Available online at www.sciencedirect.com Available online at www.sciencedirect.com
ScienceDirect ScienceDirect Energy Procedia 00 (2017) 000–000 ScienceDirect Energy Procedia 00 (2017) 000–000 ScienceDirect ScienceDirect
Available online at www.sciencedirect.com Availableonline onlineatatwww.sciencedirect.com www.sciencedirect.com Available
Energy Procedia 00 (2017) 000–000 Energy Procedia 12600 (201709) 1115–1122 Energy Procedia (2017) 000–000
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
72nd Conference of the Italian Thermal Machines Engineering Association, ATI2017, 6-8 2017, Lecce, Italy 72nd Conference of the ItalianSeptember Thermal Machines Engineering Association, ATI2017, 6-8 September 2017, Lecce, Italy nd 72 Conference of the Italian Thermal Machines Engineering Association, ATI2017, 6-8 15th International Symposium on District Heating and DiagnosticTheProcess bySeptember Using Vibrational Sensors forCooling Monitoring 2017, Lecce, Italy
Diagnostic by Using Vibrational forAutomotive Monitoring Cavitation Process Phenomena in a Getoror PumpSensors Used for Assessing the feasibility of usingPump the heat demand-outdoor Diagnostic Process by Using Vibrational Sensors forAutomotive Monitoring Cavitation Phenomena in Applications a Getoror Used for temperature function forinaApplications districtUsed heatfor demand forecast Cavitation Phenomena along-term Getoror Pump Automotive a b b Daniela Siano *,Applications Emma Senatore c a Frosina , Adolfo b c a b b , O. Le Corre I. Andrića,b,c*, A. Pinaa, P. Ferrão , J. Fournier ., B. Lacarrière a
Daniela Siano *, Emma Frosina , Adolfo Senatore a CNR – Istituto Motori, Via Marconi – 80125, Naples, Italy a Research b Técnico, b 1, Naples, IN+ Center University for Innovation, Technology - InstitutoEngineering, Superior Av. Rovisco Pais 1049-001 Lisbon, Portugal of Naples Federico II,Policy Department of Industrial Via Claudio 21 – 80125, Italy a and Daniela Siano *,Motori, Emma , Adolfo CNR – Istituto Via Frosina Marconi – 80125, Naples,Senatore Italy b b
c
Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France b University of Naples Federico II, Department of Industrial Engineering, Via Claudio 21 – 80125, Naples, Italy a Département Systèmes Énergétiques et Environnement - IMT Atlantique, 4 rue Alfred CNR – Istituto Motori, Via Marconi – 80125, Naples, ItalyKastler, 44300 Nantes, France b
Abstract
University of Naples Federico II, Department of Industrial Engineering, Via Claudio 21 – 80125, Naples, Italy
Abstract AAbstract full experimental investigation on a Gerotor pump used for the lubrication of engines is described in this paper. These pumps, Abstract as full wellexperimental known, are widely used onon engines for all hydraulic circuits and, for this they work in some conditions (such A investigation a Gerotor pump used for the lubrication of reason, enginesoften is described in this paper. These pumps, District heating networks arevalue) commonly addressed in the literature as one of the most effective solutions for decreasing the as at high speeds and pressure which are very challenging. as well known, are widely used on engines for all hydraulic circuits and, for this reason, often they work in some conditions (such greenhouse gas fromunwanted the sector. These systems require high investments which are returned through the the heat A experimental on abuilding Gerotor used the occurs lubrication of engines is described inhas this paper. These pumps, In this paper oneemissions ofinvestigation the most phenomena that for often during the pump operation been investigated: as full at high speeds and pressure value) which are pump very challenging. sales. Due to are thewidely changed climate conditions and building renovation policies, heat demand in the futureconditions could decrease, as well known, used on engines for all hydraulic circuits and, for this reason, often they work in some (such cavitation. In this paper one of the most unwanted phenomena that often occurs during the pump operation has been investigated: the prolonging thecan investment return as at cavitation high speeds and value) which are very challenging. The bepressure triggered byperiod. many multiple factors such as the sloshing in the tank (translational and rotational motions), cavitation. The main scope ofof this paper isunwanted to assess the feasibility of too using theatheat demand – outdoor temperature function for heat demand In this paper one the most phenomena that often during the operation hasthe been investigated: thea high percentage of gas dissolved in the fluid and pressure low pump suction port. Therefore, characterization of The cavitation can be triggered by many multiple factors such asoccurs thethe sloshing in thepump tank (translational and rotational motions), forecast. Theindistrict of Alvalade, located in Lisbon (Portugal), was used as a case study. The district is consisted of 665 cavitation. Gerotor pump cavitation condition is really interesting. high percentage of gas dissolved in the fluid and pressure too low at the pump suction port. Therefore, the characterization of a buildings vary both construction period and weather (low, medium, high) andofthree district The cavitation can beincavitating triggered by many multiple factors such asThree the sloshing inscenarios the tank (translational and rotational motions), In order to that replay the conditions a pump hastypology. been installed on a dedicated test bench of the Department Industrial Gerotor pump in cavitation condition is really interesting. renovation scenarios were developed (shallow, intermediate, deep). To estimate the error, obtained heat demand values were high percentage