Technology Notebook Contents

Previous ~ ~ Next

Noise & Vibration

James W. Vogel

A paper delivered at the SFTE 11th Symposium in 1980

ABSTRACT

Flyover noise testing is now an important part of the function of Flight Test organizations everywhere both because of the requirements of Federal Aviation Regulations (FAR), Part 36 - "Noise Standards: Aircraft Type and Airworthiness Certification", and of an increased public awareness of aircraft noise. Because of the large costs involved in flight testing it would be extremely advantageous to be able to predict flyover noise levels based on static tests.

This paper discusses some of the questions that must be answered in order that these predictions be made with acceptable accuracy. In addition, a test program is described which was conducted by the Lockheed-California Company and Rolls-Royce Limited in order to develop the required techniques and instrumentation and to acquire the data needed to answer the questions alluded to above. Some preliminary results of this program are also presented.

INTRODUCTION

It is almost a truism to say that noise makes news these days. Airports are awash in law suits, military bases get numerous complaints, and airframe and engine manufacturers are under great pressure to produce a quieter product. In the civil aviation field the law requires that certain categories of aircraft make no more than specified amounts of noise. For example, Federal Aviation Regulation (FAR) Part 36 sets the limits as shown in Figure 1 for subsonic transport category large airplanes and turbojet powered airplanes and in Figure 2 for propeller driven small airplanes (Reference 1).

In addition to the requirements of FAR Part 36, there are numerous airports with curfews, preferential runways, and, especially in Europe, individual airport noise requirements that may include monitoring.

All of the above is by way of saying that noise is an important part of flight testing today and promises to become even more important tomorrow. Thus Flight Test organizations everywhere are finding it neces sary to conduct expensive and time consuming programs not only to certify the airplane initially and/or meet specific customer requirements but also to develop the data base and test techniques necessary to allow for product growth and to reduce future test time and costs.

The certification requirements are specific and all of us understand, for the most part, what is required. What is not so clearly evident, however, is the necessity to develop a data base from which future models of a particular aircraft may be certified, e.g., one of in creased gross weight, one that is stretched, one that is shrunk, etc. Also, the requirement to support customer requests for expanded data relative to noise at different flap settings, climb-out procedures, etc. must be satisfied.

A further demand upon Flight Test is that of being able to certify an aircraft when the engine manufacturer makes changes in the engine. The Federal Aviation Administration (FAA), and the customer, hold the airframe manufacturer responsible for the noise (or at least most of it) but the airframe manufacturer must certify it and/or demonstrate compliance with customer guarantees. Thus, if an engine manufacturer increases an engine's thrust, or changes acoustic treatment, or makes any one of an infinite number of possible modifications, the airplane manufacturer must show compliance with FAR 36. This is not to say that this is unfair. This is just to say that the responsibility of showing compliance with FAR 36 rests with the airframe manufacturer.

It is clear then that there are many requirements for flyover noise testing today. Perhaps the easiest of these to understand is that dealing with the actual certification testing itself. FAR Part 36 delineates in great detail what is required in this case (Reference 2, for example). As previously mentioned, the testing required to develop a data base for future certification, customer requirements, and engine modifications is nowhere clearly defined, however. It is just such a flight test program involving the Lockheed L-l0ll TriStar and the Rolls-Royce RB.211 engine that will be discussed in subsequent sections of this paper (Figure 3) .

TEST PROGRAM

As the Rolls-Royce RB.211 engine has progressed over the years, our changes have been made to different components of the engine. Many of these have had an effect on the certified noise levels of the L-l0ll although these effects have not always been easy to assess. One of the reasons for this difficulty is the way in which the noise levels from the different engine components, e.g., fan bypass, core, primary jet, etc., combine to produce the total engine noise spectrum. Figure 4 illustrates the situation. If one of the lesser components is modi fied then the result will be that there is little or no effect on the spectrum. If, however, one of the greater contributors is changed then the effect can be considerable. Part of the problem is to know the contribution of the individual components. What one generally measures is the total spectrum. Hence, it would be extremely ad vantageous to be able to sort out the contributions of the separate sources. These advantages accrue not only from being able to identify those sources that significantly influence the engine's noise spectrum, and those that are weak contributors, but also from being able to identify those sources where acoustic treatment or design can be applied most effectively. This knowledge and ability would be extremely useful in discussions with the FAA vis-a-vis the effect on noise certification of engine modifications and, also, in cost- effective engine design relative to noise.

