Document
Simulation - Based Assessm ent of a Full - Scale
Installed Quiet Landing Gear
Benjamin Duda
Dassault Systèmes , D - 80637 Munich, Germany
R yan J. Ferris
Dassault Systèmes , Long Beach , California , 90802 , USA
Mehdi R. Khorrami
NASA La ngley Research Center, H ampton, Virginia, 23681, USA F ull - scale simulations of a Gulfstream G - III aircraft , performed in support of the NASA Acoustic Research Measurements flights , are presented to complemen t results discussed in earlier studies . The flow solver employed in those studies , Dassault Systèmes ’ l attice ® Boltzmann PowerFLOW , was also used during this investigation to conduct time - dependent simulations of the entire aircraft in landing configuration with a fully dressed landing gear .
The high - fidelity simulations , perfo rmed at a Mach number of 0.23 and a Reynolds number of 10.5 × 10 based on mean aerodynamic chord , capture d all relevant airframe noise sources .
The computations were used to assess the aeroacoustic performance of the main landing gear , with and without no ise reduction fairings installed , of a G - III aircraft equipped with Adaptive Compliant Trailing Edge technology and conventional Fowler flap s . To facilitate comparison of predicted noise signature s with effective perceived noise levels obtained from flight test measurements, the “ as - flown ” nose landing gear geometry, missing in our earlier studies, was added to the simulated G - III aircraft configurations. The high fidelity , synthetic data were post - processed using a F fowcs - W illiams and H awkings integral app roach to estimate farfield acoustic behavior , with pressures on the model solid surface or pressure and velocity components on a permeable surface enveloping the acoustic near field used as input.
I. Nomenclature C = Pressure coefficient p D = Main landing gea r wheel diameter Re = Reynolds number based on mean aerodynamic chord δ = Flap deflection angle f Acronyms DNS = Direct N um erical S imulation EPNL = Effective Perceived Noise Level ERA = Environmentally Responsible Aviation FDC = Flight Demonstrations and Capabilities FWH = Ffowcs - Williams and Hawkings LBM = Lattice Boltzmann Metho d MLG = Main l anding g e ar Senior T echnical Specialist, Simulia A&D.
Solution Consultant Specialist, Simulia A&D, Member AIAA .
Aerospace Engineer , Comp utational AeroSciences Branch, Associate F ellow AIAA.
NLG = Nose landing gear PNL = Perceived Noise Level PNLT = Tone - corrected PNL PSD = Power Spectral Density VLES = Very Large Eddy Simulation II. Introduction E xcess ive exposure to aircraft noise can adversely affect the qualit y of life of communities living near major airports [1] . The introduction of commercial fleets with substantially quieter and more fuel - efficient aircraft is essential if civil aviation is to expand at the rate necessary to mee t the anticipated growth in global air traffic [2] . Mitigation of aircraft noise remain s a critical goal of the NASA Aeronautics Research Mission Directorate (ARMD) . Although most efforts initially focused on reduction of propu lsion noise, the airframe component of the generated sound has received much attention in recent years due to its prominence during aircraft approach and landing.
Development and matur ation of viable airframe noise reduction (NR) technologies was vigorous ly pursued under the Environmentally Responsible Aviation (ERA) project [3] . This effort culminat ed with the evaluation of flap and landing gear NR technologies via flight testing under the Flight Demonstrations and Capabil itie s (FDC) project of ARMD [4] [5] . Equally important , advancement and validation of system - level, simulation - based airframe noise prediction methodologies that could handle the extreme geometrical complexities and the ir intricate, time - dependent flow field s – responsible for sound generation – w as initiated under ERA [6] , [7] , [8] and is being continued as a crit ic al element of the FDC project [9] , [10] .
NASA evaluated the aeroacoustic performance of several airframe NR technologies installed on a Gulfstream - III (G - III) aircraft during the Acoustic Resear ch Measurements (ARM) flight test campaign . The first test (ARM - I, 2016) evaluated the Adaptive Compliant Trailing Edge (ACTE) concept as a means to alleviate the noise produced by the wing flaps [4] , [5] . The second test (ARM - II, 2017) targeted main landing gear (MLG) and gear cavity NR treatments in combination with the ACTE technology [4] , [5] . For the third flight test (ARM - III, 2018), the ACTE flaps were removed and th e original Fowler flaps were re installed on the G - III aircraft to obtain baseline flap and landing gear data, and to assess the noise reduction capability of the landing gear technologies for conventional flaps [4] . In support of the ARM - I and ARM - II test s , our previous computational studies targeted the configurations that were being tested, namely, the ACTE - equipped G - III with and without the NR fairings instal led on the main landing gear [10] , and the G - II I aircraft with its Fowler flaps deflected at 20 and MLG deployed . This latter configuration served as a baseline to gauge the aeroacoustic performance of the ACTE technology and landing gear fairings [9] . A thorough comparison between predicted (synthetic) and measured phased microphone array data yielded very good agreement for the ir respective integrated farfield noise spectra [11] . To reduce computationa l resources and expedite time - to - solution, the nose landing gear (NLG) was omitted in our previous simulations. As a result, direct comparison s between predicted and measured Effective Perceived Noise Level s (EPNL), an important noise metric used by indust ry to determine the effectiveness of a given NR technology, would have been questionable at best .
