The influence of an adaptive nacelle inlet lip on fan noise propagation

Transcription

1 Aircraft Noise - Aeroacoustics: Paper ICA The influence of an adaptive nacelle inlet lip on fan noise propagation Frane Majić (a), Gunilla Efraimsson (b), Ciarán J. O Reilly (c) (a) KTH Royal Institute of Technology, Aeronautical & Vehicle Engineering, Sweden, franem@kth.se (b) KTH Royal Institute of Technology, Centre for ECO 2 Vehicle Design, Sweden, gef@kth.se (c) KTH Royal Institute of Technology, Centre for ECO 2 Vehicle Design, Sweden, ciaran@kth.se Abstract: The aeroacoustic performance of an adaptive inlet of a turbo-fan engine is numerically investigated in this paper. The sound intensity and directivity of the fan noise propagation to the far-field, and the sound level at lateral reference points are investigated. The investigation is performed for three Helmholtz numbers, with the influence of the mean flow included, for a single duct mode (- 8,1). The contour was defined by five movable knots at the leading edge of the inlet. The contour had to fulfil two constraints, namely it had to have a constant length and a convex curvature. The process of contour adaptation was performed in two steps. In the first step, two knots on the inner inlet side were moved in order to attain a certain shape, while other knots were kept fixed. In the second step, the rest of the knots were moved in order to fulfill the constraints. A finite element solver for the Helmholtz equation is used in the inner part of the inlet, with a perfectly matched layer boundary condition close to the inlet entrance. The propagation through the outer part of domain is solved by Kirchhoff integral method. The results show the influence of the inlet shape adaptation on the noise intensity level as well as the directivity of propagation. The maximum peak intensity level of all inlet shapes is increased by increasing the Helmholtz number. This causes the width of intensity distribution to become narrower and shifted towards the symmetry axis of the nacelle. The inlet shape with the most opened nacelle throat has the lowest peak and an intensity distribution shifted towards the symmetry axis, which indicates the influence of the mean flow. Also, the more closed nacelle throat causes a decrease of the effective perceived noise level. Keywords: aeroacoustics, nacelle inlet, duct mode, turbo-fan engine, morphing geometry

2 The influence of an adaptive nacelle inlet lip on fan noise propagation 1 Introduction The Advisory Council for Aeronautics Research in Europe (ACARE) declared the objective in its Vision 2020 report [1] of a 65% reduction of perceived noise against the baseline from the year In order to achieve such a goal, one of the possible measures is the further improvement of turbo-fan engine noise treatments for transport aircraft. A leading contributor to turbo-fan engine noise is fan noise, which is particularly dominant in the approach phase of the flight. Fan noise is a consequence of rotor-stator interaction, and is perceived mainly as tonal noise. Fan noise propagates towards both the inlet and exhaust of the engine. One noise treatment option is to use acoustic liners in order to attenuate the sound as it propagates through the inlet. Another option is to change the directivity of the sound radiation by changing the shape of the inlet. Inlets on present transport turbo-fan aircraft are non-adaptive structures. However, adaptive structures employing advanced materials could provide an improvement in the directivity. Essential features for such concepts are adaptability, strength and stability. In the last two decades, improvements in the smooth shape change technologies have been made as a result of both better adaptive skin technology and actuators for adaptive shape change. Recently, continuous camber variation on airfoils has been developed using chiral honeycomb, which offers both high deformability and structural integrity of the wing box [2, 3]. The concept was proven to work for small camber changes. The presented work a showed technological level of design and manu facture of chiral composite cores, which can provide morphing capabilities on aircraft structures. Such investigations are important prerequisites for the possibility to implement morphing capabilities in adaptive inlets. Additionally, investigations in the field of actuators have given promising solutions, which are required for the actuation of morphing concepts. The pneumatic-type actuators developed so far have been designed for small-scale demonstrators. The pneumatic artificial muscle (PAM) has become widely-used fluid-power actuator, which shows a high force-to-weight ratio, has a soft and flexible structure, and a low cost. PAM has been investigated experimentally and the results show a non-linear relationship between contraction as well as air pressure and pulling force [4, 5, 6]. In the field of morphing skins, Baier and da Rocha-Schmidt presented design and material concept of a cover structure for the gap between adjacent aerodynamic control surfaces in the deflected state [7]. It is based on large and geometrically highly nonlinear shear deformation of a suitable skin material. This material concept could be suitable for morphing inlets and is made of the shear compliant metal wire membrane with a silicone matrix. The material combination is stiff in wire direction to stand the aerodynamic loads and has a low shear stiffness to allow the shape change. Park et al. [8] have numerically investigated radiation characteristics of the discrete-frequency rotor-stator interaction noise from the scarf and the scoop inlets, respectively. The near-field solution was obtained by solving the linearized Euler equations and for the prediction of the farfield directivity pattern, the Kirchhoff integral method was applied. They have shown that a 7 db 2

