Modeling of a Hydraulic Energy Regeneration System Part II: Experimental Program

Transcription

1 A. Purmvahed Pwer Systems Research Department, General Mtrs Research Labratries, Warren, Ml 489 N. H. Beachley F. J. Frnczak Mechanical Engineering Department, University f Wiscnsin-Madisn, Madisn, Wl 576 Mdeling f a Hydraulic Energy Regeneratin System Part II: Experimental Prgram This study reprts experimental data taken with a hydraulic energy regeneratin system and cmpares the measured data with analytical results. The system tested cnsisted f tw fam-filled hydraulic accumulatrs, a variable-displacement pistntype pump/mtr, a reservir and a flywheel. During a series f experiments, energy was repeatedly transferred between the hydraulic accumulatrs and the flywheel thrugh the pump/mtr. Cmputed system variables cmpared favrably with the experimental results. At high and mderate pump/mtr swivel angles, the rundtrip efficiency varied frm 6 t 89 percent. It was significantly lwer at small angles. Experimental Prgram This sectin presents the results f regenerative cycling tests at the University f Wiscnsin-Madisn in which energy was transferred between tw hydraulic accumulatrs and a flywheel. A Rexrth variable-displacement pump/mtr in cmbinatin with tw fam-filled Parker pistn-type accumulatrs were used in the tests. During the cycling f the energy between the accumulatrs and a flywheel, a data acquisitin cmputer recrded seven parameters: the pump/mtr displacement, the pistn psitin in ne f the accumulatrs, the inlet il pressure f the pump, bth accumulatr gas pressures, the trque between the pump/mtr and flywheel, and the flywheel speed. Tests were run with different pump/mtr displacements and with tw different il temperatures. A Apparatus. The circuit used fr these tests is shwn in Fig.. Pwer fr the initial accumulatr charge was prvided by a 5 hp hydraulic pwer supply. The pwer supply relief valve was set t abut.4 MPa ( psi) t set the maximum pressure in the accumulatrs when they were being charged. Ball valve N. and an rifice islated the pwer supply frm the rest f the circuit. The tw, 8.85-liter (.6-galln) pistn-type Parker accumulatrs were bth mdel N. A6R46B69E and are rated at.7 MPa ( psi). Mre infrmatin n these accumulatrs is given by Baum (987). The rest f the circuit cnsisted f tw ball valves and a 8- liter (-galln) accumulatr used as a pressurized reservir. The accumulatrs were bth filled with elastmeric fam during all f the tests. ne accumulatr cntained 76 grams (.68 lb) f fam and the ther cntained 74 grams (.6 lb). B Instrumentatin. Instrumentatin included three pressure transducers, tw linear variable differential transfrmers 'Currently, Assistant Prfessr, Mechanical Engineering Department, GMI Engineering & Management Institute, Flint, MI Cntributed by the Dynamic Systems and Cntrl Divisin fr publicatin in the JURNAL F DYNAMIC SYSTEMS, MEASUREMENT, AND CNTRL. Manuscript received by the Dynamic Systems and Cntrl Divisin March, 99; revised manuscript received March 99. Assciate Editr: A. Akers. (LVDT's), ne trquemeter, and ne digital encder. The pressure transducers were manufactured by Schaevitz, type P5-, and are rated at -4.4 MPa ( t 6 psig). Their utputs were amplified by Himmelstein transducer amplifiers Test number Fig. REGENERATIVE CYCLING HYDRAULIC CIRCUIT Precharge pressure MPa, abs Accumulatr LVT Ball Valve The regenerative cycling hydraulic circuit Table Purnp/mtr angle (degrees) 5 5 Starting Ending Test cnditins reservir reservir pressure pressure kpa, abs kpa, abs Case drain flw (L) /Vl. 4, MARCH 99 Transactins f the ASME Cpyright 99 by ASME Dwnladed 8 ct 8 t Redistributin subject t ASME license r cpyright; see

