You are on page 1of 16
The degree reaction of an impeller having = = 2 _ Ae Fw = Haywo 1 Hayrce Oo Masta meNasr et ce Bret At the points 4. C and £, the degrees of reaction are 1 A peal C pay E pw=0 (53) From the foregoing considerations the following conclusions may be érawn: To produce a specified total head, impellers with blades swept backwards must operate at a higher peripheral velocity than impellers with blade angles = 90° or blades swept forwards (f, > 90°) and heace for pumps of the same rotational speed, impellers with blades swept backwards must be larger in diameter. Impellers with blades swept backwards display a higher hydraulic ficiency, which is partly due to the more rational shape of the impeller ‘passages from a hydraulic point of view, but chiefly on account of the smaller conversion of kinetic energy into pressure energy in the diffuser ring, so that in spite of greater leakage losses, the overall efficiency is higher. ___ At present centrifugal pumps for liquids are constructed exclusively with impeller blades swept backwards at angles z= 14° to 50"; however jy is rarely more then 30°. BIBLIOGRAPHY H.L. Cooren, Visusi experimentation with centiugal pumps. The Engineer, Vol. 172. ‘London, 194 ite No $. Frankenthal, 1963, Proc. IME, ssverau? in und hinter Laufschau H, Perea, Der St No I. Disieldare, 1961, (CHAPTER IV THEORY OF DYNAMICAL SIMILARITY OF IMPELLER PUMPS 1. Introduction The theory of dynamical similarity of rotodynamic hydraulic machines ‘enables us to describe the phenomenon of flow through the interior of a full- size machine on the basis of measurements of a similar phenomenon occurring in the interior of a model, Using the theory of dynamical similarity-—the basis of model invesrigations— it is possible to design water turbines and impeller pumps on the basis of experiments on models of less than full-size. As the results of such investig- ations apply to a whole series of geometrically similar machines, model ‘analysis constitutes a scientific foundation for the standardization of types and the production of a series of machines, whose hydraulic properties are known in advance. By making use of the theory of dimensional analysis, the sheory of dyna- ical similarity permits one simultaneously to define the specific speed of several machines and to deduce structural formulae, which are invaluable In dimensional analysis the following kinds of equations are used: quantity equa ns, numerical equations, dimensional equations. positive or complete equations, and structural ‘equations i defined as a physical equation in which the symbols denote ‘ia. products of & numerical value of a piysical quantity and a unit of Row Q must be expressed in fe of flow Q = 3600-02 = 180 102 rstusR Fowes of the research worker less investigations. They also save time, energy and mate 2. Conditions of similarity for impeller pumps 21. Condition of geometrical simil ‘Two impeller pumps are considered as geometrically similar, i the impellers, guide-vanes and stationary elements (Suction pipe, passages leading the liguid into and away from the impeller and the internal walls of the casing) of one pump are a true geometrical scale model of the same elements of the other pump. In particular, geometrical similarity is fulfilled if: (1) The ratio of all the homologous (mutually corresponding) sections of both systems is the same; since the angles contained between homologous les (2) The shape, disposition and number of impeller blades and guide vanes are the same, ) The relative roughness of the walls in contact, is the same in the two pumps. th the pumped liquid the equation Q = thesymbol denotes vhe number O4 and the symbol ©. the number 0 and the symbol Q, the numbe one corresponding dimensional equation (er =r onding dimensional equations | | 1W.0, OF DYNAMICAL SOULARTTY OF IMPELLER PCOS 103 2.2, Condition of kinematic similarity ‘The flows through one and the same pump or two geometrically similar pumps are kinematically similar, if the velocity fields in the two flows are 2 ly similar. Hence the condition of kinematic similarity of flow leads to geometrical similarity of the velocity fields. 