Evaluation of a Microporous Acoustic Liner Using Advanced Noise Control Fan Engine ^†

Featured Application

Acoustic liners for aircraft turbofan engines. The liner technology was built and tested on NASA Glenn Research Center’s Advanced Noise Control Fan low-speed test bed at the Turbomachinery Lab at the University of Notre Dame. The Technology Readiness Level (TRL) 3–4 test was successful.

Abstract

A novel microstructurally controlled graded micro-porous material was developed and experimentally validated for noise reduction through a normal incidence impedance test. Extensive parametric studies were conducted to understand the influence of test specimen size, particle size, porosity, pore size, and its distribution on acoustic absorption and transmission loss. Based on previous research, this study evaluates the application of graded microporous material as an acoustic liner technology for aircraft turbomachine engines. The liner was fabricated in eight 45° segments, assembled in an aluminum test rig, and tested on NASA Glenn Research Center’s Advanced Noise Control Fan (ANCF) low-speed test bed for tonal and broadband noise. The study demonstrates that microstructurally controlled graded microporous material is very effective in dissipating sound energy with reductions in tonal sound pressure level (SPL) of 2 to 13 dB at blade passing frequencies and reductions in broadband SPL of about 2 to 3 dB for the shaft order greater than 40. While the proposed two-layer graded liner model successfully validated the concept, additional design optimization is needed to enhance performance further. This work highlights the potential of graded microporous material as next-generation acoustic liners, offering lightweight, efficient, and scalable aircraft engine noise reduction solutions.

1. Introduction

There is a need and increase in demand in the aviation industry to improve aircraft efficiency as air travel continues to become more frequent. Noise is a serious problem, and high noise levels are experienced by those living near airports and aviation centers. In recent times, aircraft noise has been linked to numerous health issues and learning and developmental problems in the younger generation [1,2,3]. There has been a significant effort to try to reduce noise levels, and it continues to be one of the priorities for the future [4,5].

The approaches to reducing aircraft noise can be categorized into two kinds: 1. Airframe noise and 2. Engine noise. Among the two, engine noise reduction is the primary focus. The introduction of turbofans resulted in a significant reduction in conventional jet engine noise (see Figure 1a,b) [6] at that time. However, turbofans introduced a new source of noise from the compressor fan (see Figure 1b) [6]. Although the overall noise profile has been noticeably reduced, the fan noise is predominant in the broadband, low-to-mid-frequency range and produces tonal noise, which is annoying at the same frequency [7].

The current technology used in aircraft turbomachine engines is honeycomb liners with frontal perforated face sheets (see Figure 2a); another type is with wire mesh (see Figure 2b) [8]. The particular design is lightweight and strong due to the honeycomb core. This single degree of freedom (SDOF) design is a quarter-wavelength resonator, a Helmholtz resonator, or a combination and has limited ability to attenuate the broad range of noise generated by the engine. The absorption characteristic of the liner depends on the resonator depth (in this case, the liner depth) [9].

The double degree of freedom (DDOF) system (see Figure 3) [8] further increases the liner depth to achieve a second peak of sound absorption to enhance the liner’s overall performance. These conventional liner technologies exhibit spatially concentrated absorption, block lateral dissipation of sound waves, limited bandwidth absorption, and strong resonance and work best only for tones [10]. Furthermore, to achieve low- and mid-frequency attenuation, the liner depth must be increased, but with the increase in liner depth, the liner’s performance diminishes. Moreover, there are constraints on the size of the liners, and an alternative approach to liner development must be explored that has the ability to increase sound absorption over a broadband frequency range in addition to tones.

The extended reacting bulk liner is an alternative strategy to overcome the limitations of conventional liners with a zero-resistance face sheet, unblocked lateral wave, continuous distribution of resistance, attenuated internal wave propagation, reduced internal sound speed, subdued depth-related resonance, and improved absorption bandwidth [10].

Porous sound absorption materials are promising for the upcoming generation of turbine engine acoustic liners. These materials have gained significant interest in the field of noise attenuation as they possess channels, cracks, or cavities that allow sound waves to enter the material system. With complex microstructures, the sound energy is dissipated through thermal loss due to the friction of air molecules against the pore walls and viscous loss due to airflow viscosity with the material. Based on these energy consumption principles, porous materials behave like broadband sound absorbers [11,12]. They are of three types, cellular, fibrous, and granular, based on the microscopic structure [13,14] (see Figure 4).