The present computational study is a continuation of our previous efforts. T o facilitate validation of predicted farfield noise spectra and EPNL values, a hi gh - fidelity replica of the “ as flown ” nose landing gear was developed and integrated into the simulated G - III geometry . With the addition of the nose gear, the focus of the present study are those G - III configurations that were flight - tested during the ARM - III campaign : Fowler flaps deflected at 20 and 39 with and without MLG fairings installed .
III. Aircraft Model The process es used to develop the high - fidelity digital models that replicat e most of the geometric detail of the “ as flown ” G - III testbed, includi ng flap s (Fowler and ACTE) and MLG with and without fair ings, are presented in Refs.
[9 ] , [ 10 ]. Except for the nose landing gear, t he CAD model of the full - scale, full - span aircraft contained all the major components – fuselage, wing with deflected flaps i ncluding brackets, tail, flow - through nacelle with hush - kit, and main landing gear – of the actual testbed .
F or the present computational study, a highly defined CAD model of the nose landing gear, including its cavity, was developed from carefully conduct ed laser scans of the component while the aircraft was placed on jacks to have the gear in the fully extended landing configuration. Significant effort was spent ensuring that the finest details of the nose gear exterior surfaces (i.e., outer mold lines) w ere captured accurately. The complete aircraft model, with nose landing gear integrated, is shown in Fig. 1 . For an accurate repre se n tation of the aircraft in the simulation, a triangular ® surface mesh is created by tessallating the elements of the CAD mode l u sing PowerDELTA .
a) Aircraft with nose landing gear integrated b) Close - up view of nose gear Fig. 1 Gulfstream G - III aircraft CAD model .
IV. Computational Approach ® Th e numerical simulations were performed using Dassault Systèmes’ PowerFLOW solver, which is based on the three - dimensional , 19 - state (D3Q19) lattice Boltzmann method (LBM) [12] , [13] , [14] , [15] . LBM has been extensively validated for a wide variety of applications ranging from academic direct numerical simulation (DNS) cases to industrial flow problems in the fields of aerodynamics [16] and aeroacoustics [6] , [17] , [18] . At a macroscopic level, LBM uses a simpler and more general physics formulation than methods based on the Navier - S tokes equations [12] . The LBM equations recover the macroscopic hydrodynamics of t he Navier - Stokes equations through the Chapman - Enskog expansion [19] , [20] .
The local for mulation of the LBM equations allows a highly efficient implementation for distributed computations on thousands of processors. The low dissipation and dispersion properties of the numerical scheme typically produce aerodynamic and aeroacoustic results tha t are comparable to those obtained with classical CFD solvers that use higher - order large eddy simulation (LES ).
A. Turbulence Modeling The lattice Boltzmann flow simulation is equivalent to a DNS of the flow. For high Reynolds number (Re) flows, such as tho se addressed in this work, the lattice Boltzmann Very Large Eddy Simulation (LB - VL ES) approach is used to reduce computational resource requirements [14] , [21] . This means that turbulence is modeled in areas of attached flow such as boundar y layers , but resolved in wakes or regions of detached flows.
B. Wall Treatment The standard lattice Boltzmann bounce - back boundary condition for the no - slip condition or the specular reflection for the free - slip condi tion are generalized throug h a volumetric formulation [12] , [13] near the wall for arbitrarily oriented surface elements ( s urfels) within the Cartesian volume elements ( v oxels). This formulation of t he boundary condition on a curved surface cutting the Cartesian grid is automatically mass, momentum, and energy conservative while maintaining the general spatial second - order accuracy of the underlying LBM numerical scheme. To reduce the resolution requi rements near t he wall for high Re flows, a hybrid wall function is used to model the region of the boundary layer closest to the solid surfaces [16] , [22] .
C. Volume Meshing The lattice Boltzmann approac h is solved on voxels where variable resolution (VR) regions can be defined to allow for local mesh refinement by successive factors of two. Based on the facetized geometry and the local volume resolution, the model surface is discretized by planar surfels . This process allows the automatic generation of computational grids for any arbitrarily complex geometrical shapes. Except for the regions surrounding the nose landing gear and its wake, the meshes created in this study are identical to the ones used in the previous simulation s for the Fowler and the ACTE flap configur ations, cf. [8] and [9]. Fig. 2 s hows the volume mesh created in the vicinity of the nose landing gear. Care was taken to sufficiently resolve geometric and flow features.
nd Fig . 2 Volume mesh in the vicinity of the nose landing gear, only every 2 line shown .
D. Simulation Setup Like the precursor study without the nose landing gear, simulations were performed at full scale with free - air boundary conditions at a Mach number of 0.23 and a Reynolds number of 10.5 × 10 based on a mean aerodynamic chord of 13.78 ft (4.2 m) and an aircraft speed of 150 knots. Th is Re represents a value that is close to 60% of the flight Re ynolds number ; thus, it is sufficiently high to pr oduce farfield noise levels that are nearly equivalent to those obtained at full flight but at a lower cost . The simulation domain was initialized with free flow conditions except in the immediate near field region around the aircraft, where zero flow velo city and free stream pressure w ere specified .