3, Relationships between the basic physical quantities characterizing flow through impeller pump rotors 3.1. Basic assumptions ‘Assuming that the coefficients of hydraulic resistance for flow through of a pump are independent of the rate of flow, we can write down simple relations between the rotational speed of the impeller it follows that the velocity triangles at the inlet and outlet of the impeller are similar and the lengths of their sides are proportional to the square root of the total head, The same relationship of course occurs between the rotational speed of the impeller and the total head H nop H 2 3.2. Conditions for similarity of flow through the same impeller ‘Assuming that the coefficient of hydraulic Losse independent of the rate of fo flows through the same impeller. a, Rotational speed of impeller eo b. Discharge ER @ 104 -PELuEn Poses ©. Effective power or water horse power Si (HY 2QeH, ya \ Hy cu IF in both cases the liquids are of the same specific weight (4 = y,), and fe _(tt\"_ [my ae (ey - (2) iQ 33. Unit quantities a, Unit speed. Denoting the impeller speeds by n, and 1 corresponding to total heads of H, and H, we obtain the relationship ‘Hy a a The speed of rotation referred to a total head H, = H, = | m is defined by the numerical formula speed of the pump and this is called the ». Unit discharge. Denoting the rate of flow through the pump by Qe against 4 total head of H, and the rate of flow against a total head of Hf by Q, we have He = Oy F ® For H, = Hy = I m the discharge Q, = Q, and hence the uit discharge 2-0/2. ao ©. Unit po Taking into account the relationships P= 70M Pe von we have Qe, _ pts He Pym pOetts ptt Be On ay For He= Hy = 1 m. the power P,= P,: hence the unit power is HTP. Py = pa LP 12) Hy Hy ee acs gion Mine "WY OF DYNAMICAL sRALARITY OF IMPELLER PUMPS 10s 4, Specific speeds 4.1, General formulae defining specifie speeds Let us denote the diameter of a geometrically similar pump by d, whose discharge at H; = 1 m and a reduced speed m, is , and the effective power P,, Since the total heads at which both geometrically similar pumps are working are the same, (H, = H, = I m), the impeller inlet and outlet velocity triangles are congruent. The peripheral velocities 1 are equal, so the speeds of rotation of the two impellers are inversely proportional to their outlet diameters, and then the discharges and powers are proportional:to the size of the cross-sections of the flow passages, and hence to a These are expressed by the equations a) as as Inserting the values ny = n4/ 7 and 4 = V' Se ino equation (13), we obtain the quantity formula ant an By @ any 2h a" V aVo-"V wave ~V/ Zep (6) V0, \a) For Hy = 1 m, we have the mumereal formula 2 an ee "TEV Ove 106 eeieen rote [Fi ang 424 fe 3) Inserting the values ny = n/p and = Y/ pe into equation (13), wwe obtain the quantity formula vo V = VOVERT VERT For H;= 1 m, we obtain the numerical formula ” fe “Ta mat Toe qustiy 1, dened by equtions (19, (17) aud (19, (19) cb the specie seed tan pelt ump. hc Ptoecson ner te root ign i eentons (16) and (18) ae diventclss quam, he spre spats he same mension ss the speed of roton cc) fn) == 73 ‘The specific speed is most commonly expressed in r.p.m. ‘The numerical value of the specie speed depends on what characteristic quantity of «geometrically similar pump we take as unity. 42. Kinematic specific speed Assuming Q,= 1 mfjsec in equation (17), we obtain the numerical formula (20) where Q is expressed in m*/sec and H in metres. Since in the dimensions of the quantity mq Ing] = E8792 wwe have only kinematic quantities, this quantity is called the kinematic specific speed and we define it in the following manner: By the kinematic specific speed myq of a rotodynamic pump with an impeller of diameter d, pumping Q m/sec at a speed of p.m. against a head of H metres, we mean the speed nq of 2 geometrically similar pump with an impeller of diameter d,, which would have a discharge Q = 1 m/sec against ahead H= 1m, In the engineering system of units the quantity 1g is expressed by {imi J= rpm meeps 7 OF DYNAMICAL SDULARETY OF IMPELLER PMS 107 From equation (20) it follows that the kinematic specific speed of a pump does not depend on either the density or the viscosity of the pumped liquid. If the discharge of the pump is expressed in I/sec, then equation (20) will take the form nya rg = ose 2 ap 4.3, Dynamic specific speed Inserting the value P, = 1 Metric H.P. into equation (19), we obtain the numerical formula 2 where P is expressed in kp m and HY in metres. “Taking