Estimation, of the Starting Torque, of Refrigerant Rotary Compressors

Transcription

1 Prde University Prde e-pbs International Compressor Engineering Conference School of Mechanical Engineering 198 Estimation, of the Starting Torqe, of Refrigerant Rotary Compressors S. Nagatomo S. Kato Follow this and additional works at: Nagatomo, S. and Kato, S., "Estimation, of the Starting Torqe, of Refrigerant Rotary Compressors" (198). International Compressor Engineering Conference. Paper This docment has been made available throgh Prde e-pbs, a service of the Prde University Libraries. Please contact epbs@prde.ed for additional information. Complete proceedings may be acqired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at Herrick/Events/orderlit.html

2 ESTIMATION OF TBE STARTING TORQUE OF REFRIGERANT ROTARY COMPRESSORS Shigemi Nagatomo.. Consmer Prodcts Engineering Laboratory, Toshba Corporaton, kishima, Kawasaki-k, Kawasaki-city, Japan Shnichi Kato Manfactring Engineering Laboratory, Toshiba Corporation, Komkai, Saiwai-k, Kawasaki-city, Japan ABSTRACT Estimation of starting torqe is very important for motor design and for improvement in compressor efficiency. Especially, when compressors are restarted after a period of inoperation, motors sometimes meet large starting torqe. The assmed reasons are as follow. The compressors temperatre is sometimes lower than any other parts of the refrigerant cycle. In this case, the refrigerant tends to collect in the compressor and is condensed there(l). So, the oil between the lbricated srface is dissolved in the refrigerant and removed by this refrigerant liqid. Frthermore, the nbalanced magneti8 pll force in indction motors make the starting torqe larger. An experimental model has been developed for measring torqe. Using this apparats, torqe was measred at varios pll forces, varios oil conditions, and with variosly treated shafts. A mathematical model was made for analyzing the starting torqe. Based on this analysis, it is possible to predict individal torqes nder varios conditions, sing the vale of friction coefficients obtained from experimental data, and the vale of pll force. The theoretical analysis, based on the mathematical model for starting torqe, agrees fairly well with experimental data obtained from sing actal compressors. The inflences of compressor dimension, shaft srface treatment, mechanical air gap in the motor, compressors blade spring on starting torqe are fond qalitatively and qantitatively by this stdy. INTRODUCTION?air conditioning nit consists of for elements, those are, a compressor, condenser, capillary (or expansion valve) and evaporator, as shown in Fig. 1. As the thermal capacity of the compressor is sally larger than any other parts of the refrigerant cycle, the presence of snlight always cases large qantities of refrigerant to migrate into the compressor and dilte the oil. The migration path of refrigerant in the rotary compressor is shown in Fig. 2. There is no ristriction between the condenser and th.e compressor case, so the refrigerant can easily migrate into the compressor. The liqid level is often two or three times as mch as oil level in rnning. As the refrig.erant dilte the oil between the lbricated srfaces, the friction coefficients in moving parts become larger. Accordingly, compressor motors reqire large.starting torqe. Little information is available concerning the starting torqe for rotary COffiprssors. So, the starting torqe has been clarified by sing rolling piston type hermetic rotary compressors. To analyze the starting torqe mechanism in several conditions, a mathematical model was made p. This model incldes factors concerning compressor dimension, motor pll force, blade spring force and the weight of moving parts. Test apparats to measre the rotating torqe was deve1oped and confirmed to be very sefl in estimating the starting torqe for the rotary compressor Liqid Tankj c_.o_m Otdoor Unit 8 ' Fan' pressorl Fig. l Refrigerant Cycle of Air Conditioners 98