of gas dissolved in the fluid and pressure too low at the pump suction port. Therefore, the characterization Engineering of thethe university of Naples “Federico pump has been to cavitate by placing calibrated orifices onof thea In order to replay cavitating conditions a pumpII”. hasThe been installed on aforced dedicated test bench of the Department of Industrial compared with results from a dynamic heat demand model, previously developed and validated by the authors. Gerotor pump in cavitation condition is really interesting. suction side of the pump. Many decreasing diameters have been located in an aluminum connection block, to measure all the Engineering of the university of Naples “Federico II”. The pump has been forced to cavitate by placing calibrated orifices on the The results showed that when only weather change is suction considered, theon margin of error could be acceptable forripple. someofapplications In order to replay the cavitating conditions a pump has been installed a dedicated test bench of the Department Industrial working parameters like the flow-rate, pressure (at the and delivery ports), pump speeds and pressure Cavitating suction side of the pump. Many decreasing diameters have been located in an aluminum connection block, to measure all the (theno-cavitating error inofannual demandhave was lowerinvestigated than 20%II”. for all pump weather scenarios considered). However, introducing renovation Engineering theconditions university of Naples “Federico hasdelivery been forced to pump cavitate by placing calibrated orifices on the and been byThe using anand accelerometer sensor in proximity of after the pump suction chamber working parameters like the flow-rate, pressure (at the suction ports), speeds and pressure ripple. Cavitating scenarios, the error value increased up to 59.5% (depending on the weather and renovation scenarios combination considered). suction side of the pump. Many decreasing diameters have been located in an aluminum connection block, to measure all the withno-cavitating the aim monitoring phenomena in terms ofbyvibration and conditionsthehave been investigated using anamplitude. accelerometer sensor in proximity of the pump suction chamber The value of slopelike coefficient increased on average within range 3.8% up to conditions 8% per and decade, that without corresponds to the working parameters flow-rate, pressure (at of the suction and delivery pump speeds pressure ripple. Cavitating As afore mentioned, thethe pump under investigation has been the studied inofallports), operative with and cavitation with the aim of monitoring the phenomena in terms vibration amplitude. decrease inbythe number of heating hours of 22-139h during theaccelerometer heating season (depending on is the combination of weather and and no-cavitating conditions have been investigated by using an sensor in proximity of the pump suction chamber phenomena using a non-intrusive sensor like accelerometer in order to monitoring if cavitation present. As afore mentioned, the pump under investigation has been studied in all operative conditions with and without cavitation renovation scenarios considered). On the in other hand, function intercept increased for 7.8-12.7% per the decade (depending on the with the aim of monitoring the phenomena terms of vibration amplitude. More precisely, the accelerometer sensor has been located close to the pump suction chamber and vibrations have been phenomena by using a non-intrusive sensor like accelerometer in order to monitoring if cavitation is present. coupled scenarios). The values suggested could be has used to modify the function parameters for the scenarios considered, and As afore mentioned, the pump under investigation been studied in all operative conditions with and without cavitation acquired contemporarily with pressure signals (intake and outgoing discharge) and properly triggered with tachometer signal by More precisely, the accelerometer sensor has been located close to the pump suction chamber and the vibrations have been improve thebyaccuracy of heat demand estimations. phenomena using a non-intrusive sensor like(intake accelerometer in order to monitoring if cavitation is present. using a multichannel acquisition system (Siemens™). spectral vibration analysis hastriggered been used diagnostic signal tool for acquired contemporarily with pressure signals andAoutgoing discharge) and properly withastachometer by More precisely, the accelerometer sensor has been located close to the pump suction chamber and the vibrations have using a multichannel acquisition system (Siemens™). A spectral vibration analysis has been used as diagnostic toolbeen for © 2017 contemporarily The Authors. Published by Elsevier acquired with pressure signalsLtd. (intake and outgoing discharge) and properly triggered with tachometer signal by Peer-review under responsibility the Scientific Committee of The 15th International on District Heating and using a multichannel acquisition ofsystem (Siemens™). A spectral vibration analysisSymposium has been used as diagnostic tool for Cooling. * Corresponding author. Tel.: +39 081 7177108; fax: +39 081 7177108. E-mail address:
[email protected] * Corresponding author. Tel.: +39 081 7177108; fax: +39 081 7177108. Keywords: Heat demand; Forecast; Climate change E-mail address:
[email protected] 1876-6102 © 2017 author. The Authors. Published by Elsevier Ltd. 081 7177108. * Corresponding Tel.: +39 081 7177108; fax: +39 Peer-review
[email protected] responsibility of the scientific committee of the 72nd Conference of the Italian Thermal Machines Engineering E-mail address: 1876-6102 © 2017 The Authors. Published by Elsevier Ltd.