into account the relationship p= 19H 3) 7% we obtain the expression (JOH _ 4/9 typ ny/ OF = ny a4 eV pH" V asHyH oo where 7 Since force, which is quantity mp is expressed in kp/m®, Q in m*/sec and H in metres, ‘dynamic quantity, occurs in the dimensions of the ud = span the dinamic specific speed and define it as follows: mic specific speed mp represents the speed of a geometrically similar pump, which fora ft of 1 m, absorbs @ power P, = 1 Metric HP and has a discharge Q = 0.075 mse: Ta the engineering system of unite the quantity mp is expressed by true) = rpm, VP veel 44. Relationship between the specific speeds mg and mp Between the specific speeds inyg and ny there is the relationship 25) ney 108 erases ries For water with a specific weight » = 1000 kpm® and with P, = 1 Metic HP, and hence ata discharge Q, — 0075 msec, we obtain the numerical formal nyo np = 3650/2 = 365 UG — 365g VF We 26) where Q is expressed in m*/see and H in metres; en where Q is expressed in 1/sec and H in metres In pump handbooks both specific speed literature and standards mg predominat dependent of the specific weight of the liquid. The specific speed yp has the disadvantage that for the same pump it can have different values depending, ‘on the specific weight of the pumped liquid, 1p are used; in recent since its.numerical value is in- 4.8. Relationships between the values of specific speeds in the Metric, English and American systems of units Jn the formulae for rep and yg given in English (E) and Ametican (U.S.) literature the total head H is expressed in feet (I foot = 03048 m), the discharge Q in gallons per minute (1 Imp. gallon = 4546 |; | U'S, gallon = = 3-785 1) and the power in horse power. (rape = 00174 (tage (anes = 0-0159 (nos (rape = 0225 my = 0822 mq (typos = 0225 yp = O82 tg (raghe = 1295 mp = 472 mq (ages = 1413 rp = 516 Mg where yp and ng are in metric units, 4.6. The nature of specific speed The specific speed is a number characterizing the type of impeller in @ unique and coherent manner, which formerly was defined by the ratios oF dd the numerical value n, is:closely associ . the specific speed constitutes a basis of classification this number defines not only the specific speed of some given pump, bbut it also defines the specific speed of all pumps belonging to a given type. ge pumps we calculate the specific speed for one imi try impellers we take half of the total discharge of the pump. , refers to optimum operating conditions of the impeller. and hence to the peak point of the with os fica OF DYNAMICAL SIMILARITY OF IMPELLER PUMPS 109 eo mance (se2 Chapter XVI, at which the overall efficiency of the pump attains a maximum. Considering the specific speed 7g a8 a function of Q and H, we obtain sarious values of yg for various points on the pump characteristic for m = ‘= Const; yg then varies from ero with the delivery valve closed (Q = 0) up t0 mq = 0 when H = 0. From equation (20) it follows thatthe value of the specifi speed increases as the discharge and the speed increase and it diminishes with increasing total head Hence pumps with a large total head are characterized by low values of specific speed and conversely, pumps with a small H by large values of 1, For the same total heads and the same discharges Q, the speeds of the impeller are directly proportional to the specific speed, nocn,g, and hence pumps with a large mg rotate faster and their dimensions are smaller The determination of the value of the specific speed mg is the starting point in the calculation of impellers. In particular the values of the empirical velocity coefficients for the inlet Kigy and the outlet Kegs are given in relation to the specific speed (see Chapter VI) In pumps with similar impellers. n, has the same value at homologous points of the head-discharge characteristics and in particular at the best efficiency points 5. Dimensional analysis of impeller pumps $.1. Total head, power and efficiency a. Physical analysis. In flow through the rotor of an impeller pump the follow- ing physical quantities occur: ¥ = viscosity of liquid ore p — density of liquid Lr g¢ — acceleration due to gravity LT? ‘H— total head L da Q — discharge bre n — rotational speed of impeller. T™* From among these quantities we choose p, dn, Q and v as independent 1p takes place in a field of gravity. The w of wacer through the pu | energy Which, in the basic equation field is related to the potent no meLien Poses for impeller pumps, occurs in the form of the term gH= E, representing the energy imparted to unit mass of liquid by the impeller. wwe express as a function of five independent variables: gH =f, dm, 0.) 