The first ~0.60 s of the transient simulations were discarded to remove any art i facts from initialization. The simulation s were then run for another ~1.55 s of physical time, during which various data were reco rded for post - processing.
The simulation campaign replicate d flight test s for the landing aircraft . For this reason, a different angle of attack was simulated depending on the flap settings : = 5 for a Fowler flap deflection of 20 , = 2.85 for a Fowler flap deflection of 39 and = 4.65 for a n ACTE flap deflection of 25 .
E. Measurement Entities for Far f ield Noise Propagation Since the main goal of this study is to predict farfield noise l evels, Dassault Systèmes’ farfield noise solver ® PowerACOUSTICS was use d to obtain synthetic pressure signals at defined locations far away from the aircraft. An acoustic analogy approach based on the Ffowcs - Williams and Hawkings (FWH) formulation [23] was used here with the efficient and well - validated formulation developed by Farassat [24] , also known as formulation 1A. The formulation is extended to account for uniform mean flow convection effects to simulate the noise generated and measured in an ideal infinite wind tunnel [25] .
The input for the FWH computations are either a pressure dataset obtained on the geometry or a pressure and velocity dataset obtained on a per meable surface around the geometry . Since the NLG is rather isolated from other noise producing components, two different sets of permeable surfaces were cr e ated. This allows the measurement of farfield characteristics for the nose landing gear only. The r esulting surfaces are shown in Fig. 3 , where multiple endcaps are included to filter out hydrodynamic pressure fluctuations created by the wakes.
a) F ull aircraft b) NLG only Fig. 3 Permeable measurement surfaces for FWH computation , shown with multiple end caps .
In order to replicate the isolated farfield noise contribution of the NLG for the solid FWH approach, the input surfaces can be split. This is shown in Fig. 4 , where t he blue - colored surfaces of the NLG and the forward segment of the aircraft fuselage can be used as separate input to the FWH computation.
Fig. 4 S urfaces used for predicting isolated NLG contribution to farfield noise obtained with solid FWH formulation .
F. Effective Perceived Noise Level The Effective Perceived Noise Level (EPNL) was calculated using a proprietary post - processing tool developed by Dassault Systèmes. The primary reference for the definition and calculation of EP NL is a publication of the ICAO [26] . The input to the tool is a semicircular arc of synthetic pressure signals obtained with FWH. Once a trajectory is determined, it is discretized into flight segments of a given duration, in this c ase 500 ms. For every flight segment, the emission time and position of the aircraft are determined, and a ray between the observer and aircraft is traced. The ray’s intersection with the semicircular microphone array is found and a noise spectrum for the observer is formulated by interpolating the appropriate microphone ’ s narrow - band spectra. For each flight segment, t he spectrum at the observation point is corrected for distance, atmospheric absorption, Doppler shifting , and ground reflection. One - third o ctave sound pressure levels (SPL) are then computed and Perceived Noise Level (PNL) is calculated using the procedure described by the ICAO. Finally, tonal weighting and band sharing adjustments are made and EPNL is computed.
V. Numerical Results A select num ber of configurations tested during the ARM - III campaign (Fowler flap - equipped G - III with/without MLG fairings) were simulated . Additionally, t he NLG geometry was also added to the G - III ai rcraft with ACTE flap s .
A. Grid Resolution Study In our previous stud y [9] , the impact of grid resolution on the numerical solution was assess ed for the Fowler flap configuration . This wa s done in a global manner , i.e. , t he edge length of each vox el was scaled consecutively by a factor of 1.5 to obtain three resolutions termed “coarse”, “medium” and “fine”. Th at study is replicated here and now includes the NLG. The resulting mesh sizes and required time steps are summarized in Table 1 for the Fowler Flap configuration . The ACTE configuration requires slightly less computational time due to the elimination of flap side edges and brackets , which typically require high resolution. At the submission deadline of th is manuscript , the fine - resolution simulations were not yet completed . Only instantaneous aerodynamic and acoustic quantities are available for comparison because they require shorter physical simulation times for statistical convergence than the full record needed to reconstruct the farfield spectra in the desired frequency range.
Firstly, a comparison of pressure coefficients are presented in Fig. 5. Note from the figure that, a s found in previous studies, surface pressure s on the wing converge fast with increased spatial resolution and only small d ifferences at the suction peaks are visible . Attention is now turned toward the NLG, where flow separations are highlighted by iso - surfaces of total pressure in Fig. 6 . Here, small differences become visible depending on the resolution. This is expected be cause increased resolution will lead to the capture of smaller turbulent structures, which in turn can impact separation lines and mean flow.