Association. Peer-review under responsibility of the scientific committee of the 72nd Conference of the Italian Thermal Machines Engineering 1876-6102 1876-6102©©2017 2017The TheAuthors. Authors.Published PublishedbybyElsevier ElsevierLtd. Ltd. Association. Peer-review underresponsibility responsibility of the scientific committee of15th the International 72nd Conference of the on Italian Thermal Engineering Peer-review under of the Scientific Committee of The Symposium District HeatingMachines and Cooling. Association. 1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the scientific committee of the 72nd Conference of the Italian Thermal Machines Engineering Association 10.1016/j.egypro.2017.08.269
21116 2
D.Daniela Siano etSiano al./ Energy 00 (2017) et al. / Procedia Energy Procedia 126000–000 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
accurately detecting pump degradation. The results coming from the analysis have shown that in presence of cavitation accurately pump degradation. The results coming fromdiagnostic the analysis haveforshown thatpump in presence of cavitation phenomenadetecting the non-intrusive monitoring technique represent a good method assessing operability. phenomena the non-intrusive monitoring technique represent a good diagnostic method for assessing pump operability. © 2017 The Authors. Published by Elsevier Ltd. © 2017 2017 The Published Ltd. nd © The Authors. Authors. Published by by Elsevier Elsevier Ltd. committee of the 72nd Conference of of the the Italian Italian Thermal Thermal Machines Machines Engineering Engineering Peer-review Peer-review under under responsibility responsibility of of the the scientific scientific committee of the 72nd Conference Peer-review under responsibility of the scientific committee of the 72 Conference of the Italian Thermal Machines Engineering Association. Association Association. Keywords: Gerotor pumps, Cavitation detection, FFT, Vibration. Keywords: Gerotor pumps, Cavitation detection, FFT, Vibration.
Frequency originated on the inner fvan Passes Nomenclature Frequency fvan Passes Nomenclature rotor teeth originated on the inner rotor teeth p Pressure Number of inner rotor teeth Ninner teeth Pressure Number of originated inner rotoron teeth Nouter innerteeth teeth Saturated vapour pressure Frequency the outer ppv F Saturated vapour pressure Frequency originated on the outer pf0v F outer teeth frequency rotor teeth frequency fω0 RPM rotor teeth Pump shaft rotation Number of outer rotor teeth Nouter teeth Pump shaft rotation Number of outer rotor teeth ωRPM Nouter teeth 1. Introduction 1. Introduction Mechanical systems that interact with liquids, such as pumps or propellers, may present an unwanted Mechanicalcalled systems that interact with liquids, such as pumpsby or propellers, phenomenon cavitation. The cavitation, as well know, depends pressure In fact,may whenpresent a liquidan at aunwanted constant phenomenon called cavitation. The cavitation, as well know, depends by pressure In fact, when a liquid at a constant temperature is subject to a decreasing pressure, p, which falls below the saturated vapor pressure pv cavitation may temperature is subject to a the decreasing p, which falls below the saturated vapor at pressure cavitationliquid may occurs [8-11]. Therefore, processpressure, of rupturing a liquid by decrease in pressure roughlypv constant occurs [8-11]. Therefore, the process of rupturing a liquid by decrease in pressure at roughly constant liquid temperature is often called cavitation. Issues related to the cavitation are essentially the noise and the material temperature often Issues related pressures to the cavitation are essentially the noise the material damages thatiscan be called causedcavitation. by the high velocities, and temperatures. In fact, when and exposed to high damages that can be caused by the high velocities, pressures and temperatures. In fact, when exposed to high pressure, those voids tend to collapse, releasing a great amount of energy, usually in form of an acoustic shockwave, pressure, those voids tend to collapse, releasing a great amount of energy, usually in form of an acoustic shockwave, or even as visible light. or even visible light. Thoseasacoustic shockwaves can be easily detected by using a sensor such as an accelerometer positioned near the Those acoustic shockwaves can betoeasily by using a sensor such as processing an accelerometer positioned near the cavitation zone. One of the methods detectdetected the presence of cavitation, is by the time history signal of cavitation zone. One of the methods to detect the presence of cavitation, is by processing the time history signal of this sensor and performing a FFT (Fast Fourier transform) analysis, which will give an insightful point of view of this sensor and performing a FFT (Fast Fourier transform) analysis, which will give an insightful point of view of the cavitation process. As will be discussed later, when cavitation is present, a lot of the energy of the sound the cavitation As will abehigher discussed later, band, whenmaking cavitation is present, a lotthe of presence the energy of the sound spectrum tendsprocess. to shift towards