2s) P = fi(o, 4,7, 0,») (29) 1 = falo, dm, 0,9) 0) In each equation there are m= 6 physical quantities, including m = 3 fundamental quantities: length, force, time. The number of dimensionless products is therefore n—m = 6—3 —=3, We write the dimensionless produets in the form: m= pana gif t = pdt a mam Oda a ‘The dimensional equation: CET IET-ALT = Pr ro The algebraic equations for the exponents: Lo ndntnt220 naa Foouno noo Toolmey-2=0 00 y= eH a (AEP TAB TA = LPP Solving the algebraic equations for the exponents we obtain 420 y=-3 y=—l -£-4 a ay (MET)SIPT ALTA = PT 8) Solving these equations for the exponeats, we have =O ye -2 y=-1 fn Hence: per oF cas BELT OF mem EI] snd he fst structural formula Hoe ( 22) fe Se ae Ina similar way we deduce the other two structural formulae: G2) G3) numbers dy, Ag, to 2p, we call the characteristic sharacterstic discharge number, characteristic speed number impeller, at high Reynolds numbers, the influence of the viscosity of the iid is almost imperceptible. Neglecting the term in which the coefficient in the formulae: occurs, we obt G4) as) 36) Transforming equations (34) and (35), we obtain the structural formulae defining the total head H and the power P: en 5.2. Characteristic eurves From these equations the speed 1 of an impeller pump he total head H, the power P and i$ of the discharge Q: ) 6) m2 rvPeLLeR Powe The curves traced by the above functions are determined experimentally for various speeds and on this basis the complete characteristic is drawn (see Chapter XVD. ‘The dimensionless characteristic curves corresponding to the above equations, are identical for all geometrically similar puraps and independent of speed, This conclusion is only valid for turbulent flow, where an increase or decrease in Reynolds number does not alter the character of the flow, ie. at ‘At the point corresponding to the peak of the complete characteristic, Q ie at 7 = Mare the expression 2 has a constant value fn% 0) Thea Hig aw Pie fen @) where the constants 2. 2y, Zp, depend entirely on the shape of the impeller. Hence it follows from these formulae that the discharge Q is proportional to dn, the total head Hacd¥nt and the power P ccdén*, 5.3. Physical interpretation of the dimensionless products a. Assuming that the dimensionless product x, = Const, we obtain the ionshiip gH EH = Const (3) (a) b, Assuming == Const, we obtain the relation Q. Const as) “ony OF DYNAMICAL SOULARETY OF IMPELLER PUMPS 3 Hance: ‘The discharge of a pump with an impeller of diameter d is directly pro- nal to the speed of rotation, a 7 a6 Om Cal Taking into aosount the relationship ndccoeny sf we obtain (47) and so for a pump with an impeller of diameter d as) «Taking ito account the ri we obtain «9 0) ‘When the same liquid is flowing through the pump (r= 7), this equation requires that H,— Hz. Thus, the expression = f(H) may only define a eertain state of ffow through the impeller, just as Reynolds number defines the state of turbulence for flow in a closed conduit. The dimensionless product -% cannot therefore be a criterion of similarity tor liquids of the same viscosity flowing through one and the same impeller. The product =, may, however, be used as a criterion of similarity for fluids of different viscosities flowing through the same impeller (e.g. water and air): equation (50) would then be satisted 4. Assuming =, = Const we obtain the rel aship 6» trom which it follows that the effective powers attained at speeds of 1, and ‘nz through a pump with an impeller of diameter d is f(a) en ng IPeLLeR Puss Se, Characteristic numbers: head and discharge The dimensionless expression eee = Se (33) is called the characteristic head number, and the expression Q to" Ba (54) the characteristic discharge number. In the literature on impeller pumps the following characteristic numbers are encountered: characteristic pressure number ot head coefficient characteristic discharge mumber or flow coefficient (56) where the outlet section 4, = 24 5.3. Similarity numbers for flows through one and the same impeller 