Title Author(s) Citation Issue Date Performance Tests of a Centrifugal Pump Fukusako, Shoichiro; Arie, Mikio 北海道大學工學部研究報告 = Bulletin of the Faculty of Engineering, Hokkaido University, 49: 45-58 1968-09-25 DOI Doc URL http://hdl.handle.net/2115/40894 Right Type bulletin (article) Additional Information File Information 49_45-58.pdf Instructions for use Hokkaido University Collection of Scholarly and Academic Papers : HUSCAP Performanee Tests of a Centrifugal Pump“ Shoichiro FUKUsAKO** and Mikio ARIE“* Abstract The results of performance tests of a centrifugal pump in the cavitation region were presented in terms of the variation in head, driving power and eflicieney. The effect of air leakage in the suction system was also examined. lt was found, in the present experiments, that the existence of a prerotation showed its effect of im− proving the cavitation performance of a pump. 1. lntroduction The present paper presents some experimental results on the performance test of a centrifugal pump installed in the system of a cavitation tunnel recently con− structed at the Fluid Mechanics Laboratory of the Hol〈kaido University. Cavitation characteristics of the pump were first examined in terms of the reduction of total heacl and efficiency as well as a trend of increasing the driving shaft horse power in the cavitation region under a constant load of discharge quantlty. It is generally accepted, these days, that an introduction of a certain amount of air to the upper part of the draft tube of a reaction type of turbine, diminishes unfavorable features of the cavitation phenomena. On the other hand, the leakage of air in the suction system of a pump causes considerab1e hindrance in the maintainance of its function in raising liquid to the requirecl level. With special regarcl to this, the infiuence of air leakage in the suction system on the perform一 .ance of the pump was examined as a second step. lt was demonstrated that air leakage is invariably accompanied by an unfavorable effect which results in a drop in head, an increase in the required shaft horse power which leads to a low一 ・ering of efliciency. Finally, the prerotation at the suction eye of the impeller was estimated, basecl on an ideal ve1ocity triangle at the inlet of the vane to investigate the relation between cavitation characteristics and tlie valute of prerotation. The majority of the results were preparecl in acceptable dimensionless forms for the purpose of providing material for the model study of a pump. *The original paper was正,resented at the Japan−U. S. Seminar on Similitude in Fluid Mech− anics at The lowa lnstitute of Hyclraulics Research, University of lowa, lowa City on September 25, 1967. ”・ *’ Faculty of Engineering, Hokkaido University, Sapporo, Japan. 46 Shoichiro FUKusAKo and Mikio ARIE Nomenclature Ga: flow rate of air (gr.fsec) Giv: flow rate of water (kg/sec) H: total head of pump (mAq.) Ha: total head of pump when air is introduced into the suction system (mAq.) 瓦.: net positive suctive head (mAq.) 瓦: shaft horse power (horse power) 2: capacity of pump (m3/sec) g: acceleration of gravity (m/sec2) n: revolution of impeller (r.p.m.) p: density of fluid (kg・sec2/m‘) 77 : efliciency of pump (r2H!75 N,) Z{1 : peripheral velocity at the inlet (m/sec) Zt2 : peripheral velocity at the exit of impeller (m/sec) 刀?