Fig. 7 s hows farfield spectra obtained with the solid and the permeable FWH computation for a flyover microphone l ocated 120m below the aircraft . D ue to insufficient time - depend e nt records from the ongoing fine - resolution simulation , only coarse and medium resolution results could be included here . Observe from the figure that increased spatial resolution leads to add itional content in the higher frequency range. Consistent with the findings in Ref. [10] , t his trend is expected to remain true for the fine simulation , as is the assumption that overall acoustic behavior was properly captured , up to 3 kHz , with the medium resolution results to be discussed in subsequent sections. Also note from Fig. 7 that the noise levels for the two FWH surfaces are similar in the mid - frequency range , but results obtained with the permeable FWH formulation sh ow an earlier cut - off . This is a consequence of the lower spatial resolution of the flow field out to the permeable surface compared to the solid surface. The presence of tones can be observed , whose characteristics are dependent on the FWH formulation and the resolution. The lower frequency tone is due to a cavity mode that will be discussed in detail in section IV.B. The higher frequency tone stems from a hollow MLG post, which was described in Ref. [9] and identified to be se nsitive to local flow and spatial resolution.
Table 1 Simulation metrics used to study spatial resolution effects on Fowler flap configuration with f lap deflection angle δ f = 39 °.
6 6 - 4 6 Resolution Voxels [10 ] Surfels [10 ] Min Edge [10 m] Timesteps CPUh [10 ] Coarse 1 711 41 7.2 2 293 760 0.2 Medium 5 436 72 4.8 3 440 640 1.2 Fine ( not finished ) 17 436 126 3.2 5 177 344 ≈ 8.5 a ) I nboard flap edge b) M id - flap b) O utboard flap edge Fig. 5 Pressure distribution at three different spa nwise locations for coarse, medium and fine simulation for Fowler flap deflection angle δ f = 39 °.
a) C oarse resolution b) M edium resolution b) F ine resolution Fig. 6 NLG wake comparing resolution levels a) S olid FWH formulation b) P ermeable FWH formul ation Fig . 7 Power spectral density obtained with FWH formulation for coarse and medium simulation s for flap deflection angle δ = 39 °.
f B. Impact of NLG on A erodynamics T he aerodynamic impact of the NLG on the downstream components of the aircraft is assessed in this section . A snapshot of the flow is given in Fig. 8 , where r esolved vo rtical structure s are shown as iso surface s of vorticity based on criterion . As expected, the unsteady simulation resolves a large spectrum of turbulent structures as the flow passes over the nose landing gear. Since the mesh resolution is maintained downstream of the gear, turbulent wake structure s are convected with the local flow and sustained over long distance s . O bserve from the figure that these structures impinge on the leading edge of the wing close to the fuselage junction and that the majority pass above the wing.
For a quantitative compar ison of the flow field, two coarse simulations with a Fowler flap deflection of 20° were conducted at the same angle of attack; one with out NLG and one with NLG. The additional computational cost for the latter is in the order of 10%. The resulting pressur e distributions are shown in Fig. 9 for three different spanwise locations. The impact of the NLG is hardly discernable throughout the wing . To examine NLG effect s on the flow field upstream of the MLG , c ontours of s treamwise and spanwise velocity componen ts on a spanwise plane 1.5 wheel diameters (D) upstream of the MLG are plotted in Fig. 10 . The extracted velocity fields reveal that the p ortion of the NLG wake that does not pass over the wing occupies only a small region on the aircraft’s underside . Due to the large distance between left and right MLG s , these components encounter a rather undisturbed flow. The same holds true for the inboard flap side edges, indicating that the impact of the NLG on the acoustics of the MLG and high lift system is minimal .
Velocity magnitude m/s Fig . 8 Snapshot of resolved turbulent structure for Fowler flap deflection δ = 20 configuration with nose and f main landing gear deployed , medium resolution .
a) I nboard flap edge b) M id - flap b) O utboard flap edge Fig . 9 Pressure distribution at three different spanwise locations for configur ation with NLG and without NLG and Fowler flap deflection angle δ = 20 °.
f a) S treamwise contours b) S panwise contours Fig . 10 V elocity contours for simulation with (top) and without (bottom) NLG o n a plan e 1.5 D upstream of MLG for a Fowler flap deflection δ = 20°.
f C . NLG Cavity Noise Even though the NLG geometry was carefully scanned to reflect the flown geometry, the NLG cavity had to be geo metrically approximated and sub systems inside the cavity (e.g ., major linkages and hydraulic lines ) have been omitted in this study. Similar to the MLG bay discussed in Ref. [8] , an idealized (empty) cavity can lead to large resonances in a simulation. Certainly, this is a meaningful computational result for the modeled bay , but nonetheless undesirable since the real cavity either does not produce such tones or , if it does , they will be smeared and have low amplitudes due to the presen ce of the sub systems . For this reason, a sponge zone was applied to the ceiling of the cavity to dampen acoustics reflections for all configurations with landing gear deployed . The impact on farfield noise is shown in Fig. 11 , which depicts coarse resolution results for a flyover microphone located 120m below the aircraft .
The noise levels were computed with the solid FWH formulation and on ly the blue surfaces, cf. Fig. 4, were used as input to highlight the impact on NLG noise. A very strong cavity mode is seen with the first harmonic at about 270Hz.
