all over the cross section of the pipe\dnd any one pressure reading would equal the average of all readings, hence the velocity due to average pressure would be equal to the average of velocities. | But if there is any inequality in the readings the nu- merator VY (if p)n becomes greater than the denominator aN’ >D by an amount varying with the inequality of readings and the number of readings. AN Thus the average of pressures always gives results too large. The error is usually within 2 per cent inpractice, however, There is still another method of calculating average velocity. A velocity pressure reading is taken at the cen« ter of the pipe and the corresponding velocity computed. This center velocity is multiplied by a factor to give the average velocity. . This method is proper if the factor be correct. For the Taylor tube a factor of .91 is used, However, it appears that this factor varies for different sizes of pipe and velocities of air. Also it gives results somewhat small even if applied to the Taylor tube, with or- dinary sizes of ducts. If applied to the A.B.C. tube it gives results too large in any case, since the Taylor tube as usually made gives a higher static and therefore a lower velocity reading than the A.B.C. tube. The factor .91 was derived for use with the Taylor tube. (ad) Static Pressures are observed at the same time as velocity pressures, and simiaarly averaged. After reading the pressures in inches of gasoline re- duce them to inches of water by multiplying by the specific gravity of the gasoline, (e) Brake Horsepower is best recorded by a transmis- sion dynamometer. Following is a drawing and description with analysis of the type of transmission dynamometer used at the American Blower Co.'s laboratory, and there styled *"Treat's Dynamometer" (See blueprint page inserted follow- ing this. ) ° Belt and journal friction can be observed by removing the fan wheel from its shaft and tunning bare shaft at test speeds, If electric motor drive is used, a test can be run on the motor by means of a Prony brake in connection with a watt meter, and the effeciency of themotor determined at various loads, In the fan test, the watt meter readings multiplied by the corresponding motor efficiency will give the horsepower output of the motor,(when reduced from watts to horsepower. ) Another good method is to support the motor at two points in its shaft axis, or practically, on small well oil- ead trunnions in line with the shaft. A scale beam is ate tached to the motor frame or its support and provision made that both belts shall run horizongally at themotor pulley. (See blue print page inserted.) By summation of =mononts about the shaft center, (P-p) x r & Wa or Pep c= . Also (B-p) X belt speed = foot lbs. per minute. Of course the motor and all connecting parts are first bal- anced, with the belt off, and the weight W at zero on the scale beam. " Hy ka(P-p) =WerGob aes 3 We+G2O ; rae at mg), KT SE N= FRU LLL OLS ALLL Lilt Gd he 2 : ye erer ee 7 EMAL OL ELLA C4+G = Los weighed , he above ay be siimplitied 11 pracsice. see oe nce SaaS = oa ee on a | a a FP pen | me ° a a) Taj MOTOR LDYNAMOME TER. IWF Pye. Oas3S 7 & LO. Look out for belt slip. ‘(f) Barometer Reading. This may be obtained from a U.8. Weather Bureau station if mear-by, or it may be taken from an ordinary barometer accurately enough. hugr- (eg) Relative Humidity is determined by means of a bare ometer, whose readings referred to a table, show the rela- tive humidity and the absolute humidity of the air. The barometer and hygrometer readings are for the pur- pose of computing the density of the air handled by the fan. This density can be obtained from tables such as those in “Instructions for Calculating and Testing Ventilating Sys- tems" which is Appendix 8 of General Specifications issued by the U.8&. Gov't printing office. Unless the relative humidity is 100%, which never oc- currs in practice, the water vapor mixed with the air is not saturated vapor, but superheated, which complicates any formula which might he used instead of the tables, Howeve- er, the following explanation with formulae developed and example, will enable the observer to compute the air density, tables lacking, Steam tables will be the only necessary reference, Considering the atmosphere a mixture of a gas and a vapor, Dalton's law applies, i.e. the pressure of a unit volume of the mixture (barometer reading) ie the sum of the pressures of a unit volume of air and of a unit vole ume of vapor at the same temperature. Thus if Pm =e unit pressure of 1 cu. ft. of mixture in lbs. per sq. in. and Pa = unit pressure of 1 cu ft. of odoxtesmondk dry air in lbs per sq. in. and Pv = unit pressure of 1 cu ft of vapor in lbs. per sq in. then Pm sa Pa 4+ Pv and Pa = Pm .. Pv (equation (a) ) Now whatever the relative humidity observed, it is necessary to assume 100%, or saturation, as a starter. Then Pm being observed from the Barometer, Pv read from the tables for saturated steam for the observed temperature, Pa may be found from equation (a). Remember that neither Pa nor Pv is the observed barom- eter reading (reduced to lbs. per sq. in.). . Each is less than that but their sum Pm equals that. Note that Pv He- pends not at all on Pa or Pm, but is the independent varia- ble of the three. Pv does depend only upon the fact# of saturation, (As now assumed) and upon the temperature. In a cubic foot of this saturated mixture there is NOT a cubic foot of dry air, plus a cubic foot of saturated vapor, as is sometimes stated. Fach possesses its own portion of the volume, at the common pressure Pm, but if we could from one cubic foot of mixture sepatate the water vapor (which occupies a fraction of a cubic foot, and is under a pressure Pm) and let it expand to one cubic foot at a constant (the abserved) temperature, then it would expand under pressure Pm to a pressure Py. The same is true for the air in the mixture which would expand to a pressure Pa. Then in order that both these cubic feet of gases may be.packed back into one cubic foot each must be compressed isothermally back to Pn. . li. So Pa and Pv are not actual pressures at all, but im- aginary ones, which would occur if air or saturated vapor in turn were freed from its forced companion and left free to fill the cubic foot at the same temperature. The amount (weight) of saturated vapor in a cubic foot of mixture is, for a given temperature, constant regardless of pressure. If this pressure stays constant with a rise of temperature some increase of saturated vapor crowds some other weight of air out of the cubic foot or if not, then the pressure rises, but the same increase of saturated va- por weight crowds in just the same, squeezing itself and the air in so doing. The above remarks are in the hope of killing geome common errogs and’ “make plainer what follows. The weight of a cubic foot of saturated vapor which has a pressure Pv at the given temperature can be found in the steam tables. Call it Wv. Remember that when this cubic foot is squeezed in with the air it still has the same weight. The welghy.of t ater opgetion :can be had from this for- la W Vo mu a 459.2 + temperature. Then the weight of the mixture (per cubic foot) i.e. its density Wn — Wv Wa. But sup»vose the hyeroneter shows the migture to be only partially saturated, as will always be the ease in all prob- ability. Then Pv is proportional to the per cent of sat- uration, Find Pv from the steam tables as before. Pao a Pm % Pv Weight of vapor = % Wv. Wao 1.5255 x Pag x 2.05 459.2 4 temperature. Wm =x Wao yy. Example = Barometer 29" hygrometer wet bulb is 56°: dry bulb 70° = air temperature, = temperature of the mixture, ws vapor temperature. . Find Wt per cubic ft (density) of the atmospheric mix- ure, From humidity chart (Kent's Handbook) this is 40 per cent relative humidity. From Peabody's Steam tables ™ PV w .5602 lbs. per sq. in. 40% Pv = .144 lbs. per sq in. Wve .001130 lbs. per cu faot. 40% Wv = .000452 lbs.- per cu ft. Pm = 29 x .4925 = 14.282 lbs. per sq in. (reducing bar ometer inches to lbs. per sq in. Paces Tesi Pv Ss 14.282 = 44 = 14. 158 lbs. per sq in x x Wm = Way + %£ Wom .07189 + .000452 = .07235 lbs. per cu ft, density. 12. This checks with the government tables, but if Davis and Marks! steam tables are used, it will not check. 4. Principdés Underlying Applicability of Test Data. There are two main facts. First, Similar Fans produce similar results. This means that all fans of a certain type give results proportional to their sizes, if all the fan parts and di- mensions are proportional to their respective fan sizes, yee that all are run at the same peripheral velocity, P.V. e Thus, if a #6 fan with a 3 ft. wheel 18" wide in its own standard housing delivers 7950 CFM at a total pressure of 1/2 oz. per sq. in. at a P.V. of 2520 ft. per min., then at the same P.V. a #12 fan with a 6 ft. wheel, 3 ft. wide in a housing twice as large in each dimension will have the same outlet velocity, the same total pressure, and the same mechanical efficiency. Of course the outlet will be four times as large so at the same velocity the CFM will be four times as great, and so will be the horsepower, (The above statements are true for the usual sizes. Strictly, however, the larger fans of some designs show bet- ter efficiency than the smaller. One reason for this may be that frictional and eddy losses are proportionally great- er for the smaller fans. ) This principle of similarity makes possible the appli- cation of a test fan's performance to similar fans of other sizes. Note that the word "similar" is used in a geometri- cal sense. The truth of this law is well demonstrated by experience. Second, For a certain given condition or "ratio of opening* for any given type of fan there is a relation among the es- sential fan characteristics,a ratio each to each, which is peculiar to, and constant for that ratio of opening and is not altered by any variations of these characteristics, if the “ratio of opening®™ be enlarged. The essential characteristics are :- Peripheral Velocity Pressure P.V.«P. Outlet or Inlet Velocity Pressure —P. or V. Static Pressure P. or &. Dynamic or Total Pressure —P. or T.P. = D. or T. Mechanical Efficiency MB. or E. In a suction (exhausting) test sometimes the terms Friction Head or Pressure replaces Static,and Static corresponds to Total, since the static is numerically larger than Velocity or Friction, The usefulness of the above principie of constant ratios is in the fact that if any two of the above characteristics can be determined in a practisal: case, their ratio is a key to the equivalent ratio of opening on the test fan, and all the other characteristics bear certain ratios to the two known character istics, which ratios are_established on the test fan for that ratio of opening. Thus, the performance of the given fan can be computed. See later under heading 7. There are certain relations between speeds, pressures and power for a given ratio of opening, which are due to this fixity of the characteristics! ratios. ~ or P, YaAt | 48 reeieeteiatedt Ot oe see pls se tantn® J ane ae JP bhi wwnwwwoinwhwnyny 3 a ee From the tabulated results the characteristic curves are @rawn as here displayed. BLOWING Og As aan Ada ide fet KERN 93? yy ae ys F FADOLE WHEEL FAN Aaya Ul F fA nop 7 V4 A My ap tos Ma FOYT Similarly for any type of fan,characteristic curves can be drawn from test data. 7. Peculiar Usefulness of Characteristic Curves. The peculiar value of such a chart lies in tne fact that from it can be computed with practical exactness the per= formance of any size of this type of fan for any given conditions within the fans capacity. Example:= Required to deliver 10000 c.f.m. S.P. 1 i.e, against a resistance equivalent to 1" water gauge. Temperature 67 degrees F. Barometer 29.92 Air very dry. Steel Plate Fan Blowing. Solution :=- The velocity constant = 4000 ft.per minute. (This is from the formula V = V2 ch. V = Velocity in ft. per second. 26 = 64.4 hs head of air equivalent to W.G. in feet. Density of water 62.5 Density of air =p, Therefore h 2 WG x 62.5 12 D Ha 19. V = square root of (64.4 x 62.5 x WG ) 12 D If WGs 1” for a velocity constant of unit value, V 2 square root of 335 - 18.3 D . VD Velocity in ft. per minute » _60 x 18.5 = /O77 = AC D . VD In this case K s 1097 - 4000 ). Vl0755 For ordinary purposes with elevations up to 1000 ft. above sea level, the humidity may be neglected, so the fol- lowing table may be used, without corrections if merely practical working results are desired. Velocity Constants for Dry Air at 29.92" Barometer. Temp. Const. Temp. Const. Temp. Const. 0 3734 110 4161 240 4604 10 3776 120 4192 260 4669 20 3817 130 4232 280 4734 50 3855 140 4263 500 4797 40 5892 150 4303 3550 4956 50 3931 160 4333 400 5081 60 3971 170 4374 450 5248 70 4007 180 4403 500 - §390 80 4045 190 4441 550 5830 90 4081 200 4471 600 5664 100 4119 220 4538 Now to select the size of fan to use. Since the fan is blowing its outlet fixes the VP. There follows a table of commerical sizes of this type of fan, with outlet areas. It would be feasible to s@lect sizes at random, calculating theperformance of each, and finally choose the most desirable, Or, a certain outlet velocity (e.g. 2750 ft. for quietness) might .be chosen, thus approximating the outlet size. But the”“éxperienced will find use for a table such as the following which is based on an average working con- dition or ratio of opening. 20. Speeds, Capacities and Horse-Powers at Varying Pressures _ ot . | : : | : 7 | | .FF | Fan Diam. Static | ae ” ar! yo 4e?.| No. ! Wheel — Press. mw" 1 lye" 2 | 2% 3” | 3%” | 4” ” .F.M.) 3840 5425-6640 = 7650-8595! 9400-10110 10810 P.M. 471.665. 8169451060, 115012501330 LP. 88 2.48 4.55 7.00 9.81) 12.85 16.20 19.75 2.425 60 30 M. 5475, 7740 »=—- 9460-10900. 12250, 13400 14410 15420 Ms. 393 555 681 786 880. 