3 noise reduction at downward peak radiation angle can be achieved by using the scoop inlets with scoop angle of 15, and a 5 db noise reduction by the scarf aero-intake with the scarf angle of 15, when compared with by the axisymmetric inlets. Also, the peak lobe radiation angle toward ground was shifted upward by both the scarf and the scoop geometries. Efraimsson et al. [9] have investigated propagation of acoustic waves in the inlet of a turbo-fan engine using a Navier-Stokes solver. The upstream propagation of the fan noise in a semiburied engine inlet, with the focus on a potential shielding effect of the S-duct, was investigated. A strong influence of the flow and the curved geometry was identified, yielding a focusing of sound waves to the middle of the duct. A transmission loss of the acoustic power from the fan plane to the inlet plane of around 5 db was identified for the first and the second radial modes and for acoustic powers in the interval db. Also, a shielding effect of the supersonic regions was identified. Manoha and Mincu [10] have investigated the influence of the fan installation position in the semi-buried inlet. The inlet was installed in the airframe in order to attain a shielding effect. They used the non-linear Euler equations in perturbation form. The results showed that noise levels radiated by more buried inlets are globally lower than the noise radiated by less buried inlets. In the framework of the Morphelle project supported by the European Commission, under the Environment Theme of the 7th Framework Programme for Research and Technological Development, numerical aeroacoustic investigations on the adaptive nacelle concept have been performed. This paper presents part of this work. Here numerical aeroacoustic computations are performed in order to investigate the performance of the adaptive nacelle inlet. The adaptation is made on the front inner side of the inlet lip, by changing position of single contour knot. 2 Methodology The reference geometry of the nacelle is described first. It is used as a starting point for further geometry adaptations. This geometry is predicted for a mid-to-long-range aircraft in year 2025 by the project partner Bauhaus Luftfahrt[11]. The reference nacelle geometry is axisymmetric and is shown in Figure 1. The global dimensions of the reference geometry are given in Table 1 and sketched in Figure 2. Each global dimension of the nacelle is scaled to the dimension of fan diameter D FAN. The fan diameter of the reference geometry used in this paper is D FAN = 3.3m. This was predicted according to the engine thrust required for the forecasted mid-to-long-range aircraft, simulated in commercial software GasTurb TM by project partner Bauhaus Luftfahrt. The inlet contour is composed of three parts. The first part is the inner contour composed of a spline curve. The second part is front part of the outer contour, which is defined by NACA-1 series profile coordinates. The third part is the rear part of the outer contour. The spinner geometry is arbitrarily chosen, with a parabolic cross-section. The reference contour was selected as a best guess contour. In the frame of the project there was no available contour 3

4 Figure 1: Reference nacelle - 3D view Table 1: Global dimensions of the reference nacelle geometry Scaled length L NAC /D FAN L/D FAN D MAX /D FAN D EXIT /D FAN D SPIN /D FAN Scaled value Figure 2: Nacelle cross section with global dimensions from the recent real engine nacelle and there was neither resources for the performance of optimization of the reference contour. The aim of the simulations was to investigate the effect and performance of the limited leading edge adaptations. The adapted geometries are derived from the reference geometry. The extent of the inlet shape displacement corresponds to the displacements achievable using the technology level of advanced materials and actuators described in the introduction. The achievable structure displacements are confirmed by the finite element simulation of the inlet structure deformation [11], performed by the project partner at Technical University Munich. The investigation is concentrated on the adaptable leading edge of the inlet presented in Figure 3. The adaptable contour has five movable knots (P2 - P6) and two fixed knots. The contour had to fulfil two criterions, the constant contour length and convex curvature of the contour. The process of contour adaptation was performed in two steps. In the first step, the knots P2 and P3 were moved in order to attain certain shape, while other knots were kept fixed. In the second step the knots P5 and P6 were moved in order to attain original contour length. 4

5 Figure 3: Adaptable leading edge knot positions 3 Numerical method The commercial software LMS Virtual.Lab was used to numerically calculate the fan noise propagation through the inlet and radiation to the far-field. A finite element method (FEM) was used to solve convective Helmholtz equation in the near-field domain. The FEM domain contained the space inside the inlet and space outside and vicinity of the inlet. The FEM domain was bounded by the fan plane, the inlet walls and the boundary, which was half sphere surface with radius 3 m and centre at the fan centre. The outer boundary, the fan plane and the inlet walls are shown in the Figure 4. The half-sphere surface was the boundary of the finite element domain on which a perfectly matched layer (PML) was applied [12]. The PML region should be sufficiently thick in order to rapidly decay propagating waves so that their reflections at the outer boundary of the PML region are very weak. A layer of the five elements was assumed to be enough for this technique. In LMS Virtual.Lab this technique is implemented as an automatically matched layer (AML). Figure 4: AML side - transparent As the boundary condition of the outer boundary was prescribed as an AML, the solution on this boundary of the finite element domain was automatically set as a source surface for the calculation of the far-field noise radiation. The fan noise radiation from the half-sphere surface to the far-field was computed by the Kirchhoff integral method. The finite element mesh was generated in the LMS Virtual.Lab based on the surface mesh 5