With inclusion of the sponge zone , the cavity mode is still present but pollution of the sp ectrum by higher harmonics in the mid - frequency range has been removed successfully.
Fig . 11 NLG contribution to farfield noise levels at flyover microphone for coarse simulation with and without sponge zone in NLG cavity.
Since both solid and permeable FWH computations can be performed for the entire aircraft and the nose landing gear only, the corresponding spectra for the flyover microphone are s hown in Fig . 12 . Two prominent tones are visible, the lower one stem ming from the above - mentioned cavity mode and the higher one from the hollow MLG front post .
The striking difference between these two tones is that the former is visible only in the solid FWH formulation , while the latter is present independent of the FWH formulation.
Fig . 12 Farfield noise levels at flyover microphone for full aircraft and isolated NLG obtained with solid and permeable FWH formulation.
The reason for this behavior has been investigated by creating the pressure dilatation field. Two sna pshots with a time difference Δt equivalent to half the period of the cavity mode are presented in Fig. 13 . Observe that strong and large pressure fluctuations are present inside the NLG cavity. Fig. 13 b reveals , as expected , that the cavity mode has chang ed its phase by 180°. While this cavity mode leads to strong pressure fluctuations inside the cavity surface, the mode is trapped and hardly any fluctuations propagate outside. This means that the permeable surface, which is located away from the NLG , does not see these pressure fluctuations.
a) Streamwise contours at time t b) S treamwise contours at time t+Δt Fig . 1 3 Pressure dilatation field on aircraft symmetry plane at two different times showing a trapped NLG cavity mode .
D . Impact of NLG on F arfield A eroacoustics In order to assess the impact of the NLG on farfield acoustic s , results from tw o medium resolution simulations are compared in this section. The first one is a Fowler flap configuration with a flap deflection δ = 20° with both MLG f and NLG deployed. The second simulation, discussed in Ref. [8], is geometrically identical to the first one except for the NLG being retracted. A difference in angle of attack of 0.75 ° exists between the two simulated configurations , but it has been demonstrated that variation s of a few degrees ha ve little impact on farfield noise signature s [27] .
Directivity contour maps are presented in Fig. 14 . They have been computed from 360 microphones located on a semicircular arc with a radius of 120m centered on the aircraft in flyover direction . For consistency, only the solid FWH formu lation has been used to compute the microphone signals. The left directivity map contains the results from the simulation with the NLG and the right directivity map contains the results from the simulation with out the NLG .
Note from Figs. 15a and 15b that adding the NLG to the configuration increases the SPL almost throughout the entire frequency and directivity ranges, as expected. The largest differences in SPL are seen in the low - to mid - frequency range a t forward directivity angles. The global maximum a round 300Hz between 0° and 40° for the simulation with NLG can be attributed to the NLG cavity tone, which is not present in the aft directivity angles. For a quantitative assessment, the differences in SPL between these simulations is plotted in Fig. 15 . An increase of approximately 2 dB in the mid - frequency range is observed , which decreases for higher frequencies . Only for the low frequency region below 200Hz , a decrease caused by a small shift in cavity tones is visible .
Sample spectra are shown in Fig . 16 for the overhead position at 90 ° , highlighting that addition of the NLG increases SPL values in the mid frequency range by about 2 dB . Possible reason s for the reduced impact in the higher frequency range are insufficient spatial resolution in our sim ulations and/ or the masking effect of MLG or flap side edge noise.
a) With NLG b) W ithout NLG rd Fig . 1 4 Directivity contour maps in 1/3 octave bands for Fowler flap s deflected 20° .
Fig . 1 5 Di fferences in dir ectivity contour s f or configuration with Fowler flap s deflected 20 , with and without NLG geometry .
rd a) Power spectral density b) 1/3 octave SPL Fig . 1 6 Spectral compari sons for Fowler flap configuration with and without NLG geometry .
As implied in section IV.E, in the FWH computation we can numerically remove the noise contribution of the NLG from a flow simulation where the NLG was geometrically present. Th e motivation is to compare the obtained noise levels (NL G1) with those of a flow simulation that did not contain the NLG geometry at all (NLG0) . If these two results are quantitatively similar, then the NLG can be regarded as an isolated noise source with little interaction with MLG and flap side edge noise . Th e differences that result ed from subtracting NLG1 from NLG0 , are shown i n Fig. 17 for the configuration with Fowler flaps deflected 20° . The differences between these two results are within ±0.5 dB except for the low frequency range below 200 Hz. The large r differences could be caused by slight changes in the MLG cavity modes induced by the NLG wake. Overhead spectra are shown in Fig. 18 , where the previous observation is confirmed: very small difference s in the mid - to high - frequency range and larger diffe rences in the very low - frequency range. Th us, the hypothesis that the NLG can be regarded as a n isolated noise source for mid to high frequencies for this configuration and flow conditions is supported .
Fig . 1 7 Di fferences in dir ectivity contour s between a flow simulation where the NLG was present but its contribution removed in the FWH computation and a flow simulation where the NLG was not present.
rd a) Power spectral density b) 1/3 octave SPL Fig . 1 8 Spectral compari sons of a flow simulation where the NLG contribution was removed in the FWH computation and a flow simulation where the NLG was not present.