Among these types of porous sound absorbers, we focused on microstructure-controlled granular porous materials. They are rigid, macro/microscopic particles with dimensions much larger than the interconnected pores, making them suitable for structural applications. The rationale for selecting granular materials is that they allow for the design of porous sound absorbers of engineered acoustic response by controlling the porosity, pore size, and their distribution through the thickness through appropriate selection of particles, and the surface imperfections of the particles improve fluid flow resistivity and tortuosity (the ratio of the actual sound path to the sound path taken). Furthermore, there are acoustic models in the literature that link the microstructural/physical properties (fluid flow properties) of porous materials to acoustic performance. In the frequency domain, all of the acoustic models are based on these fluid flow properties to give a physical basis for describing sound propagation through porous media. The commonly used acoustic models, namely Johnson–Champoux–Allard (JCA) [15,16], Johnson–Champoux–Allard–Lafarge (JCAL) [17], Johnson–Champoux–Allard–Pride–Lafarge (JCAPL) [18], and Pore Size Distribution (PSD) [19], are used to predict the acoustic response, i.e., the absorption coefficient versus frequency and transmission loss versus frequency.

We selected natural granular ceramic hollow microbubbles, i.e., Cenospheres, commercially available and supplied by Sphere One—called Extendospheres (commercial name)—which are the mineral waste from coal-burned electric power plants. The rationale for selection is that they are lightweight ceramic particles filled with air, characterized by size, nearly spherical, with a surface roughness that can improve sound energy dissipation by enhancing friction (sound energy converted to heat) and tortuosity. A new particle surface coating technique with a minimal polymer high char binder was developed to mimic the pore properties of the loose granular system to maintain near-similar porosities. First, this concept of material development was demonstrated on solid glass beads (model materials) [20,21]. Then, processing and fabrication were expanded to lightweight hollow microbubbles, and this new material system is addressed as a microporous hollow bubble (μPHB) composite material described in detail in references [22,23]. The developed material is multifunctional, offering thermal, mechanical, and acoustic properties [23]. An extensive experimental and analytical study was conducted on microporous specimens in terms of particle diameter, pore size and their distribution, and specimen thickness [21,24]. Several different particle size microporous material specimens of 100 mm diameter and 25 and 50 mm thickness were fabricated and impedance tested. Firstly, the particle size must be less than 1 mm to achieve effective acoustic responses, as demonstrated [21,24]. Then, we concluded that for a given frequency, the absorption is higher for larger-size particles and, alternatively, it is lower for smaller-size particles. Transmission (sound energy) loss directly increased with a decrease in pore size (particle size) and with the increase in specimen thickness. The two extreme responses from the microporous materials of specific particle size groups lead to the conclusion that both high sound absorption and transmission loss (energy dissipation) cannot be achieved by single-size particle distribution through the thickness. An alternative approach of different size particle distribution, in other words, a microstructural controlled graded pore size distribution approach using different size particles, is needed. This was demonstrated through both experimentation and predictive acoustic analyses in references [21,24]. Also, the specific particle size distribution was established [24]. A graded (large to small pore size) μPHB composite material (layers of different thicknesses but with a total thickness of 50 mm) was designed and developed to achieve an absorption coefficient (α) ≥ 0.50 and transmission loss (TLn) ≥ 20 dB in the broad frequency range above 500 Hz. A novel concept of graded μPHB composite material development, acoustic testing, and validation was demonstrated in the normal incidence case, but fabricating an acoustic liner and its assessment in the case of duct or grazing flow inside an aircraft engine needs to be tested and validated.

The current study reports the fabrication of a microstructurally controlled graded μPHB composite material liner that was designed and validated for normal incidence impedance and met the design requirements, showcasing successful results. The material technology was subsequently evaluated as an inlet liner in an Advanced Noise Control Fan (ANCF) engine. The NASA Glenn Research Center’s ANCF low-speed test rig [25] was chosen. The ANCF engine is well known and has been used for many years in numerous research works.

The liner specs include an inner diameter of 1219 mm (48″), an outer diameter of 1321 mm (52″), an axial length of 229 mm (9″), and a liner thickness of 50 mm (2″). We chose a two-layer graded material system with each layer being 25 mm (1”) thick; the innermost layer was made of large-size particles with a mean particle diameter (d_m) of 625 μm, and the outer layer was made of medium-size particles with a mean particle diameter of 160 μm. The details are provided in the next section. The liner was tested at the Notre Dame Turbomachinery Lab Whitefield [26] test facility. The decibel (dB) loss in the sound pressure levels from this liner technology relative to the baseline configuration was evaluated to assess the impact of the proposed concept.

