Enhanced Modeling of Aircraft Taxiway Noise, Volume 2: Aircraft Taxi Noise Database and Development Process (2013)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

6-1 CHAPTER 6. PROPELLER AIRCRAFT TAXI NPD CURVE GENERATION USING EMPIRICAL DATA A process has been developed for the creation of Propeller Aircraft Taxi NPDs from empirical data. All of the propeller aircraft in the INM data base are either identified as prop or turboprop based on the field ‘ENG_TYPE’ in the aircraft.dbf file. There are 28 aircraft with prop or turboprop designations for which NPD and spectral class data were developed. These will be referred to in this report as props. The process developed for the commercial jet aircraft fleet was leveraged and applied to the Propeller NPD development procedure to the greatest extent possible. The prop NPD approach utilizes a single thrust-noise sensitivity relationship for all propeller aircraft, developed from a composite of the available empirical data in lieu of having suitable ANOPP prop predictions from first principles. This prop NPD technique is a hybrid of the jet methods described in Chapter 4. 6.1. Propeller Taxi NPD Procedure Step 1. Process the Empirical Acoustic Taxi Data and determine the taxi NPD noise metrics (SEL, EPNL, Lmax, PNLTmax) as a function of distance for the reference atmospheric conditions for the measured taxi condition. Step 2. Assign a “nominal” taxi thrust (Fn/delta) to the acoustic data based on an assumed taxi thrust derived from the aircraft maximum certificated takeoff gross weight. Step 3. Develop a thrust-noise sensitivity from the empirical data as a function of metric and distance. Step 4. For a given prop aircraft determine the baseline (anchor point) taxi noise based on either of the following (provided in priority order): • Empirical data utilized in Step 1; • Approximated taxi thrust level based on the aircraft takeoff gross weight. Step 5. Evaluate NPD levels at other power settings based on the relationship between taxi noise and thrust derived from empirical data across all prop data (Step 3). In step 3, the empirical thrust-noise relationship serves as a surrogate for the ANOPP sensitivity used in Method I for the Turbofan (Jet) aircraft (Chapter 4). We were unable to obtain suitable prop ANOPP datasets which matched certificated noise values from which we could develop analytical thrust noise sensitivities. This procedure is described in further detail in the Empirical Thrust-Noise Relationship section. 6.2. Prop Spectral Class Assignment Spectral classes were derived from the empirical data wherever possible based on the spectrum at the point of maximum A-weighted noise level for a taxi operation and normalized to 70dB at 1000 Hz. This is consistent with the procedures outlined in Chapter 2, Section 2.3. 6.3. Empirical Datasets Table 6-1 itemizes the seven aircraft for which empirical taxi noise data is available. Three prop aircraft (ATR-72-500, Fokker-50, and DHC-8Q3) were included in the Madrid report (Lopez et al., 2004) which was also utilized for development of the Commercial Jet NPD dataset. Procedures (Section 3.1) to

6-2 convert sound power to spectral noise source data were also applied to the props. Additional empirical data for military propeller aircraft with equivalent commercial variants was found in the NoiseFile database from NoiseMap. NoiseFile data includes (for many aircraft) static idle acoustic data for a variety of metrics, plus directivity and spectral data. TABLE 6-1 Empirical Datasets for Prop Aircraft used for Prop Thrust-Noise Sensitivity Aircraft Type Data Source Notes ATR-72-500 Madrid Not in the INM/AEDT database. Used to expand measured database in finding noise vs. thrust curve fits Fokker-50 Madrid Not in the INM/AEDT database. Used to expand measured database in finding noise vs. thrust curve fits DHC-8Q3 Madrid DHC-8Q300 measured in Madrid L188 NoiseFile L188C/ALL 501-D13 NoiseFile Surrogate P-3A C130 NoiseFile C-130H/T56-A-15. NoiseFile C-130H C130E NoiseFile C-130E/T56-A-7. NoiseFile: C-130E 1900D NoiseFile Beech 1900D/PT6A67. NoiseFile 1900D 6.4. Taxi Thrust – Weight Relationship A graph of the maximum gross takeoff weight (ICAO/CAEP8, 2008) versus the static thrust and the relationship between these parameters, as was used in the generation of the NPD database for the fixed wing jet fleet, is shown in Figure 6-1 and provided in Equation 6-1. The prop aircraft data is overlaid on the fixed wing fleet to show the suitability of utilizing the same Taxi Thrust – MTOGW relationship. FIGURE 6-1 Max takeoff weight vs. Taxi thrust/1000 (lb) for INM Commercial Jet and Prop aircraft. Trend line was developed based on INM Commercial Jets as reported in Chapter 4. y = -2.1852E-06x2 + 1.5114E-02x R² = 9.7640E-01 0 2 4 6 8 10 12 14 0 100 200 300 400 500 600 700 800 900 1000 To ta l T ax i T hr us t/1 00 0, lb f MTOW/1000, lbm Taxi Thrust Correlation 5% of ICAO BPDB T/O Thrust ANOPP Aircraft Cases INM Props Poly. (5% of ICAO BPDB T/O Thrust)