961 1040 1110 .P.” 1,95 3.53 6.49 9.94 14.00 18.35 23.10 28.10 oe J.a/ 60 = 36 nm | wRO my oom | or a \ a wmo|emO| prOl ero] wr M., 7100 10020 12280 14150 15900 17400. 18700. 20010 M. | 336 + 475 583° 675 (90 825 890 950 4./6\| 70 42 P. 1.62 4.58 8.35 12.93, 18.19, 23.80 29.90 36.60 M., 8640 12200 14950 17200, 19350! 21150 22800 24350 M. 204 416 511,590,660, 722780” B30 P| 1.97 5.57 10.20 15.71) 22.10! 28.90 36.50, 44.50 M. 11000 15540 19000 21900 24600} 26950 29000. 31000 M. 262 370 454 525 587 641 693 749 P. 2.52 7.08 13.00 20.00 28.10 36.85 46.40 56.50 5.06\| 80 437 mus | mv 6476| 9 £x” | —_———————_. M. 14050 19850 24300 28000-31450 34400 37000 39600 M. 236 333 409 473 520 578 625 665 _P. 3.21 9.05 16.65 25.60 35.95 47.10 59.10, 72.30 | \ M. 16600 23500 28800-33100 37200-40700 43800 46900 M. 214 308)sB371)~s 4B0——«iB0— eH C85BCWGSOS P.. 3.80 10.75, 19.70 30.25, 42.50 55.60. 70.00; 85.60 G20) 100 60". mos | mo 9.76| 110 667 \ I . C.F.M. 20300 28700 35100 40500 45500 49700. 53500' 57300 J/GS | 120 722) R.PLM. 196-278 BHO BOE HOB 520,555 B.H.P. 4.64 13.10 24.00 37.00 52.00 68.00, 85.50, 104.50 C.F.M. 27400 38700) 47400 54500-61300! 67000! 72200, 77250 /6.00\ 140 84° R. P.M. 168 238 22 337 378 41:3, 445 475 B.H.P. 6.25 17.75 32.40 49.80 70.00, 91.70 115.20 140.9 | | —__ FM. 34500 48900 59800 68900-77300 84500-91000 97500 P.M. 147 208. 256 296 331 362 390-416 H.P. 7.88 22.30) 41.00) 62.90 88.40) 115.5 145.4. 178.0 2O0.20\ 160 06" eee eee eee — samen ~~ ace C. R. B. C.F. MM... 42600 60300 73800 85000 95500 104300 112500: 120000 KG{00\| 180 U8" R.P.M.! 131 185207 262 293/320 3463389 B.H.P. 9.75 27.55 50.50 77.60 109.0 143.0 180.0 219.0 _ C.F.M.) 51600 = 73000 $9400 103000 115700: 126500 136100 145800 3O0485\| 200 120" R.P.M. 118 166 04 236 8H 812! 339 B.H.P. 11.8 33.30 61.20 93.50 132.1 173.0 217.50 266.0 C.F. M. 61400 86800 106000 122200 137400, 150200 162000! 173000 36.00\ 220 1327 R.P.M. 107 151 185 214 240-262 283-302 B.H.P. i 14.0 39.60 72.50 111.50 157.0, 206.0 259.0 316.0 . C.F. M.. 72000 101800, 124500 143500 161000! 176000 189500! 203000 4f.£0\| 240 144. RL P.M. 98 139 170 197 220: 241 260, 377 B.H.P. 16.5 46.50 85.00 131.00, 184.0) 241.0 303.0 370.5 NOTE—Any of the above fans. when running at the speed and pressure indicated, will deliver the volume of air and require no more power than given in the table. ; From the table a #70 fan gives nearly the required c.f.m. at the given static pressure. This fan has an outlet area of 4.16 sq.ft. 10000 = 416 = 2400 ft. velocity. 24aooy" es ecb" Vere 1®° = 8.P,. 2 a6" VePe +> 1* 8.P. o 4000 1.56" D.P. vP SOG DP.” 1°36 265 On the curve of VP & p.P, dind tle “ratio of pressures®. .265. This is at 70% ratio ot opening. 21, At this ratio of opening VP -# PVP m= .215. Therefore, PVP = .36 divided by .215 = 1.675". 4000 V 1.675 »w 5175 ft. per min. =P.V. Tne fan wheel is 42° in ddameter. Therefore this PV means 470 RPM ° 10000x 1.36 The Brake HP is 50 x ME MB = 47.5 % for 70% ratio of opening. Brake Horsepower (B.H.P.) m 4.5 These results are seen to agree very closely with the table which is fact is made up for 70% ratio of opening which represents average pressure ratio for heating and ventilating work for this type of fan. (The horsepower formula is thus derived:- Theoretical H.P. =x ft.lbds. = cfm x lbs. per sq. ft. ~ 33000 33000 1" WG - 1/12 x 62.5 lbs. per sq. ft. = 5.2 lbs. per sq. ft. Theoretical H.P. = cfm x 5.2 x t.p. — cfm x DP blowing. SOOO 00 6350 Actual HP (blowing) = cfm x DP ) ° 6350 x ME Suppose now the #70 fan proves to be more expensive than one a competitor is offering. Find out what a smaller fan would do. Problem:- Find the performance of a #60 paddle wheel fan at 10000 cfm 1" SP and 67 degrees F., Solution: Outlet = 3.2sq. ft. Veloctty «= 10000 # 3.2m 3130 ft. per min. VP = (3130 2 " " (50007 - .61 DP = 1.61". VP = .379 80% ratio of opening. A tabulation is convenient. Characteristics Ratios Pressures Results. PVP 1.90 2.37" 4000 V3.3% = 6160" per minW SP 42 1.00" 36" Wheel = 652 RPM VP 26 61" DP 68 1.61" pup = 3,61 x 10009, 6.04 6050 x .42 ME .42 eX Re So this smaller fan would have to run at 652 rpm re- quiring 6.04 brake HP. It would have been impossible to get this information from the printed table, hence thepeculiar usefulness of the characteristic curves. The speed and HP. of any size fan could thus be fig- ured for the given cfm and SP. The operations are performed on a slide rule but een so they consume considerable time. There is a shorter way, sufficiently accurate. (8) Derived Curves (Performance Curves) If in the above examples we divide cfm and HP by out- let area we get results per sq. ft. of outlet. If these results be multiplied by the outlet area of any other fan at the same ratio of opening the performance of that fan can be determined, Thus a fan having twice the outlet area of a No. 60,would if running at a PV of 6169 feet, deliver 20000 cfm against 1° SP. and require 12.08 Brake HP. Its outlet velocity would be the same, i.e. 3180 ft. therefore its VP its DP and a would all equal those of the No. 60 Fan. Therefore,it would be working at the same ratio of opening. The converse is also true. If then, we can construct curves giving outlet velocity (i.e. cfm per sq ft of outlet) and HP per sq ft. outlet, in terms of PV and SP we can compute fan performance by the following simple calculations. Unit HP x Outlet area = B.H.P, Velocity x *® « = cfm PV ¢# wheel circumference = RPI. Such a curve chart must adhere to one velocity constant but that can be corrected for as it affects only RPM. Since for any ratio of opening the pressure ratios are fixed and interdependent it follows that if one pressure be given (as 1" SP) then each of the other pressures and the ME can have but one value peculiar to that ratio of op- ening. Therefore the velocities represented by these other pressures, and the Unit HP dependent on the outlet velocity, and the ME, can have but one value each for that ratio of opening. Therefore, lines representing the assumed pressures, the velocities, and the unit HP must intersect in one point upon a line representing a certain ratio of opening. Conversely then, all points on a line representing a ratio of opening would indicate simultaneous values of the other variables which would bear a set of constant ratios each to each and if one point be chosen it would fix the actual values of all those variables, Therefore, whatever variables are chosen to occupy the co-ordinates, the lines of ratio“ opening will be straight slanting lines, each making a different angle with the co- ordinates. (since the pressure ratios are constant for each opening, but different for different openings. ). So this Performance Curve chart is to be made thus:- ZS Figure for each ratio of opening the ratio of P.V. to outlet velocity. Then with P.V. and outlet velocity as co- ordinates draw the lines of "“opening®. Then on these "“open- ing® lines plot the points of static pressures and of unit horsepewers as derived from the Characteristic Curves, For example:-= The 80% ratio of,opening line has a slope of 3130 = outlet vel 6160 e es The 1" static pressure occurs at the intersection of 3130 and 6160. The 2" static pressurefor this same ratio of open- ing is at the intersection of 3130y7 3 and @60\"Z etc. The unit horsepower needs further derivation to get it into a simple form for plotting. BHP ow cfm x DP 6350 x ME Unit BHP = outlet vel. x DP 6550 x ME Call outlet velocity - y bhp es v x DP 6350 x LE But v= PV xv PV and DP = DP PVP x SVP Therefore bhp = PV x PVP xv_ x DP 6050 x MB PV Fp But, PVP = (PV) 2 (4000)” Therefore, bhp = (PV)® x v x DP (1016)(10)8ame = PY SPP and Pv [ bhp x 1016 (10)® x ype 1/3 _v¥ xX _DP PV PVP The values of the dependent variables and the PV for one H.P. per sq.ft. of outlet for different ratios of opening are shown in tabular fpgem thus. it 24. —_— - | | re a ro - woes WFalto Qrenc 10 f, £6 30. SO | IO E60 | SC §0 | FO \7oOK | Me 255 WS 51 515 \5Y3\. S97 | 42 | SS | 2H BE OM WIS ARS AU 3440 | HSE | 1510 | 56S |, 608 tf L21b | 1263 AES: 1135 \ 102 | 205'\. 795\ 660 IIS 13,70 iso} iF WBS GEG sud gu" i IEF es i131 28 Rs vse s: ! [abore[$ 6500 4/80\5]20" 5520: D550 se20 S0EC\ YES YE V0 : 75S. The last line gives the PV for 1 H.P. per sq.ft. For 2H.P. per sq.ft. outlet, multiply by the cube root of 2 . For 3} «6 oe " @ L « @ * 0 @ of 3. etc. The completed curve chart follows: Jee perl PAGE. CVE T (MOR WITT Hp WETS fae L218, MAME Ge Alii a) Za) eae ELIE LAL ALIEN AO a ae Pa rs MCE EL EEA a) yj OEE IMMA EME LL / PI LA i | i a 4 PALO) 26. Problem for Performance Curves. Wanted, to deliver 6000 cfm at a peripheral velocity of not over 4000 feet per minute against 3/4" static pressure, water gauge. Dry air about 67 degrees. Find fan size, speed and brake HP. Solution: Notice on performance chart that the less the BV for 3.4" SP, the less the OV (outlet velocity) and hence the larger the fan. Usually in competition we would select the smallest fan that would fill the specifications, hence keep- ing the price down, regardless of HP. So, assume the high- est PV allowable, i.e. 4000 feet. Its intersection with the 3/4" SP lire shows an OV of 1625 feet per minute. 6000 ;4 1625 wm 6.7 sq. ft. fan outlet area. The commercial size next larger has 4.17 sq. ft. 6000 24.17 sz 1440 ft. actual OV. Following 1440 OV to the SP 3.4" line, we find 3800 ft PV and about .375 BHP. The wheel diameter is 34 ft., the circumference is 11 ft. 5800 #¢ 11 = 345 rpm. It is best to make a liberal allowance for the motor whiéh is to drive the fan. While a 1 3/4" HP motor would do here, a 2 HP motor would be safer. Allow about 10% for direct con- nected rigs and 15% for belt driven ones. The main reason for this is that if the resistance has been over-estimated, the fan will handle too much air at that speed, and overload the motor, while if the resistance (S.P.) has been under-esti- mated, then enough air will not be delivered at that speéd, and a speeding up of the fan will be required, taking still more horsepower than in the first contingency. Get the motor too big, unless bidding in close competition under strict spec- ifications. In that case the responsibility is with the en- gineer. A larger fan would take less horsepower. This is often a good talking point, and may sell the larger fan at a better price if power is expensive. Fan performances at different temperatures. Take theproblem first given; 10,000 cfm 1" SP. Assume that the temperature is 300 degrees, so that the ve- locity constant is practically 4800 (see table of velocity constants). Assume a # 60 fan. Find performance. The outlet and the OV are the same as before, i.e. 5.2 sq. ft. and 3130 feet per minute, but now the VP is *} 27 « 231307 : VP = .298 (Zs00) = «3 425°" The DP is 1.425 and Zz: This is 73% open. Note how the ratio of opening has drop- ped with the decrease in density of the air, just as would have been the case if the cfm had been decreased, or the S.P. in- creased. Tabulating for 73% open. PVP 1. 1.875 ' 4800 V-ITB7S = 6570' PV — 700 Yom SP .536 1. VP .227 .425 DP .76 1.425 BHP = 1.425 x 10000 _ 4.88 E le 6350 = .46 Any such case can be solved very readily with the char- acteristic curves. Now to use the Performance Curves for it. It will be remembered that the Performance Curves were made out for a velocity constant of 4000 ft whereas we now wish to use 4800 ft. To get the proper corrections, study what happened to the characteristics. First, the S.P. was given, eo that remained constant in the two problems. Next, the V P was changed in the inverse ratio of the vel congtants, squared, thus for 67° yv P = .61" and K = 4000' Bor 500° V P =.425" and K = 4800', i.e. .61" x {4000}; = 425" 4800)“ Note that if the temp. were 67° and K = 4000 thena V P of .425" would represent an outlet vel.of 4000 2425 = 2610 ft per min. 0O.V. Notice that 2610 _ 4000. so that the V P for 500° is : 3130 4800 ° the same as if the temperature were to remain at 67 , but the outlet velocity were to be reduced in proportion to the ve- locity constant. Likewise all the other characteristic& are just nthe same as if the problem had been for 2610 ft. O V i.e. 8353 cfm at 67. So the P V P would have been just the same and as would the W.E. But notice that VP ve has to be multiplied by 4800 to get P V , and that tne horsepower is likewise pro®& portional to the cfm. So the following rule is applied to the Performance Curves, for a K other than 4000. Divide the required cfm by the ratio of the new K to 4000. From the Performance Curves find the P V and the BH P for this new cfm, the quotient. lultiply this P V and this B H P by the ratio of the new K to 4000. The results will be the correct P V and BHP Example. 10000 cfm 1" SP. K = 4800. Wo.60 fan. Find RPM and BHP. Solution. 4800 . 1.2 4000 ~ 10000 ¢ 1.2 = 8333 cfm 43.2 = 2610 ft O V (fictitious) For this OV and 1" SP the PV on chart equals about 5450. 59450 x 1.2 — 6550 ft PV 6550 + 9.4 = 697 RPM, which checks 700 very closely. The brake HP = 1.3 as nearly as can be estimated. 1.5 x 3.2 x 1.2 = 4.95 BHP, not quite so close aw a check, This shows the need of more BPH lines, for the spaces are not proportional to their value, which makes interpolation difficult. , Now a brief consideration of the Berformance of a Fan when Bxhausting. It is not the purpose of this thesis to reproduce characteristic curves for all types of fans, blowing and exhausting, but to show enough to illustrate their prep- aration and useful application. The characteristic curves for the Paddle Wheel fan, exhausting, will not be shown, but the readings will be recited as needed, Take the case of the number 60 fan which blew 10000 cfm againct 1° SP at 652 RPM and used 6.04 BHP. Suppose the same fan to be exhausting 10000 cfm through a collece- tion system whose frictional amé resistances amounted to 1" W.G. at the fan inlet. Find the RPM and BHP. Solution. The inlet is 415 sq in. — 2.88 sq dt. 10000 # 2.88 = 3470 ft inlet velocity ( IV) K = 4000 VP= 5470 — 75 7" p 1.75 4000] = =e VP_ .428 On the characteristic curve, (not shown) this is 8344 open. PVP 1. 