2. Laboratory Experiments—Normal Incidence Impedance

Based on our normal incidence test target design requirements of both absorption coefficient (α) ≥ 0.50 and transmission loss (TLn) ≥ 20 dB in the broad frequency range above 500 Hz [24], we designed a three-layer graded μPHB composite material liner (50 mm total thickness—typical acoustic liner) without any interface between the layers based on the concept shown in Figure 5a,b. Here, L—large (d_m = 625 μm), M—medium (d_m = 223 μm), and S—small (d_m = 95 μm) particle sizes, with subscripts representing the percentage of the total thickness.

However, we faced fabrication challenges while fabricating the three-layer graded system liner in the lab-scale environment. Therefore, we decided to fabricate a two-layer graded μPHB composite material system, as the priority was to evaluate the liner technology concept. But before we made the modifications in the design of the liner technology, we fabricated a two-layer graded μPHB composite material cylindrical specimen of 100 mm diameter and 50 mm total thickness using two different size groups of Cenospheres with normal incidence impedance tested to measure the absorption coefficient versus the frequency using a two microphone setup according to ASTM E1050-12 [27] standard and transmission loss versus frequency using a four-microphone setup according to the ASTM 2611-09 standard [28]. All of the test details and smoothening of the raw data are provided in reference [23]. In the two-layer graded system, the front layer (25 mm) with large-size particles (d_m = 625 μm) was retained, whereas the second layer (25 mm) with a medium particle size distribution (d_m = 160 μm) was designed to replicate the particle size distribution of the second and third layers from the three-layer system. This redesign aims to reproduce the acoustic performance of the original three-layer graded system. A detailed analysis of specific particle size distribution selection is provided in references [21,24]. Therefore, for verification, the two-layer graded system results were compared with the three-layer absorption coefficient and transmission loss results. Figure 6 shows the absorption coefficient versus frequency for the two- and three-layer graded μPHB composite material; here, we see that there is a small difference in the response between the two. However, the two-layer graded μPHB composite material system met our design target condition of a sound absorption coefficient ≥ 0.50 above 500 Hz. Similarly, Figure 7 shows the transmission loss versus frequency for both systems; the two-layer graded system shows a lower transmission loss of about 1 dB throughout the frequency range relative to the three-layer graded system. However, the two-layer graded μPHB composite material system met our design target condition of transmission loss ≥ 20 dB above 500 Hz.

3. Design of the Mold and Ram

The liner was envisioned to be made of eight 45° segments. The mold was designed and made of aluminum for ease of handling. Therefore, an aluminum mold of (see Figure 8a) a 45° segment with an inner diameter D_i = 1219 mm (48″), outer diameter D_o = 1321 mm (52″), thickness t ≈ 50 mm (2″), and axial length l = 229 mm (9″) was designed. Similarly, the top/bottom plate (Figure 8b) and two rams were designed with adjusted angles and respective radii for the two layers (one ram shown in Figure 8c). The local industry fabricated the mold and the rams. The mold was tested to validate the geometry. This represents one of the eight segments that go around the liner cavity.

4. Fabrication of a Liner Segment

Eight liner segments of 45° each were fabricated to form a full liner. The fabrication of a liner segment consists of seven steps, as shown in Figure 9. Each of these steps is described below.

• Step 1

The required amount of hollow microbubbles of large- and medium-size groups were surface-coated with SC1008, a phenolic resin, and prepared before fabrication. The mold setup is shown in Step 1 of Figure 9. Tap screws were used for the mold assembly and were finger-tightened during fabrication. A steel grid with an open area fraction of 55% and a wire diameter of 0.016″ was used to separate the two layers. Two ash composite supports, each of ≈25 mm (1″) thickness, were placed at five locations to hold the steel grid in place during the fabrication process. These supports split the mold cavity into four parts marked (a), (b), (c), and (d), as shown in Step 1 of Figure 9.

• Step 2

Each of the four split cavities on the outer diameter portion was filled with medium-size coated particles up to one-fourth of the height of the liner segment, and then the inner diameter portion of the liner segment was filled with large-size resin-coated particles up to one-fourth of the height of the liner segment, as shown in Step 2 of Figure 9.

• Step 3

Once all four split cavities were filled with the particles (Cenospheres), a wooden rod of ≈25 mm (1″) diameter was used to ram the material. First, the medium-size particle cavity (a) was rammed, as shown in Step 3 of Figure 9, followed by cavities (b), (c), and (d). Next, the large-size particles were rammed in the same order.

Step 2 was repeated, followed by Step 3. The ash composite supports were then removed one after the other from all five locations. Additional material was added at the locations where the ash composite supports were placed using a small scoop, ensuring material continuity in both layers.

• Step 4

An aluminum ram of a 44° segment was used to ram the entire cavity, as shown in Step 4 of Figure 9. The same sequence of ramming, i.e., the medium-size first and then the large-size particles, was followed.