6 generated in the Star-CCM+ commercial software, where the CAD model for different inlet shapes was generated. The boundary condition in the finite element domain was set as follows: nacelle and spinner walls were prescribed as hard wall, at the fan plane the fan noise source was prescribed as a specific duct mode and at other surfaces (half-sphere and circular plane) the AML condition was prescribed. The numerical aeroacoustic calculations were performed for the duct mode (-8,1). The case was described by specific duct mode which has certain number of the lobes in circumferential and radial direction. In the rest of the paper the duct mode will be described in parenthesis as follows: (m, n). The first number m represents number of circumferential lobes and the second number n represents number of radial lobes. 4 Results For the evaluation of the aeroacoustic inlet performance, the influence of shape adaptation on the noise intensity distribution in the front field was investigated. The intensity was investigated at the line of microphones positioned on the circular line in the front field and symmetry plane of the inlet with centre in the middle of the inlet highlight plane. The probe line was part of the circle with 180 extent. Because of symmetrical results, only results on one half of the probe line, on 90 extent, will be presented. The probe line is shown in the Figure 5, where the nacelle is positioned inside the half sphere. Figure 5: Probe line angle postition The aeroacoustic calculations included the mean flow. The aeroacoustic performance was evaluated using lateral reference noise measurement point presented in Figure 6 [15] and noise intensity distribution at the probe line shown in Figure 5. The noise level at the lateral reference point is the average of the two maximum noise levels measured during take-off at two points located on lines parallel to and at a distance of 450 metres from runway centre line. The noise level is expressed as an effective perceived noise level (EPNL), which consists of instantaneous perceived noise level corrected for tones and duration [16]. In Figures 7-9, the intensity distributions for three Helmholtz numbers and a series of nacelle adaptations were presented. Helmholtz number is defined as He = 2π f R FAN /a, where f is frequency, R FAN is fan radius and a is speed of sound. Each curve in the figures presents the distribution for one nacelle 6

7 adaptation, where the smaller radial position of point P2 (r P2 ) represents a bigger bump in the leading edge contour. The intensity distribution is shifted toward the lateral direction (lower angles of probe line) and it is more spread at the low Helmholtz number. The noise propagation is dependant on the Helmholtz number. At He = 15.7 the distributions have very little difference for the presented nacelle adaptations. The maximum peak for the adaptation with the highest r P2 is shifted slightly toward centre line and the same adaptation has higher intensity in lateral direction, which is important for the noise level at the lateral measurement points. Figure 6: Lateral reference noise measurement point r P2 = r P2 = r P2 = Intensity [W/m 2 ] Angle of line probe [ ] Figure 7: Noise intensity distribution, for He = 15.7, at the probe line for a series of adaptations and fixed position of point P3 (x P3 = r P3 =0.740) and fixed x-coordinate of point P2 x P2 = At increased Helmholtz number He = 19.8 presented in Figure 8, the adaptation with highest r P2 (smaller bump in contour) has slight shift of maximum peak toward the centre line and increased intensity level on small angles of probe line with respect to the other two adaptations. 7

8 Intensity [W/m 2 ] r P2 = r P2 = r P2 = Angle of line probe [ ] Figure 8: Noise intensity distribution, for He = 19.8, at the probe line for a series of adaptations and fixed position of point P3 (x P3 = r P3 =0.740) and fixed x-coordinate of point P2 x P2 = r P2 = r P2 = r P2 = Intensity [W/m 2 ] Angle of line probe [ ] Figure 9: Noise intensity distribution, for He = 25.12, at the probe line for a series of adaptations and fixed position of point P3 (x P3 = r P3 =0.740) and fixed x-coordinate of point P2 x P2 = The increase on lower angles of probe line affects the noise level at the lateral measurement points. At the highest Helmholtz number He = presented in Figure 9, the adaptation with highest r P2 has also shifted maximum peak toward centre line with respect to the other two adaptations. From Figures 7-9 it can be seen that the noise propagation depends on the Helmholtz number. Also, comparing distributions for cases with and without mean, it can be concluded that the mean flow influences the noise propagation, especially close to the walls and leading edge where high velocity is attained. In Table 2, the effective perceived noise levels for three adaptations presented in Figures 7-9 8