ノ己0: avera.cre velocity of fluicl in suction pipe (m/sec) ¢: capacity coeflicient definecl at the inlet (v,.o/zti) ipft : fiow rate ratio between air and water (G./G,.) φ,‘。乙; capacity coethcient for the mixed liquid of water ancl air (¢,,.,,==¢.・¢> g/T : head coefficient (H!u:/2g) gba : head ratio (H./H) c: shaft horse power coefficieRt (Ni/PA2u;/75) ・・縄薦・ suction specific speed Z. Experimenta1 apparatus A schematic view of the circulation type of water system is given in Fig. 1 to show the arrangement of water tank, centrifugal pump to be tested, electric dynamometer, vacuum pump, pipe lines and manometers. The capacity of tank is 6.63 m3, which is suthcient to fill the tap water available at the laboratory for the whole system. The water level in the tank was carefully watched to maintain a free surface in such a way that the vacuum pressure there evacuated by a wet type vacuum pump could be adjustecl to give various values of tlie net positive suction head to the pump. BaMe plates are provided in the tank in order to avoid the direct infiow of the cavitation bubbles into the suction system. The specifications of the 150 mm single suction type of centrifugal pump are H:==7m, Q :2・8m3/min, n=960 r.pm, and n,,(m・m3/sec)=47.6. The shroud of impeller and the casing cover at the suction side were transparent so that the cavitation bubbles in impeller coulcl be observed by the aid of a stroboscopic flash light. The main dimensions of the impeller are shown in Table !. All pressures were read on the mercury co!umns of various manometers, and the capacity of pump was measured on an orifice meter. Fig. 2 shows the main dimensions of the impeller tested. 2 3 47 Performance Tests of a Centrifugal Pump MO VACUUM PU}([P o噂 1860 1 寸 ou OR工F工CE BAFFIiE PULTE 肛晶 一il 一一一 o酷 DELIVERY VA工,VE 頃 ゼ d oい o oり ひ .副68・tiI TA.eHO)E]EWER i MESM PUMP 250 :E}玉EC田R工C DYNAMO琵:E歴ER Fig. 1. A schematic view of experimental apparatus. Table 1. Mean dimensions of the impeller testecl. exist diameter 1)?. == 270.0 mm inner di”c meter ( g: lllg 111 Sta:”.,S:ii:,ShrO”d Die= 152.0 mm Dl,・ =:: 1OO.O mm urjdth of exist’ b2 一一 45.0 mm width at inlet bi =65.0 mm vane angle at the exist β2=20。0/ ノ量鑑∵h…d 25。30/ 32027, 39。25/ number o/f vane 6 ijト…45.謂 i slxh.y..{. mp. e / ノ ;/一 一一、 1 イ/ t i 1 1鼠 ト .卒.... ’ ttt 狽狽狽煤@ttttttttl o .巾一 @トGD qD oc LA l l E i l Fig. 2. 鵡モ oo −uへト マー マ楢 ト…9Q_…’T Dimension of impeller. s. 48 4 Shoichiro FuKusAKo and Mikio ARIE 3. Cavitation performance Six series of experiments were first undertaken for different constant speeds of 700, 800, 900, 1000, 1100 and 1200 r.p.m. by changing the discharge quantity to examine the nature of similarity in performance curves in the cavitation−free region of practical operation. The results are plotted in Fig. 3 in terms of head coefficient, capacity coeMcient, power coefficient and pump ethciency. As may be O.6 ,1’ サ ? ’sO. O.5 1。轟麟・紳麟 O.06 e 鴎璽寧。 厨・ββ e8 O.4 O.05 麟需螺曽 評 を〆 調ヂ e Oe5 慰6Cb .6 ノ o n = Oe2 O.1 O.04 礒讃講 Zkaps” が e e o ・癖 700エ・.pe】臼コ。 800 900 1000 0.0う M OO 1200 ee O.02 包 O.o o Fig. 3. o.ol Oel .o Oe2 0。