• Step 5

After ramming manually in Step 4, the aluminum ram was hit with a hammer to achieve better compaction, as shown in Step 5 of Figure 9. A partially filled cavity with two layers of medium and large-size particles, separated by the steel grid, is shown in Figure 9. The mold was refilled with the material, and Steps 4 and 5 were repeated. This process of filling the material and compacting continued until the mold cavity was filled completely.

• Step 6

The mold was covered using the top aluminum plate, and the material was further compacted using the hammer by hitting the top plate. Then, the top plate was bolted to the mold and all of the tap screws were tightened, which were finger-tightened initially. A further three C clamps were attached and tightened, as shown in Step 6 of Figure 9, to avoid lateral expansion during the curing process.

• Step 7

The mold was placed inside the oven for curing. The curing steps were the same as those specified in references [21,24].

A fabricated liner segment and fully arranged liner segments are shown in Figure 10a and Figure 10b, respectively. The fabrication steps are reported in our previous work in reference [29] and are reiterated herein to provide the complete details of the process.

5. Experimental Test Setup

This section focuses on the description of the Advanced Noise Control Fan (ANCF) engine, the installation of an acoustic liner on the engine, and details on the testing and far-field acoustic measurements.

5.1. Advanced Noise Control Fan Engine (ANCF)

The ANCF was envisioned as a testbed for fan noise attenuation liner technologies when it was designed in the early 1990s. However, the ANCF has also served to improve the understanding of aeroacoustics physics and the validation of codes. The ANCF is currently at the University of Notre Dame Turbomachinery Laboratory [25]. The ANCF is a duct fan that has a 1219 mm (48”) internal diameter. The duct can be modified to multiple configurations with the ring sections that can be moved to obtain the required duct geometry or have rigid hard wall sections replaced with liner technology for testing. Figure 11 shows the schematic of the ANCF. The operating speed range of the ANCF is between 100 and 2400 RPM, with the nominal speed being 1800 RPM. At the nominal speed, the fan tip speed is about 115 m/s, and the inlet Mach number is about 0.15. The fundamental blade passing frequency (BPF) is about 500 Hz.

The ANCF rotor comprises 16 blades with the nominal pitch angle being 28°, though it can be modified to 18° or 38°. The ANCF can run in the rotor alone or with a stator attached downstream of the rotor configurations. Note that the counter of the stators is variable, and it can be set to any angle.

Furthermore, an inlet control device (ICD) has been installed on the front side of the ANCF, the purpose of which is to condition the airflow into the duct during the static ground tests. As specified, the purpose of the ICD is to condition the airflow by breaking up large vortices and reducing the size of the structures in the flow, which may potentially become an additional source of noise in the duct. This ICD is a honeycomb layer that has been sandwiched between two fine mesh layers; it is installed in 22 longitudinal sections [25]. From the previous tests, the ANCF’s performance in the tests is considered to be within the linear range. The maximum Mach number of the ANCF is 0.3 [25].

5.2. Installation of the Acoustic Liner on the Engine

The liner segments were taken to the University of Notre Dame Turbomachinery Lab for testing. Each segment was carefully slid inside the holder with four liner segments on top and four on the bottom half, as shown in Figure 12a. The liners were positioned inside the holder so that the surface with the large particle/pore size distribution faced the sound flow. The liner segments were covered with a perforated sheet with an open-area fraction of 0.40 and screwed onto the holder (see Figure 12b). This covering was for surface protection of the liner segments, and it would not alter the performance of the liner (verified before). Finally, the liner portion of both the top and bottom halves of the holders was covered with aluminum tape to acquire baseline test data (see Figure 12c).

The bottom half of the liner holder was assembled onto the engine first, and then the top half of the holder was assembled with the help of a crane, as shown in Figure 13. These details are reported in reference [29] and detailed in this work as well.

5.3. Testing and Far-Field Acoustic Measurements

The test was conducted as an outdoor test at the University of Notre Dame, Turbomachinery Lab White Field Facility [26]. The test conditions involved measuring baseline data (sound pressure level (SPL)) with aluminum tape on the perforated sheet of the liner holder, followed by measurements of the liner with the aluminum tape removed. To assess the repeatability of the test measurements on the ANCF, the standard deviation was determined across four data sets for both the microporous liner and baseline conditions at each microphone location. The tests were completed in one day. In the test run, four data sets were successively recorded by varying the speed from 2000 RPM down to 1500 RPM in 100 RPM intervals and subsequently increasing the speed back up to 2000 RPM and repeating the process one more time.