2 Table Flywheel lss cefficients Unit x ~ x (T x (T " N»m N-m/rpm) NTn/(rpm) N-m/frpm) Table Lss cefficient estimates fr Rexrth AV pump/ mtr size: 7 cmvrev Date surce 9 C s 5 C sl Manufacturer UW-Madisn Cmbined,5 5,47 9, FLYWHEEL SPIN DWN 6.P/M Angle = Deg., Cld il 5 ffl lt. 5 Time, (s) 5 Fig. and calculated flywheel speed histry during the spindwn test Fig. Flywheel speed histry fr Test number mdel 6-, which have an utput f =t 5 vlts. These pressure transducers were calibrated with a dead-weight tester. The accumulatr pistn psitin was measured with a Schaevitz LVDT mdel HR with a range f ± inches and 4% linearity. A 6.5-mm (/4-in.) diameter rd cnnected the accumulatr pistn t the LVDT cre. The pump/mtr displacement was measured using a Schaevitz LVDT mdel HR with a range f ±5.8 cm (± in.) and.5 percent linearity. A 6.5-mm (/4-in.) diameter rd cnnected the pump/mtr t the LVDT cre. The LVDT utput signals were cnditined using Himmelstein LVDT amplifier mdels 6-A, which have an utput f ±5 vlts. The LVDT's were calibrated using a vernier caliper. The trque was measured using a Himmelstein MCRT trquemeter mdel 9-T with a range f 45 N»m (4 in*lb) and. percent linearity. The trquemeter utput signal was cnditined by a Himmelstein transducer amplifier mdel 6-, which has an utput f ±5 vlts. The flywheel speed was measured using a Hewlett Packard ptical encder mdel HEDS-6. An encder interface cnverted the signal frm the encder t a binary cde which was read by the cmputer. The seven signals were cntinuusly sampled during each test using an ISAAC 4, a -bit A-D and binary data acquisitin system manufactured by Cybrg Crpratin, and an IBM cmputer, mdel PC. A secnd IBM PC cmputer with a Data Translatin Crpratin data acquisitin and cntrl card was used t cntrl the displacement f the Rexrth pump/mtr. C Prcedure. Fr each test, the fllwing prcedure was fllwed: Ball valves Ns.,, and were pened and il was circulated thrugh the system until it was at the desired test temperature. Ball valve N. was then clsed and il allwed t flw int the lw pressure reservir. The reservir was charged with il until the pressure was kpa, abs ( psig), the amunt f bst pressure required by the Rexrth pump/mtr when running as a pump. At this pint the flywheel was braked t a stp and the tw high pressure accumulatrs were charged t.79 MPa, abs ( psig). When the relief valve f the hydraulic pwer supply started t relieve, the data acquisitin was started. Ball valve N. was clsed and the brake released. This allwed the flywheel t accelerate while the accumulatrs were discharging. When the accumulatrs were nearly discharged, the displacement f the Rexrth pump/mtr was changed by ging "ver center.'' This turned the unit int a pump s that it culd begin charging the accumulatrs using the energy in the flywheel. When the flywheel stpped spinning, the accumulatrs started the Rexrth pump/ mtr turning in the ppsite directin, since the strke f the unit was nt changed at this pint. The prcess was then repeated, cycling the energy between the flywheel and the high pressure accumulatrs until it was dissipated thrugh lsses in the system. Tw cmpnents shwn in Fig. have functins which may nt be readily apparent. These are the rifice and ball valve N.. The rifice's nly functin was t allw fr a safe discharge f the il in case f prblems during the test. Since the rifice is nt part f the regenerative circuit, its inclusin des nt affect the efficiency f the system. Ball valve N. is attached directly t the lw pressure reservir and is smetimes clsed when the system is cnfigured differently. The pump/mtr displacement, accumulatr pistn psitin, pressures, trques, and speeds measured during the tests were stred n flppy disks. Table shws the cnditins fr all nine tests. The reservir used in this investigatin was a 8-liter (- galln) hydraulic accumulatr. Due t its large vlume, the gas pressure variatin was small. Fr example, in test N., the reservir pressure varied frm t 48 kpa, abs ( t 55 psig). A cnstant reservir pressure (the average value) can, therefre, be assumed. D Flywheel Spin Tests. As a prelude t the regenerative cycling tests the flywheel lsses were determined s that their effect culd be taken int accunt when analyzing the regenerative cycling data. The hardware was cnfigured slightly differently fr this test than fr the regenerative cycling tests. The majr change was that the pwer t the flywheel was prvided by a Vickers Jurnal f Dynamic Systems, Measurement, and Cntrl MARCH 99, Vl. 4/6 Dwnladed 8 ct 8 t Redistributin subject t ASME license r cpyright; see