6-3 ܶ ൌ െ2.1852ܧ െ 06 ∗ܹݐଶ ൅ 1.5114ܧ െ 02 ∗ܹݐ Eq. 6-1 Where Wt is the aircraft takeoff gross weight/1000 in lbs. and T is Total Taxi Thrust/1000 in lbs. 6.5. NPD Curve Development Using AAM An input file for the Advanced Acoustic Model (Page, et al., 2009a) was created which simulated the aircraft taxiing by at 16 knots on a long track placing the noise source representing the airplanes at the height of their engines. Point of interest locations were placed 4 feet above ground at slant distances equal to those of INM’s NPD curves. The ground was modeled as flat earth with a flow resistivity of 150 Pa s/m. The atmosphere was modeled with a temperature of 77oF, 70% relative humidity, and one atmosphere of pressure. The output of AAM for each modeled source contains the integrated metrics and a spectral time history at each of the point of interest locations for the entire pass-by that was modeled. The integrated metrics at each of the point of interest locations for a given aircraft are used to create the NPD curves. The NPD curves for the three metrics: A-weighted Lmax, A-weighted SEL, and EPNL, are plotted in Figures 6-2 through 6-4, respectively for the Madrid prop aircraft. FIGURE 6-2 A-weighted maximum level versus distance for taxiing aircraft.

6-4 FIGURE 6-3 A-weighted SEL versus distance for taxiing aircraft. FIGURE 6-4 EPNL versus distance for taxiing aircraft. 6.6. Empirical Thrust-Noise Relationship Linear fits of the various noise metrics versus thrust parameter ln(Thrust/2) were developed for the seven aircraft with empirical taxi noise data as Step 3. The linear fits are plotted for different distances and overlaid with the empirical data in Figure 6-5 through Figure 6-10. In a few instances, the integrated metrics (EPNL and SEL) go slightly negative, but only at the very largest distance (25,000Ft) and for very low thrust levels. The limit of applicability of the correlation is approximately 100lb total thrust, or a thrust parameter (ln(thrust/2)) of ~4. Note that the linear fits of the data improve as the distance L increases (Figure 6-13). However, at the very largest distances, the goodness-of-fit deteriorates somewhat, which is to be expected, considering the probable inaccuracies of the extrapolation methodology for very large distances. The standard deviation (sigma) and the goodness-of-fit (R2) of the

6-5 linear fits vs. distance L are shown in Figure 6-14. These figures show that above 10,000 ft, the fits do deteriorate somewhat but are still better than the very close-in fits. There is more non-linearity in the noise vs. thrust behavior at close-in distances, and with larger propagation distances the air attenuation reduces the influence of higher frequencies and the metrics behave more linearly. For each metric and thrust correlation parameter the slope and y-intercept were determined. These are displayed in Figures 6-11 and 6-12. These provide the thrust-noise sensitivity employed in the prop NPD development. FIGURE 6-5 Taxi noise metric vs. thrust parameter for 400 ft. distance, all props. FIGURE 6-6 Taxi noise metrics vs. thrust parameter for 1000 ft. distance, all props. 40 50 60 70 80 90 100 110 0 1 2 3 4 5 6 7 8 9 N oi se  M et ric , d B Thrust Correlation Parameter ln (thrust/2) Taxi Noise vs. ln(Thrust/2) Data vs. Linear Fits  Turboprops ‐ L = 400 ft. EPNL dBA‐Max PNLT‐Max SEL EPNL FIT dBA‐Max Fit PNLT‐Max Fit SEL Fit 40 50 60 70 80 90 100 110 0 1 2 3 4 5 6 7 8 9 N oi se  M et ric , d B Thrust Correlation Parameter ln (thrust/2) Taxi Noise vs. ln(Thrust/2) Data vs. Linear Fits  Turboprops ‐ L = 1000 ft. EPNL dBA‐Max PNLT‐Max SEL EPNL FIT dBA‐Max Fit PNLT‐Max Fit SEL Fit