2.408 PV = 6200 = 657 RPIf SP . 415 l. VP .311 275 TP 729 1.75 HP = 1.75 x 10000 — 6,18 Brake € » WHYS" 6550 x 1.445). This shows that when exhausting, the speed is .765 of 1% greater and..the HP is about 2 1/3 % greater than when blowing. This relation is not the same for all ratios of opening, but for ordinary work, with the proper and nece- essary allowance for motive power, it is permissible to add together the inlet and outlet resistances and treat their sum as a static pressure on theoutlet, using the blowing characteristic curves, 29 APPLICATIONS. II, From the foregoing it will be noted that the problem in centrifugal fan application is- first to determine the needed cfm and static pressure and next to apply to these conditions the proper sort and size of fan; the speed and HP then to be determined from the fans charace- teristic curves. l. Héating and Ventilation. (a) How to figure the amount of air required. (bo How to figure the static pressure. °C How to select the sort and size of fan. d How to figure fan performance. The subject af heating and ventilation is so large that many books have been written about it, but without solving satisfactorily most of the problems relating to heating and ventilating vty means of blowers. Even write- ers in the current engineering journals are often serious- ly in error in statements which the experience and standing of the writers should make authoritative. This shows that there is considerable lack of knowl- edge of the subject even among those who write about it, and this thes#s will not attempt in this limited space to deal conclusively with it. In fact, there can be no conclusive dealing with some of the questions involved, for several of these are still being vigorously agitated. We shall endeavor to avoid discussion and confine our efforts to an exposition and explanation of methods which have giv- en undoubtedly successful results when properly executed and followed up by intelligent management of the systen. (a) Amount of Air Required. There are rules, but they must be followed with good judgment. Experience is the teacher, and observation must be digested by an analytical sgnse of the influence each factor has on the result. Most can be learned from failures if properly analyzed. A successful installation is hard to analyze, and to deter- mine how much less it might have been and yet have been en- tirely satisfactory; but a failure advertises its own short- comings and is a guide to betterment. So the following rules, while based on experience, must be considered flexible, They will not make a student a heating and ventilating engineer. He must learn to bend them to fit each case, However they have been successful without great modification in most cases, and that is their justification. 50 Fixst, Heating, with ventilation of secondary impor- tance, The amount of air required for heating is merely the answer to a problem in heat unit exchanges. Air introduced into the room at a high temperature loses heat to the contents and the contours of the room and departs with considerable heat left in its possession. What becomes of that heat and that air afterward does not concern us at all. We need to know how hot the air en- tered the room and how cold it departed. The heat thus released is passed off through the enclosing surfaces of the room, and if the room temperature is to be kept cone stant, heat received equals heat passed off. Heat received = lbs. of air x drop in temperature x specific heat of air. Therefore, Lbs. of air —= heat passed off drop in peacorature” x SP heat of air ~ The heat passed off will be called exposure loss, The drop in temperature of the air supplied will be diffusion The specific heat of air is taken as ,2575 in this work. Diffusion has a value usually fixed by an assumption, It is assumed that the air leaves the room at the average temperature of the room and this is very reasonable assump- tion. The temperature of the entering air is a matter of assumption. It may be assumed directly or may be the re- sult of other fixed conditions. At any rate, it fixes the amount of air for a given exposure loss and room temperature, Exposure loss is a function of the building materials and the difference between inside and outside temperatures, Success has followed the rule that this loss is directly proportional to the temperature difference, The follow- ing table shows the number of heat units per square foot of surface per hour per degree difference in temperature, The values are called Bxposure Constants. They dif- fer in some cases from other published values,but offer the justification of successful application in blower work. Constan$s for Ordinary Brick Walls. Thickness. Constant. Thickness. Constant. 4" - 68 24" «20 a 46 28" 18 12" 235 32" 16 16® ear 36" ~ 15 20* 020 40* ~13 Brick Furred and Plastered, Thickness. Constant Thickness Constant. 4" 28 , 20* 216 8* e200 24" 214 12" e2l | 28" eld 16" 019 32" 212 Sl. Add 1/3 to either of thd above for stone wall, 1/2 for cement. Other Walls. Carrugated iron on wooden or steel bracing «84 Cement Plaster on wire lath 14" to 24" thick ©615 ° “ “ ” « 2" to 33" * 0492 Outside Walls of Frame Buildings, lath and plaster, and with outside, coverings, as follows:- Ordinary overlapping siding about #" thick 44 Same with paper under eol “ “ and sheathing e200 “ * Sheathing, no paper. - 28 Inside Partitions. Ordinary Studding, Leth and Plaster, one side only «60 " on bam sides 204 Roofs, Slate, Exposed -80 #® over wood —~d0 Iron, exposed | 1.32 " over wood el? Composition ower wood .30 Tiling 3/8" to 1® thick .80 6" hollow tile, 8" concrete and tar and gravel cover .36 aR ® @ 1* a 0 0 " e 40 4*® coficrete, cinder fill * *® ° « ° “ 60 6* # a 0 0 0 0 ] " 10 e 54 Floors, Cement or tile ool Dirt .23 Wood on ground ~10 Wood on cement .08 wood 3/4", no plaster beneeth jaiste 045 ._ 8 lath and plaster * 026 " double, 14" thick no plaster ool 0 " lath & plaster beneath - 18 Glass and Doors, Single window 1.20 Double window 2 56 * skylight 1.50 ® Skylight 62 Monitor 1.55 Wood Door 042 Additions. For north, or windy exposure add 10% If heated by daytime only 10% * * and greatly exposed add 30% e " only at long intervals, add 50% S2 Judgment must be used in applying the additions. If the building is to be heated up in a given time, allow- ance must be made for the extra heat units required. This addition should be the heat necessary fo raise, in the given time, the contents of the room from 30 to the desired tem- perature, not from the outside temperature to that desired inside, This is a matter of experience, if the assumed outside temperature is zero. If more or less than zero, make corresponding change above or below 30°. Such addi- tion is often unnecessary because usually the temperature quickly rises above zero in the morning. Poorly constructed buildings must have an allowance for leakage, termed infiltration. Most of these points are taught only by experience. The usual assumption in localities where the tempera- ture in winter goes as low as 10° below zero quite often in winter nights, is to figure on zero degrees outside, The inside temperatures have been assumed about as follows, with zero outside; Foundries, steel mills, etc. where the labor is vig- oroug, 50° to 55°. Machine shops, carpenter & pattern shops etc. where the workmen are stang ing and moving, but not vigorously much of the time, 60° to 65° . Clothing shops or rooms where the work is sedentary, 70°, Offices are usually figured 70° to 75°, Paint shops, 80° to help painting and drying. Many other special processes besides painting re- quire their own special temperatures. Many also re- quire a definite degree of humidity. It appears that the temperatures may be reduced 5 or even 10 degrees with comfort if the proper humidity be, maintained. The fig- ures given above are for a simple heating system with air entering the rooms at about 120°, with no attempt at hu- midifying. The comfort of workmen, and their healthfulness can be increased by blowing a 1/4" jet of steam into the air, preferably just after it leaves the heater coils. This moistens the air and makes it much softer and pleas- anter. This is not so desirable for an office or drafting room, as there is sometimes a stale smell to the steam. Having estimated the heat unit loss for each room, with proper allowances, settle upon the temperature of the air entering each room, This temperature would be the same for all rooms in a building if no heat were lost from the conducting pipes. Some of the heat thus lost helps a little, but it is not efficiently placed for heating, usually. The allowance commonly made for such loss is 10 degrees per 100 feet. This seems to do fairly well for galvanized iron ducts run along the upper part of a yoom, or for underground ducts under a basement floor. If the ducts run outdoors,. the loss may be twice as great 53. with fair insulation, to very much more if insula- tion is poor. For a larger room such as a foundry, measure the distance in feet, from the fan and heater half way to the farthest outlet measuring along the pipe. Divide the distance by ten, the result will approximate the average temperature loss. 0 Air should not enter any rooms at over 120 wun- less in the case of foundries,erecting shops etc. where the construction is loose and there is considerable ine filtration of outer and more humid air. The difference between the entering air and the room temperature is the Diffusion. Los, air per min.w Exposure per min. Diffusion x .2575 Los per min x specific volume at any temperatures cfm at that temperature. For example, if the Exposure per minute is 10000 H.W. and the indoor temperature is 60° while the enter- ing air is 120 then the lbs, air required is 10000 | 60 x. 700 lbs. per min. At room temp. this is 700 x 13 = 9100 cfm. At outlet temp. " *® 700 x 14.6 = 10200 cfm. If the temperature loss is 10° in the piping, the temperature at the heater is 130°. If the fan draws through the heater, it handles the air at 130° and 700 x 14.85 = 10400 cfm. @ fan eapacity..- If the fan draws air from autdoors and blows through the heater the fan should be figured for the highest out- door temperature at which it will be used. If it recir- culates the air from the room its capactty should be fig- ured at room temperature, ite. 9100 cfm, if it blows thro! the heater. If a test is to be made of the fan's eapacity by means of an anemometer, the readings will be taken at the outlets, and 10200 cfm should be expected; if by " means of a Pitot tube, the readings will be taken in the pipe, and if near the heater, the cfm should be 10400 cfm, Second, Ventilation, with heating necessary but sec- ondary. : Where ventilation is the prime requisite, but the air is also used to warm the rooms the "air change" is decid- ed upon by some other means than the consideration of heating. Usually an arbit#zary air change is assigned, as guided by experience, or a certain number of cfm is provided for each occupant of the roon. There is also a method of figuring ventilation where by a certain percentage of COo is not to be exceeded, as- ne | 34 suming a certain percentage in the entering air. After the cfm has been determined by any of these methods, the necessary temperature of the entering air to keep the room warm can be determined thus:- Diffusion w= Exposure per min =ah78 x Ibe per min, Diffusion flue room temp. = entering temperature. Since this entering temperature would probably be dife- ferent for each room unless all had similar exposures, there would be necessary some means of providing these various temperatures to the entering air. This is accomplished by means of a by-pass under the heater, the fan blowing through by-pass and heater. Thus cold or hot air may be taken into the ducts. This necessitates a double set of ducts for each room, one for cold, one for hot air. Of course these two ducts join into one just after leaving the hot-eand-cold air chamber (called the "plenum chamber") and have suitable mixing dampers whereby the cold or hot air is admitted to the duct as desired, Sometimes where this plenum system will not be con- sidered, it is possible to so trim the figures. that fair average results can be obtained in all rooms, with the same entering temperature for all. Of course line loss must be considered. Such a compromise should result in a proper temper ature for all, with for some a greater cfm than was in- tended, but there should not be a reduction of cfm to any room at the expense of good ventilation, for the sake of keeping the entering temperature up and the total cfm small. The arbitrary assignment of air change should ree- quire the best experience, The following rules have been used with success and so have others, somewhat dif- ferent. Much depends on the method of distribution, on the temperamermt of the occupants, the humidity, etc. Foundries, a change of air every 20 to 30 minutes. When natural ventilation is poor and smoke is plentiful, there must be a more frequent air change, Workshops where workers move around some, such as machine shops, 20 minutes. Workshops such as shoe factories and garment fac- tories, 15 to 20 minutes. Same loft buildings are very crowded, A 10 to 15 minute air change is then neces- sary. Churches,15 to 20 minute. Theatres, where seating capacity is known, 20 to 50 cfm per person, otherwise 15 to 20 minute air change. Schoolrooms 50 cfm per pupil by law in most states, and 15 sq ft floor space per pupil. This is 2 cfm per sq ft of floor. Hospitals, 70 to 100 cfm per patient and 30 cfm per attendant. Drafts must be avoided, t e . ., ewai kh 55 Hotel dining rooms 15 min. Cafes 10 