う 005 Oo4 ¢ Head−capacity, power−capacSty and eesciency−capacity curves in normai operation, seen in this figure, the similarity of performance can be accepted for various speeds of pump now selected. Therefore, any further results to be obtained for these pump speeds by changing the parameter concernecl was expected to show the effect of this particular parameter. The cavitation and its effect on the function of a pump may be defined in several ways: the production of cavitation bubbles somewhere insicle the pump, the generation of cavitation sound and vibration, the degree of cavitation erosion on the material, the drop in head and capacity, etc. However, there is no doubt that the value of suction head governs a larger part of the initiation o’f these cavitation features. Fig. 4(a) and Fig. 4(b) are give to illustrate the change of Performahce Tests of a Centrifugal Pump 5 49 15.O H(mAq) 2 1う。0 11.0 1.0 9.0 O.8 7.0 O.6 5.0 O.4 0.2 4.O o.o 1.0 2.0 5.0 4.0 5.0 6.0 7.O Hsv (mAq) (a) 9.0 Ni(HP) 8.0 D 7.0 」瓦。 c B 6.0 1 5.O O.O 1.0 2.0 5.0 4.0 5.0 6.0 7.O Hsv(田Aq) (b) Fig. zlt. Definition of the chac rac cteristic points in cavitation region (ノド1!00r.p.m.,(2=32.5331it/sec>. characteristics, based on the interests of mechaRical performance of the pump, obtainable by loacling the same capacity of discharge for a constaBt speed of impeller and by changing only the suction head for the purpose of clearly exag− gerating the effect of cavitation. Fig. 4(a) is a plot of heacl against various values of net positive suction head. As may be seen in this figure, the head starts to lower at A when the suction heacl increases to a certain value and finally it falls suddenly at B where the pump fails to lift water any more. Since all the operating parametrs except the suction head are kept constant, AH. may be con− sidered to be the head loss due to cavitation. ln the same way, the shaft horse power begins to increase at C in Fig. 4(b) and attains its maximum value at D, and the pump efficiency begins to decrease at E. AN. in Fig.4(b) may be considered to be the increase of shaft horse power due to cavitation in the same way as AH.. A series of experiments were undertaken to reveal the general trend of these characteristic pojnts for four different speeds of 900, 1000, 1!00 and 1200r.p.m. (The experiments for 700 and 800 r.p.m. are discarded, because these 50 6 Shoichiro FUKusAlsco and Mikio ARIE 15.0 H(m) 15.0 Q = 17・469(lit/$ee) 19.808 22.819 26.227 11.0 29e157 32.359 55eO77 9.0 37.641 7.0 5.0 o.o 1.0 2.0 5eO 4.0 5.0 6.0 7.O H。v(皿Aq> Fig. 5・ HeadTINPSH relation (n = !200 r,p,m.). 10.0 mat37641(lit/sec) Nl(HP) う5.077 52.159 9.0 29.15 26e227 8.0 22p819 19.808 7.0 17.64 E 6.O o.o 1.0 2eO 3.0 4eO 5.0 6.0 7.0 H..(mAq) Fig. 6. Shaft horse pox・ver−NPSH relation (n := !200 r.p,m.). characteristic points could not clearly be detecteet in the present installation). The results for !200 r.p.m. are shown in Figs. 5, 6 and 7, as an example. Two characteristic natures were founcl in these experiments in the cavitation region produced by decreasing the net positive suction head. One is a decrease in total head AH. in Fig.4(a) that gives the loss of power AN.h wt raH. Q/75 in the cavitation region, ancl the other is an increase in shaft horse power inclicated by dN. in Fig.4(b>. The sum of these two powers dNi may be considered to be the loss of power due to cavitation: 7 51 Performance Tests of a Centrifugal Pump 1.0 Q 冨 う7.641(1it/sec) 35.077 ? @ う2.159 O.8 O.6 Q = 29。157(:Li七/sec) 26.227 22.819 19.808 O.4 O.2 @ 17.469 o.o o.o 1 1 1.0 2.0 5eO 4.0 Z5 ,O 6.e 7.o Hs.(mAq) Fig. 7. Efficiency−NPSH relation (n=1200 r.p.m.). ¢= O.008 A NI A・茎/75 O.1427 0.1259 0.1644 0.1890 0.2う17 O.006 0.2527 0.2712 O.004 O.002 o o o.ooo Q.o O.2 O.4 O.6 O.8 1.0 1.2 1.4 Hsv/(u?