6-6 FIGURE 6-7 Taxi noise metrics vs. thrust parameter for 4000 ft. distance, all props. FIGURE 6-8 Taxi noise metrics vs. thrust parameter for 10,000 ft. distance, all props. 20 30 40 50 60 70 80 90 0 1 2 3 4 5 6 7 8 9 N oi se  M et ric , d B Thrust Correlation Parameter ln (thrust/2) Taxi Noise vs. ln(Thrust/2) Data vs. Linear Fits  Turboprops ‐ L = 4000 ft. EPNL dBA‐Max PNLT‐Max SEL EPNL FIT dBA‐Max Fit PNLT‐Max Fit SEL Fit 0 10 20 30 40 50 60 70 80 0 1 2 3 4 5 6 7 8 9 No ise  M et ric , dB Thrust Correlation Parameter ln (thrust/2) Taxi Noise vs. ln(Thrust/2) Data vs. Linear Fits  Turboprops ‐ L = 10000 ft. EPNL dBA‐Max PNLT‐Max SEL EPNL FIT dBA‐Max Fit PNLT‐Max Fit SEL Fit

6-7 FIGURE 6-9 Taxi noise metrics vs. thrust parameter for 16,000 ft. distance, all props. FIGURE 6-10 Taxi noise metrics vs. thrust parameter for 25,000 ft. distance, all props. 0 10 20 30 40 50 60 70 80 0 1 2 3 4 5 6 7 8 9 No ise  M et ric , dB Thrust Correlation Parameter ln (thrust/2) Taxi Noise vs. ln(Thrust/2) Data vs. Linear Fits  Turboprops ‐ L = 16000 ft. EPNL dBA‐Max PNLT‐Max SEL EPNL FIT dBA‐Max Fit PNLT‐Max Fit SEL Fit 0 10 20 30 40 50 60 70 80 0 1 2 3 4 5 6 7 8 9 No ise  M et ric , dB Thrust Correlation Parameter ln (thrust/2) Taxi Noise vs. ln(Thrust/2) Data vs. Linear Fits  Turboprops ‐ L = 25000 ft. EPNL dBA‐Max PNLT‐Max SEL EPNL FIT dBA‐Max Fit PNLT‐Max Fit SEL Fit

6-8 FIGURE 6-11 Metric and thrust correlation parameter slopes, all props. FIGURE 6-12 Metric and thrust correlation parameter y-intercepts, all props. 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 100 1000 10000 100000 Lin ea r Fi t Sl op e m Taxi Distance from Source, ft. Linear Fit Slope m for Noise vs. Thrust Turboprops EPNL dBA‐Max PNLT‐Max SEL ‐60.0 ‐40.0 ‐20.0 0.0 20.0 40.0 60.0 80.0 100.0 100 1000 10000 100000 Lin ea r Fi t  I nt er ce pt  b Taxi Distance from Observer, ft. Linear Fit Intercept b for Noise vs. Thrust Turboprops EPNL dBA‐Max PNLT‐Max SEL

6-9 FIGURE 6-13 Linear fit standard deviation sigma for props. FIGURE 6-14 Linear fit goodness of fit “R2” for noise vs. thrust for turboprops. 6.7. Spectral Class Data The point of interest 1000’ from the taxi track was used to capture the spectrum at the time of the maximum A-weighted sound level. This spectrum represents the free-field emissions to the point of interest. The ground effect has been removed and only the atmospheric absorption and spherical spreading are accounted for in the propagation. The slant range at the time of maximum A-weighted sound level was not necessarily 1000’ due to the directivity of the aircraft, but by using the slant range, the spectrum was adjusted to 1000’ by subtracting the extra spreading and air absorption past 1000’. All 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 100 1000 10000 100000 Li ne ar  Fit   St d.  De vi at io n S ig m a,  dB Taxi Distance from Observer, ft. Linear Fit Standard Deviation "Sigma" for Noise vs. Thrust Turboprops EPNL dBA‐Max PNLT‐Max SEL 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 100 1000 10000 100000 Lin ea r Fi t  G oo dn es s‐o f‐F it R ^2 Taxi Distance from Observer, ft. Linear Fit Goodness of Fit "R^2" for Noise vs. Thrust Turboprops EPNL dBA‐Max PNLT‐Max SEL

6-10 three spectra were then normalized to 70 dB. The three spectra for the Madrid aircraft are plotted in Figure 6-15. FIGURE 6-15 Spectra of taxiing aircraft at time of A-weighted maximum level 1000’ away corrected to 70 dB at 1 kHz. As can be seen from the previous figures, the NPD curves of these three aircraft are very similar while their spectra are not. Based upon visual comparison, these three spectra should be used as three separate taxi spectral classes. The significant differences in the spectral shapes are in the region below 160 Hz. This could possibly be attributed to propeller Blade Passing frequencies being different. It might be possible to create a common spectrum with adjustments for blade passing frequency differences in the future, however this was not explored. In INM, the ATR-72-500 aircraft uses the HS478A substitution for flight operations. The flight departure and arrival spectral classes are plotted with the ATR-72-500 spectrum in Figure 6-16.