These measurements were taken in the rotor alone (16 blades @ 28 deg pitch) and with 14 vanes at 1 chord spacing. The same configuration was used to acquire liner data by taking off the aluminum tape.

The plan view of the ANCF test rig and the microphones (dots) is shown in Figure 14. Far-field acoustic measurements were acquired using an array of 30 ground-based microphones placed at a 12-foot radius from the duct centerline and 10 feet high. Fifteen (1–15) of these were in an arc around the inlet exit plane (0–90° measured from the inlet axis, Fwd array) and fifteen (16–30) were in an arc around the exhaust exit plane (90–160° Aft array with 180° being the exhaust axis). Figure 15 shows the ANCF test setup with all of the microphones arranged to acquire far-field data for baseline and liner configurations. These details are reported in reference [29] and detailed in this work as well.

Data acquisition was performed using an HMB GEN2i Data Recorder, which records the time series data from the 30-microphone array and the ANCF shaft rotation. For example, Figure 16a shows the time history (original spectral response). The time histories were processed using ensemble averaging to analyze the tone and broadband components separately by extracting the tones generated by the fan (shaft orders and harmonics). This can be performed exactly since the data are acquired synchronously to the shaft rotation [30]. The broadband spectral response separated from the tonal response at blade passing frequencies is shown in Figure 16b [30].

6. Test Results

This section discusses the test results for tonal and broadband noise separately. A comparative study has been conducted between the two recently tested acoustic liners on the ANCF engine and the current graded μPHB composite liner technology in terms of SPL reductions.

6.1. Impact of Microphone Locations on the SPL—Shaft Speed of 2000 RPM

This particular study was carried out to understand how the microphone’s location impacts the reduction in the sound pressure level for the liner relative to the baseline. As discussed earlier, Farfield acoustic measurements were acquired using 30 microphones placed at a 12-foot radius from the duct centerline and 10 feet high. Fifteen (1–15) of these were in an arc around the inlet exit plane (0–90° measured from the inlet axis, Fwd array) and fifteen (16–30) were in an arc around the exhaust exit plane (90–160° Aft array with 180° being the exhaust axis) (see Figure 14).

In this testing, the dynamic test data were recorded as MATLAB 9.13 scripts. By running the ANCF code, the data were extracted as a spectral file in MATLAB format. The spectral file was run to extract the tones and the broadband sound pressure level (SPL) separately. This process of data extraction was followed for both baseline and liner cases. First, the shaft speed of the 2000 RPM test case was chosen for the initial result analyses as this shaft speed was of primary interest to NASA. Later, the results for other shaft speeds, i.e., RPMs 1500 to 1900, are reported.

6.1.1. Forward Array Microphones

In microphones 1–15, which were placed at the inlet exit plane, we noticed that there is insignificant or no difference in SPLs between the baseline and liner, and this is expected as the microphone locations are at the inlet ahead of the liner setup, i.e., before engine sound flow takes place through the duct.

6.1.2. Aft Array Microphones

The Aft array microphones (16–30) recorded tonal and broadband SPL for both the baseline and the liner configuration. The SPL reduction between the baseline and liner (dB loss = SPL_baseline − SPL_liner) was significant in microphones 19 to 23. This region of microphones is highlighted in Figure 14. From the analysis, we noted that the liner was effective beyond 500 Hz. For a more detailed analysis, first, we chose a single microphone case, i.e., microphone 21 in the region of the significant SPL reduction for 2000 RPM, to compare the baseline and liner configuration results and understand, smoothen, and analyze the data. The analysis was performed for tonal and broadband noise separately. Then, later, we considered microphones 19 to 23, where the noise reduction was significant. However, Aft array microphones 16–18 and 24–30 showed about 1 to 1.5 dB loss between the baseline and liner configurations. Some of these details are reported in reference [29] and detailed in this work as well, with additional results.

6.2. Comparison of Test Results Between Baseline and Liner Configurations

This section will discuss the test results for tonal and broadband SPLs and dB loss response between the baseline and liner configurations.

6.2.1. Tonal Sound Pressure Level Response

Figure 17 shows the tonal SPL (dB) versus frequency for both baseline and liner configurations for microphone 21 for 2000 RPM. There are 8 x Blade Passing Frequencies (BPFs) in addition to tonal noise at other frequencies. From the results, we did notice a reduction in sound pressure level between the baseline and liner beyond 500 Hz.