/2g) Fig. & Dimensionless plot of power loss vs, NPSH (n= 1200 r,p.m.). ∠1!>}=∠12\L,+∠1八1,i、,=∠3Nc+γ、イ.H,,(2/75 where, AN. is positive when it increases and r(2AH./75 is positive when it decreases. The power loss of cavitation thus computed for n== 1200 r.p.m. is shown in Fig. 8, as an example. lt may be seen in this figure that the loss in the cavitation region shows a trend to increase with the decrease of the net posi− tive suction head vLThen the load of discharge is positively kept constant. This fact may be unclerstood in two ways : 52 8 Shoichiro Ful〈usAI〈o and Milcio ARIE (a) The production of cavitation bubbles resulted in an increase of the apparent specific volume of the fiuid, and hence the velocity triangles at the inlet and exit of the impeller were deformed to be accompanied by a reduction of the hydraulic efllciency. (b) lnsofar as insistance of the experimental technique of keeping the dis− charge of the pump constant, the increment of the power exceeding the drop in head must be consumed in some way directly related with the production of cavitation bubbles. 2000 10000 F ・穫 可,。。。 礪 1000 d 1100 0 1200 e 5000 1000 ct sv 1100 e l200 e 500 n ; gOOエ㌧p,皿・0 朝 n = 900r.p.m. 0 2000 1000 200 500 O.5 1.0 1.5 2.0 2.5 O.2 Hsv/(u?/2g) 0。う O.5 .0.6 O.7 O.4 Hsv/(u9/2g) (a) The point where the head beging to drop.(A) (b) The point where the pu:ロping ac七ion s七〇ps。(B) ・凡 可 sv 2000 1000 n = 900r.p.m. 0 礪 1000r.p。皿. d sy 10000 BS/.4 n; 900r.p.m. 0 5000 1000 づ 2000 llOO e 1200 e 1100r.p.m. c 500 o 1200r.p.m. e 1000 200 O.5 1.0 1.5 2.0 2.5 H../〈u?/2g) 500 O.4 O.6 O.8 1.0 1.2 1.4 1.6 H,./(u?/2g) (c) The point where the shaft borse power begins to increase.(C) (d) The point where the shaft horse pover attains i t$ maximum value.(D) 蠣 藩 2000 n = 900r.p.m. 1000 500 o 1000 d 葡100 o 1200 e o 200 O.5 1.0 1.5 2.0 2.5 Hsv/(u?/2g) (e) The point where the effieieney begins to lower.(E) Fig. 9. Suction specific speed vs. characteristic points, 9 Performance Tests of ftc Centrifugal Pump 53 Perhaps, these two processes of energy dissipation may be combined together to increase the loss of power in the cavitation region w1th the decrease of the net positive suction head. The distribution of the characteristic points indicated by A, B, C, D and E is shown玉n Fig.90n a plane of dimensionless net posi伽e suction head.砿,ノ(凶2σ) and suction speci五。 speed /2へ/一瓦7(私の髪。 since the characterist三。 po玉nts on the characteristics iR the cavitation region represented by A, B, C, D and E are fairly well distributecl respectively on one curve, as may be seen in these figures, the mode of the plots may be useful in the estimation of the limiting value of the suction height to be given to a centrifugal pump. 4. 〕B任ect of air leakage in the suctioh system of a centrifugal pump on its performance Fig. 10 shows a schematic view of the air injection clevice at the suction pipe. Thirteen holes of 4 mm diameter were drMecl as the air injection ports in a brass tube of 9.8 mm diameter inserted in£o the suction pipe along its diameter. The clischarge load and the speed of the pump were again kept constant as in the previous case of cavitation test ancl a certain amount of constant weight of air was given through these holes from the surrounding atrnosphere into the pump for each seyies of experiments to measure the total head and the driving horse. power for various values of suction head. The experimental results obtainecl for n==!200 r.p.m. are sbown in Figs. 11, !2 ancl !3, as examples with respect to the variations of head, shaft horse power .and ediciency, respectively. The relation between the head ratio H,,/H and the 瓶 。 