frequency possible to identify of this problem. spectrum tends to shift towards a higher frequency band, making possible to identify the presence of this problem. One disadvantage of this method, is that in practical situations, the accelerometer is subjected to secondary sources One disadvantage of the thisdiagnose method, of is the thatcavitation in practical situations, accelerometer is subjected to secondary sources of vibration, making problem very the complex [2]. of vibration, making the diagnose of the cavitation problem very complex [2]. In the case of this article, an experiment to detect cavitation was conducted in a Gerotor Pump, used for In the case of this article, an experiment detect cavitation was conducted Gerotorspeeds Pump,ofused automotive applications. The pump in question to is an oil pump, which were subjectedintoarotation 1000for to automotive applications. The pump in question is an oil pump, which were subjected to rotation speeds of 1000and to 6000 RPM in 1000 RPM steps, with three different sizes of orifice on the inlet side of the pump of 18mm, 5mm 6000 RPM in 1000 RPM steps, with three different sizes of orifice on the inlet side of the pump of 18mm, 5mm and 3mm, to induce the cavitation phenomena [2]. An accelerometer was fixed on the pump, near the cavitation zone, 3mm, to instantaneous induce the cavitation accelerometer wasinlet fixed ondelivery the pump, near the cavitation zone, and two pressurephenomena transducers [2]. wereAn used to measure the and pressure of the pump, inside and two instantaneous pressure transducers were used to measure the inlet and delivery pressure of the pump, inside the chamber, near the rotor. The pump rotation was also measured. The present work will be a discussion about the the chamber, the rotor.of The rotation was also measured. present work will be a discussion about the detection and near identification thepump cavitation problem on this GerotorThe pump. detection and identification of the cavitation problem on this Gerotor pump. 2. Reference pump and experimental setup 2. Reference pump and experimental setup The pump under investigation is an internal gear pump used for the lubrication system of internal combustion The pump underconsists investigation is anrotor internal gearteeth pump for the lubrication engines. The pump of an outer with15 andused an inner with 14 teeth. system of internal combustion engines. The pump consists of an outer rotor with15 teeth and an inner with teeth. range of [0 ÷ 6] bar. A relief The pump works in a rotation speed range of [1000 ÷ 7000] rpm and a14 pressure The pump works in a rotation speed range of [1000 ÷ 7000] rpm and a pressure range of [0 ÷ 6] bar. A relief valve is included in the pump with an opening pressure of 4.5bar. valve is included in the pump with an opening pressure of 4.5bar. As already said, the pump has been tested on a test bench of Department of Industrial Engineering of University already said, theII”. pump has been tested a test layout bench of Department Industrial Engineering of in University of As Naples “Federico In figure 1a the testonbench is shown. The of bench can drive the pump the real of Naples “Federico II”. In figure 1a the test bench layout is shown. The bench can drive the pump in theofreal operation condition reaching a maximum shaft rotational speed of 8000 rpm. A hydraulic axial piston motor 12 3 operation condition reaching a maximum shaft rotational speed of 8000 rpm. A hydraulic axial piston motor 12 cm3/rev powers the pump, (see figure 1a). The variable displacement axial piston motor is driven by a PowerofUnit cm /rev powers the pump, (see figure 1a). The variable displacement axial piston motor is driven by a Power Unit that consists of an oil tank with a nominal capacity of 100 l, an axial piston pump with theoretical displacement of 3 that consists an oil tank electric with a nominal capacity of 100 l, anof axial piston pump with theoretical 71 cm /rev, aofthree phase motor with 4 poles, capable delivering a maximum power of displacement 18.5 kW andofa 71 cm3/rev, a three phase electric motor with 4 poles, capable of delivering a maximum power of 18.5 kW and a
Daniela Siano et al. / Energy Procedia 126 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
1117 3
heat exchanger air – oil. The pump shaft rotation speed has been varied, during the tests, through an electrohydraulic regulator handled by a dedicated card. A potentiometer gives the input signal to the card managing the displacement of the axial piston pump of the Power Unit [3-7]. Looking at figure 1a, it is possible to locate all the transducers installed on the test bench. These sensors can monitor and acquire the Gerotor operating parameters. In particular there is a flow-mater to measure the oil flowrate, two instantaneous pressure transducers (P1) (P2) located in the suction and delivery chamber volumes, a mean pressure transducer (P3) on the pump delivery and a torque meter. The thermostatic tank, with an oil volume capacity of 13 dm3, is enable to heat oil up to a temperature of 150°C. In the tank has been also installed a separation baffle (figure 1c) to separate suction and delivery ports of the pump.