Tonal noise constitutes the major part of the overall fan noise, especially the blade passing frequency (BPF), which is generally the dominating component. Therefore, a separate tonal BPF response analysis was performed to understand and interpret the effectiveness of the liner. Figure 18. shows the mean values of the tonal BPF sound pressure level of the baseline and liner configurations, along with their respective standard deviations. The standard deviation is minimal at all BPFs, indicating consistent material performance. Additionally, with a confidence interval of 95%, the margin of error was calculated and tabulated for all eight BPFs, as shown in

Table 1. Margin of error of microphone 21 (2000 RPM) data for baseline and liner configurations.

Blade Passing Frequency, Hz	Margin of Error—95% Confidence Interval
	Baseline Configuration	Liner Configuration
533 (1st BPF)	87.09 ± 0.15 (±0.18%)	86.54 ± 1.26 (±1.46%)
1067 (2nd BPF)	83.44 ± 1.78 (±2.13%)	78.82 ± 0.49 (±0.62%)
1600 (3rd BPF)	77.29 ± 1.66 (±2.14%)	74.73 ± 0.52 (±0.69%)
2133 (4th BPF)	80.62 ± 1.02 (±1.26%)	72.93 ± 1.88 (±2.58%)
2667 (5th BPF)	78.31 ± 0.69 (±0.88%)	71.09 ± 0.45 (±0.63%)
3200 (6th BPF)	68.78 ± 0.98 (±1.43%)	63.87 ± 0.95 (±1.48%)
3733 (7th BPF)	65.97 ± 2.93 (±4.44%)	63.21 ± 2.24 (±3.55%)
4267 (8th BPF)	67.25 ± 2.10 (±3.13%)	60.31 ± 0.41 (±0.67%)

. From Table 1, the margin of error range is quite reasonable and within industry standards; also, it involves environmental factors like sustained wind and highway vehicle noise during the time of testing. Furthermore, from Figure 18, the liner effectively reduces the SPL (SPL_baseline − SPL_liner) at all eight BPFs. A minimum of 0.6 dB at 533 Hz (1st BPF) to a maximum of up to 8 dB at 2133 Hz (4th BPF) as recorded by microphone 21 can be noticed.

The standard deviations and margins of error at a 95% confidence interval for microphones 19 to 23 were similar to mic 21 at 2000 RPM and other RPMs. To calculate the dB loss (SPL_baseline − SPL_liner), the mean values of the respective sound pressure levels of baseline and liner configurations were used.

For the five Aft array microphones 19 to 23, wherein the reduction in SPL was significant, the dB loss (SPL_baseline − SPL_liner) at the 8× BPFs is shown in Figure 19. We see that all five microphones, 19 to 23, record significant dB loss with the liner relative to the baseline configuration at all 8× BPFs. The tonal BPF dB loss and maximum dB loss for respective frequencies, as recorded by corresponding Aft microphones (between 19 to 23), are listed in

Table 2. Tonal BPF dB loss recorded by Aft microphones 19 to 23 [29].

Frequency, Hz	dB Loss (Aft Microphones 19–23), dB	Maximum dB Loss, dB	Microphone
533 (1st BPF)	0.2 to 0.6	0.6	21
1067 (2nd BPF)	0.4 to 8	8	20
1600 (3rd BPF)	2 to 5.5	5.5	22
2133 (4th BPF)	6 to 13	13	19
2667 (5th BPF)	1.5 to 12	12	20
3200 (6th BPF)	3 to 7	7	19
3733 (7th BPF)	0.2 to 4	4	19
4267 (8th BPF)	2 to 7	7	21

. A maximum dB loss of 13 dB was recorded by microphone 19 at the 4th BPF (see Table 2).

As mentioned, tonal noise constitutes the major part of the overall fan noise, especially the BPF noise, which is generally the most dominant component. Figure 20. shows the tonal BPF dB loss versus shaft order of microphone 21 for 8× BPFs at turbomachine shaft speeds of 1500 to 2000 RPM with increments of 100 RPM. We see that the dB loss is different at different shaft speeds for all 8× BPFs, but the liner is effective in all cases at the microphone 21 location. The maximum dB loss is about 9 dB at 1900 RPM shaft speed for the 6th BPF (shaft order 96–3040 Hz). Furthermore, we noticed that the liner was effective with similar recorded noise levels (dB loss) by microphones 19 to 23 (a highlighted region in Figure 14) at shaft speeds of 1500 to 2000 RPM.

6.2.2. Broadband Sound Pressure Level Response

In the case of broadband response, the data are in the form of signals, so these measurements need to be smoothened to eliminate the noise and analyze the data. The raw broadband response of SPL versus frequency for baseline and liner configurations for microphone 21 for 2000 RPM, as shown in Figure 21, requires smoothening.