In addition to the problem of the contribution of the several pro pulsive sources to the noise spectrum, there are airframe effects themselves to consider. Of considerable interest today is the aerodynamic noise (sometimes called "self-noise") of the airframe, i.e., the noise generated by the airflow over the surface of the wings, fuselage, flaps, landing gear, etc. (Reference 3, for example). This non-propulsive noise represents a so-called "floor" below which the noise cannot be reduced without an extensive redesign of the air frame. This can be seen by referring to Figure 4 which shows how the noise from individual components combines into the resultant noise spectrum. Even if the propulsive noise were reduced to less than the aerodynamic noise we would still have the problem of reducing the noise of the airframe/engine system, in this case dominated by aerodynamic noise. The state of the art is such that it could not be done, e.g., we don't know how to make the landing gear quiet. This hints at the importance of knowing the value of this floor in as much as there is little to be gained by setting noise requirements that are lower than this floor because they could not be attained even by the total elimination of the propulsive noise contributions and we certainly have not accomplished this yet.

Related to the self-noise problem is that of the impingement of the primary jet flow upon the flaps. Specifically, one wants to know if, in addition to the aerodynamic noise undoubtedly produced by the flaps, there is an additional noise source caused by the high veloc ity jet flow striking the flaps. If so then this is yet another area of the total noise signature that will be unaffected by acoustic treatment.

One further item requires mentioning and that is the effect of forward velocity on the noise (Reference 4). Engine manufacturers, out of necessity, acquire most of their data from static tests. These data are then input to the next design cycle and so on. It is impor tant then that these static results carry over to the flight testing that will ultimately occur. If the effect of forward velocity pro duces noise signatures significantly different from those obtained statically then one could make erroneous decisions based upon such static tests. Thus one wants to know if there is a significant in fluence due to forward velocity on propulsive noise and, if so, how to predict it.

Based on the above problems and considerations it became clear that both Rolls-Royce and Lockheed had much to learn from a comprehensive flight test program investigating some of the aforementioned problems utilizing the L-l0ll airframe and the RB.211 engine. Accordingly, such a program was devised and carried out. The details follow. It should be noted that in addition to the questions that the program was designed specifically to investigate, advantage was taken of the opportunity presented to study the effects of meteorological conditions and microphone placement on the data. Meteorological data were acquired by a tethered balloon complete with telemetry, an instrumented light airplane, and the test L-l0ll itself. Temperature, relative humidity and wind speed and direction were also acquired at the conventional ten meters.

Microphone placement can effect the results primarily through the can cellation/reinforcement phenomenon due to the direct and reflected acoustic ray. To investigate this, microphones were mounted at ground level; at 1.2 meters (FAR 36 height) and at ten meters (Reference 4).

Tests Conducted

The test program itself was as shown in Figure 5 . Note that Runs 1-14 and 17-25 were planned in order to determine the effects of aircraft speed on engine noise over the approach-power-to-takeoff-power range of the engine. Airframe noise was minimized by flying clean, i.e., flaps, slats, and gear up. Altitude over the center of the microphone array was also held constant (except for Runs 22 and 23) in order to minimize corrections to the data. Runs 22, 23 and 24 (conducted at a constant airspeed) serve as a check on the validity of the atmospheric absorp tion corrections applied to the data. Data measured at 300 and 600 feet can be corrected to 900 feet, for example, and compared with the results actually measured from the 900 foot flyover.

Runs 15 and 16 were designed to investigate the influence of flaps on the noise signature. Note that Run 16 was performed with all engines at flight idle and hence is a first measure of the contribution to the aerodynamic noise signature made by the flaps.

Runs 16, 26, and 27 had as their purpose the investigation of the self-noise generated by flaps and/or landing gear. Again, note that all engines were at flight idle.

Run 28 is a typical approach configuration.

It should also be pointed out that from Runs 24 and 25 an estimation of the installation effects can be made. The number two engine is tail-mounted with an S-duct on the L-l0ll and the number one and number three engines are mounted in pods under the wings (Figure 3) .

Test Procedures

Each condition was flown at least twice, i.e., until we had obtained two acceptable runs from the standpoint of the acoustical data, the aircraft speed and altitude, the engine power setting, and the meteorological conditions speed. Note that the restraint on speed was +/- 5 knots and on altitude +/- 50 feet. The weather conditions had to comply with the requirements of FAR 36, Amdnt 9, plus no temperature inversion. These are as shown in Figure 6 .