3 P/M Angle = Deg., Cld il P/M Angle = Deg., Cld il g- A A - A A h^ rk rv d. a> E h- 5 ^- - Fig. 4 / V / Pump/mtr trque histry fr Test number [H I -. t. nset f Turbulence Fig. 7 line pressure lss histry fr Test number P/M Angle = Deg., Cld il P/M Angle = Deg., Cld il..8 t p-.6 CB C Fig. 5 Accumulatr gas pressure histry fr Test number C Fig pump/mtr trque efficiency histry fr Test number P/M Angle = Deg., Cld il. P/M Angle = Deg., Cld il rn.8 9 Q) C CD Fig accumulatr gas temperature histry fr Test number hydraulic mtr rather than the Rexrth pump/mtr. This change was dne t allw the mtr t be quickly disengaged frm the flywheel. In rder t facilitate the disengagement, the Vickers mtr was munted s that it culd slide back and frth. The mtr was cupled t the flywheel and the tr- (B E > T Fig. 9 pump/mtr vlumetric efficiency histry fr Test number quemeter by means f a face plate munted t a flexible cupling which engaged tw pins munted n the trquemeter cupling plate. Instrumentatin cnsisted f the IBM PC cmputer, ISAAC 4, HEDS-6 ptical encder, and interface. The test was cnducted as fllws. The mtr was slid int psitin t engage the pins n the trquemeter cupling plate. The mtr was used t accelerate the flywheel t 5 RPM 6/Vl. 4, MARCH 99 Transactins f the ASME Dwnladed 8 ct 8 t Redistributin subject t ASME license r cpyright; see

4 ~.78 Spec Gas T c P/M Angle = Deg., Cld il Fig. gas specific heat (c r) histry fr Test number I 5 <D 5 D" -5 H " #jfl* P/M Angle = Deg., Ht il &w M :~HN W s is \h Fig. Pump/mtr trque histry fr Test number i" m 8 CD f 6 LL 4 P/M Angle = Deg., Ht il Fig. Flywheel speed histry fr Test number 4 CL a CD 8 <D 4 P/M Angle = Deg., Ht il 6 Fig. Accumulatr gas pressure histry fr Test number 4 at which time the mtr was slid back t disengage it frm the flywheel and the trquemeter. At this time the data acquisitin began. The speed f the flywheel was sampled at a rate f Hz until the flywheel stpped. Three identical tests were run t ascertain repeatability. The experimental spin-lss data btained fr the flywheel used in this study resulted in the numerical values shwn in Table (see Eq. (7), Part I). T test the accuracy f Eq. (7) in predicting the flywheel spin lss, this equatin was integrated t calculate the flywheel speed histry during the spin-dwn test. Figure shws the calculated and the measured flywheel speed indicating excellent agreement, thus validating the regressin results. Estimatin f the Pump/Mtr Lss Cefficients. A nnlinear least-squares methd can be used t estimate the lss cefficients in Eqs. (6,,) and ( p ) fr a pump/mtr (see Part I). Fr the unit used in this investigatin data were available frm tw surces: the manufacturer (Rexrth) and the University f Wiscnsin (Baum, 987). Separate estimates fr each data surce as well as cmbined estimates based n all the data were btained. The results are given in Table. Examinatin f the estimated lss cefficients reveals a significant difference between the data frm the tw surces. Since the tw sets f data give different results, it is difficult t knw which data t use r whether either surce f data can prvide adequate infrmatin. Cmparisn f Experimental and Analytical Results Figure shws the experimental and the analytical flywheel speed histry fr Test number l. It is seen that there is gd agreement between the tw curves during the first few cycles, but as time ges n the accumulatin f errrs causes a nticeable difference and phase shift between the experimental and the analytical results. These differences can almst entirely be attributed t the uncertainties in the pump/mtr efficiency data. Similar cmments apply t the pump/mtr trque (Fig. 4) and the accumulatr gas pressure histry (Fig. 5). Accumulatr gas expansin in the first cycle resulted in a calculated drp f.5k in the gas temperature (Fig. 6). This drp is relatively small because during this expansin prcess a substantial amunt f heat transferred frm the fam t the gas. Figure 7 shws the calculated ttal pressure drp in all cnnecting lines between the accumulatrs and the reservir. Transitin frm laminar t turbulent flw during the first cycle is evident frm the sudden rise in the pressure drp. Figures 8 and 9 shw the calculated trque and vlumetric efficiencies as functins f time. As the pump/mtr speed appraches zer, the leakage terms in Eq. (6) becme large. Under these cnditins (small S and cr), this equatin des nt In all analytical cmputatins, the tw accumulatrs were treated as a single accumulatr twice as large, and with the same ttal fam mass. The input data are shwn in the Appendix. Jurnal f Dynamic Systems, Measurement, and Cntrl MARCH 99, Vl. 4/6 Dwnladed 8 ct 8 t Redistributin subject t ASME license r cpyright; see