As mentioned, there is a need to smoothen the data to analyze and interpret the results. Hence, the measured data were smoothened using a robust locally weighted regression (rLowess) method [31] in MATLAB. This powerful method was applied to the fatigue model for data reduction [32], as adapted here.

Figure 22 shows the smoothened broadband response of SPL versus frequency for baseline and liner configurations. The gray shading represents the standard deviation for each configuration, which is minimal in both cases. The overall margin of error in the measurements is small (<2%). Due to the large volume of the data set, individual values are not reported. From the results, we see that the liner configuration reduces the noise level beyond the frequency of 500 Hz, and the reduction in noise level increases with the frequency. The broadband dB loss (SPL_baseline − SPL_liner) is noticeable beyond 500 Hz, increases with the frequency, and varies between 2 to 3 dB beyond 1300 Hz. The clear distinction between the baseline and liner test results shows the effectiveness of our microporous liner technology. To calculate the dB loss (SPL_baseline − SPL_liner), the mean values of the respective sound pressure levels of baseline and liner configurations were used.

A dB loss for a broadband response versus frequency between the baseline and the liner; the configurations are as shown in Figure 23. A typical region of 105 to 125 deg (microphones 19, 20, 21, 22, and 23 are plotted, respectively) was analyzed. From the result, we see that for a shaft order of 20 (667 Hz), the dB loss is >1 dB and increases with a further increase in the frequency with a maximum dB loss of 2.70 dB. Also, a significant dB loss was recorded in the region of microphones 19 to 23 (highlighted in Figure 14). Based on this test result, we can design (control microstructure) the acoustic liner through appropriate selection of the particle size, distribution through the thickness, layer thickness, and total thickness to improve the broadband dB loss.

The broadband dB loss for microphone 21 at different shaft speeds of 1500 to 2000 RPM is shown in Figure 24. From the results, we see that for the shaft order 40 and beyond, for all of the shaft speed cases, the dB loss is at least 1.5 dB and increases with an increase in the shaft order varying from 2 to 3 dB. From this, we can conclude that the liner shows similar dB loss as recorded by microphone 21 at all of the tested shaft speeds. We noticed that the liner was effective with similar recorded noise levels by microphones 19 to 23 (the highlighted region in Figure 14) at different shaft speeds of 1500 to 2000 RPM.

The present results are for a two-layer graded μPHB composite liner of 50 mm thickness. This paper demonstrates the idea that a passive system of 3D microstructure controlled graded porous media (bulk material system) can be effective in mitigating noise. While effective, further optimization can enhance noise attenuation.

Future research can focus on the following:

• Layer thickness ratio—Exploration by varying the layer thickness ratios to maximize sound absorption and transmission loss.
• Particle size gradient—Refinement in the transition of particle sizes through the material thickness for improved broadband noise reduction.
• Porosity and pore size distribution—Controlling and tailoring the pore connectivity and distribution to achieve enhanced acoustic impedance matching.
• Alternative graded configurations—Re-design of other graded material configurations as the manufacturing method improves.

As the current study validates the effectiveness of the two-layer system, future work can focus on applying the proposed strategies and optimizing key parameters for improved noise mitigation.

6.3. Comparison of ANCF Test Results of Acoustic Liners

In this section, a comparison of the acoustic liner performance of two technologies recently tested on the ANCF with the same test conditions as the current work is provided.

The two technologies tested on the ANCF engine are as follows:

• 3D-printed acoustic metamaterial liner: A multi-degree-of-freedom acoustic metamaterial that is a modification of the traditional perforated panel over a honeycomb liner called a T-liner or Trinity liner that was manufactured using additive manufacturing (3D-printed) and was tested on the ANCF [33].
• Stepwise gradient metal foam liner: Stepwise gradient metal foam liners of two different configurations were tested using the ANCF at the Notre Dame Turbomachinery Laboratory. The details of the assessment are provided in reference [34].

Table 3. Comparison of acoustic liners’ performance in ANCF tests.