Fig. 1. (a) Test bench layout; (b) Test bench; (c) Suction and delivery view of the pump, (d) and (e) Calibrated orefices. As already mentioned, three pressure transducers (P1, P2, P3) are installed in the hydraulic circuit. (P1) and (P2) are both instantaneous. These sensors are located inside two chambers (respectively suction and delivery chambers) close to the rotors. Both of them are type AVL LP11D, in particular the measuring ranges are [0 ÷ 10] bar for P1 and [0 ÷ 30] bar for P2. Thanks to the clock signal of an incremental optical encoder installed on the pump shaft, these sensors can measure the pressure ripples in a pump shaft rotation. Both sensors have a frequency response of > 50 kHz. (P3) is a mean pressure transducer. It is located at the outlet side of the pump. It is a type Burkert 8314 with a measuring range of [0 ÷ 10] bar and is based on ceramic technology measurement principle. The lamination valve (R2) regulates the pressure at the delivery side of the pump. The valve is located at the delivery port of the pump after the flow meter (Q). The flow meter (Q) works with a functional principle based on the controlled generation of Coriolis forces. It is able to measure also the temperature and density of the oil. Different signals are acquired by the hardware system NI PCI16-E1-MIO (12-bit ADC converter resolution). All the signals are routed to the board NI PCI-MIO 16-E1 through a 68-pin shielded desktop connector block NI SCB-68. A finite - machine study, developed with National Instruments LabVIEW® 2015 software, thus to control, manage and capture data. As said, the goal of this analysis is the study of a Gerotor pump forced to cavitate. Cavitating conditions have been achieved reducing the diameter of the suction port with three calibrated orifices (R1). These orifices are shown in figures 1d and 1e, where the first one (18mm) corresponds to the regular diameter of the inlet port of the pump. The second orifice has a diameter of 5 mm, while the last one is of 3mm. By reducing the diameter of the suction port the performance of the pump changes as well. The described test bench is able to perform three acquisition methodologies: Low sampling rate for main bench parameters (flow-rate, mean pressure, rpm); External clock variable sampling rate (instantaneous suction and delivery pressure); Time dependent sampling frequency (10 kHz) with a fixed time window (1 ÷ 10 s). The experimentation has been done in all the pump operating condition with oil: PETRONAS SYNTIUM 7000 0W-20. In particular the pump has been tested varying the delivery pressure in the range [0 ÷ 6] bar and with an oil temperature of 30°C, 60°C and 90°C. In figure 2 experimental data are shown. In particular, figure 2a shows the pump characterization with no restriction. Figure 2b presents the delivery varying the pump speed by changing the suction diameter with orifices of 18, 5 and 3mm. By the end, in figure 2c the measured pressure ripples are shown, the four spikes are due to geometry of these four teeth.