3 mechanism. Frthermore, it was fond that the friction coefficients for shafts and motor pll force are very important in determining the starting torqe. Their details are discssed in the following. MATHEMATICAL MODEL A schematic model of the starting torqe for a rolling piston type rotary compressor is shown in Fig. 3. Following assmptions are sed for simplicity to calclating the starting torqe. (1) Inertia forces for moving parts are neglected, becase interest is focsed on the instantaneos movement at low speed in starting. (2) The force needed-to compress the refrigerant is neglected, becase the sction pressre is nearly eqal to the discharge pressre jst before starting. (3) A shaft is assmed to be rigid and has no deflection. (4) Friction forces of roller and blade are neglected since they are very small. (5) There is no force considered de to gravity between the roller and the crank. (6) The shaft contacts two points, the pper end of the mainbearing (B in Fig. 3) and the lower end of the sbbearing (A in Fig. 3). The motor pll force (Fmol is larger than the blade spring force (Fvl - On the above assmptions, the starting torqe is calclated by sing the motor pll force, spring force and gravity of the motor and shaft. Spring force (Fv) is shown as a fnction of blade displacement (y)- where: k Fvo ky + Fvo (1) y = e (1 +cos e) ( 2) e = 8 "' spring constant spring force at the bottom dead center eccentricity of the crank crank angle from blade position The motor pll force acts pon the center of the motor in radial direction of the smallest gap. The reaction forces for mainbearing (Rl) and sbbearing (R2) are gained by solving the force and momentm eqations in each of two mtally perpendiclar direction. L Rl = L- 1 R2 Fmo /1-2Clcos8p + c1 2 (3) 1 L- 1 Fmo / l + 2C 2 cos8p + c 2 2 (4) From Condenser Fmo Fig. 2 Migration Path of Refrigerant Fig. 3 Schematic Model of Starting Torqe 99

4 where: c ;2 Fv L Fmo 2 + 4;2 Fv 1 Fmo L e,l distance from mainbearing to rotor center 2, 3, 4 lengths of mainbearing, roller and sbbearing ep angle of pll force direction from blade position These eqations can be written in simple form in the following typical cases. (1) where motor pll force is zero R2 FvU-3 + 4/2) L - 1 Fv(Q,2 + 4/2) L - 1 (2) where spring force is zero L Rl = Fmo. L- 1 (5) (6) (7) ( 8) The spring force acts pon the crank and the roller. Weight of the rotor and the shaft (Wrsl acts pon the thrst srface (D in Fig. 3). Accordingly, the torqe concerning to friction force (Tf) can be determined from the following eqation. Tf JJlRl rs + JJ2R2r s + 1J3Fvrc + 1J4Wrsrt (9) where: 1Jl,1J2,1J3,1J4 friction coefficients for mo:ving parts rs,rc,rt revoltion radis of moving parts The blade spring force acts pon the crankshaft and the roller, so that the momentm force () de to crankshaft eccentricity can be determined as a fnction of rotation angle. Tm -efv sin 8 (1) Ultimately, torqe to rotate (Ts) is the sm of the torqe de to the friction force and the momentm force. Ts JJlRlrs + JJ2R2rs (rnainbearing) (sbbearing) + JJ3Fvrc + 1J4Wrsrt (crankshaft) (thrst) CHARACTERISTICS OF MOTOR It is known that the radial magnetic pll (RMP) occrs in eccentric direction de to rotor eccentricity(2). RMP stdies have mainly concentrated pon (1) Intensity of strctre, that is, shaft difrection, gap beat etc. (2) Vibration, noise and drability of bearing. However, it was fond to strongly affect the starting torqe for compressors. Using a strain gage pt on the mainbearing, the motor pll force was measred. The relation of RMP and ro tor eccentricity gained by this method is snown in Fig. 4. It is known that the pll force is proportional to the rotor eccentricity. When rotor eccentricity for an indction motor is small, these relations are shown by the following eqation. where: Cm: constant 66: eccentricity o : mechanical air gap (12) As the rotor eccentricity occrs de to parts inaccracy and monting inaccracy, RMP acts pon the rotor when compressors start. 4-l tj>,.'.; (\) 1-l 1'4 r-1 r-1 o.. ()....j.) (\) r:: b' ru ::.: Relative Eccentricity efvsin8 (blade spring) (11) Fig. 4 Effect of Eccentricity on Magnetic Pll Force 1