Feature	Graded Microporous Liner	T-Liner Metamaterial Liner [33]	Stepwise Gradient Metal Foam Liner [34]
Noise Control Mechanism	Microporous controlled viscous and thermal losses	Resonance-based attenuation	Large, interconnected pores
Tonal Noise Reduction at Blade Passing Frequencies	Effective with 0.6 dB to 13 dB reduction across multiple locations (mic 19 to 23). Maximum = 13 dB	Up to 18.6 dB (mic 13) and 9.5 dB (mic 16) as reported by the authors Maximum = 18.6 dB	Not reported
Broadband Noise Reduction	2 to 3 dB reduction over multiple locations (mic 19 to 23). Maximum = 3 dB	3.7 dB at mic 16 and 2.9 dB at mic 13, no effect reported elsewhere Maximum = 3.7 dB	No or marginal reduction; enhanced noise levels with the liner at some locations, as reported by the authors.
RPM Effectiveness (Tonal and Broadband)	Effective SPL reductions at all tested RPMs (1500–2000)	Effective at 1800 RPM, as reported by the authors	Reported only 2000 RPM where SPL increased with the liner relative to baseline configuration.
Performance Consistency	Effective, reliable, and consistent noise reduction across multiple frequencies and locations	Highly location-dependent (effective at a few points)	Unstable and can enhance noise levels due to impedance mismatch
Effect on Overall Acoustics	Attenuates noise effectively, aiming to improve the noise environment	Good but localized noise control and not widespread	No or marginal reduction along with enhanced SPLs with the liner at some locations

summarizes a comparison of the current work microporous liner, 3D printed metamaterial T-Liner, and stepwise gradient metal foam liner.

Thus, the comparative study clearly demonstrates that the graded microporous material offers superior and consistent SPL reductions across multiple locations and is effective at all of the tested RPMs (1500–2000 RPM), making it a highly effective and promising solution for aircraft noise mitigation. In contrast, the metamaterial T-liner, while showing a higher tonal noise reduction at a single location, was effective only at 1800 RPM; this may limit its broad applicability. The step-wise metallic foam was not effective in reducing the SPLs and enhanced the noise levels at some locations for the reported 2000 RPM and may be unsuitable for these applications. Among the three, the graded microporous material stands out as the most effective noise mitigation technology. Therefore, we conclude that a microporous material with controlled porosity, pore size, and its distribution through the thickness to achieve effective static airflow resistivity and tortuosity is critical in noise attenuation.

7. Concluding Remarks

Eight 45° segments of two-layer graded microporous (μPHB) acoustic liners (the innermost layer is made of large-size particles with an mean particle diameter of 625 μm and the outer layer is made of medium-size particles with an mean particle diameter of 160 μm), with each layer having a thickness of 25 mm and a total thickness of 50 mm, were fabricated to evaluate the microstructurally controlled graded material liner on a real ANCF engine. The liner was assembled in the aluminum rig and tested on an ANCF engine at a University of Notre Dame facility. The test conditions involved acquiring far-field acoustics data for baseline and liner configurations. Far-field acoustic measurements were acquired using 30 microphones placed at a 12-foot radius from the duct centerline and 10 feet high. Fifteen (1–15) of these were in an arc around the inlet exit plane (0–90° measured from the inlet axis, Fwd array) and fifteen (16–30) were in an arc around the exhaust exit plane (90–160° Aft array, with 180° being the exhaust axis). The tone and broadband SPL response data points were extracted separately by running MATLAB scripts.

The evaluation of the impact of the proposed concept led to specific conclusions:

• Fwd array microphones at the inlet exit plane recorded insignificant or no difference in SPLs between the baseline and liner, and this is expected as the microphone locations are at the inlet ahead of the liner setup.
• Aft array microphones 19 to 23 placed at an angle of 105 to 125 deg at the exhaust recorded a significant reduction in tonal SPLs of 0.6 to 13 dB and broadband SPLs (dB loss) of about 2 to 3 dB for shaft speeds of 1500 to 2000 RPM.
• Aft array microphones 16–18 and 24–30 recorded about 1 to 1.5 dB broadband dB loss for shaft speeds of 1500 to 2000 RPM.
• Experimental ANCF testing was repeatable and consistent, with the standard deviations of the measurements being small and the margin of errors with the 95% confidence interval being quite reasonable and within industry standards. Note that the measurement also involved environmental factors like sustained wind and highway vehicle noise during the time of testing.
• The current liner design is effective and showed a noise reduction of about 0.6 to 13 dB for tonal noise at blade passing frequencies and about 2 to 3 dB for the broadband noise.
• The graded microporous hollow bubble composite material offers superior and consistent SPL reduction across multiple microphone locations and at all tested RPMs (1500–2000), making it the most effective noise mitigation technology relative to the T-liner and stepwise metal foam liner. In contrast, the T-liner is only effective at 1800 RPM at a few locations, and the metal foam liner showed marginal or no SPL reductions and sometimes enhanced the noise levels.

The current liner technology, as tested on the ANCF, highlights the fact that controlling the porosity, pore size, and its distribution through the thickness to achieve effective static airflow resistivity and tortuosity is critical in noise attenuation. The present work demonstrates that a 3D microstructurally controlled graded microporous material acoustic liner can be built and effectively reduce the noise of an ANCF engine.