1、_6$(7HFKQLFDO6WDQGDUGV%RDUG5XOHVSURYLGHWKDW7KLVUHSRUWLVSX EOLVKHGE6$(WRDGYDQFHWKHVWDWHRIWHFKQLFDO and engineering sciences. The use of this report is entirely voluntary, and its applicability and suitability for any particular use, including any patent infringement arising therefrom, LVWKHVROHUHVSRQ
2、VLELOLWRIWKHXVHUSAE reviews each technical report at least every five years at which time it may be revised, reaffirmed, stabilized, or cancelled. SAE invites your written comments and suggestions.Copyright 2017 SAE InternationalAll rights reserved. No part of this publication may be reproduced, sto
3、red in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE.TO PLACE A DOCUMENT ORDER: Tel: 877-606-7323 (inside USA and Canada)Tel: +1 724-776-4970 (outside USA)Fax: 724-776-0790Ema
4、il: CustomerServicesae.orgSAE WEB ADDRESS: http:/www.sae.orgSAE values your input. To provide feedback on thisTechnical Report, please visithttp:/standards.sae.org/ARP6068AEROSPACERECOMMENDED PRACTICEARP6068Issued 2010-06Reaffirmed 2017-02Gas Turbine Engine Test Facility Vibration MeasurementRATIONA
5、LEARP6068 has been reaffirmed to comply with the SAE five-year review policy.TABLE OF CONTENTS 1.PURPOSE AND SCOPE 22.APPLICABLE DOCUMENTS AND REFERENCES . 23.VIBRATION MEASUREMENT 33.1General information 33.2Accelerometers . 33.3Velocity Transducers 33.4Mounting Locations and Brackets . 43.5Thrust
6、Stand Effects 43.6Vibration Filters . 43.6.1Broad-Band Vibration Measurement 43.6.2Narrowband Vibration Measurement 53.7Trim Balancing the Engine 53.8Spectrum Analysis 63.9Calibrating Vibration Measurement Systems 63.9.1Manual references 64.MEASUREMENT UNITS. . 65.DATA TYPES 96.EVALUATION AND ANALYS
7、IS 101. PURPOSE AND SCOPE The testprocedure per the applicable Engine Manual does require a vibration check for the low/intermediate and high speed rotorsystems. Release of an engine with high vibrations can result in: x On-wing vibration complaints, with subsequent troubleshooting x Rotorsystem fai
8、lures x Premature engine removals Limits are provided for transient conditions and steady state data points. Troubleshooting recommendations are limited to verification of the proper signal input and tracking. This practice provides recommendations for: x Correct cable and transmitter installation a
9、nd connections x Calibration x Recorded data interpretation and data analysis 2. APPLICABLE DOCUMENTS AND REFERENCES The following publications form a part of this document to the extent specified herein. The latest issue of SAE publications shall apply. The applicable issue of the other publication
10、s shall be the issue in effect on the date of the purchase order. In the event of conflict between the text of this document and references cited herein, the text of this document takes precedence. Nothing in this document, however, supersedes applicable laws and regulations unless a specific exempt
11、ion has been obtained. 2.1 SAE Publications Available from SAE International, 400 Commonwealth Drive, Warrendale, PA 15096-0001, Tel: 877-606-7323 (inside USA and Canada) or 724-776-4970 (outside USA), www.sae.org. AIR5026 Test Cell Instrumentation 2.2 Other References Bently D.E., Fundementals Of r
12、otating Machinery Diagnostics, 2002 Bentley Nevada, Minden Vance, J.M., Rotordynamics of Turbomachinery, 1988 John Wiley, Texas Ehrich, F.F. Handbook of Rotordynamics, 2004 Krieger publishing company, Malabar Florida SAE INTERNATIONAL ARP6068 Page 2 of 11_3. VIBRATION MEASUREMENT 3.1 General informa
13、tion In principle, the vibration of an engine under test can be measured by means of transducers which measure displacement, velocity, or acceleration. In practice, only velocity transducers, which are sometimes called vibration pickups, and accelerometers are used. Engine manufacturers specify vibr
14、ation limits in units of velocity or acceleration. Even with a complex vibration signal containing many frequency components, an acceleration signal can be converted into an equivalent velocity signal by means of a single integration, or into an equivalent displacement signal by means of two integra
15、tions. These integrations reduce the sensitivity of the system to high-frequency vibrations; conversely, they increase its sensitivity to low-frequency vibrations.It is not possible to make such conversions by simple calculator multiplications, except for pure sine-wave signals of known frequency. M
16、easurements made with narrowband filters (see 3.6.2) are an exception, as the output of a narrowband filter approximates a sine wave. Complex vibration signals require actual integrations.Although in theory it would be possible to convert a velocity signal to an equivalent acceleration signal by mea