1118 4
Daniela Siano et al. / Energy Procedia 126 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
Fig. 2. Experimental data: (a) Pump caracterization, (b) Infuelce of restriction on the pump performance, (c) Pressure ripples. 3. Experimental vibration analysis setup As discussed before, the experiments were realized on a Gerotor automotive oil pump, settled on a custom test bench. The pump was powered by an axial piston motor powered by an axial piston pump which was finally powered by a three phase electric motor, capable of delivering 18.5 kW of power. The Gerotor pump was equipped with two instantaneous pressure transducers inside the inlet and delivery regions of the chamber, directly on the oil flow. The transducers used were the type AVL LP11DA, with the inlet one being capable of measuring pressures in the range of [0 - 10] bar and the delivery one on the range of [0 - 30] bar. The transducer accuracy is in the order of < 0.1% FSO, and it is capable of performing measurements at rate of 50 KHz. A ceramic shear ICP® accelerometer PCB Piezotronics, 10 mV/g, 5 to 60 kHz, model 352A60 was placed outside of the pump near the cavitation zone, as indicated on figure 3. Its sensitivity is low at very low and very high frequencies, with a reasonably flat response ( ±3 dB variations) over the frequency range 5 Hz to 60 kHz. Also, on the pump shaft was installed an incremental optical encoder, to measure the shaft speed rotation. All signals were acquired by using an LMS SCADAS XS 12 channels scope, with the sampling frequency of 52.2 KHz for each channel. The complete experimental array used, is shown on the figure 3 below.
Fig. 3. Experimental Layout. All the experiments were performed with a constant oil temperature of 40 ºC, for which 54 different conditions with and without cavitation were tested, by mixing the following states:
Daniela Siano et al. / Energy Procedia 126 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
1119 5
1000 to 6000 RPM with 1000 RPM steps 0, 1 and 3 bar of delivery pressure 18, 5 and 3 mm of diameter of the intake orifice Those tests generated 216 data streams in total, composed by, tachometer, accelerometer, pressure-in and pressure-out data. All of this data was post processed on the Siemens LMS.Lab Signal Acquisition software, mainly by applying a FFT to obtain their FRF’s (Frequency response functions). 4. Results and Discussion The results obtained by the experimental setup described in last section will be exposed here. They will be classified in two classes: no cavitation and developed cavitation. One clue to identify if cavitation is present, is that it will tend to shift most of the energy in the frequency spectrum into higher frequency bands, between 7 KHz and 14 KHz, when compared to low frequencies. On the low band of the frequency spectra, some important regions that usually contains a lot of energy are located around the fundamental frequency - eq 1, that originates from the shaft rotation, the van passes frequency – eq 2, that originates from the inner teeth’s of the rotor and the outer teeth frequency – eq3, which, as the name implies, is originated on the outer teeth’s of the rotor. The theoretical value of those frequencies can be calculated from the equations below. � �� = ����60 (1) ���� ������ = �� ∗ ������ �����
������ ����� = �� ∗ ������ �����
(2)
(3)
On the following figure 4, is possible to see the spectrum the response frequency of the accelerometer, for a first essay realized at 2000 RPM, with an orifice size of 5mm, and various delivery pressures (0, 1 and 3 bar).
Fig. 4. Accelerometer response at 2000 RPM. In this spectrum is noticeable that no cavitation is present, because of the lack of any important activity on the high frequencies bands. Most of the energy is located on the lower end of the spectrum, meaning the system is being mainly excited by conventional mechanic sources of vibration, such as misalignments and manufacturing defects, which does not tend to present high frequency responses. In next figure, the spectrum shows the frequency response of the accelerometer for a pump rotation of 3000 RPM, an orifice size of 5mm, and the same three delivery pressures from last case.
Fig. 5. Accelerometer response at 3000 RPM. It is evident that, when the delivery pressure reaches around 3bar, an important amount of energy gets concentrate on the frequency band between 7 KHz and 14 KHz, meaning the cavitation phenomena is starting to develop. It is not possible to confirm surely that it is really present, because the lower frequency band also contains a lot of energy, mainly around 1400 Hz, which represent two times the van passes frequency at 3000 RPM, and on 3500 Hz and 4500 Hz, which are caused by the fluid recirculation inside the pump chamber.
1120 6
Daniela Siano et al. / Energy Procedia 126 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
In next figure, the spectrum shows the response of the accelerometer for a pump rotation of 4000 RPM, an orifice size of 5mm, and two delivery pressures (1 and 3 bar). The delivery pressure of 0 bar cannot be achieved at 4000 RPM.
Fig. 6. Accelerometer response at 4000RPM. Now it is possible to confirm that cavitation fully developed with 3bar of delivery pressure. It is explained, mainly because the energy density of the spectrum, between 7 KHz and 14 KHz is much greater than the lower end of the spectrum, which is usually excited by conventional mechanical sources. The peek of vibration amplitude of around 3500Hz is caused by recirculation inside the pump chamber. In this case, the FRF method to detect cavitation is perfectly suited. In next figure, the spectrum shows the response of the accelerometer for a pump rotation of 5000 RPM, an orifice size of 5mm, and various delivery pressures (1 and 3 bar). The delivery pressure of 0 bar cannot be achieved at 5000 RPM.