min. Bar rooms and smoking rooms 10 min. Paper mills require special treatment. All rooms but the machine rooms are treated as a heating propo- sition, but the machine rooms must have an 8 min air change if there are no hoods over the machines, and a 12 min. if there are hoods. The hoods are connected to exhaust fans. The supply outlets blow air along the under side of roof, over the machines. This is to pre- vent condensation which otherwise drips on and spoils the paper. Dye houses should have a 5 to 10 min air change, depending on the volume of the room in relation to the steam given off. An exhaust system as well as a sup- ply of hot air is necessary to keep the air even reason- ably clear, in the vat rooms. Other places will be mentioned under Ventilation. (bo) How to figure the Static Pressure required. Since in the test of the fan, Static Pressure was the difference between total pressure and outlet veloce- ity pressure, in practice the same ie true, and we have to estimate the resistances of the system before the fan performance can be determined. There is still much data lacking in regard to pres- sure losses: in pipes, ducts, and various portions of a distributing system. It has been customary to assume that air friction is analagous to water frictinwy and to deduce results for air from experiments with water flow, This has been done because of the difficulty of getting true static readings, perhaps, but with the Pitot tube — herein described, such readings can be had and it is hoped that more data along this line may be collected. The information which follows is the result of tes to a large extent, modified or approved by observations of actual installations for a number of years. Friction in Round Gaivanized Iron Ducts. If K w .0002 and S= inner surface of pipe in sq. ft. and v= air velocity (average) in ft. per second, and @e cross sectional area of pipe in sq. inchgs, Then Frictinn (in inches WG) = KS v a For example, a 12® round G.I. pipe is carrying 1000 cfm. The pipe is 100 ft long. Find the friction. Solution: S= 3.1416 x 100% 314.16 sq. ft. &wmw 113° = .7854 sq ft. v= _1000 21.2 60 x .7854 ™ v = 450 K8v" = )= . 0002_ x 314.16 x 450 _ .252" loss of head. a liz 56. The above constant, .0002 has been questioned at times. A careful laboratory test will probably show a constant about two-thirds as great, for the pipe will be very smooth, round, and straight, but for actual working conditions the constant given comes very close to the true friction. If you ever run a test and get a smaller friction try putting just a few dents into the pipe. It will make a big difference. The accompanying chart is very useful in computing pipe sizes and frictions, cfm and velocities. It is made out for 100 ft lengths, and the friction is proe- portional to the length. See sah poche, for char’. If no chart is at hand, compatation y the formula may be much shortened if the formula be shown in another shape. Thus, a length of forty diameters of straight round G.I. pipe presents a friction equal to the veloci- ty nead. Same example. 100 ft 12" pipe 1000 cfnm. Solution: 100 ft of 12" pipe = 100 diameters Velocity= 0)2 #; .7854 m 1280 ft per min. VP = eee = 1025" If 40 diameters length would present .1025" fric- tion, then 100 diameters present .252". Where elbows occur their resistance can be equated to an equivalent length of straight pipe. Since 90 elbows are usually the sort encountered, the following chart will be helpful. Note twat there is no advane- tage in making the throat radius more than twice the pipe diameter, See bach pocker. If a square elbow is encountered, the loss will be that of 80 diameters of pipe, or twice the velocity pres- sure, Avoid the square elbow. Friction in Rectangular G.I. Ducts. The same formula holds good, but the chart can be used by the aid of a little computation. Thus, aneq- uivalent round duct's diameter can be computed, which will give the same friction per humdred feet at the same velocity as the actual rectangular duct. Let the equivalent diameter — 4d « the sides of the rectangular duct be b and oc, Then dad = 2 bc b+ Example; 100 ft of 8* x 16" G.I. dauct carties 890 cfnm. From the chart for round pipe, find friction. Solution: The velocity is 1000 ft per min. The equiv- @ lent diaméter is 2x8x16 . 10.658 24 - From the chart, the frictfon of a 10.65" pipe at 1000 ft 57. velocity is practically .17" per 100 ft. Derivation of above equation. F= KS8 v“ for both duct and pipe. a F is desired to be the same for both, v is assumed the same for both, and K is the same, for the same material, G.I. Then 81 85 Al~ Ag But for the round pipe 8s] ~ 4 a, ™ d And for the rectangular duct § 2b + 2¢ a= “bo Then if 4. _— Ab +yAc _ a = be ad — 2bec b+c Q.E.D. Elvows in rectangular ducts can be allowed for as equivalent straight lengths of equivalent round pipe. Rémember that the formula above is true only if the velocities are the same, For the same c.f.m. the re- lation is more complicated. Constants for Other Surfaces. For brick or concrete ducts, K= .000565 ‘Tunnels in rock K = .00073. G.I. Pipe very smooth K - .000185 or less « ~~ © commercially smooth K= .0002 " " rivets & lap joints K = .00023 " * with stiffeners K = .00027 K-= .00035 Tile, if well laid, about Other Sources of Resistance, Heaters offer resistance in nearly direct ratio to their depth, for a given kind, and in proportion to the square of the velocity of the air through then. The friction for cast iron Vento sections is given in the Vento data book issued by the American Radiator Co. Some of the values are here inserted for convenience, Regular Section Vento, 5" Spacing. Velocities figured on volumes at 70°, in ft per min. Friction loss in inches W.G. 58. Velocities. 1 section. 2 sec. 3 sec. 4 sec. 5 sec. 6 sec, 800 ~040 2072 - 104 156 - 168 200 1000 065 e115 e165 e215 ° 265 e515 1200 090 ~ 162 0254 «506 0378 ~ 450 1400 0 L22 0220 -518 416 514 ~612 1600 «160 288 © 416 544 0672 ~800 The American Blower Co. "A.B.C." pipe coil heater with vertical 1° pipes set on 2 #" centers presents frictions partly as follows; Thesf four pipes deep to a heater section: Regular section A.B.C. Coils. Velocities are for average volumes, i.e. about 70 F. Velocities. 1 section. 2 sec. 3 sec. 4 sec. 5 sec. 6 sec. 800 ~05 09 2135 18 0°22 » 26 1000 08 o 15 021 28 035 041 1200 e12 021 50 «40 ~ 50 59 1800 2 26 047 «68 ~90 1.11 1.32 Fresh air Intakes, or openings for recirculation should be of such size that the air need not exceed 1000 ft. velocity through them. If these orifices open into a chamber in which the fan sits, so that the air velocity drops great- ly between the air intake and the fan inlet then there mst be added to the pipe friction and heater friction, the ve- locity pressure due to the intake velocity. This is for the reason that a suction must be maintain- ed in the chamber to induce the intake velocity, and if this velocity is lost before it reaches the fan inlet, it is of no help to the fm. So it must be allowed for. The same is true of chambers in the distributing system. If there is a ptenum chamber, it must contain a stat- ic pressure equal to the sum of the greatest duct friction and the cprresponding velocity pressure. To this must be added heater and intake resistances, to get total static. With nicely made smooth nozzles of the proper shape velocity pressures would be convertible into static pres- sures without loss other than friction losses but such noz- zles rarely exist in practice and it is safest to assume, in their absence, that the fan outlet velocity pressure must be added to these other pressures to get the total pressure, If at any point in the system a highér velocity be carried than at the fan outlet, without the velocity loss due to a plenum chamber or other chamber, then the pressure due to this high velocity should be added to the sum of resistances to give a total pressure from which the fan outlet velocity pressure should be subtracted to get the equivalent static pressure for use with the performance Curves. Example: If a fan have an outlet velocity of 2000 ft. while at a point in the pipe from the fan outlet there ey 59. is a velocity of 3000 ft and the pipe friction 48 020", then the total pressure would be .25" 4+ {S5000]" gim, 000] — The 0 V P is 2000) 2 ase 4000) ™ TheS P= .81" -,25" = ,56*, Understand this is equivalent to allowing .31" in addition to the pipe friction for the purpose of boost- ing the velocity. This is a case of conversion of stat- ic to velocity pressure. The system should be so designed as to avoid this. All other velocities should be less than fan outlet (or inlet) velocity. The velocities which are to be maintained in the ducts are usually arbitrarily assigned,with regard to the noise caused by high velocities and to the cost of large ducts, on the other hand. In figuring duct friction it is a common error of the beginner to add the frictions of all branches, Of course the friction of that air route which presents the greatest friction should be the ruling friction. If the branches present less, they will receive more air than designed to, but this is remedied by dampers and déflectors, which equal- ize the friction and divert the air. -In factory heating where noise is no objection, the shorter brarches can be designed for higher friction so as to equalize the flow, This is of course the most econ- omical way. After all resistances are added, there should be a further allowance of about 10% to cover inaccuracies in the computations. Note, however, that this allowance must be met vy a similar allowance in power, for if the resist- ance be less than figured, and the speed of the fan is fixed, either hy specification or by a constant speed mo- tor, then the power will be greater than figured, for more air than required would be handled, — By the aid of the friction chart, many variations of this sort of problem may be solved, . First, velocities and sizes may be assumed for the ducts’ and the friction of the worst run may be figured there- from. Second, a static pressure may be assumed, and the pip- ing designed to fit it. In this case, a guess must be made at elbow sizes in order to allow a certain length of straight pipe for equivalent friction. Third, a certain size of fan and some other charac- teristic, such as horsepower, or P.V. may be assumed and the piping designed to fit. Thus in one instance, a boiler was installed for heat- ing only, and the piping was to be so designed that the engines driving the fans should need to be large enough to furnish all the exhaust steam for heating the air. This was computed by repeated approximations, but it would have been much easier if the friction chart had been in 40. use at that time, See poches in bach forfrichion Chart. Sometimes a specification reads "The fan shall deliver at least 35000 cfm through the system of ducts as shown" (or as installed). This makes it hard for the salesman who has to fig- ure the fan, especially if the cfm or velocities are not marked om the drawings, for each portion of the piping. Such data lacking, it is usual to assume equal ve- locities at all outlets and apportion the total cfm in proportion to outlet areas. This is a hardship which can easily be avoided for the maker of the plans had to figure the cfm for each branch and section and he might as well put it on the drawing. He is paid to figure the system out, and he should always state the estimated static pressure, To omit it is to attempt dodging responsibility. The average blower salesman is not able nor has he time to figure the static pressures. It is an engineer's job. The estimated pressures should be specified, just as is the estimated c.f.m. Another sort of problem sometimes comes up. It is desired to use a certain duct system and through.it to de- liver as much air with a given fan as a given brake horse- power will deliver. This could be worked by "cut and try", but the char- acteristic curves help us to a shorter solution. Assume a cfm and apportion it either by equal outlet velocities or by equalization pressures. The former is usually the method to be used, Then for this assumed cfm compute the static pressure, and figure the ratio of opening and the fan performance for this assumed cfm. Thus a fictitious HP will be found which by comparison with the allotted HP will show the allowable rpm and the possible cfm. It will be remembered that for a certain ratio of op- ening the rpm and cfm vary as the cube root of the HP. Note that for a given distributing system a fan is at the same ratio of opening whether the fan is running slow or fast, delivering much or little air, for the re- sistances vary as the square of velocities, and so does the V P so that the ratio of VP to SP is constant and therefore the ratio of opening is unchanged. | This makes possible the assumption of cfm, and the rest of it. Certain standards with regard to velocities in ducts have come to be set up. They are formed by experience. The idea is to avoid noise in some places, in others to save power which high velocities would waste; sometimes to cheapen first cost by using high velocities, however, Always it is sought to have the outlet velocities low enough to avoid unpleasant drafts, and to have the duot velocities gradually increase from outlets to fan. The following are the maximum velocities recommended. Velocity at Kind of Bldg. Fan Outlet. In Mains. In branches. At Outlets Schools, Chur- ches, Theatres, 25650 ft 1200 600 ft 500 ft & wherever it per min. to 800 ft to 350 ft. must be quiet. Factories about 2500 ft 1500 ft see below 4000 ft 4l. There is more latitute in factory heating than in public building work. The outlet velocities depend on the distance air has to be blown before reaching the floor or the workmen. Velocity through Outlets for Factory Heating. 