い s一“ 69 ORIFICE 10 璽 」乙_。A工R SUPPLY 4夢 Fig. 10. Air injection device. 54 10 Shoichiro FUKusAI〈O and Mikio ARIE 15.O 1 H(mAq) 15.0 ,ノノ 11.0 G = o.ooo(g/s) a O;079 o 0.147 0.185 9.0 0.255 一 7.0 0.399 0.480 0.546 0.629 〇。う24 0.754, 5.0 o.o 1 1.0 2.0 5.0 4.0 5.0 6.0 7.0 Hsv(mAq) Fig. 11. Head−NPSH relation with air injection (n =: 1200 r,p.m., G!v =: 29,157 kg/sec). 10.O Ga 7.8:80,8(g/sec) Ni(Hp) O.147 9.0 o 8.0 Ga 二 8:;}含;(9/sec) 0.う24 0.599 0.480 0.546 0.629 7.0 6.0 o.o 1.0 2.0 5.0 4.0 5.0 6.O 7.0 Hs.(mAq) Fig. 12. Power−IXTPSH relation with air injection (7t 一un !200 r.p.m,, Gzv == 29.!57 kg/sec). weight ratio of air and vLTater G.IG,. is plotted for n nt!200 r.p.m. to show the effect of air leakage on pump characteristics. The introduction of air in the suc− tion system shows no sign of improving the nature of cavitation at least on the characteristics of a pump. However, it can clearly be seen that even the same quantity of air leakage gives a different value of effect on the pump performance depending on the value of the net positive suction head required. This result must be caused by the difference in volume occupied by the air bubbles−when the suction head is large, the air bubbles may be considered to expand under the 11 55 Perforrnance Tests of a Centrifugal Purnp O.7 Ga = 7 W:898(g/sec) O.1.4 O.6 !ノ 一 b/O Ga ” 8:Ellg(g/sec) O.5 O.う24 0.599 0.480 0.546 0.629 0.754 O.4 O.5 o.o 2.0 5.0 4.0 5.0 6.O 1.0 7.0 Hsv(mAg) Ecaciency−NPSH relation with air injection Fig. 13. (n= 1200 r.p,m., G?v= 29.!57 kg/sec). 1ow pressure of the surrounding water. The head rat圭。耳、/H obtained in the present investigation is given圭n Figs. 14and!5 to show the effect of air leakage on a pump performance in a piot of き (玩/H)vs.(G、/G,v)and(n>G,、φ騙/私冨)・(凡/H)vs.φ,、“、. As may be seen in thls figure, the effect of air leakage in the suction system of a centrifugal purnp can be given rather pert三nently in a customar三ly acceptable dlmensionless form. Since all the parameters to describe the pump operation wlth an a圭r leakage are taklen into cons量derat圭on, thls manner of expressing the characteristics of a pump can be used for the estimation of pressure drop when a mixture of air and water is pumped. 1.00 Ha/H glsv .= 6・25(mAq) .90 5.54 .80 4.7 o ,9P .70 .60 .60 2.84 唱.16 1.481 .66 1.94 .50 .40 2.62 2.48 5・4S 3・18X 5・35 30 . o. O.005 O.OIO O.OG 5 O.020 O.025 Ga/Gw F圭9・14・ Head ratio vs. fiow rate ratio (n=12eO r.p.m., Gv = 29.157 kg/sec). O.050 56 12 Shoichiro FuKusAKo and Mikio ARIE 150 o e el 120 ・摩 sv 譜睡憩幕轡●. H 論 H o e e 100 e e g oe e 蕊 e ¢= 0.14う e O.210 o 0.