5 Minimm voltage reqired to start cornpressors is sally measred to check the starting torqe. Figre 5 shows the relation of the starting voltage and torqe. In this example, to start at 85 volts, mechanism torqe needs to be below 5.6kg ern with Starting Capacity (SC), bt 3.6 kg cm withot SC. When friction coefficients for moving parts are.25, moto pll force for starting shold be below 5 kg-ern and rotor eccentricity shold be below.14 rnm. As shown in the above example, rotor eccentricity inflences the magnetic pll force and the starting torqe for rotary compressors. with oil. However, when oil is not sed these coefficients have a tendency to be large. In case of shaft A, coefficient from.25 to.3, are 5 percent larger withot than with oil. On the other hand, shaft C gives the best reslts, which are as low as.15, with and withot oil. In addition, coefficient of friction vales in the other moving parts of compressors were measred and the same tendency as that for shaft A was obtained. COEFFICIENT OF FRICTION IN COMPRESSOR PARTS In estimation of the starting torqe, one of the important factors is coefficient of friction in the moving parts of compressors, particlarly between the shaft and the bearing. In order to estimate it, two kinds of shaft treated with phosphoric acid chloride (Specimen A, B) and a shaft coated with solid film lbricant (Specimen C) have been prepared. The former is MnHP 4 generally sed for protecting from abrason. The latter is Mos}3), The test shafts were made from cast iron with srface finish of abot 1.6 m R and these srfaces were treated. SEMmfiotographs and profiles of test shaft srfaces are shown in Fig. 6. For these shafts, coefficient of friction vales were measred with the bearing combined with and withot refrigerant oil (viscosity is 5 cp at 4 C). Figre 7 shows the test reslts nder several loads. The coefficients of friction in either shaft is mostly less than.2 (a) Specimen A (b) Specimen B (c) Specimen C Fig. 6 Test Shaft Srfaces 8 () Lf-.1 tj'i 6 (\) ;:I b' 4 With sc E-< IT> j2 :... rrj.j.j U) Withot sc 5 1 Starting Voltage (V).3,.j.IJ () rl D 1-J II<.2 Lf-.1.j.J GJ -1.1 t t Specimen A (li thot Oll) Specimen A (With Oll) rl Specimen B (Withot Oi 1. l Lf Specimen B (\Vi th 'h l) il) t. Specimen c (Withot Oil)... Specimen c (Wlth Oil) l Load (kg f) Fig. 5 Starting Characteristics of Motor Fig. 7 Coefficient of Friction in Test Shafts 11

6 EXPERIMENTAL APPARATUS FOR TORQUE AND TESTING METHOD An experimental apparats for measring the rotating torqe, as shown in Fig. 8, has been prodced. A rotary compressor mechanism that consists of a cylinder, a roller, a blade, a shaft and bearings is set pside down differing from actal setting in center of the apparats. The rotating torqe is measred by a load transdcer as reaction force of a driving motor. Generally, it is known that friction force depend on rotation speed. However, driving motor is set p at constant low speed of 1 r.p.m., becase it seems that the starting torqe concerns with static friction force(4). Moreover, a spring force is sed instead of radial magnatic pll on the center of axial direction. Using this apparats, the rotating torqe is recorded at the rotation angle obtained by a potentiometer on an X-Y recorder. A schematic diagram of the torqe measring system is shown in Fig. 9. EXPERIMENTAL RESULTS OF TORQUE Figre 1 shows typical torqe crves for Specimen A, where (a) is with and (b) is withot oil, and theoretically calclated reslts, where coefficients of friction in every moving part are assmed to be.25 or.16 constant. As the experimental crves agree fairly well with theoretical analysis, it seems that the torqe can be predicted by the theoretical analysis. Then, the torqe vales were measred nder several conditions. The measred torqe variations, with or withot oil, are shown in Fig. 11, plotted against the specific pll force. Vales obtained are compared with theoretical reslts. In this case, the maximm vales are selected as the torqe. Individal experimen- tal reslts are in proportion to coefficients of friction in shafts and the pll force, which are apparently important factors in estimating starting torqe. The plotted date are fn limits from.1 to.2 as coefficients of friction with oil, verss from.2 to.3 withot oil. Accordingly, the torqe is effected by the oil condition. s 3 Experimental Analytical 6 kgf Fig. 8 Experimental Apparats for Torqe I'ot.entimnet.er 4 Rotation Angle (degree) (a) Specimen A (With Oil) Pll Force = 6 kgf y Rotation Angle (degree) (b) Specimen A (Withot Oil) Fig. 9 Schematic Diagram of Torqe Measring System Fig. 1 Typical Torqe Crves 12