17、ns of a differentiation, this is not done in practice because differentiation enhances high-frequency noise. 3.2 Accelerometers An accelerometer contains a seismic mass mounted on a piezoelectric element, typically quartz or a special type of ceramic material, which produces an electrical charge in
18、proportion to the force exerted on the element. Acceleration of the accelerometer body produces a force on the seismic mass, and therefore on the piezoelectric element, generating an electrical charge, typically of the order of tens of picocoulombs per unit of acceleration due to gravity. This charg
19、e is a relatively weak, high-impedance signal, which therefore is subject to electrical noise interference. Special low-noise cable should be used between an accelerometer and the first amplifier in the system, to prevent mechanical vibration of the cable from generating spurious electrical signals
20、caused by triboelectrical effects. Relative motion between mating connectors in this circuit can produce spurious signals similarly. Great care is needed in grounding accelerometer systems. Ground loops may inject excessive noise into the circuit. The amplifier should have a differential input, able
21、 to reject large amounts of common-mode noise. Typically the first amplifier is a charge amplifier, which accepts the electrical charge that was developed by the accelerometer and produces a low impedance voltage output that is accurately proportional to the input charge, and therefore to the accele
22、ration. When using single ended microdot accellerometers, which are used represent approximately 50% of all accelerometers in use, care must be taken to avoid creating ground loops in the wiring between the accellerometer and the amplifier. Accelerometers (as well as velocity transducers) need diffe
23、rential amps for high CMRR (Common Mode Rejection Ratio). Some manufacturers now offer accelerometers which contain one or more stages of amplification within the accelerometer housing. This avoids the triboelectrical effects which were discussed above, permitting use of ordinary electrical cabling.
24、 In many cases, however, the amplifier will not tolerate as wide a temperature range as the accelerometer will, limiting the application of the instrument. 3.3 Velocity Transducers A typical velocity transducer contains a magnetized seismic mass sliding on a guide rod, and restrained axially by spri
25、ngs. When the body of the transducer is vibrated along the axis of the guide rod, the seismic mass tends to remain in a constant position, producing relative movement between the mass and the transducer body. That movement generates a voltage in an electrical coil which surrounds the magnetized seis
26、mic mass; the voltage is proportional to the velocity of movement, typically of the order of several volts per meter/second. This relatively strong and low-impedance signal is less vulnerable to electrical noise and triboelectric effects than the weak high-impedance signal from an accelerometer. It
27、is still desirable to avoid ground loops and to use an amplifier with a differential input stage to reject common-mode noise. Velocity Transducers (as well as accelerometers) need differential amps for high CMRR (Common Mode Rejection Ratio).SAE INTERNATIONAL ARP6068 Page 3 of 11_3.4 Mounting Locati
28、ons and Brackets A few aircraft engines have accelerometers permanently mounted within the engine, usually on a major bearing. For most measurements, however, the vibration transducers will be mounted on the engine case. Aircraft engine designers go to great lengths to keep the engine as light as po
29、ssible. In consequence, the engine case is relatively flexible, and its vibration patterns will vary substantially at different locations of the engine. In order to make reproducible vibration measurements it is essential to mount the vibration transducers in precisely the locations specified by the
30、 engine manufacturer, and to use mounting brackets that are correctly designed to transmit the engine vibrations without modifying them. It is best to use the types of vibration transducers which are recommended by the engine manufacturer, and to purchase mounting brackets from the engine manufactur
31、er. To avoid introducing problems such as bracket resonance. If the vibration transducer is to be mounted in a location where the temperature exceeds the rating of the transducer, thermal insulation may be incorporated in the mounting bracket to protect the transducer from the engine heat, and some
32、form of cooling must be provided for the transducer. 3.5 Thrust Stand Effects The thrust of an engine has small rapid fluctuations which are discussed in 4.2. The fluctuations shake the engine and the thrust stand. The resulting vibrations look like low-frequency random noise, and are often spoken o