7 to 11 ft above floor 800 to 1000 ft per min. 12 to 14 ft above floor 1000 to 1100 ft per min 15 to 17 * « ° 1100 to 1300 * *® to 22 * " ad 1600 ft per min. It is not well to try to blow hot air down to the floor from a height greater than about 25 ft. (c) How to Select the Kind and Size of Fan. For this sort of work there are three types of cene- trifugal fans to shoose from, the multibladed, the paddle wheel, and the cone fan, The multibladed fan is true to its name. It has many blades. In general it has a better mechanical efe- ficiency, but as we have seen , thés does not always mean less horsepower. However, it will déliver more air for the same sized wheel at given P V. or HP so where space is scarce or valuable, it is the fan to use, It is best for public building work because it is quiéter, there being a smoother air flow. There are many different trade names for this sort of fan. Here are a few of them, Maker, Name, American Blower Co. Sirocco B. F. Sturtevant Co. . Multivane Buffalo Forge Co. Convidal Mass. Fan Co. Squirrel Cage N.Y. Blower Co Seri-vane, etc. There is a wide varietxy in the performance of these makes, and it is best to demand a guaranty of results with- in reasonable limits, rather than to specify a size. A multibladed fan is usually best for motor drive if the current is purchased from outside, if it is gen- erated in the plant, and the exhaust steam can be utili- zed, there is not so much loss in driving a different type of fan. The paddle wheel fan is well adapted to engine drive and to factory work where the exhaust can be used for heat- ing. It is less efficient, usually takes more power and always more space than the multibladed fan, and is noisier. It is also cheaper, The cone fan should be used mainly for exhausting air. It is not a good blower, It is much like a paddle wheel fan without any casing. It is hung in a chamber from 42. which the ducts radiate, when it is used as a blower. It is cheap but inefficient and it will not overe- come much pressure, It has been used in public buildings where the ducts are large and velocities low. The size of fan to use for public buildings is lime ited by the 2350 ft per min. outlet velocity. Less should be allowed with paddle wheel fans. Also the PV should not exceed 4000 ft per min. Both these requirements are easy to keep within, for multiblade fans, but not for the pad- dle wheel sort, or the cone type, for a given S.P. | The size is further limited by a specified B HP, or by economic considerations of cost of power. The prop- er analysis of every job ought to balance running expense against the interest on first cost, voth of fan and drive er, as well as of heater and ducts. Usually after some experience, two trials will suffice for selecting a size to meet almost any reasonable set of requirements in the most favorable way, (d) How to Figure Fan Performance. This has been well explained under the general head- ing of characteristic curwes and Performance Curves, Re- member to make the proper computations for the temperature or density at which the fan handles the air. 2. Ventilation. Where ventilation only is required, and heat is sup- plied from some other source, the method of application depends largely upon what elements are present, which make the ventilation necessary. Thus some sorts of smoke must be exhausted from the teiling, others from the floor, while in some cases no exhauster is used for smoke, but it is displaced by fresh air blown in. Where ventilation is required because the air is vitia- ted by the breathing of occupants, there are several methods of ventilation. If warm air is used for heating, and an exhaust system is used for aiding ventilation as is often the case in school houses, then the hot air is usually ine troduced 8 ft above the floor, and the vent opening is at the floor line and on the same wall as the hot air inlet. Where direct radiation is the source of heat, sometimes a forced supply of warm air is used for ventilation, with vent openings placed as before, but with no exhaust fan. In this case the air is vented by natural draft and by fan pressure from within the room. Sometimes with direct radiation the air inlets are beneath the radiators. In this case the vent registers should be so placed as to draw the ftesh warmed air across the room, i.e. short cir- cuiting should be prevented, There are some cases where vent openings at both floor and ceiling are desirable. This is true of assembly rooms which are to be ventilated on account of heat in the sum- 43. mer, as well as for vitiated air when the windows are closed in winter. If the heating is done by hot air, the lower vent. registers are to be used when heating is necessary, thus deflecting toward the floor the warm air which would atherwise rise, This keeps the room warmer near the floor, and keeps the fresh air nearer the breath- ing line. In summer , if cool air is introduced through what was the heating system, the upper vent openings should be used, drawing off the hot vitiated air which has risen to the ceiling, and defaecting upward the cool entering aiy which naturally falls to the floor. This prevents drafts and gives better distribution than there would be if the lower vents were used. Theatres should be vented largely from the backs of galleries and balconies, or from the ceilings of same, where pockets of hot vitiated air are fommed,. Some air should ve withdrawn near the floor at the front of the main floor also, to prevent stagnation there, No air should be blown or exhausted behind the cur- tains. ‘It waves the scenery. The air changes before noted apply to exhaust venti- lation also, wxcept where air is used for heating, with forced supply. Then there should be withdrawn about 2/3 as much air as is supplied, This keeps the leakage oute- ward. For toilets and other rooms where an outward leakage would annoy, ventilation should be entirely exhaust. Toilet rooms require a five minute air change. Smoking rooms should have ceiling vents, with equal exhaust and supply. A five to 10 minute air change is used, depending on the expected amount of smoke, Kitchens should have a two to one minute air change. The exhaust should be through hoods over the ranges, boil- ers, and steamers, The dish washers sometimes give off much steam and then should be hooded, A fan exhausting from a kitchen range hood will in time have its blades caked with grease. For this reason a paddle wheel type is better than a multi-blade, for in the latter, the baades might cake nearly full spoiling the suction. THis grease will take fire at times. In- sulate the ducts. | The best way to design a hood for any purpose is to watch where the smoke, steam, or other vapor naturally goes, and then design the hood to catch it. This is true in exhausting materials also. Where steam is to be remewed from a room it is best to catch it at itssource if possible. Otherwise, it is best to admit warm air to equal that withdrawn, for cold air entering will chill and condense the moisture and make a fog. This is the case in dye houses, Wherever acid fumes are to be handled through a fan it must be acid resisting. Wooden fans of few blades have been used, but are inefficient. Fans made of cop- per last quite well. Some of the alloys are longer lived. Monel metal which is néckel 84% and copper 16% is good, 44, but it will not resist chlorine gas. A cheap and good protection is asphalt paint in several coats. If there is not much resistance to flow of air, a blower can be arranged to blow air into the flue above the acid fumes inductng a current by ejector action. Experi- ments by the writer indicate that in an 8" tile duct the induced flow from the lower end will equal the capacity of the blower which has a 2$* outlet. The blower outlet must have a pipe bent in the direction of the desired air flow, inserted into the duct at a sharp angle, the nearer paral- lel to the duct the better. This avoids passing acid fumes through the fan. The fan nozzle should be painted where it is in the flue. Cooling can be accomplished by blowing in outdoor air which is cooler, or by inducing its inward flow by ex- hausting out the hot air. The former produces a more no- ticeable effect on account of the drafts it may cause and is best for local ventilation, 1.e. when some small part of the room is to be cooled. Exhausting is best where the whole room is to be cooled without drafts. A combina- tion of the two is common. The cofling effect can readily be figured on a heat unit basis, reversing the operations of a heating problem. An air washer is highly desirable. See heading 6. 3. Mechanical Draft. While natural draft is much used on account of sup- posed cheapness of operation, there are many cases where a mechanical draft would be better. A thesis could be written on this subject alone. There are two ways to apply mechanical draft. lst. Forced; 2nd. Induced. Forced Draft is the application of air under pressure beneath the fire, It has the following advantages: : The fan is smaller, for less air and no gases are hand - led by it. The power is less for the same reasons. There is no leakage of cold air into the combustion chamber, The fan need not be so heavy nor the bearings be pro- tected from heat. The fan can be made double inlet which is more efficient for the paddle wheel type. It has the following disadvantages: Leakage of smoke and gas into the boiler room. Blast must be shut off when door is opened. Therefore, it is not so well adapted to hand firing. It is comparatively difficult to install on an old job, on account of the separate duct tfequired to each ashpit, usually beneath the flooring. Forced draft is well adapted to automatic stokers, and is usually used with them... It lends itself well to auto- 45. matic regulation. (a). The amount of atr required is a matter of chemis- try, combined with experience. The theoretical amount of air necessary per lb. of combustible is easily figured. Chemically, ordinarily good coal needs about 12 lbs. of air per lb. of coal, but practically much more is needed. There is some loss from the ducts and connections and out of the ashpit, and there is much of the air that does not properly combine. The method of firing makes a difference, Usual as- sumptions are:- Hand firing, chimney draft 24 lbs. air per Lb of coal * mechanical *" 18 *®* " Automatic Stokers * * 15 * " 0 e o 0 a e chimney @ 18 ” oe @ 0 a @ The number of lbs. of coal per boiler H.P, depends on the quality of the coal, but it is true that on the average a better efficiency of the coal is had with mechanical draft because the proper amount of air for complete com- bustion can be constantly maintained, as it is independent of atmosphereic conditions. So there will be less incom- plete combustion on one hand and less chilling of the fires on the other. The amount of coal that can be burned per sq. ft. of grate is greatly increased with the mechanical draft on ac- count of the increased draft. A; formula for draft required is as follows: Rate m= ib of coal per.sq. ft of grate required to be burned per jour. Rate + 1 x .03542 draft needed at fires in inches W.G. [Rate +2} 5.2 The above formula applies fairly well to anthracite coals of size No. 2 and larger, but given too much draft for soft coal. (3): The static pressure is found by adding to the draft required, the resistances of ducts. stack and all air passages. The loss of pressure in water tube boilers - and horizontal tubular boilers is about .3" to .4" W.G. Strictly, the friction is less for hot gases than for cooler air, but it is sufficiently accurate to employ the formula and charts already given. (6) The sort of fan to use is the same as for heat- ing and ventilating, though there is an apportunity to use for forced draft a double inlet fan, thus bettering the efficiency if a paddle wheel fan is used. The size is de- termined from considerations of horsepower, since noise. is no objection. So smaller wheels can be used than for heating work, but at the expense of power. (d) The performance calculations will be shown by an example, Suppose a static pressure of 1° is required, and 18,000 cfm at about 67° average. Use a double inlet paddle wheel fan. Try a No. 90 fan. The outlet is 6.45 sq ft. The O.V. is 2785' and the V.P. is .485". 46 DP w 1.485" VP ~526. On the characteristic curve for double inlet Pins this is equivalent to 75% ratio of opening. PvP 1.00 1.52 4000 V1.52 = 4940' PV SP ~ 655 1.00 Wheel is 4.5* diam. R.P.M.= 350 VP ~320 ~ 485 E - 575 B.H.P. = 18000 x 1.485 _ 7.9 6550 x .975 By referring to the Performance curves it will be seen that for 1" SP and 2785 ft OV the PV is 5400 and the bhp 1.5. This means 381 RPM and 1.5 x 6.45 = 9.67 BHP. A Sirocco fan (many-bladed type) of the same outlet size would have PV = 3300 and bhp 1.1. For this size of outlet the Sirocco wheel would be 3.8 ft diameter. §8o0 the R.P.™. would be 276 RP?. The B.H.P. would be 1.1 x 6.45 = 17.1 BHP. Note that the Sirocco fan has a much smaller wheel, ae would occupy less space, Note also that while the fan handles air only, yet it muet force through the boiler flues and stack the prod- ucts of combustion, so the friction for these parts must be figured on more cfm. Add to the weight of air per min the weight of combustible per min. and the sum, times the specific volume of the gases will be the cfm from which to figure these frictions. Induced draft. Here the fan must handle the products of combustion and the cfm will be that figured just above , for the fric- tion in stack, etc. for forced poratt. The temperatures will be perhaps as high as 880° for which the velocity con- stant is about 5500 ft per min. There should be an allow- ance of 5% to 10% for leakage inward. These induced draft fans should be made of very heavy plate, and the wheels should be overhung if possible; if not, the bearings should be placed beyond the hot gas in- take, and water jacketed. Example:- Assume the same boiler as before. The fan must now handle more gases, The coal burned is 77 lbs per minute, 10% nonscombustible. The air is 1350 lbs. per min. The SP is 1", The temperature is 550°. 