8. Product m. Since the product dn* is proportional to the square of the mean velocity of flow, the reciprocal of the product =, 7) constitutes the frst form of Fro1 rae (58) for impeller pumps b. Product =. Since the quotient of flow and simultaneously we expression id* is proportional to the mean velocity we the relationship nd cce acy/gH, the (9) corresponds 10 two new forms of Froude's mumber for impeller pumps Fe (0) ne 1) (2) wee 3) er = Be wo) 4. Product my. Taking ino account the relationships given in a, b,c, we may define effective power output of the pump by the structural formula P= pdsiG, ®) 6) « Conclusions. From the preceding considerations, the following conclusions may be draven (1) The criterion of similarity for fows of liquid of the same viscosity through the impeller of one and the sume pump is Froude's numb (2) The criterion of similarity for liquids of difering viscosity flowing ugh the same impeller is Reynolds number; @) For flows through the impeller of the same pump, the two cond- tions, F, = F, and R, = 2p cannot be satisfied simultaneously. 5.6, Kinematic specific speed iminating the diameter d from equations (40) and (41) gives the fort ng? yO fen 6) defining the kinematic snecifc speed 5:7, Dynamic specific speed Comparing equations (41) and (42), gives the expression ae o called the dinamic specifie speed of an impeller pump. 16 ELLER puss 58. Similarity conditions for flows through geometrically similar impellers Considering two geometrically similar pumps with impeller diameters dy and da, we obtain the following relationships for two similar flows of iquids with densities p, and p, and viscosities » and ¥ a. Discharge (65) ». Total head Hy _ dint a (6) © Power (70) 4. Rotational speed mond wy ad a Eliminating the diameters from equations (68) and (69), leads to re- lationship mV _ nav Ae a oe ‘equivalent to the condition of equality of the kinematic specific speeds (tgs = (tae 73) Eliminating the diameters from equations (70) and (68), we have ) (73) the simitarty of the dynamic specife speeds (rh = Gur) 8) The expressed in the conditions for flows through impeller pumps, may be lowing way Two flows of liquids of different densities through two geometti similar impeller pumps are dynamically 1. if the Kinematic specis speeds mq in both flows are equal and the dynamic specific speeds mp are proportional to the square roots of the densities of the liquids. If both pumps deliver liquid of the same density, the dynamic similarity of the flows is characterized by the equality of both specific speeds. kinematic and dynamic Eliminating the impeller diameters from equations (68) and (71) gives \ 7 alam Hy ens @) ‘The correctness of these relationships may easily be verified by inserting the ratios Qx/Q and H,/H, into equation (72). When comparing flows of liquids of different densities p_ and p, through a model pump M and a real, full-size pump R, the equations (73), (75) and (76) take the form: (n= (aa 9) (tm _ (tar aa Vex Ver (2a = Case ci) 39. Discharge 4 Physical analysis. In order to deduce the structural formula, defining the relation between the discharge of a pump, the nominal diameter and speed Of the impeller and the total head, may be taken as the starting point the general physical equation: FE. d, Q,m.1) =0 (2) where » denotes the kinematic viscosity of the liquid. b. Dimensional analysis. It may be assumed that the quantities E and d are independent variables and set up the three dimensionless products n= E8a%Q 183) ay = EX (84) a= Bay 5) 118 ernie Pwr the algebraic equations for the exponents, we obtain the dimension- less products in the form (86) en) 8) aE “The physical equation 82) corresponds to the dimensionless equation (89) Inserting the values (00) and hence the discharge onevie. aa) oor in the equivalent form * ) o The product 94) is proportional to Froude's number. r 1 OF DYNAMICAL SIMILARITY OF IMPELLER PUMPS 1 The product dey? _Q_ _@ cd? _ed 7 =LatituR 09 7 aygh de dv if mains js proportional to Reynolds number. The product 9 represents new dimensionless specific speed called the shape mumber. Equation (90) may thus be written in the form nx)=0 on cor in the equivalent form Qa AY GH Erp, P) (98) Neglecting the influence of Reynolds number on the character of the flow through the impeller leads to nj) =0 99) lw “) = VETO) (100) S.11. Relation