33、f as “thrust stand noise“. Vibration instruments for use with turbojet engines usually have high-pass filters cutting off at a specified engine dependan frequency to suppress that noise, which is acceptable since the noise is not related to engine performance. For new generations large Turbofan appl
34、ications operating at Idle speeds below 10 Hz a lower cutt off freqency is required. Some of the vibration characteristics of the engine under test are influenced by the way in which the engine is mounted. Engines which are tested with nonresilient supports, like the RB211, may be more affected by d
35、etails of thrust stand and adapter design than engines which have supports with controlled spring constants, like the JT9D and the CF6. 3.6 Vibration Filters Most vibration is related to unbalance in one of the engine rotors, although occasional trouble is caused by bearings or engine accessories. S
36、ome means is needed to identify the part of the engine that is producing excessive vibration. This is normally done by means of filtering techniques. 3.6.1 Broad-Band Vibration Measurement In testing older engines, band-pass filters are commonly used in the vibration instruments to separate the vibr
37、ations caused by the different rotors, since they turn at quite different speeds. In comparing vibration measurements made in different test cells with different vibration instruments, a frequent source of difficulty is the filter specification. For example, a bandpass filter for a particular engine
38、 might be specified to pass vibrations from 40 to 115 Hz, with a 3-pole - asymptotic to 18 dB/octave - attenuation curve above and below the passband.Unfortunately, some manufacturers specify the pass frequencies at the 95% gain point (5% attenuation, or 0.45 dB down), while others specify them at t
39、he 71% gain point (29% attenuation, or 3 dB down). Also, when dealing with multipole filters, the designer may choose to use Butterworth, Tschebyscheff, elliptical, or other filter designs, with an extensive variety of parameters. The effect of such variations depends on the amount of vibration ener
40、gy in the affected frequency regions. It may range from negligible to 50% of the reading or more. Another difficulty is the detection specification. In reading peak-to-peak values, older vibration instruments, and some current ones, use averaging detectors with the output meter calibrated to read 3.
41、14 times the average voltage. This gives the correct peak-to-peak voltage only if the signal is a sine wave. Some newer instruments use rms detectors, with the output meter calibrated to read 2.83 times the RMS voltage. This is more satisfactory than an averaging detector, but still is not precise.
42、It is also possible to measure the true peak-to-peak value directly with a relatively elaborate detector. With a typical complex vibration waveform, the three types of detectors will give readings that vary substantially. SAE INTERNATIONAL ARP6068 Page 4 of 11_If the filter specifications and detect
43、or specifications of the vibration instruments in the post-overhaul test cell are equivalent to the specifications of the instruments in the manufacturers baseline test cell, the broad-band vibration measurements should agree reasonably well. If the specifications are quite different, the two instru
44、ments, both working correctly, may differ by as much as 2:1 in measuring a given vibration signal. Without detailed knowledge of the vibration frequency spectrum and the instrument specifications, it is not possible to predict how the two readings will compare. 3.6.2 Narrowband Vibration Measurement
45、 The manufacturers of many of the newer engines specify that vibration measurements are to be made with a narrowband filter. Such measurements are less affected by details of the filter and detector design than are broadband measurements, so that better agreement with the manufacturers data can be e
46、xpected. Many of the narrowband filters can be automatically tuned to the rotation frequency of one of the engine rotors, following it as the rotor speed changes. Such a filter is called a tracking filter. Tracking filters are very convenient, making it simple todetect rotor unbalances, but they can
47、not be used in studying any significant vibration peaks that do not correspond to a rotor speed. For such studies the operator must tune the filter manually over the frequency range of interest. The operator must use care in working with narrowband filters, which typically have bandwidths of the order of 5 to 10 Hz. Some such filter designs have constant bandwidth, others have constant Q, which means that the ratio of center frequency to bandwidth is constant. With any such filters i