90% of 77 is 69 lbs. of gases from the coal per minute. | 1550 * 69 = 1419 lbs. gases #4 10% = 1560 total lbs. gases per min. At 550° the density is .04. Therefore the cfm is 1560 # .04 = 39000 cfm, twice as much as with forced draft Here it will be better to use a single inlet fan. Try a No. 110 paddle wheel. The ottlet is 9.75 sq. ft. The O.V. is 4000 ft. The P.V. +9 (4000) .53. D.P.=.1.53 VP .53_ _—s «346. 5500 DP = 1.53 = 47 On the characteristic curves this is 78% open. PVP 1. 2.22 5550 Vee = 8200 ft. PV. SP -45 1.00 VP 25 .55 54 ft wheel = 475 rpm DP 70 1.55 E 043 BHPe 1.55 x 39000 _ 22.2. 6350 x . ~ Of course a larger fan could be used, with less power. Forge and Cupola Blowers. Forge Blowers are of a different type from the paddle wheel or multiblade. The reason is that more pressure is required in proportion to the amount of air. In forge blower work it is still the custom to speci- fy the performance of the blower, as the total or dynamic pressure required, and the size is chosen according to the number of forges. The size of tuyeres or air outlets should be also stated, Then the proper size of fan oute- let is chosen to correspond to the number and size of tuy- eres, from a manufacturer's table such as will be shown later. Then from another table the size of blower and its speed and capacity for the assumed pressure are found, Such a method is evidently no more than a rough guess, for the total pressure depends not only on the draft, needed, but upon the length and crookedness of the ducts from fan to forges. The pressure asked for varies from 2 oz. to 10 oz. according to the number and size of the forges. The cfia. is rarely stated and is usually not known. One reason for this lack of definiteness is that the blower is usual- ly driven from a line shaft and no one knows how much pow- er it takes, Another reason for not specifying the amount of air is that each forge is provided with a blast gate which is often closed and seldom wide open, so that no one knows just how much air would really be required. The demand is for pressure, and probably there will not be a change to more scientific specification very soon, because the present way works fairly well. Much depends on the area and arrangement of the small air outlets in the nozzle of the tuyeres, | However, it has been found that for a small forge with a 2" tuyere, 90 cfm at a static pressure of 2" at the forge was an ample allowance. Probably this pressure would be sufficient for even heavy and large fires, except for the fact that the black- smiths are used to higher pressures and think them nece- essary. The proper allowance for duct friction must be added. Such a method would save in horsepower, for it would permit the use of a fan of the most favorable size, whereas the old method would fix the fan size by the sum of the tuyere areas. . 48 For comparison of the old and new methods take an ex- ample, The style of fan used will he like that shown in the accompanying cut. NO 9. aT eg peTRorT, MICH 2 Lt is called a "Steel Pressure Blower". 49, Others are cased in cast iron for the same purpose, and called Cast Iron Pressure Blowers. They are comparatively thin, and the wheel is narrower in proportion to its diameter than This is true of all fans built more for pressure than volume. the wheel of any other type. Example: tuyere. them. There are 10 forges, each having a 2" Find the size, power and speed of a blower for Old method: Making a rough allowance for the duct friction, assume 3 oz. total pressure at the blower. From the "table below, the fan outlet size for ten 2" tuyeres is 8". second table). This is the outlet of a #3 blower (see For 3 oz. pressure (total) this would run at 1785 rpm. at 1.47 HP and deliver 789 cfnm. This is less than the cfm at 90 cfm per forge, but they will seldom be working full blast all at once. (see third table) NUMBER OF FORGES DIAMETER oe Tuyere | 1|2/3/4/5]6|7|8| 9/10 —_|— - y” | 1g] 1g] 2 | 2 | 24) 24/3) 3130 3 1 14} 2 | 24] 3. | 3 | 34) 34) 4) 4, 4 1¢ | 2 | 24] 3 | 34] 4] 4] 4d) 5] 5 | 5 1h 2/3 | 34 4 | 43/5 |6)6] 6) 6 1} 23| 34) 4 | 4415/6) 6) 7! 7 | 7 2 3/4] 435/6/7/7/8 {8/8 2} 3/4/5|/61/7/7]/8|9]9)9 2h 34/5] 6/7) 8|8| 9/9 |10 10 23 4|5|6 | 7] 8] 9 {10 |10 11 11 3 4/6|7|8|9 |10 |12 {11 [12 12 3h 4317] 819 ]10 [in [12 [13 [14 14 4 6 | 8 | 9 [it |12 |13 [14 [15 [16 17° Fable 7 y 5 4 5 é 5} te eG wey = x 23 vy |SS)eF7e}] of < = 22) 82 |Ze|S2e] 2] =] = y=a| Ze jee |/Sea| 5} =] = 1 145] 13 6 7h] 224) 5 2 17 13 7 S4] 26 ao} 3 19g | 13 8 Qt} 3] 64 4 22 2 9 11 | 35 74 5 24h | 24 10 124] 39 8 6 27 24 LOG | 3h} 43 s} 7 32 3} 128 | 16 ] Sr) 108 8 37 34 144 | IS3] 59] 11! 9 42 43% 16) | 21 ] 67 | 133 10 47 +4 Sh | 234) 76] 14; 11 22 oF 214 26 | &8O iti Dimensions are in inches. T@D/€ bf S7€el forge Blowers. Capacity Table |: Ounces inches | 3. 46, 5.19 } 2 HP. | 2.6) F 9 7.36 13.5 4030. 4870, 5610) 6850 YQ. 06. 13.90 24.55 = 2 EPCon! = S{ at 1000 1.242 7 = 1Cu. Ft. | | a C.F, 361) | HP. 0.45 UR. PML! 1673 - CLR Fags | HP. 0.62 RPM. 160 CLF, 655 H. P. | O.s2 as I" ML | 1292 C.F. SOS IP. tO. 04 OPM! Lie : Cor. 1040 H.-P. | 1.30 ROD, Mw “1055 6. C.F. 1262 HOP. | 1. 37 ROPM sag 7, C.F. 1705 | HP 22 RPM 769 8) C.F. o 2140 Te Pp. M. | 679 9 C.F. 2S00 H. PL 3.48 RPM. 606) 7 0 CLF 3350) Il. P. 4 7) ROP. vn 548 ioOGLE. HP. 5.03) K P M. 500 12 FE. 4820! ti P, 6.00 10. 78116. 62 15821 1825 Te 3 |_*_|_! loa R. Pp. M om 1960 2400 2770 3390 3915 434500) «G10 70S O.st) 1.24 2.28 3.9] 2050) 2862 2895 3 340 000 ol Sd. 3 OTs L120 1.72 3.15 >, 4.84 1785 2060 2: 520, 2010 789) G10 1] 10) 1286 1.47 2.26 4.15) 6.36 (9935 2585 10060 E162) 1415) 1643 1. So 2.88 “122, 1640 2010) 2220 1250 1442 1760) 2040 10 12 | 14 16 8.92 10.38 13.83.'17.28 20. 75,24. 22 27. 66 e248 4.73 4.956.20 744! 8.69 9 92 _| 3740 4090 1003 1196 6. as: 8.90 3955 3370 8955 4120 M40 1575 170) 1820 P08 TE, 72 14.75 18.05 2890) 3163 3420 3650 IM40 2012 2175 23R5 9.28] 8.14 11.40 14.96 1s. 30. 23.10 055) 2845 3073 3200 2280) 2500 2700 | INS35 2.33) 3. 38 6.97/10.10 14.13 18 60 23.45 28.66 1240 “1490 1825} 2105 2355 2480 2790 2ASQ 1520) 1750 2135) 2475 2770 30383 2280 3500 2.83 4.54 7.96 12.25 17 JOST 1255 1535) 1775 2055 23668 28901 3350 (18 22.60 28.50 a. 70 1083 2170 2345 2510 3750) ALLO 4430 4730 3.83. 5.80 10.78) 16.60 23.25 30.60 38.50 47.00 210 1085 1328 1523 1715 1850 2030 2170 2575 2470 8620) 4200 4700) 5150 5560) 5040 Sol O58" 17 “1355 337 Q 3SSO: 4730 6.27 9. 6317, 65/27 742 a lows} 1210 4025 4610° S660) 6570 C8 TS QE A282 55 45 67 *() aT) O47 1003 en 705 5800, 6700 S116 62180. 45/46. 85) #4 SADIE 3. 20.80 29. 15! 38.338 48.50 OM. OU “1515! 1660 1792 1916 5500 H150 6730 7270 £760 BS. 1), 50.15 63.20 77.00 1352 “1480 1600 «1710 73800) S055 S700 9°00 ».60 60.00 75.60 92.25 1222) 1340) 1447 146 7450 S900; 9750 10520 11220 39.88 55.20 72.50 91.50 111.33 S63 “oon 113 1220 1318 1410 $160) 9460 10. a0, 11600 12520 13380 65. 50! 86. 33 10900 132. 75 50. New method: At 90 cfm each, 1080 cfm would be required . 24" SP at forges. Design the ducts for 1" loss = 3.5% at blower S.P. Choose a #2 blower. Outlet .2485 sq. ft. OV= 4520 VP= 1.13" DP = 4.63" Vf = .243. This is 63% open on characteristic curve not shown. PVP 1. 4.80 PV = 4000 V-t-8 = 28750 SP 673 3.50 Wheel 17" diam = 1970 rpm VP £23 1.13 DP .96 = 4,63 HP = 1080 x 4.63 _ 1.37 E 575 es 6550 x .575 = When some of the forges are shut off, the power will be reduced, In this second method, a larger fan could have been used, taking less power, but note that the #2 takes less power than the #3 figured by the old method, This shows that the possession of proper data will lead to more efficient applications. Another type of blower can be used for forges. It ie called a Volume Blower. It has a wider wheel than the same diameter of Pressure Blower wheel. It works well up th 4 ounces DP, It would handle five times the air that a pressure blower would, at the same total pressure, taking nearly five times the power. The air would be so throttled by the forge blast gates, however, that the cfm would be only what was required, For this reason, the manufacturer's capacity table is not here given for volume blowers, as it is misleading, They should be selected by outlet siz- es and then their desired performance in cfm and SP should be worked out from a characteristic or performance curve, A cut and list of outlet sizes is here given. No. of Fan Price Outlet a) a e c i] ~ | sh 10¢; 6) 54] 4/3 2 a os 9| 8% | 103 | 7 13} 124 | 8 | 15 | 144 4/17] 164 | 10 |! SC@neunawn— | a 51 CUPOLA BLOWERS. Here experience has noted the cfm required per ton of iron melted, but the pressure required depends on the depth of charge in the cupola, i.e. on the resistance to air flow. However, a certain size of cupola is intended to melt a certain number of tons of iron per hour, and the depth of charge and manner of charging must be about uniform to accomplish this so it may be said that each size of cu- pola has its own cfm and SP. Of course, duct friction must be added. If a cupola be charged lightly, so that the rated capacity is reduced,it will call for less pressure and less air. But if equipped with the blower adapted jo its maxi- mum needs, its lowered resistance will cause that blower to deliver too much air at standard speed. Therefore ev- ery blower should have a blast gate to regulate the amount of air. Small charge, close the blast gate somewhat, re- ducing cfm and SP at cupola and vice versa, The type of blower for cupolas is the same as that for forges. Following is a manufacturer's cupola blower chart. (From American Blower Co.) | ! 4 | 4 : BS ND Oo Oe RS ave nS yy 6@] es Seuikal aMhe SF oW LSB Qe. SIF ECE SS & ee 8 SES SNity Bw Ss >: : W. . x tN 2 S SS eN x&s RUS SA Sk SLRR SETS RAS LE SE SEVHS NTST WYN 2 1 5/ ax 4. WS Specrat ee £3 4 | €3 | 3070, IP) 8.5. ; 4 {4/50 Ha 5h | GF_. | £7 | 9 | M5 | Y420 nb, 42| 2 oT 69 67Y\ 70 _. | $81 F 461 6480 to. 10 | 4 ee 4 CMY) 13 $71 & | 2/5) F980 £246 109 5 am M1 Ly AF! 42 | 12,277 | 1960. 2990, MS 6 24070) 199 10: He, Le YS | [2 | $18 | (3960.3490: N49 7 £000 9 12?\ 19 4 th. | $62, 6120. 4040. as) 7. 2/70 306.6: [ats ‘q S4| /2 | 4582950. 5260; (3. | © SMS YES /Y | Be 40112 | Sbb 26690, 6650. (37, 0 M28 Sb 16 | 24 (4661/2 695 ss000 G250: Ms | jo | [§¥0: 677 We | 26 2/8 | G74: YolS0' foez0' 1S, BIS GL. Me at OF | 18. |95F\ Yo0e| (2000 1 a. | alt of 127 | 32_ BY | (2 | 1109 $6750| /¥200 /6,2| sz. | 14/0 | 138, al | 3S £7 | 16 (1189 \b1600\ /5¥00\ (6,5 | Sp. | 13%0| 186, ‘gat IS U2 4. Moving of Materials. This is a branch of fan application which is as much in the darkness of rule-of-thumb methods as is forge blow- ing. The field is wider and of much greater importance, and the need for scientific methods and accurate data is greater. It has been the custom to guess at the ounces (total) pressure at the fan. Those experienced at it have become good guessers, but they often fail, and the only sure way is to have knowledge of the velocity required for various materials, and of the frictional resistance in ducts to the flow of air carrying these materials, From tests reported to the writer, it appears that the friction is about the same as if the air were clear, but the case is quite otherwise with a. centrifugal pump, where dredging operations show that if the water carries sand the HP of the pump is increased enormously over that required for clear water. The old practice like that for forges, was to assume some dynamic pressure at the fan for moving the material. These assumptions were founded on success and failure and have the virtue of experience in their favor. Thus the following duties and pressures have come to be associated, Buffin@ wheels .. . ° ° © 2 02 Rag threshers & cutters in paper mills 4 oz Light shavings or dry sawdust. ... ... 202 to 2} oz Ashes. . . « « «© © © © © «© © © © © oe © 016 02 Seed cotton... oe ew ew ew ww tw ww OME OF Hay , straw, and feed o 0 eo oe ew ew ew ew wl wl BO OCOOB tO 5 OZ In some states there are laws governing the vressures for exhaust systems. For instance the Michigan law re- quires for buffing wheels 9000 ft per minute air velocity which amounts to about 4 oz total suction at the fan. Wet sawdust. . ... . oe ee tw wl wl lw OD COZ Crushed ore & stone dust © © © ew ew wl wl wlCUK COZ Cotton & Woolen mills ........e. OD OF Sandpaper dust, powder, etc. .....e 2 02 Ensilage, 3/4 lengths, wet. ....-.. 4 0% Tanbark . . 2. « « e «© © © « © «© « «© « e e & OF HOMB . . 2. © © © © © © © © te ee ew tl ltl ele SB COZ Grain . . 2. 