between similar ty mambers and specific speeds Writing the general dimensional formula (89) in the form 1 = fC Me) ives the relationship F = fm) (102) en the similarity numbers of pumps with different specific speeds, ince bet ny = Ne fen in the form F =H, me) (104) Equations (102) and (104) are expressed in the following way: Ws through geometrically similar impellers of rotodynamic Froude’s and Reynolds numbers are connected by a univocal function with the specific speed. equation (102) may also be wi 120 apetuem Puss For two pumps of the same specific speed F=lR (405) Froude's number is connected by a univocal function with Reynolds number, 5.12. Relations between the characteristic aumbers ng and mp and the shape number my ‘The dimensionless shape number was defined in equation (96) ne Hy in which the speed n is expressed in sec~!, Q in m*/sec, g in m/sec* and H in metres. Ta order to avoid very small values of my this equation may be writen in the form _ 1000 ny "=e Ge (106) where i in cpm Introducing the standard value of the acceleration due to gravity g, = = 980665 msec, we obiain the practical formula ng = 05S = 3g (107) hence Ing = 0822 rap (108) fap = 1215 ray (108) nig = 0333 my io) In the English system of units: Je = 365 (we ay The value of m, for given parameters of Q, Hf and 7 isthe same in any system of units, ie. ia the mettic or any English or American system, pro- Applying for instance the MKS ‘system, the discharge is expressed i m/sec, H in metres, m in see"? and g in m’sec*, Similarly, in the English system, the discharge is expressed ia fi8/sec (not in gallons per minute), H in ‘and g in ft/sec 1e number also I to standardization Day OF DYNAMICAL SIMILARITY OF IMPELLER PUMPS Ra 5.13. The nature of similarity numbers in the theory of impeller pumps It should be noted that there is a fundamental difference between the iy numbers for flow in pipes of constant cross-section and the similarity numbers occurring in the theory of rotodynamic hydraulic machines. For flow in open channels, Froude’s number is the criterion of transition 1m tranquil flow to rapid flow and for flow in closed conduits, Reynolds mnber univocally defines the state of turbulence, the critical value of Reynolds smber being the criterion of transition from laminar flow to turbuleat flow In the theory of impeller pumps, both these similarity numbers are of conventional nature, since their numerical values depend on what impeller imetet is taken as characteristic, In particular, the Reynolds number delined by equation (63) or (64) does not define the type of flow through ‘a pump, but is an independent parameter to which we can refer the results assuming conventionally that the outlet iameter of the impeller. BIBLIOGRAPHY H, Davts, Influence of Reynolds number on the performance of turbo-machinery. Trans ASME, Vol. 73, 191, G.G. MacDosato, Dimensional analysis of centrifugal pumps. The Enginer, Vol. 167. 938. The principle of dynamic similarity, with special reference to model exper- r.Yol. 117, London, 1924. E, Hastorzet, Das Aholicheitsgesots fe Steémunssprobleme. Selweserische Bauzetmne. Bd, 65, Zurich, 1947, A.J, SterANorr, Centtfusal pump performance as a function of speciic speed. Trans. ASME, Vol. 65, New York, 193." A.T, Troskotassr, Uber die Vereinheitichung der Schnelitufgkeitszah! von hydraulic Schen Kreiselmaschinen. Moschinenbautelnik, Heft 11/62 CHAPTER V IMPELLER SHAPE 1, Geometrical velocity fields tn medyfow trough a mls pump the flow ete pend ty on the shape ofthe impeler and ifs: pusngs and onthe shaping of the i tho depen onthe ped wih te mpfr at gut ey Bow tag i ise te stem sis ae fandom sa, bt ms fim a connon Stem of eam suas i ofthe blader in sich aye, we regard the sues ofthe la as stream surfaces bounding the flowing liquid. = (a) 0 Fr. 1, Steam sic in ota me pumpin (i ea pump, (0) naga pump) in pe ler pumps (H, AoDison), Figure 1is a diagrammat representation ofthe shapes of 7 in the impale oft (@) cata! pumps, (9) hele! es pumps, (4) propeller pumps. ft The steam surfaces inthe region beeen the inlet and the blades may bave one ofthe following shaper (@) plane surface witha ciculatoulne (Fg La) perpendicular to the impeller axis (eentrifu low it as igal pumps of low specific speed with blades of single (b) a bellshaped helical surface (Fig. 1b), whose axis of : ig. 1b), axis