F . Flyover Noise T he noise generated by the air frame was also analyzed from the standpoin t of an observer on the ground as the aircraft pass es overhead. W e will discuss the tone - corrected Perceived Noise Level (PNL T ) as a function of time as well as the overall Effective Perceived Noise Level (EPNL) in decibels (EPNdB). The calculation procedu re for these two quantities can be found in Section IV . F. W e will make determinations of differential PNL T and EPN L as a result of configuration changes in th e following ways : 1) addition of the NLG, 2) use of both flap and landing gear noise reduction con cepts in landing configuration, 3) aerodynamically equivalent configurations , and 4) use of MLG fairings for baseline flap configuration . T he data presented here are based on the full aircraft solid measurement surface, as shown in Section IV . E, in medium resolution.
S imulations were also performed for the ACTE equipped G - III , δ = 25° , with MLG deployed and porous MLG f fairings installed , with and without NLG deployed. Fig. 19 shows the change in tone - corrected PNL (PNLT) values as the aircraft passes over the observer for these two configurations . The overall shape of the PNL T curve is similar , with a n increase in noise levels throughout the aircraft’s trajectory caused by the deployed NLG . This indicates that the additional noise content generated by the NLG is not necessarily dominant in a single directivity , as was also show n in the contour maps of Fig. 15 . Looking at the change in PNL T between the two configurations, an increase max of approximately 3.7 dB is recognized with the addition of the NLG and its cavity to the aircraft geometry. EPN L also increase d by 3.2 EPN dB. Alt hough aircraft noise reduction is a system - level endeavor, the component - level results of the previous sub section showed that the overall noise contribution from the NLG is rather isolated. Thus, we may infer that the increase in EPNdB from NLG deployment can be offset by addi ng the porous MLG fairing s : the decrease in flyover noise associated with the fairings [10] is approximately equa l to the increase from NLG deployment shown here .
Of interest to the present study was the determination of overall total noise reduction of the G - III in landing configuration as a result of the flap and MLG NR technologies. The Baseline G - III aircraft considered is that with the Fowler Flap δ = 39°, MLG/NLG deployed, and without porous MLG fai ring s installed. The G - III with NR concepts f considered is that with the ACTE Flap δ = 25°, MLG/NLG deployed and with porous MLG fairing s installed. From f the PNLT curve in Fig. 20, we observe that most of the noise reduction occurs as the aircraft is appr oaching and is directly overhead, with PNLT having been reduced by approximately 7.2 dB. A corresponding decrease in EPNL max was also about 7.1 EPNdB. The largest change in PNLT is identifiable immediately before and after the baseline G - III aircraft trave ls overhead, indicating that the NR technologies not only affect the overall sound level but also alter the characteristic of the sound reaching the observer.
Fig . 19 Effect of NLG deployment on the PNLT time history of a G - III in landing with NR concept s installed (ACTE δ = 25 ° with MLG deployed and porous MLG fairings).
f Fig . 20 Effect of NR technologies on the PNLT time history of a G - III in landing. Baseline: Fowler δ f = 39 ° , MLG/NLG deployed, without porous MLG fairings; treated: ACTE δ f = 25 ° , MLG /NLG deployed with porous MLG fairings.
The change in flap system from Fowler to ACTE led to a reduction in aerodynamic performance with flaps deployed, as established in [10] . Since airframe noise from lifting surfaces, such as flaps, is proportional to the magnitude of lift generated, a brief investigation was undertaken to quantify noise from aerodynamically equivalent configurations. T o reduce the lift created by the Baseline G - III aircraft (Fowler δ = 39°, MLG/NLG deployed, and f without porous MLG fairing s installed) , the Fowler flap deflection was decreased from 39° to 20° while leaving other settings the same (now labeled Baseline G - III Aircraft – Aero Equivalent) . The PNL T time histories for th is configuration and th e equivalent ACTE with flaps deflected 25 and porous MLG fairings installed are shown in Fig.
21 . Observe that the flap and MLG NR concepts tested reduce d PNL T by 4.2 dB while reducing EPN L by 4. 6 max EPN dB.
The PNLT time histories fo r the two configurations of Fowler flap s deflected 20 , MLG/NLG gear deployed with and without porous MLG fairings are compared in Fig. 22 . The plot demonstrates the acoustic performance of MLG fairings in conjunction with a conventional Fowler flap system . As shown in the figure, applica tion of the MLG fairings decreased PNLT values between 1 and 2 dB over the entire period , resulting in an EPNL reduction of 1.2 EPNdB .
Fig. 21 PNL T t ime histories for the aerodynami cally equivalent Baseline G - III (Fowler δ f = 20 ° , MLG/NLG deployed, without porous MLG fairing s) and the acoustically treated G - III (ACTE δ = 25 ° , MLG/NLG f deployed with porous MLG fairing s) Fig. 22 PNLT t i me histories for Fowler f lap 20 , MLG/NLG deployed with and without porous MLG fairings.