6 « «© «© © © © © © © 0 0 ew wo tw 4 OF Probably the above figures are proper for those mater- dials when the piping is of usual length, not over 100::ft, say. For more piping or high lifts, say over 20 ft. an extra allowance must be made in pressure, Some materials can be blown long distances, Shavings were blown 1672 ft. with two fans in Albuquerque N.i. and the owner reported that it was later done with one fan. This was a 50* fan on a 19" pipe. 53. The pressure is not recorded but the static pressure for the necessary velocity figures to 14.2 ‘for straight pipe. At Kenosha, Wis. shavings were biown 1100 ft and at Cadillac, Mieh. were sucked 250 ft. A separator is used to free the material from its carrying air current and to deposit the material quietly when wanted. The resistance of the separator should be equal to the velocity pressure in the pipe. The above figures took no account of the cfm, appar- ently. Actually this was fixed by the pipe size and cor- responding fan size for the pressures assumed, just as with forge blowing. In the memory or notebook of the practical "blow-pipeé® men, each machine had its appropriate size of pipe, and the areas of these pipes added together, gave the inlet size of the proper fan. Then the fan was speeded up to a tabular speed to give the pressure required, and it was taken for granted that the power consumed was that men- tioned in the same table. Whenever the fan failed to de- liver the material, it was speeded up until it would do so. Since the drive was usually from line shaft, the great in- crease in power was not worried about. The use of motor Grives for these fans made the difference noted, and fans are coming to be better applied. There is some information regarding the velocities needed for materials and it is here given without the writer's sure knowledge of its accuracy, or universal ap- plication. Tests and observations on some materials seem to show that the following formula applies. Smallest Velocity to carry material = 1700 Vueight per cu.ft. As a criticism of this formula, it will he noted that it takes no account of the nature, form, or state of solidi- ty of the material, Thus, a heavy powder such as pulver- ized rock can readily be moved by a velocity of 4000 ft. per min. but broken rock, weighing less per cubic foot be- cause of more voids, could not be so moved, though the for- mula would call for less velocity. Apparently the formula applies to dry, light, and to finely divided materials, The following notes have been taken regarding veloci- ties necessary to keep materials in motion in pipes. Material Air Velocity. Lbs. per sq in permin Ashes, 8000 to 9000 ft per min 5. Seed cotton | 5220 * 4.5 Sawdust 4310 "* * # 4.54 Hay 28530 * * “ 075 Shavings 4000 * *® “ 3 Data is also needed regarding the weight of material which can be moved per square inch cross section of pipe per minute at the proper air velocity. The old practice took no account of this except to speed up the fan if the material was not moved away fast enough. The third column 54. above records some notes on this point. Evidently with other velocities the weight handled should vary as the ve- locity, or as the square root of the pressure. For exhausting materials either the Volume Exhauster which is the Volume Blower with only one inlet, or the steel plate exhaust fan is used. The latter is shown in the following cut. For a known cfm and SP it is figured in the same way as the paddle wheel type. Different wheels are used for different materials, as long shavings, etc. would clog in wheel No. 1 which is used for fine shavings, sawdust, etc.;while wheel No. 2 is for hay, long shavings, rags, etc. Wheel No. 3 is for wool, cotton, etc. which have fine fibres that would cling to wheels No. 1 and No. 2. Each of these wheels shows a different characteristic curve. Wheel No. 2. shows itself most efficient. 2 3 Example: Required to move 100 lbs of dry shavings per min- ute, a distance of 300 ft with a lift of 30 ft. and dis- charge into a separator. Select fan and figure its per- formance. Solution: At.3 lb per sq in per min. 100 lbs per min would require 333 1/3 sq in. pipe area in cross section. The nearest size is 21" pipe = 345 sq in = 2.4 sq ft. The velocity to handle g .3 lb per sq inch per min is 4000 ft. 4000 x 2.4 = 9600 cfm. At 4000 ft velocity the friction for 300 ft of 21" pipe is 4.35" Add the velocity pressure which is 1",:>for the separator. To lift 100 lbs. 30 ft = 3000 ft lbs exerted per minute At an air velocity of 4000 ft per min this means a suction 55. of .75 lbs. on the area of 2.4 sq ft. or .06" W.G. (1" We = 5.2 lbs per sq ft.). The static resistance then is 4.3 +1 .06 = 5.36". The velocity pressure is 1°, making a total suction at fan inlet, or a DP at fan outlet, depending on whether fan is mostly sucking or blowing, of 6.36". Now as to fan size, the old rule wouidbe to have the fan inlet the same diameter as the pipe. It has later been found that a larger fan gives better efficiency. Trying a #50 fan whose inlet is 21" diameter, its outlet is 2.25 sq ft. Its OV is for 9600 cfm 4250 ft. From its performance curves for SP = 5.36 and OV = 4250 eq ft., its PV is 11000 and its bhp is 10. Its wheel is 35" diam, so its rpm is 1275 RPM and its BHP is BLD A #60 fan would have an outlet of 3.21 sq. ft. and an OV of 3000 ft. = :56" VP. 6.56 - .56 = 5.86" SP, The performance curves show for 3000 ft. OV and 5.86" SP that bhp = 6 and PV = 9750. Wheel diameter is 40* s0 RPM = 930. BHP = 6 x 3.21 = 19.3, which is quite a saving in speed and power. 5. DRYING. Here empirical methods come more to the fore than in any case yet considered. While it is easy to figure out how much moisture a pound or cu. ft. of air could "Pick up" if entering the dryer at a certain absolute humidity and leaving at another known humidity, it is hard to say how thoroughly saturated the air is likely to become, Much depends on the management and charac- ter of the material. While it is usually necessary to know the wet and dry weight of the material and the amount of it per hour, 80 as to determine the weight of moisture to be removed per hour, yet that is not a guide, and there is no case where full saturation of the air can be expented. Neither is there any rule by which to calculate the percentage of full saturation to expect. Besides, there is the further fact that the drying would vary each day as the absolute humidity of the entering air changed. : This last variation can be avoided by using an air washer in which the absolute humidity is kept constant by managing the water temperature. The water cools ( or warms) and saturates the entering air to a pre-dete ined temperature and relative humidity, which fixes the spare humidity. In this way uniform results are secured, and in many industries it would be worth the expense, where it haesnot yet been tried. There is a good field for such applications. Up to the present rule-é6f-thumb methods are largely employed. The safest way is to follow the ideas of the operator or superintendent in charge of the drying opera- tions. Then he will be more inclined to make the dryer work, to vindicate his opinion, than if he had been oppos- ed, In many operations, there is a limiting temperature 96 for drying, above which the material would be spoiled. In most cases the material is arranged in a room through which hot air is blown or drawn. In the following notes, the “air change” means the number of times the cfm is containe& in the volume of the room. This 1000 é€fm would be a 12 min. air change for a room of 12000 cu ft. volume. Glue is dried on wire trays on cars in tunnels. Teme- perature 90° at hot end, 75° at cooler end. 1.5 min air change in tunnels. Glue 1/4" thick 36 to 48 hours to dry. Thick glue, 4" and up, 3 to 5 days. Raise temperature gradually. Lumber is dried in so many different conditions and of so many kinds that its treatment would form material for a thesis all by itself. It will not be entered into here, Apples are dried on trays or canvas belts. The fruit is sliced. Contain about 80% of wet weight as water. Dried on trays in 3 days (presumably 72 Brs.). Will stand any temperature attained with steam coil heater. Fish are split and hung on hooks on racks. 24 hrs to dry. ¢ min air change. 110°. Wool spread 5" to 6" deep on wire netting. 140° temp. 2hrs to dry. YVoisture 53% of wet weight. Cocoa 159° temp. 24 hrs. to 30 hrs to dry. 58% moisture, Dried on trays. Casein can best be dried on trays in sheets 2" thick. Temp. 120° maximum. 24 to 3 hrs. Chicle, used for chewing gum, d in trays in sheets 2" thick. 100 temp. . 36 hrs. 1/3 min air change. Burlap, temp. 1509, continuous dryer, 6000 yerds in LO hrs. 2500 cfm. Air blown across traveling cloth. Soap. Différent BOaps will stand different tempera- tures without melting. 190° usually is the working tem- perature. Continuous wire mesh belt dryer, soap 4" to 3~" thick on belt. Dried inl hr. Only a small portion of the moisture is removed, varying in different soaps. School crayons will stand 140° when wet, 120° when dry. Dried on trays, in tunnels. Allow 234 cfm per thous- and crayons per day. 24 hrs to dry. Enamel on .bedsteads etc. 14 min. air change. White enamel 165° Otner colors, any temperature, and 34 hrs to dry. White takes longer. Air Washer Applications. Air washers can be used only where fans are used to move the air, hence they are closely connected with fan applications. Air washers are coming into greater demand for ventilation work and the results justify their being used wherever the air needs to be cleand@d or humidified, or de-humidified. For cooling also, an air washer is of- ten the best thing, taken in connection with its other features. The only points considered here witl be those which affect the performance of the fan. The resistance of the air washer to the flow of air 57. is its main feature as regards the fan performance. The old style of air fiiter made of cheese-cloth was objectionable in this respect because it soon clogged up and was not properly nor frequently cleaned. It then stopped the air supply. Its successor, the coke filter, with a rain of water over the coke choked up very soon, andwas even harder to clean or renew, Such makeeshifts should not be used, In the modern air washer the greatest resistance comes from the baffle plates called eliminators, whose duty it is to whip the free water drops out of the air after it passes through the sprays. To -be effective these baffles must change the direction of air flow abrupt- ly and repeatedly. To avoid excessive pressure loss, the velocity across the gross area of the washer should not exceed 350 ft. per minute. With such a velocity the resistance may be safely al- lowed for as .25" W.G. Some washers are claimed to have less resistance, and some have fully this. It is well to get a statement from the manufacturer as to what frice- tion his washer sets up. This low velocity is desirable to prevent water being carried past the eliminators or flipped from their drip- ping edges or angles into the air current. If possible set the fan inlet away from the washer a distance equal to half the washer's width, to permit of even air distribution across the washer. Do not try to blow air through a washer. The air washer affects fan performance also by raise ing the relative humidity of the air. The air is usual- ly almost 100% saturated leaving the washer. If it is handled by the fan in this state, the proper velocity constant for the density should be used. 7. Organ Blowers. Multi-stage blowers are largely used for this purpose but as usually made, they are inefficient. The organ manufacturers are almost always able to specify the cfm and SP wanted at the wind box, For ore- dinary installations it is best to allow 4" more to get SP at fan. Then make the motor amply large, say twice the power figured, as the fluctuations of demand are large and sudden. Belted blowers are convenient because they can be speeded up if necessary, by a change of pulleys, but as spacé is often limited, direct connected blowers are frequently used. Quiet running is essential, and felt or cork bearing pads, flexible couplings, and the canvas connections in the air pipe line are features of this work. Often a damper is set in the line so that when the bellows is full, it closes the damper, and vice ver- sa. Of course there is a flap valve which prevents the bellows from deflating when organ and blower are stopped. The pressure and cfm vary from 3" static and 300 cfm 08. for small 3-manual organs up to 8" and several thousand cfm for large ones. The blower to use is a pressure blower the same as for forges or cupolas. If a multi-blade blower is used, it must be very narrow and have a smaller inlet than usual. The pressure blowers should be of such size that the re- quired cfm and SP will be had at about 20% ratio opening. 8. Gas Blowers and Boosters. Gas blowers handle air to supply the gas generators. They are usually double inlet blowers and aside from es- pecially heavy construction to withstand the high speeds and severe duty imposed by the high pressures and contin- uous runs, they are no different from the other air blow- ers. The pressures run from 8 to 16 oz., and proper al- lowance must be made for the compression of the air. Usually the customer will specify the pressure and cfm of free air. The blowers are ordinarily cast iron housed, for rigidity and look like the volume Blowers. The gas boosters or exhausters draw the gas from the hydraulic main and deliver it to the condensers. Some=- times boosters are inserted in a gas main to increase the pressure, In either case the fans are of special construc- tion. The pressures run about the same as the blowers, and since the gas is of about .5 the density of air, the speed of the fan must be Breatly increased to give the desired cfm, This is calculated from the proper char- acteristic curves, by using the velocity constant corres- ponding to the density. Here also, the compression of the gas must be allowed for,if the cfm is to be measured at the high pressure. Since these fans receive the gas under some pressure, their inlets must be air tight, and of course the cases must be also. Therefore, the shaft must run ina stuf- fing box,- whencit enters the casing. Brass surfaces are provided wherever parts might strike, to avoid esparke and consequent explosions. | Because of their special features, these boosters have characteristic curves different from the blowers they resemble, COOLING TOWERS. Hot water from a condenser may be coéled by evap- oration, and used over again. The cooling tower is a de- vice for presenting large surfaces of hot water to a blast of air. The amount of water required is a matter for the cone denser manufacturer to state, also the temperature at which the water will be delivered to the cooling tower, 59. Most of the cooling effect comes from evaporation and not by contact with the air, so the amount of cool- ing depends upon how much moisture can be evaporated by the air as it passes over the water. This of course, depends largely upon the completeness of contact between air and water. o With the usual arrangement, and water at 110 at top of tower and air in proper quantities at 70° and 70% relative humidity, the air can take yp 14 grains of water per cubic ft, departing at about 100 e almost fully saturated. The water will be cooled 76°, approxi- mately. Since the latent heat of water at atmosphezic pressure is 965 H.U. and the temperature drop is 35 ‘ then 55 of a 1b of water must evaporate to cool 1 1b. 355°. 