The testing was accomplished at the Lockheed Flight Test Center in Palmdale, California. Figure 7 is an aerial view of the site. The passes were all made in one direction to simplify the procedure, as well as to aid in the integration of the light airplane (used for meteorological purposes) into the pattern. Two "marks" were given to the aircraft (via radio) from the ground:

1) One at the point of overhead of the microphone array in order to allow the crew to assess speed and altitude re lative to the target values.

2) One at the end of each test run at which point engine power settings, aircraft ground track, etc. could be changed.

Speed and power were set by a geographic landmark, visible from the aircraft, and a mark was transmitted from the aircraft at this point. On the ground a stopwatch was started upon receipt of this mark. At the overhead mark the elapsed time between the marks was noted and this amount of time was allowed to transpire prior to transmitting the end-of-test mark. This insured that there would be sufficient data at constant speed and engine power, both before and after over head, to perform the planned analysis.

Because the effect of airspeed, i.e., forward velocity, on flyover noise was under study so it was necessary that it remain constant for each individual pass. The effects of landing gear and flaps on noise were also under investigation and, thus neither device could be used to control speed. Hence, to avoid speed changes, it was necessary to allow the airplane to climb or descend once the speed and power were set. The only other way would have been to load the aircraft to the exact weight that would have allowed straight and level, unaccelerated flight. Since at least 56 passes were planned, all to be accomplished within stringent meteorological restraints, this was obviously impractical.

In order to determine where the starting point of the climb or descent should be (to bring the aircraft over the microphone array at the desired altitude) plots of rate of climb or descent versus aircraft gross weight were made. This was done for each desired airspeed for different flap settings, gear up or gear extended, and different engine power settings. One typical plot is shown in Figure 8 . With the aid of these plots, plus one or two practice flights, it was possible to select one starting point and vary the altitude over this point in such a way so as to arrive over the array at the proper height. One other device that aided in determining the height at which to begin the run was the use of a so-called "calibration run". For each new configuration tested a level flight pass was made, in that configuration, over the array. This run was conducted using the thrust required for level flight at the desired altitude and airspeed. This enabled the pilot to develop a good "feel" for just how the airplane really would fly at that particular weight. If the power settings called out in the flight card were greater than that required for level flight then the airplane would climb. If the power settings requested were less than that required for level flight then the airplane would descend.

Instrumentation and Set-Up

As mentioned, the test site is shown in Figure 7 . A schematic of the instrumentation set-up is given in Figure 9 . The test headquarters was in a specially equipped trailer shown in Figure 10 . This housed most of the instrumentation and served as the base of operations for test coordination, communications, etc.

Note that the source location array (Figure 11) consisted of twenty-two microphones flush-mounted on a smooth asphalt surface. A close-up of this pad is shown in Figure 12 . It was painted white to reduce the heat absorbed by the asphalt and thus to insure that the microphones that were lying on the surface were not exposed to too high a temper ature and, also, to avoid large temperature gradients at the surface. The design of this asphalt surface was such that there were no parallel sides in order to prevent any standing waves between edges. The edges were made flush with the surrounding terrain to the extent possible. Note that the size of the pad was approximately 6500 square feet.

All microphone data were recorded on a thirty—two track tape recorder with signal conditioning (Figure 13) at the recorder in the trailer located about 230 feet from the microphone array. Thirty microphones were used with one track for voice annotation and one track for IRIG time code. Through this time code all data on the ground were syn chronized with the airplane parameters. All microphones were cali brated using a pistonphone and both pink noise and white noise were recorded through each microphone system for response corrections.

The ten-meter microphones, i.e., those mounted ten meters above ground level, (Figure 14) were erected using telescoping masts with guy-wires for support. The 1.2 meter microphones were mounted on more conventional tripods.

The primary tracking instrument was a Rolls-Royce ground-based 70mm camera positioned on the expected ground track of the airplane, as shown in Figure 15 . This system had a separate three-track tape re corder dedicated to it. TRIG time code and voice annotation were recorded on one track, the shutter pulses on another, and the elevation and tilt angle of the camera (i.e., left or right of the expected ground track) on the third channel. This latter was accomplished by transforming voltages that were proportional to the two angles into two separate and distinct frequency bands. On playback, the merged signals could then be separated. From these inputs and scaling techniques involving the aircraft image the space position as a function of time could be obtained.