9 are shown. These numbers show the influence of the adaptation in the radial direction on the noise on ground around runway. The decrease of radial position of knot P2 (bump increase) in the leading edge contour causes the decrease of the effective perceived noise level. Table 2: Effective perceived noise level for adaptations from figures 7-9 r P2 =0.719 r P2 =0.724 r P2 =0.729 EPNLdB Conclusions The aeroacoustic performance of adaptive turbofan engine inlet was analysed. A coupled approach was employed, using finite element method for solving Helmholtz equation in the near field and Kirchhoff integral method for the radiation to the far-field. The adaptation was made axisymmetrically on the front part of the inlet. The maximum peak intensity level was increased by increasing the Helmholtz number. The inlet shape with the most closed nacelle throat had the lowest peak intensity at all Helmholtz numbers and also at all tested duct modes. By increasing the Helmholtz number, the width of intensity distribution became narrower and shifted towards the symmetry axis of the nacelle. Also, the more closed nacelle throat caused a decrease of the effective perceived noise level. Acknowledgements This work was done in the framework of the Morphelle project ( supported by the European Commission under the Environment Theme of the 7th Framework Programme for Research and Technological Development. The authors would like to thank the project partners Technische Universität München, Bauhaus Luftfahrt and University of Bristol for their support. The computations were performed using LMS Virtual.Lab software on the resources provided by the Swedish National Infrastructure for Computing (SNIC) at PDC Centre for High Performance Computing (PDC-HPC), High Performance Computing Center North (HPC2N) and Center for Scientific and Technical Computing at Lund University. References [1] European Commission, Advisory Council for Aviation Research and Innovation in Europe, European Aeronautics: A Vision for 2020, Belgium, [2] Martin, J.; Heyder-Bruckner, J.-J.; Remillat, C.; Scarpa, F.L.; Potter, K.; Ruzzene, M. The hexachiral prismatic wingbox concept. Phys. Status Solidi B, Vol 245 (3), 2008, pp

10 [3] Bettini, P.; Airoldi, A.; Sala, G.; Di Landro, L.; Ruzzene, M.; Spadoni, A. Composite chiral structures for morphing airfoils: numerical analyses and development of a manufacturing process. Composites: Part B - Engineering, Vol 41 (2), 2010, pp [4] Chen, G.; Li, J. Design and manufacture of three dimensional integrated braided composite tube with flange. Applied Mechanics and Materials, Vol , 2010, pp [5] Wereley, N. et al. Pneumatic artificial muscles for aerospace applications. AIAA , 50th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, Palm Springs, USA, May 4-7, [6] Wickramatunge, K.,C.; Leephakpreeda, T. Study on mechanical behaviors of pneumatic artificial muscle. International Journal of Engineering Science, Vol. 48 (2), 2010, pp [7] Da Rocha-Schmidt, L.; Baier, H. Morphing skins and gap covers for aerodynamic control surfaces. Deutscher Luft- und Raumfahrt-Kongress, Stuttgart, Germany, [8] Park, Y.; Kim, S.; Lee, S.; Cheong, C. Numerical investigation on radiation characteristics of discrete-frequency noise from scarf and scoop aero-intakes, applied acoustics, Vol. 69 (12), 2008, pp [9] Efraimsson, G.; Forsberg, N.; Nordström, J. Simulations of acoustic waves in a turbo-fan engine air intake, AIAA , 16th AIAA/CEAS Aeroacoustics Conference, Stockholm, Sweden, 8-9 June, [10] Manoha, E.; Mincu, D.,C. Numerical simulation of the fan noise radiated through a semiburied air inlet. AIAA , 15th AIAA/CEAS Aeroacoustics Conference (30th AIAA Aeroacoustics Conference), Miami, USA, May, [11] da Rocha-Schmidt, L.; Hermanutz A.; Baier, H.; Seitz, A.; Bijewitz, J.; Isikveren, A.T.; Scarpa, F.; Allegri, G.; Remillat, C.; Feuilloley, E.; Majić, F.; O Reilly, C.; Efraimsson, G. Progress towards adaptive aircraft engine nacelles. 29th Congress of the International Council of the Aeronautical Sciences, St. Petersburg, Russia, [12] Berenger, J.,P. A perfectly matched layer for the absorption of electromagnetic waves. Journal of Computational Physics, Vol 114 (2), 1994, pp [13] AERO2k, Global Aviation Emissions Inventories for 2002 and 2025, website: cited May [14] Official Airline Guide (OAG), Historical Data [15] Noise Certification Database, General information on the noisedb database. website: cited October [16] Sperry, W., C. Federal Aviation Administration, Aircraft noise evaluation. Technical Report FAA-NO-68-34, USA,