VI. Conclu ding Remarks and Outlook A simulatio n campaign was conducted for a full G - III in landing configuration where , in contrast to previous simulations , a highly detailed NLG was included . Th e simulations were done for the aircraft equipped with conventio nal Fowler and ACTE flaps set at various deflection angles. Flight conditions were chosen to match flight tests. Although fine resolution simulations had not been completed by the time this report was submitted, a spatial resolution study indicated that ae rodynamic quantities converge d fairly quickly.
The impact of the NLG on both aerodynamics and acoustics was investigated. The aerodynamic effect was minor, since the wake from the NLG pass ed mostly over the wing without interacting with the MLG due to the relatively large distance between its port and starboard components . We showed that including an empty NLG cavity can lead to high - amplitude tones and that those can be alleviated with numerical damping inside the cavity without impacting the noise charac teristics of the NLG itself. Different strategies for obtaining farfield noise were investigated , with the outcome that the noise contribution of the NLG can be isolated from other prominent airframe sources . This is due to the large distance that exists b etween the NLG and the MLG and flap side edges. By adding the NLG , an increase of about 2 dB is observed for all directivity angles in the mid - frequency range. The impact decreases for higher frequencies due to insufficient spatial resolution and partial m asking of NLG noise by MLG and flap side edge sources .
The acoustic results generated from the present simulations are compared to flight test data in our companion paper [28] to ascertain the capability of the computational a pproach in predicting the farfield noise signature of a full - scale aircraft with and without airframe noise reduction treatments .
Acknowledgments This work was supported by the Flight Demonstrations and Capabilities (FDC) project under the Integrated Aviat ion Systems Program (IASP) of the NASA Aeronautics Research Mission Directorate. We would like to express our sincere appreciation to personnel at the NASA Armstrong Flight Research Center for their assistance with development of the full - scale G - III aircr aft geometry , particular ly Daniel Nolan and Michael Yandel l of Jacobs Engineering for meticulously creating a high - definition CAD model of the nose landing gear from the laser - scan files .
The authors also gratefully acknowledge the invaluable contribution of Scott Brynildsen of Craig Technologies for providing geometry modifications and CAD support. All the simulations were performed on the Pleiades supercomputer at the NASA Advanced Supercomputing (NAS) facility at Ames Research Center. The logistical supp ort provided by NAS staff is greatly appreciated.
References [1] M. Basner, C. Clark, A. Hansell, J. I. Hileman, S. Janssen, K. Shepherd and V. Sparrow, "Aviation Noise Impacts: State of the Science," Noise & Health, vol. 19, no. 87, p p. 41 - 50, 2017.
[2] Boeing, [Online]. Available: http://www.boeing.com/resources/boeingdotcom/commercial/about - our - market/assets/downloads/cmo_print_2016_final_updated.pdf. [Accessed 17 October 2017].
[3] M. R. Khorrami, J. W. M. Humphreys and D. P. L ockard, "An Assessment of Flap and Main Landing Gear Noise Abatement," in AIAA Paper 2015 - 2987 , 2015.
[4] M. R. Khorrami, D. P. Lockard, J. W. M. Humphreys and P. A. Ravetta, "Flight - Test Evaluation of Airframe Noise Mitigation Technologies," in Paper to be presented at the AIAA/CEAS Aeroacoustics Conference in Atlanta, Georgia, June 2018 , 2018.
[5] E. Baumann and E. Waggoner, "Flight and Ground Operations in Support of Airframe Noise Reduction Tests," 2018 AIAA/CEAS Aeroacoustics Conference,, Vols. AIAA paper 2018 - 2970, 2018.
[6] M. R. Khorrami, E. Fares and D. Casalino, "Towards Full Aircraft Airframe Noise Prediction: Lattice Boltzmann Simulations," AIAA Paper 2014 - 2481, 2014.
[7] E. Fares, D. Casalino and M. R. Khorrami, "Evaluation of Airframe Noise Reduction Concepts via Simulations Using a Lattice Boltzmann Approach," AIAA Paper 2015 - 2988, 2015.
[8] M. R. Khorrami and E. Fares, "Simulation - Based Airframe Noise Prediction of a Full - Scale Full Aircraft," AIAA Paper 2016 - 2706, 2016.
[9] J. Appelbaum, B. Duda, E. Fares and M. R. Khorrami, "Airframe Noise Simulations of a Full - Scale Aircraft," AIAA Paper 2018 - 2973, 2018.
[10] R. J. Ferris, J. Appelbaum and M. R. Khorrami, "Simulations of a Full - Scale Aircraft with Installed Airfra me Noise Reduction technologies," AIAA Paper 2018 - 2974, 2018.
[11] M. R. Khorrami, P. A. Ravetta, D. P. Lockard, B. Duda and R. Ferris, "Comparison of Measured and Simulated Acoustic Signatures for a Full - Scale Aircraft with and without Airframe Noise A batement," AIAA Paper 2018 - 2975, 2018.