う08 e’ eo 80 eee @o e 60 40 e o O.OOI O.00う O.002 O.004 O.005 O.006 ¢am ヨ Fig. 15. (ノz》GαφノH謁)(Ha/H) VS, (PS a7ie一 5+ Qualitative nature of the effect of prerotation on cavitation characteristics of a centrifugal pump An lmpeller of pump is usually designed to give ai==900 at its rated capacity so that the flow of fluid onto the vane would not have the componet of prerota− tion from the standpoint of an ethcient momentum transmission between the impeller and fluid. However, there exists a certain amount of prerotation when the oper− ating point deviates from the design point as will be expected on the velocity triangle at the lnlet of the impeller: the component of prerotation is ’in the direction of the rotation of the impel!er when the capacity is less than the rated discharge, and it is in the reverse direction when the discharge demand exceeds DES工GN POINM FOR CAPAC工{PY OPERA〔P工NG PO工N[r FOR SMAL工,ER C.A:PAC工{PY ・{量 Cl Clα1 wl c血雪 F FOR A SMAエ江ER CAPACI[PY U1 THAN RATED DISCHARGE uT c己1』. ヶL1\ COMPONEN〕摩 OF PRE PRIEROTA or Z ON ltLIHEN THE DISCHARGE・工S 五ESS 〔EHAN:DES工GN CAPACITY Fig. 16. Velocity triangle at the inlet. 13 57 Perforrnance Tests of a Centrifugal Pump 1.0 ’1.0 .. O’9 O.9 〈rO.8 eNNgoo 7e 0づ O.7 oS ミ。・8 や・々P 蕾。.7 や・ta ・7罐・.. O.6 O.6 ’・鴫・ O.5 O.5 O.4 O.5 O.4 t.o 0.2 1.5 2.0 2.5 OJ H../(u?/2g) O.4 O.5 O.6 H,./(u?/2g) (a) [the point where the head begins to (b) rhe point where the pumping action drop.(A) stops.(B) 1,0 1.0 rO.9 .一〇・9 ×” .go.s k’Ngoo 黶Z.8 評 O.7 O.7. O.6 O.6 O.5 O.5 O,4・ O.4 鳥%・ ’e 1/Obさ 1 O.7 1.0 1.5 2.0・ 2.5 O.7 1.0 ’カ 1.5 ・¢ 2,0 2.5 H.,/(u?/2g) H・v/(u?ノ29) (’c) Mhe point where the shaft horse power begins to increase.(C) (d) The po±nt where the efficiency begins to lower.(E) Fig. 17. Effect of prerotation on characteristic points in cavitation region. the rated capacity as is shown in [Fig. 16. The existence of these types of pre− rotation can easily be detected by an observation of the fiow in transparent suction pipe. Since there might occur a certain amount of shock loss at the inlet when the operating point is different from the design point, a quantitative consideration, of the prerotation can not be expected in the present experiment. However, a qualitative effect of prerotation on the pump performance is derivable when the component of shock velocity at the inlet is neglected. By constructing the velocity£r三angle at the inlet, the peripheral component of inflow c.i was estimated for each series of experiments, and the results are shown in dimensionless form in Fig. 17 to show the relations with the characteristic points A, B, C and E on cavitation performance. As may be seen in this figure, the increase of prerotation results in a movement of the characteristic points to a smaller value of the net positive suction head. Therefore, it may be permissible to conclude that the existence of a prerotation in the clirection of ui has a trend to improve the cavitation performance of an impeller of a pump. 58 Shoichiro FuKusAI〈o and Mikio ARIE References 1) 2) Krisam, F. : “Neue Erkenntnisse in Kreiselpumpenbau,” V.D,1.一Z., Vol. 95, 320 (1953>. Kasai, T. and Takamatsu, Y.: “Studies of cavitation aspect and suction performance of centrifugal pumps, 3rd, 4th and 5th reports,” Gn Japanese), Trans. 」.S.M.E., Vol. 29, 1294 (1963). 3) Minami, S., Kawaguchi, K. and Homma, T.二“Experimental study on cavitation in centri− fugal pump impellers,” (in Japanese), Jour. J,S.M.E., Vol. 62, 881 (1959). 4) Gonger, G. A. : ‘CA theory of cavitation flow in centrifugal pump inducer,” Trans. A.S.M.E., Vol. 63, 29 (194!), 5> Wislicenus, G. F. : “Test stand for centrifugal and propel}er pump,” Trans. A.S.M.E., Voi. 64, 619 (1942). 6) Siebrecht, W. : “Untersuchungen ttber Regelung, theoretische und wirkliche F6rderh6he von Kreiselpumpen,” V.D.1.一Z., Vol. 30, 87 (1930). 14