of symmetry «¢ incides with the impeller axis. (centri feed ng mrlle axis Cemetual pumps with blades of double t edges of Y, IMPELLER SHAPE 123 (oa diverging conical surface (Fig. 1e), whose axis coincides with the impeller axis (diagonal pumps), {da eylindrical suface (Fig. 1d), whose axis coincides with the impeller axis (propeller or axialfow pumps). In view of the direction of How through the ims names (1) radial-ow impeller (Fig. 12), (2) mixed-low impeller* (Fig. 1b and Fis. 1c) G) axiahfiow impeller} (Fig-16) Both the stream surfaces illustrated in Fig. 1 and the directions of flow throush the impeller are of an approximate nature, their purpose being to te the classification of impeller pumps. In reality, the velocity fields eviate somewhat from the diagrams shown in Fig. 1. ‘Depending on the construction, impellers may be divided into the follow- ing types: (1) closed impellers, occurring in centrifugal pumps and some types of helical and diagonal pumps; these impellers have two shrouds, front and back, between which the blades are fixed (Fig. 4 Chapter XIN), (©) open impellers, occurring as a rule in propeller pumps and in almost all types of helical and diagonal pumps; these impellers bave only a hub fand a back shroud, to which the blades are attached or in the ease of propeller pumps, there are no shrouds at all and the blades are only attached to the hub (Figs. 2 and 3 Chapter X1ID. ‘Open impellers are also used in centrifugal pumps for special. purposes er, we use the following 2. Evolution of impeller shapes It follows from considerations in previous chapters, that the shape of the impeller passages depends on the basic parameters Q, H and n, which ace represented by the specific speed of the impeller (see Chapter 1¥)- Let us consider what changes of shape the impeller will undergo, if the speed of rotation 1 is altered while keeping Q and H constant. In these cit- cumstances, the value of the approach velocity ¢, and the outline of the inlet edge remain unchanged. + The term “mixed-slow pump” doesnot refer to the random turbulent mixing of radial meridional velocity through the impeller has both radial ing is designed 50 as to 5 i nos i widely used in English irate, one encounters the ambiguous term * Wt swiat components and the bi * The term axialfiow German 24 sneccen rents Figure 2 shows the profile of @ ra the characteristic dimensions d,, 4 From the fundamental equation (III, 26) Hy =a D ; Oy) A Fic, 2, Evolution of centrifugal impeller shapes whil inversely proportior it follows that the velocity 1, depends on speed of rotation n. The outlet diameter independent of the ton eller passage ably shorter, the suction performance and efficiency of the pump will be reduced. In order to correct necessary to lengthen the blade in the direction of the inlet, so that the inlet edge takes up the po- the broken line ——— represents the h a high specific speed). wwe reduce the total head Hi, the impeller ‘ain Fig. 3 (helical pump) or Fig. 4 (propeller vs wrstase sire 125 The properties of impellers with profiles shown in Figs. 2.3 and 4 may = 30-90) — low rota head (Fig. 2; blade edges shown tional speed, small discharge and large to by full lines) ee: Fro, 3. Diagram of helicaltype impeller Fio, 4 Diagram of propellertype impeller. Il, Francis-typel impeller of medium specific speed (4p = 90-150) — medium rotational speed, medium discharge and’ medium total head (Fig. 2: blade edges shown by the line —-—-—-—) f high specific speed = 150-240) — np = 405-640) — bi total head (Fig. 4) Table I shows the pr and propel je performance characteristic curves H = £(Q) and P = (0) As can be seen from T: the smaller the ratio of ie speed, charac- angles and the higher the specific speed of the xt and inlet diameters dyfd, becomes. name is used on account ity ofthe impellers of types I and TIL runner of the ine invented by J.B, FRANCIS. 