[12] H. Chen, "Volumetric Formulation of the Lattice - Boltzmann Method for Fluid Dynamics: Basic Concept," Physical Review E, vol. 58, no. 3, pp. 3955 - 3963, 1998, doi: dx.doi.org/10.1103/PhysRevE.58.3955.
[13] H. Chen, C. Texeira and K. Molvig, "Realization of Fluid Boundary Condition via Discrete Boltzmann Dynamics," International Journal of Modern Physics C, pp. 1281 - 1292, 1998.
[14] H. Chen, S. Kandasamy, S. Orszag, R. Shock, S. Succi and V. Yakhot, "Extende d Boltzmann Kinetic Equation for Turbulent Flows," Science, no. 301, pp. 633 - 636, 2003.
[15] S. Chen and G. Doolen, "Lattice Boltzmann Method for Fluid Flows," Annual Review of Fluid Mechanics, vol. 30, pp. 329 - 364, 1998.
[16] E. Fares and S. Nölting , "Unsteady Flow Simulation of a High - Lift configuration using Lattice - Boltzmann Approach," AIAA Paper 2011 - 869, 2011.
[17] D. Casalino, S. Noelting, E. Fares, T. Van de Ven, F. Perot and G. Bres, "Towards Numerical Aircraft Noise Certification: Analysi s of a Full - Scale Landing Gear in Fly - Over Configuration," AIAA Paper 2012 - 2235, 2012.
[18] D. Casalino, A. F. P. Ribeiro, E. Fares, S. Nölting, A. Mann, F. Perot, Y. Li, P. - T. Lew, C. Sun, P. Gopalakrishnan, R. Zhang, H. Chen and K. Habibi, "Towards La ttice - Boltzmann Prediction of Turbofan Engine Noise," AIAA Paper 2014 - 3101, 2014.
[19] H. Chen, S. Chen and W. H. Matthaeus, "Recovery of the Navier - Stokes equations using a lattice - gas Boltzmann method," Physical Review A, no. 45, pp. R5339 - R5342, 1992 .
[20] Y. H. Qian, D. D'Humières and P. Lallemand, "Lattice BGK Models for Navier - Stokes Equation," Europhysics Letters, vol.
17, pp. 479 - 484, 1992.
[21] V. Yakhot and S. Orszag, "Renormalization Group Analysis of Turbulence," Journal of Scientific C omputing, vol. 1, no. 2, pp. 3 - 51, 1986, doi: 10.1007/BF01061452.
[22] E. Fares, "Unsteady Flow Simulation of the Ahmed Reference Body using a Lattice Boltzmann Approach," Journal of Computers and Fluids, no. 35, pp. 940 - 950, 2006,.
[23] J. E. Hawkin gs and D. L. Ffowcs Williams, "Sound Generation by Turbulence and Surfaces in Arbitrary Motion," Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, vol. 264, no.
1151, pp. 321 - 342, 1969.
[24] F. Fara ssat and G. Succi, "The Prediction of Helicopter Discrete Frequency Noise," Vertica, vol. 7, no. 4, pp. 309 - 320, 1983.
[25] A. Najafi - Yazdi, G. Brès and L. Mongeau, "An Acoustic Analogy Formulation for Moving Sources in Uniformly Moving Media," Proceedi ng of The Royal Society of London A, vol. 467, no. 2125, pp. 144 - 165, 2011.
[26] ICAO, Environmental Protection, Annex 16 to the Convention on International Civil Aviation, Vol. 1: Aircraft Noise, Appendix 2.
[27] M. R. Khorrami, D. P. Lockard, W. M. J. Humphreys and P. A. Ravetta, "Flight - Test Evaluation of Landing Gear Noise Reduction Technologies," Paper to be presented at the AIAA/CEAS Aeroacoustics Conference in Delft, The Netherlands, May 2019.
[28] M. R. Khorrami, P. A. Ravetta, D. P. Lockar d, B. Duda and R. J. Ferris, "Measured and Simulated Acoustic Signature of a Full - Scale Aircraft with Airframe Noise Reduction Technology Installed," Paper to be presented at the AIAA/CEAS Aeroacoustics Conference in Delft, The Netherlands, May 2019.
[29 ] B. König, E. Fares, P. Ravetta and M. Khorrami, "A Comparative Study of Simulated and Measured Main Landing Gear Noise for Large Civil Transports," AIAA Paper 2017 - 3013, 2017.
[30] S. Kota, P. Flick and F. Collier, "Flight Testing of the FlexFloilTM Adaptive Compliant Trailing Edge," AIAA Paper 2016 - 0036, 2016.
[31] E. Fares, B. Duda and M. R. Khorrami, "Airframe Noise Prediction of a Full Aircraft in Model and Full Scale Using a Lattice Boltzmann Approach," in AIAA Paper 2016 - 2707 , 2016.
[32] E . J. Miller, J. Cruz, S. - F. Lung, S. Kota, G. Ervin, K. - J. Lu and P. Flick, "Evaluation of the Hinge Moment and Normal Force Aerodynamic Loads from a Seamless Adaptive Compliant Trailing Edge Flap in Flight," AIAA Paper 2016 - 0038, 2016.