6-11 FIGURE 6-16 Taxi spectrum of ATR-72-500 plotted with the INM spectral classes used for it. 6.8. INM Propeller Database and Substitution Assignments The full INM propeller database contains 28 aircraft. For each aircraft type a judgment was made based on the aircraft size, configuration, thrust class, engine types and where appropriate substitutions were made from the existing empirical data. For other props the general taxi thrust-weight curve was made to assign a taxi thrust, and the taxi noise metric vs. thrust parameters (both slope and y-intercept) were used to create the NPD. This in essence has the effect of assigning a single “generic” taxi NPD curve to those prop aircraft for which no empirical data is available. While the NPD values are the same, the spectral classes followed a substitution process described above so the data entries should not necessarily be combined within INM / AEDT. Table 6-2 describes the specific aircraft processes applied in the development of the INM NPDs and spectral class data.

6-12 TABLE 6-2 INM Prop Aircraft and NPD / Spectral Class Development Notes (PropNPDs_rev6c.xlsm). 6.9. INM Propeller NPD Summary Noise Power Distance curves were developed for each of the 28 unique INM propeller aircraft using procedures defined in this Chapter. Each of these aircraft was assigned a spectral class that represents the character of their noise emissions during taxi operations. A composite database with INM/AEDT Aircraft Info, NPD and spectral class data for both the propeller fleet and the Turbofan (Jet) fleet is provided in Appendix F. INM ACFT_ID NOISEFILE DESCRIPTOR ACFT_DESCR Source ATR72500 Madrid FOKKER50 Madrid DHC830 Use Madrid DHC8Q30 DASH 8-300/PW123 Madrid DHC8 Use Madrid DHC8Q30 DASH 8-100/PW121 Madrid Substitute 1900D Same BEECH 1900D / PT6A67 Noisefile C130 C-130H C-130H/T56-A-15 Noisefile C130E Same C-130E/T56-A-7 Noisefile DC6 C-118 DC6/R2800-CB17 Noisefile L188 P-3A L188C/ALL 501-D13 Noisefile BEC58P U-4B BARON 58P/TS10-520-L Noisefile Substitute CNA172 U-4B-3dB CESSNA 172R / LYCOMING IO-360-L2A Noisefile Substitute CNA206 U-4B-3dB CESSNA 206H / LYCOMING IO-540-AC Noisefile Substitute CNA20T U-4B-3dB CESSNA T206H / LYCOMING TIO-540-AJ1A Noisefile Substitute CNA441 U-4B CONQUEST II/TPE331-8 Noisefile Substitute COMSEP U-4B-3dB 1985 1-ENG COMP Noisefile Substitute CVR580 T-29 (CV240) CV580/ALL 501-D15 Noisefile Substitute DC3 T-29 (CV240) DC3/R1820-86 Noisefile Substitute DHC6 Use Beech 1900 DASH 6/PT6A-27 Noisefile Substitute DHC6QP Use Beech 1900 DASH 6/PT6A-27 RAISBECK QUIET PROP MOD Noisefile Substitute DHC7 Use Beech 1900 DASH 7/PT6A-50 Noisefile Substitute EMB120 Use Madrid ATR-72-500EMBRAER 120 ER/ PRATT & WHITNEY PW118 Noisefile Substitute GASEPF U-4B-3dB 1985 1-ENG FP PROP Noisefile Substitute GASEPV U-4B-3dB 1985 1-ENG VP PROP Noisefile Substitute HS748A T-29 (CV240) HS748/DART MK532-2 Noisefile Substitute M7235C U-4B-3dB MAULE M-7-235C / IO540W Noisefile Substitute PA28 U-4B-3dB PIPER WARRIOR PA-28-161 / O-320-D3G Noisefile Substitute PA30 U-4B-3dB PIPER TWIN COMANCHE PA-30 / IO-320-B1A Noisefile Substitute PA31 U-4B PIPER NAVAJO CHIEFTAIN PA-31-350 / TIO-5 Noisefile Substitute SD330 Use Beech 1900 SD330/PT6A-45AR Noisefile Substitute SF340 Use Beech 1900 SF340B/CT7-9B Noisefile Substitute