126 seein PuMes ‘TABLE I. Proms oF SINGLEANERY IMPELLER PUMPS n areal eroties ety gis SS ae = ay = 30-20 Draraceiis| gm J0-50 p20 = 050 ‘ma 50-80 p= 200-300| ny =180-280 ‘ng 80-150 ‘p= 300-550 ae 0-450 (ag =185-900 e—_ | trp fe 500-1200 sr) ee aa pellers connected in series the specific speed ed on the basis of the head per stage. When impellers are connected is taken, remembering at the same les, It is therefore necessary to alu itz for the complete pump. The specific speed of a pump consisting of # impellers of specific speed rn, connected in series, is defined by the formula @ inmate sane 127 3. Choice of number of impeller blades not possible to calculate the number of impeller blades exactly. In wwe use approximate formulae, Fra. 5 streamline 4,4, and fy the mean angle of the blade corresponding to line Ay 4y ‘AS an approximation, we take By and fy are the inlet and out and hence o 128 ELLER Pues In propeller pumps = ae in By (8) where r denotes the outer radius of the blades and k = 45, These formulae make it possible to calculate the number of impeller blades for clean liquids. For liquids conveying solid m between I and 3, in order to obta + the number of blades usually lies large flow-sections. 4. Range of application of impeller pumps AS the specific speed 1, decreases, the impeller passages get progressively longer and for small discharges, they also get too nat y large. As a result, the hydi mn of the shrouds rotating of the input power of the pump and lower its efficiency. The lowest value of the specific speed m, is thus dictated by economie factors. ‘On the basis of experiments carried out on low specific speed impellers, it has been established that the lower limit of specific speed is Hence, the range of apy specified by the in iq EB 510 (20) Vave hence ay a2 3) aa) Ex. from equation (14) and for Q= head Fingy = 20 m = 0001 msec, the maximum rots! \y, neLLER SHAPE 129 _ Relationship between the overail efficiency », the hydraulic efficiency % and the specific speed , efficiency overall efficiency ‘a the basis of measurements of the discharge ,, taken during a pump test, In order to cal fhe motor of a projected pump, an experimental value Jency from a pump of the same type, and having a similar specific speed and discharge. The values of » with a discharge of Q 1 I/see, up to ~ 90% depending on m,, the type of pump, its si ingl almost invariably determined experimentally bead H, and input power fe the power required for assumed for the a volute casing. This Jide the pump and the small number of changes in the direction tage pumps generally have lower efficiencies. ‘The overall efficiency rises as the specific speed increases, mainly due to ‘a reduction ‘of the friction losses in the shrouds and the drop in volumetric TT T5830 idiae 30 see 130 ernie PuMeS ins a maximum at a certain value of n,, which is diferent for each type of pump (centrifugal, helical, diagonal or propeller). Increasing n, still further produces a reduction in 7 due to the rise in hydraulic losses. Figure 6 represents the relationship 7 = f(g) for single-entry impellers of different types. ‘5.2. Hydraulic efficiency The hydraulic efficiency ny in contradistinction to the overall efficiency 1 cannot be determined directly from test measurements on a given pump. Tt may be calculated from the overall eficiency wwe know the volumetric and mechanical losses. The maximum values of hydraulic efficiency 1, the limits ~ 65 to 96%, depending on the value of n,, the size, type of pump and quality of manufacture. A.J. STEPANOFF gives the following maximum values of 1 for large modern single-stage double-entry pumps. 18 3 zi | 28 | as | soo 38 and small pumps and less carefully made of course be lower. 92 For older types of me pumps, the values of 7 BIBLIOGRAPRY F. Knssau, Die Grenzen der Verwendbarkeit der Kreselpumpen. KSB-Sonderabsnick Frankenthal. a ‘aisles. CHAPTER VI IMPELLERS WITH BLADES OF SINGLE CURVATURE 1. General remarks on the design of impellers follows that the total head generated by 2 pump depends on many variables, such as the peripheral velocity up and the meridional velocity cy at the impeller outlet, the blade angle A, the number of blades 2, the ratio Gels and the rat a The same total head may be at pel bbut having a greater angle 8; and a gre problem of calculating the dimensions of an imy hhead may have several sol when considered from the point From Euler's fundamental equati but they are not of efficiency and production costs. There is only Euler's equation to help in solving this problem so it is ary to assume the values of some of these variables. Optimum resul ‘on the specific speed n, and the head H (as in the design of bine). 12 meridion: inlet and outlet may be chosen mt = Kowa) 22H @ ich are discussed in Chapter XXIL. 13

You might also like