5 S H (/> xs t ra 9 P/M Angle = Deg., Ht il Fig. 4 accumulatr gas temperature histry fr Test number 4 Efficiency metri Vlu Calcul iated P/M Angle = = Deg., Ht il n ( ' Au \ \ I i : ; : Fig. 7 pump/mtr vlumetric efficiency histry fr Test number _,.8 I 6 jjj.4 P/M Angle = Deg., Ht il.77 ra a> S. 76 a. w <A a P/M Angle = Deg., Ht il I.. Fig. 5 line pressure lss histry fr Test number Fig- 8 gas specific heat (c,) histry fr Test number 4 P/M Angle = Deg., Ht il, Cld.. >. c >.8 s= UJ a.6. ra. Fig. 6 ber 4 4 f 5, Ht _.-.>" ^ ^^/ ^, ~~", Ht Cld ^r'~~" Small Angles Large Angle Mderate Angles Cycle Number, i pump/mtr trque efficiency histry fr Test num- ^9-9 Rund-trip efficiency versus the cycle number fr varius pump/ mtr swivel angles and il temperatures accurately predict pump vlumetric efficiency. Frtunately, this ccurs nly fr a few time steps in every cycle and hardly affects the verall results. The trque efficiency curve des nt exhibit such behavir. The variatin in the gas specific heat (c ) is shwn in Fig.. Due t the "isthermalizatin" effect f the fam, the 64/Vl. 4, MARCH 99 variatin in c v is less than percent. Fr a cnventinal accumulatr (withut fam), hwever, the gas specific heat variatin may be significant and must be taken int accunt. Figures thrugh 8 shw similar results fr Test number 4 where the swivel angle was small ( deg) cmpared t the prir data. Transactins f the ASME Dwnladed 8 ct 8 t Redistributin subject t ASME license r cpyright; see

6 Rund-trip efficiencies were calculated frm Eq. (), Part I, based n the experimentally-measured flywheel speed. Figure 9 shws this efficiency versus the cycle number fr eight tests. The rund-trip efficiency culd nt be calculated fr Test number 5. Fr this test, the pump/mtr swivel angle was degrees and due t excessive lsses in the pump/mtr the flywheel accelerated nly nce. By examining Fig. 9, ne can bserve that at high and mderate pump/mtr swivel angles, the rund-trip efficiency varied frm 6 t 89 percent. At small angles, hwever, this efficiency was as lw as 4 percent. Large variatins in the rund-trip efficiency frm ne cycle t the next ccurred in sme tests. This was caused by an abrupt change in the pump/ mtr angle which happened a few times during the tests. Cnclusins. The analytical mdels develped fr all system cmpnents are valid.. If prperly selected, the cmbined lsses f the cmpnents used in the hydraulic energy regeneratin system are nt prhibitive fr use in a hydraulic hybrid vehicle.. At high pump/mtr swivel angles, the flw f hydraulic fluid ut f the accumulatr can becme turbulent. 4. The specific heat f nitrgen gas in a fam-filled hydraulic accumulatr may be assumed cnstant. References Baum, S. A., 987, "Investigatin f a Dual Pump/Mtr Hydraulic Accumulatr Energy Strage Autmbile," MS thesis, Mechanical Engineering Department, University f Wiscnsin-Madisn. Brandrup, J., and Immergut, E. H., Plymer Handbk, nd Editin, Jhn Wiley & Sns, p. III-5. APPENDIX The Input Data The fllwing numerical values were used as input data fr all test runs: The Accumulatr Fam Specific Heat:. kj/kg»k (Brandrup and Immergut) Mass f Gas:. kg Mass f Fam:-.496 kg Maximum Gas Vlume: 5.7 L Thermal Time Cnstant: s Frictinal Lss: 4 percent f Input Energy The Pump/Mtr Displacement: 7 cmvrev Cefficient f Frictin:.48 Hydrdynamic Lss Cefficient: Laminar Cefficient f Slip:.4 x " 9 Turbulent Cefficient f Slip:. X ~ 5 Cefficient f Viscus Drag: 5,47 Maximum Swivel Angle: 5 degrees Lines, Fluid, Flywheel Equivalent Hse Length:.96 m Hse Internal Diameter:.5 m il Density: 869 kg/m Flywheel Inertia:.98 kg-m Initial Flywheel Speed: RPM The fllwing values were used fr Tests and 4: Test Test 4 Accumulatr Wall Temperature (K) Fluid Cnditin Cld 5 Ht Fluid Kinematic Viscsity (cst) 6 4 Initial Gas Temperature (K) Pump/Mtr Swivel Angle (degrees) ± 5 ± Reservir Pressure (MPa, abs) Jurnal f Dynamic Systems, Measurement, and Cntrl MARCH 99, Vl. 4/65 Dwnladed 8 ct 8 t Redistributin subject t ASME license r cpyright; see