965 This in grains is 7000 x 35. Since each cfm of air takes 14 grains, then 965 the cfm per lb of water is 7000 x 35 , or for the conditions above named, 18.15 14 965 cfm per lb. of water. Where the relative humidity of air is greater its evaporative power is less, but it is usually cooler, so say, 20 cfm per 1b. water is a safe figure for 35° cool- ing. The resistance of a cooling power is from 3" to }", An average of 3/8" SP. is a good assumption. Disk fans are usually used, because of the large cfm and low resistance. A special disk fan called a tooling tower fan" is shown in the accomvanying cut. Lows: Size. RPM cfm capacity. Brake H.P. 48 610 25000 7.06 54 542 29200 8.95 60 490 55900 11.05 72 406 51700 15.90 84 550 70400 21.60 96 305 92000 28.20 108 271 116000 | 55.60 120 244 144000 44.20 60. A table for capacities for 3/8" static pressure fol- Of course the above table is only for a certain ra- tionof opening, that which gives good efficiencies. At the same static pressures the fans would deliver more air with more HP at higher speeds, and vice versa, 10. Transformer cooling. The cooling of transformers and generators is an im- portant but limited field for fan application. The great- est demand is direct from electrical manufacturers and those making a specialty of electrtcal machinery installation, These users have their own requirements worked out by experience, The paddle wheel type of fan has been gener= ally used, with two inlets. The air ducts exb&nd beneath the floor of the transformer room and blow up through the transformer the cooling air. The cfm varies with the re- quirements, but it has been noted that 1800 cfm per KW would cool a transformer to within 1 degree of the air temperature and 150 cfm would bring it to 12° above air temperature, The humidity of the air is_an important factor, as the warming of a jb. of water 1° is the same as warming 4.2 lbs. of airl. In this connection, and for generator cooling, an air washer and humidifier could very well be applied. The dynamic pressures are from 2 to 4 oz, depending on the resistance of the ducts and the free &£% of the transformer, For generator cooling, generators are provided with blades which normally draw enough air for the purpose. AJncase of overloading an extra supply may be needed, or when the air is too warm for the normal quantity to do its duty. The cfm required depends on the cooling effect de- sired, of course, but the usual amount is 5 to 74 cfm per KVA. The pressures vary With different makes from 168 to 4 oz DP. The Westinghouse require the most and the Crocker-Wheeler the least pressure. 61 DISK FANS. Disk fans, in which the air enters, and is discharg- ed, parallel to the shaft by a wedging not a centrifugal action, may often be used to better advantage than cen- trifugal fans. Wherever large amounts of air at low pressures are to be moved at comvaratively low velocities, disk fans should at least be investigated as their fitness. For ventilating, cooling, drying and heating work where there is little or no duct friction, these fans are usually preferable. They can sever be used in connection with considera- ble duct work, if properly computed as to size, speed, and horsepower. The cross section of the duct should at no place be less than that of the fan circle, A peculiarity of disk fan performance is that its HP rises as the delivery is restricted. This is just the opposite of a centrifugal fan. For this reason, the disk fan when set so as to encounter the adverse force of a heavy wind will sometimes burn out its motor. For ine stance, a straight bladed disk fan of the type to be il- lustrated takes nearly twice the power: when entirely closed off, i.e. at zero opening, that it requires at full opening. Thus, if it were delivering air into still atmosphere at 1000 ft per min. velocity, then a wind of 11 1/3 miles per hour would counteract the fan, and double its power consumption. Use a larger motor than required or protect the fan discharge by a hood or by an elbow turned up or down, or by a cowl that will turn from the wind. The disk fan is cheaper than the centrifugal fan, so that larger wheels can be used at a less price, which is really, the reason for its popularity, for its efficiency even at full opening, is less than that of the paddle wheel at full opening, but it would take a large paddle wheel to handle air at low pressures, on account of its outlet velocity being comparatively high. The following cuts and characteristic curves show the style and performance of this sort of fan. Far) SIZES. Wheel diam. In 1aChHES. 24 JO 36 AZ 48 54 60 7a 4 96 708 420 132 Outler Mea (Wea Jé@s7 FUE, SY Ft V.E65 Jh7 SH 746 985 43.) 76,5 2033 £9./ IGA IALS 646 79.8 IES Me. . 7 j ‘ata 4 / 4 i“ 7 SIMNSSIAS $0 SOUOL PUO fF YIIAW Sn LA 2 a Ad MEL PEE ae wie SIAM DISUA IOI] Weve e, i Or O02" Tae / LL Vn “a 7 ™S § S +. s) > ¥ Ss G QS bes ww RS XN Nj b Ss ) >] 4 ' , : ATOM AGE AAU Lao bd ) a, Comparison of performance of Disc Fan and Paddle Wheel Fan. Problem: Handle 20,000 cfm at 67°, 1/8 " SP. Fan must be quiet running. Compare disc and paddle wheel fans for PV, RPM, and HP. Solution: Disc Fan. Quiet PV is about 3750 for the limit. From the performanee curve, for 1.8" 8P, the outlet velocity is 725 ft. per min. 20,000 #725 s 27.5 sq ft. outlet area. The nearest size is a No. 72, whose outlet area is 29.1 eq ft. 20000 #; 29.1 s 688 ft 0O.V. This, on the performance curve, gives 3600 ft. PV. and .05S BHP. -05 x 29.1 2 1.45 BHP 5600 ft PV for a 6 ft wheel is 191 rpn. PADDLE WHEEL FAN Quiet PV is about 4000 for the limit. From the performance curve it is seen that at 4000 ft. PV and 100% ratio of opening i.e. at zero SP the OV is only 2450 ft. The limiting OV for noise is about 2750, so there is no fear of noise from this source, since at 1/8" SP the OV will be less than at full op@m for the same PV. 1/8" SP is not shown on the performance curve. Go to the Characteristic curve. PV e 4000 PVP = 1” SP ¢ PVP = .125 This is 95% open, for a trial. P 1.00 1.00 8 ~125 ~125 V 545 545 4000 V .345 is 2350' 0O.V. D 47 — @ AT 20000 # 2550 is 8.5 sq. ft. outlet If we used a smaller outlef, it would bring the PV too high, i.e. over 4000 ft. The next larger standard outlet is | that of No. 110, which is 9.75 sq. ft. 20000 #¢ 9.75 equals 2050 ft. OV. The square of 2050 is .263" VP SP es .125* 4000 DP = .388" VP is .677" This is 934% open. DP Pp 1.00 - 780 5530 ft PV 5 - 160 e125 Wheel 5.5 ft diam. gives 204 RPM V 23356 «263 D 2495 088 HP s 2088 x 20000 = 3.83 HP Thus the paddle wheel fan takes 2.64 times as much power. Notice that the disc fan at a quiet tip speed, cannot handle air at much more than 1/4" SP. Often it is possible to so locate the fan that its whir- ring will not be objectionable. Then it can be run at high- er speeds than 3750 ft PV. ADDENDA. Suggestions for Thesis work with Fans. Since the limitations of the mechanical equipment of any school will confine the student to experiments with a very few fans or blowers, perhaps only one of a kind, the follow- ing suggestions will indicate general principles, which can be investigated with almost any size of fan, rather than de- tails, which might require special or eleborate apparatus, or comparison of the performance of different fan sizes. There is much need of better and more data regarding the flow of air in ducts. For the flow of air in round ducts, it would be a great convenience to have a reliable factor by which to multiply pome one velocity reading, to get the average velocity read- ing. For instance, find what the velocity at the center of a circular pipe should be multiplied by, to get the av- erage velocity. If the velocity is retarded at the sides of the pipe and is greatest in the center and if this variation follows some law, it should be possible to find some point at which the velocity is an average, and so one reading at this point would suffice, Thus, if has been suggested that the veloci- ties vary as the ordinates of a parabadda. If this is true, the average velocity would be found at a point 2/3 of the radius, measured in from the side toward the center. Per- haps this velocity variation takes some other form, depend- ent tipon the center velocity. Compare results obtained for the velocity in round pipes by Treat's method of concentric rings, by Taylor's method of 91 x the center velocity, by the “contour and planimeter ” method and by some new method developed perhaps as suggested above. Use the A.B.C. Pitot tube, as it gives best results. Determine the error in taking an average of velocity pressures instead of an average of velocities. Data is needed regarding friction in elbows, especially in square elbows. It would be of value to have corrobora- tives data regarding friction in rectangular ducts, but the expense would be considerable, However, rectangular elbows can be with a wooden platform for a bottom and another for the top, and thus many variations of dimensions can be tried with one width of galvanized iron side pieces, and as many more with each new width, while any curvature desired may be laid out on the boards, and conformed to by bending the sides. If the boards are smooth they need not be cov- ered with galvanized iron sheets. The elbow should be attached to the end of the fan test pipe. Perhaps a linseed oiled canvas connection would be the cheapest and most flex- ible means of fastening the different elbows to the pipe, or a small rectangular wooden chamber of box could be made, into one sidé of which the test pipe would run and to the oppo- site side the elbows could be attached, Static readings in the box would indicate the different elbow resistances and the side of the box could be made large enough to accomodate all elbows tested. Another useful investigation would be the matter of con- veying materials, Find out what velocity will just suffice to sarry dif- ferent materials, and what maximum quantity of each can be handled thus per square inch of pipe eross section. Determine how the friction in the pipe varies with the amount or density of the material carried. See how the horsepower varies with the different quan- tities. Perhaps it would be better to drop thematerial into the current of air in front of the blower rather than pass it through the blower, but probably the investigation of this point would be rather expensive for college thesis work. Forge blower work needs further study. Determine the maximum amount of air used by each forge in the college shop, and also the maximum amount used per minute by all the forges while they are in ordinary use. Al- so find out the average amount used by all and by each forge. Make similar observations for static pressures at the forges and at the blower. Make observations of power required under different con- ditions above outlined. Find out the size of air ducts and compare the actual friction Ioss with the friction as computed from the tables or formula before given herein. | Try to make a good working statement of the cfm and §P required, for the sizes of tuyeres and forges commonly used. Try throttling the forge blower, or showing its speed, to see if less pressure and cfm would do the forge work just as well. It is seldom that the blast gates at the forges are used fully open. Make a report on the fan equipment of the forge shop, both blower system and smoke exhaust system. Acid fumes have to be removed from places and sometimes natural draft is not strong enough. The fumes eat a fan. Acid resisting fans are expensive. A jet of air at a high velocity driven into the vent flue in the direction of air flow would produce an ejector. action. Then the fumes would not touch the fan, which would merely handle a small amount of air for the jet. Determine the best velocity, angle, form of nozzle and relative cfm to use. Find aut how much air can be drawn up the flue compared to the cfm handled by the fan, for differ- ent nozzle velocities and different sizes of flue. Find what velocity can be induced, and what resistance overcome. This is a very practical problem. It could be applied to induced draft and to ventilating problems, Perhaps a disc fan could be made to handle much more air by meane of a proper nozzle on the fan discharge, and an annular air passage around the circumference of the fan. See if the fan horsepower is decreased or increased by this induced air flow. The writer has found that a small fan blowing into a short duct somewhat larger than the fan outlet will handle much lees air through the duct than it will if the fan oute let be set back a short distance from the entrance to the duct, There is an induced flow into the duct entrance. Apparent- ly the blower takes less power in the second position. It is possible that less air passes through the blower. All these points are worth investigation. There is a small vacuum cleaner which uses two centrifugal fans, one blowing into the inlet of the other. There is not much data as to what pressures or suctions can thus be obtained with larger fans. It might be of great commercial value, to know the laws of such tandem work. Maybe several fans in series could overcome difficulties now ¢onsidered too much for centrifugal fans. The writer will be glad to give any further suggestions, references etc. of which he ie capable, W. E. PIPER. 434 Rialto Bidg. San Francisco, Calif. WEP/N. - 2 Gp ph a 59 wii All TE UNIVE - © we vz .- ~ - o oO 7) om iN _ IGAN i 3 i