7 Moreover, in order to simlate the migration of refrigerant, it was attempted to sink a compressor mechanism filled with oil in refrigerant Rll3. The relation between the torqe and time was investigated. As shown in Fig. 12, th'e torqe, which was as large as the vale with oil at first, increased and reached a vale as large as that withot oil. These phenomena show that the oil between the lbrcated srfaces is easily removed by :ef1gerant. In starting a compressor, it 1s 1mportant to consider--the migration of refrigerant sfficiently. CONCLUSION Several conclsions have been reached as follows: (1) The starting torqe is effected by the coefficient of friction in the moving parts of compressors, the lbricant condition and the radial magnetic pll for motors. (2) The theoretical analysis, based on the mathematical model for starting torqe, agree fairly well with the experimental date sing actal compressors. STARTING TEST ON ACTUAL COMPRESSORS Finally, in order to prove whether the developed method to estimate the starting torqe is proper or not, the starting test on actal compressors have been carried. So, the effect of motor eccentricity on the starting voltage was investigated, sing the coefficient of friction and oil condition as parameters. As shown in Fig. 13, starting voltage tends to increase in proportion to the motor eccentricity and is effected by the coefficient of friction and oil condition. Accordingly, a characteristic of motor can be determined by the dimensions, the coefficient of friction in the moving parts of compressor and the lbricant condition. to: 5 Magnetic Pll Force= 7.7 kgf o() ,..----o 5.3 o o- o kg l Time (Hors) Fig. 12 Effect of Sink Time on Torqe (with Specimen A) to: 5 4-l tn4 <J) 3., ;::J ' 2 E-< 1 Specimen A Specimen A Specimen B Specimen B Specimen c Specimen C (hli thot Oil) (hli th Oil) (Withot Oil) (\Vi th ''Ji.l) (!Vi thot Oil) (With Oil) xo «/.., /od / 'b.. Vt>. : o-ll s--9--lo- Magnetic Pll Force (kgf) <J) tn l1l 15.j.J 1...; g tn.:: rl.j.j 5.j.J U) Specimen A e Specimen A Specimen B Specimen B.. Specimen C.6. Specimen C Relative Eccentricitv Fig. 11 Experimental Reslts of Torqe Fig. 13 Effect of Motor Eccentricity on Starting Voltage 13

8 (3) The starting mechanism, when compressors are restarted after a period of inoperation, is concer ned with the migration of refrigerant. This has been confirmed experimentally. (4) It has been confirmed experimentally that the radial magnetic pll increases approximately in proportion to motor eccentricity. (5) It is abvios that, sing the estimating method developed in this paper, the starting mechanism in actal compressors is well explained. (6) This method is effective to design motor and compressor mechnisms. ACKNOWLEDGEMENT The athors wish to thank Masao Oz of Fji Works, Toshiba Corporation for his motivation and valable advice REFERENCES 1. Richard E. Cawky, Sidney A. Parker: Evalation of Crankcase Heat and Its Effectiveness with a Hermetic Compressor, Proc Prde Compressor Technology Conference, Jly A. Cobo: Unbalanced Magnetic Pll in Indction Motors with E"ccentric Rotors, AIEE Transactions, December M. Kawamra, K. Yoshida, I. Aoki: Rnning-in Effect of Bonded Solid Film Lbricants on Conventional Oil Lbrication, ASLE Proceedings 2nd International Conference on Solid Lbrication, A. Hibi, T. Ichikawa: Torqe Characteristics of Pressre Balanced Type Vane Motor at the starting and at Low s peed, Transaction of the JSME, Vol.42, No.356,