Fig. 7. Accelerometer response at 5000RPM.
Fig. 8. Accelerometer response at 2000 RPM, in cavitation and no caviation conditions.
Daniela Siano et al. / Energy Procedia 126 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
1121 7
Notice that the main behaviour of this scenario at 5000 RPM is very similar to the scenario earlier described at 4000 RPM. The main differences that can be spotted are related to the overall amplitude of vibration in the frequency bands around 3500 Hz and the cavitation zone between 7 kHz and 14 kHz. At 5000 RPM, those amplitudes are much higher than at 4000 RPM, and one more time, as expected, cavitation is present. In case in figure 9, the three spectres show the Gerotor pump at 2000 RPM in different discharge pressures (0.1 and 3 bar), in cavitation and no cavitation conditions. Cavitation is induced by installing a 3mm orifice on the oil inlet, as oppose to the no cavitation condition, in which the pump is operating normally, with it’s default 18mm inlet orifice. In all three spectrums, it is visible that there is a lot of energy on the lower band of the frequency spectrum. Which is explained again by ordinary mechanical excitations, the van passes frequency and some recirculation, mainly when the lamination valve is open at 0 bar. In all cases also, when cavitation is present, for the green curve, the overall behavior is the same. The main variation that occurs from case to case is related to the overall amplitude of the response of the accelerometer on the cavitation zone, which is bigger on the lower discharge pressure. In the next case (figure 9), the two spectres show the Gerotor pump at 5000 RPM in different discharge pressures (1 and 3 bar), in cavitation and no cavitation conditions. As on last case, cavitation is induced by installing a 3mm orifice on the oil inlet. The pressure value of 0 bar on the pump discharge cannot be achieved at 5000 RPM, for this reason this test was performed only in 1 and 3 bar of pressure.
Fig. 9. Accelerometer response at 5000RPM, in cavitation and no caviation conditions. Notice that in all two spectrums when cavitation is present (green curve), the overall behavior is the same. It is important to notice that the cavitation amplitude at 5000 RPM is much higher than at 2000RPM, at least 5 times as high. In addition, it is noticeable, on the second spectrum, at a discharge pressure of 3bar that the pump starts to cavitate even on the normal working condition, as can be seen on the red curve, on the frequency range between 7 KHz and 14 KHz. 5. Conclusion In this paper a typical Gerotor pump used for the lubrication of Internal Combustion Engine has been studied in cavitation conditions. The pump has been tested on a test bench installed in the Department of Industrial Engineering of the University of Naples Federico II. Cavitation has been forced by reducing the suction port of the pump with calibrated orifices with diameters of 18mm, 5mm and 3mm. For each diameters, the Gerotor has been characterized measuring the delivered flow-rate, pressure in the suction and delivery chambers, the mean pressure at the delivery port the absorbed torque and noise with an accelerometer. The detected cavitation is stronger with lower diameters and higher pressure. As discussed earlier, there are some secondary sources of vibration and noise, other than the cavitation itself, like radial misalignments, impeller pressure variations, manufacturing defects of seals, bearings and impellers, and even
1122 8
Daniela Siano et al. / Energy Procedia 126 (201709) 1115–1122 D. Siano et al./ Energy Procedia 00 (2017) 000–000
hydraulic sources of noise, such as transient and unstationary flow causes. All of those secondary vibration sources add to the final measurement captured by the accelerometer, and their identification is important to the quality of the final analysis of the pump. As the identification of some noise sources may be very difficult, such as the hydraulic sources, the fault identification by using only the FRF of the accelerometer is not always obvious. The main difficulty of implementing a method to detect fault (in this case cavitation) by using the FRF method relies on the threshold that must be set to differentiate cavitation from normal operation. As can be seen on the results section, mainly on figure 5 (3000 RPM-5mm orifice), is not always obvious when the cavitation problem is present or not, and it gets even more difficult to make conclusions, because