The back-up tracking system utilized an airborne camera on the test aircraft itself photographing surveyed ground targets.

The weather balloon (Figure 16) had its own paper tape recording system on which pressure, height, temperature, and relative humidity were recorded for later analysis. These parameters were telemetered to the recording station from the balloon as it ascended or descended and real-time read-out was available to determine that meteorological conditions above the surface were, or were not, acceptable for testing.

In addition to the balloon, a fully instrumented light aircraft was utilized to make spiraling descents through the test area immedi ately preceeding the test airplane itself in order to obtain temper ature and humidity data as a function of height above the surface for use in subsequent analysis. Real-time read-out of these parameters was again available for use in determining the suitability for testing. The data from the weather balloon and the light airplane were used in deciding whether to launch the test aircraft. The test airplane also carried instrumentation to determine temperature and relative humidity as a function of altitude.

Winds aloft were determined by pilot balloons (pibals). These were launched from the ground at the test site and tracked to obtain wind speed and direction as a function of altitude. Winds at ten meters were obtained from the ten meter weather tower.

Aircraft and engine parameters such as airspeed, altitude, total air temperature, engine pressure ratios, shaft speeds, fuel flow, and turbine gas temperatures were recorded onboard through the airplane data center for subsequent reduction and analysis, and correlation with the noise data.

RESULTS TO DATE AND CONCLUSIONS

The flight test program described in this paper has been completed. The principal areas associated with flyover noise that it was designed to examine were:

1) In-flight source location determination.

2) The effect of forward speed on the noise produced by coaxial jets over a power range from flight idle to takeoff.

3) The influence of airframe noise on the engine/airframe noise signature.

4) Engine/airframe installation effects.

5) Atmospheric effects on acoustic propagation. The data obtained have been verified to be of high quality.

To date the data have been reduced only in terms of overall sound pressure level and for the test day atmospheric conditions, i.e., no corrections have been made for meteorological variations between test days. In addition, no corrections have been made as yet to normalize the data to a common altitude.

Although the data are still being analyzed, some preliminary conclu sions can be drawn. One such conclusion is that in the rear arc engine noise dominates but in the forward arc the noise levels are airframe dominated. This can be seen in Figure 17 where the noise data for flight idle thrust are plotted for several different air speeds. It can be seen that in the rear arc the data for the dif ferent airspeeds all come together, indicating that in this direction they are not a function of airspeed but, rather, of engine power which here is constant. However, in the forward arc they separate quite nicely as a function of airspeed. Figure 18 indicates that this is true for engine powers up to approximately 75% Nl, i.e., throughout the approach power range.

Another tentative finding from this program is that deploying either the flaps (to 33 deg) or the landing gear increases the overall noise level by approximately eight dB relative to the level associated with a clean airframe. Deploying both simultaneously adds about ten dB to the clean airframe overall noise level. These effects are shown in Figure 19 .

There is also some evidence to indicate that there are less static-to-flight effects relative to the center engine than for the wing engines.

In addition, it appears likely that there is indeed a jet/flap interaction effect on noise. This appears to be true regardless of engine power setting and acts to increase the noise levels signifi cantly, especially in the forward arc.

The investigation, primarily by Rolls-Royce personnel, is continuing into these latter two phenomena as well as all the other aspects of the program. It is anticipated that more definitive results will be available by the end of this year.

Acknowledgement is hereby given to Rolls-Royce Limited for their conception and implementation of this program. Appreciation is also expressed for their support in the preparation of this paper.

1. Anon., Federal Aviation Regulations, Part 36, "Noise Standards: Aircraft Type and Airworthiness Certification", 1 December 1969.

2. Vogel, James W. "Flyover Noise: Measurement and Analysis", Society of Flight Test Engineers Eighth annual Symposium Proceedings, August 1977.

3. Fethney, P. and Jelly, A. H., "Airframe Self-Noise on the Lockheed L-l0ll TnStar Aircraft", AIM Paper 80-1061, 6th Aeroacoustics Conference, June 1980.

4. Szewczyk, V. N., "Coaxial Jet Noise In Flight", AIM Paper 79-0636, 5th Aeroacoustics Conference, March 1979.

Table of Content ~~ Previous ~~ Top ~~ Next