the response induced on the accelerometer by cavitation, in g amplitude varies for each shaft rotation speed. Future developments could regard in implementing an effective method to predict failure, some other approaches may be used, mainly the ones that permit to set a constant threshold to determine if cavitation is present or not. One possible approach is the use of the ARMA (Auto Regressive Moving Average), [12-14] mathematical model that applies directly on the time history of the accelerometer signal, and permits to set a constant threshold, or even the use of an ANN (Artificial neural network). Acknowledgements The authors would like to Gennaro Stingo, Giuseppe Iovino and Umberto Cesaro of University of Naples “Federico II” for the technical support, and at last, but not least BADAN Joao Pedro de Campos student at Supmeca (Public national engineering School of Paris) for technical support in editing the article and in preparing experimental tests. References [1] C. E. Brennen, 2014, “Cavitation and Bobble Dynamics”, Cambridge University Press, [2] D. Buono, D., Siano, A. Senatore, E. Frosina, 2017, “Gerotor pump cavitation monitoring and fault diagnosis using vibration analysis through the employment of auto-regressive-moving-average technique” Simulation Modelling Practice and Theory, 71, pp. 61-82, [3] D. Buono, Schiano di Cola, F. D., A. Senatore, E. Frosina, Buccilli, G., Harrison, J., 2016, “Modelling approach on a Gerotor pump working in cavitation conditions” 70th Conference of the Italian Thermal Machines Engineering Association, ATI2016, Energy Procedia, [4] M. Pellegri, A. Vacca, E. Frosina, D. Buono, A. Senatore, 2016, “Numerical analysis and experimental validation of Gerotor pumps: A comparison between a lumped parameter and a computational fluid dynamics-based approach”, Institution of Mechanical Engineers Part C Journal of Mechanical Engineering Science 1989-1996, pp. 203-210, [5] A. Senatore, D. Buono, E. Frosina, L. Santato (2013), “Analysis and Simulation of an Oil Lubrication Pump for the Internal Co mbustion Engine”. In: Fluids Engineering Systems and Technologies. San Diego, 15-21 November 2013ASME, Proceedings (IMECE) vol. 7B, ISBN/ISSN: 9780791856321, doi: 10.1115/IMECE2013-63468, [6] A. Senatore, D. Buono, E. Frosina, A. De Vizio, P. Gaudino, A. Iorio (2014), “A Simulated Analysis of the Lubrication Circuit of an In-Line Twin Automotive Engine”. SAE 2014 World Congress and Exhibition; Detroit, MI; United States, [7] E. Frosina, A. Senatore, D. Buono, L. Santato, “Analysis and Simulation of an Oil Lubrication Pump for Internal Combustion Engines”, ASME J. of Fluids Eng., 2015, Vol. 137, Issue 5 [8] Siano, D. , Panza, M.A., D’Agostino, D., “Knock detection based on MAPO analysis, AR model and discrete wavelet transform a pplied to the in-cylinder pressure data: Results and comparison”, SAE International Journal of Engines, Volume 8, Issue 1, 1 January 2015, Pages 113, ISSN: 19463936, DOI: 10.4271/2014-01-2547. [9] De Bellis, V. , Teodosio, L. , Siano, D., Minarelli, F., Cacciatore, D., “Knock and Cycle by Cycle Analysis of a High Per formance V12 Spark Ignition Engine. Part 1: Experimental Data and Correlations Assessment”, SAE International Journal of Engines, Volume 8, Issue 5, 6 September 2015, ISSN: 19463936, DOI: 10.4271/2015-24-2392. [10]Siano, D., Severi, E., De Bellis, V. , Fontanesi, S., Bozza, F., “Employment of an auto-regressive model for knock detection supported by 1D and 3D analyses”, Energy Procedia, Volume 81, 1 December 2015, Pages 704-714, 69th Conference of the Italian Thermal Engineering Association, ATI 2014, ISSN: 18766102, DOI: 10.1016/j.egypro.2015.12.076. [11] Siano, D. , Bozza, F. , D'Agostino, D. , Panza, M.A., (2014), “The use of vibrational signals for on-board knock diagnostics supported by in-cylinder pressure analyses”, Volume 2014-November, 11 November 2014, SAE/JSAE 2014 20th Annual Small Engine Technology Conference and Exhibition, SETC 2014. [12]Wilkes, D.M., Cadzow, J.A., ARMA Model Order Estimation Based on the Eigenvalues of the Covariance Matrix, (1993) IEEE Transactions on Signal Processing, 41 (10), pp. 3003-3009. [13]Ning Zhang, Minguan Yang, Bo Gao, and Zhong Li, (2015), “Vibration Characteristics Induced by Cavitation in a Centrifugal Pump with Slope Volute”, Shock and Vibration Volume 2015, [14]Akaike, H., (1970), “Statistical predictor identification “, Annals of the Institute of Statistical Mathematics, 22 (1), pp. 203-217.
