Development of a Metasilencer Considering Flow in HVAC Systems

Abstract

Although the driving noise of electric vehicles has been reduced compared with that of internal combustion engine vehicles, a new interior noise problem is emerging. It is crucial to reduce the noise of the heating, ventilation, and air conditioning (HVAC) system, which is one of the main causes of interior noise. Therefore, in this study, a metasilencer with an acoustic metasurface structure is presented. The metasilencer was designed to restrain the travel direction of the sound wave of the target frequency into a U-shaped configuration using an acoustic metasurface while considering the flow noise effect of the HVAC system. Acoustic analysis confirmed the noise reduction frequency range and refraction effect of the metasurface. The speaker test confirmed the noise reduction effect of the silencer. The same was also confirmed via HVAC tests, even in the presence of a flow.

1. Introduction

In recent years, the e-mobility industry, which is a means of transportation utilizing electricity (considered to be an environmentally friendly power source), has expanded globally, and the use of electric vehicles, which is one of the representative fields of the e-mobility industry, has been rapidly increasing. Although the level of noise of electric vehicles is lower compared with conventional internal combustion engine vehicles, now it seems that noise, vibration, and harshness (NVH) problems associated with electric drives as well as heating, ventilation, and air conditioning (HVAC) systems are emerging. Thus, it is crucial to reduce the noise of the HVAC system, which is one of the main causes of interior noise.

Various studies have been conducted to reduce noise in HVAC systems. Neise et al. [1] utilized design changes such as fan casings and blade spacing in HVAC systems to reduce noise, whereas Wang et al. [2] changed the internal shape of the air conditioner. Singh et al. [3] studied sound absorbing material (SAM) attached to HVAC systems using natural materials such as jute felt. These studies focused on how to change the internals of heater and evaporator systems. Another way to dampen the noise level of an HVAC system is to install a silencer on the HVAC duct. Case et al. [4] produced a silencer with a Helmholtz resonator structure to reduce the noise of an HVAC system. A Helmholtz resonator was installed and tested in the HVAC system, but the silencer was developed without considering the flow noise. To reduce duct noise in HVAC systems, there is a need for a method of reducing noise using structures other than Helmholtz resonators.

One method of reducing noise by attaching a silencer to a duct involves the use of acoustic metamaterials. Among acoustic metamaterials, there is a method that can manipulate the direction of sound waves called a metasurface [5,6,7,8,9]. The metasurface may have a structure similar to those of Helmholtz [10,11,12], quarter wave [13,14], or hook-like [15,16] resonators. Other metasurfaces include labyrinthine structures such as tapered [17] or space-coiling [18] labyrinthine types. These metasurfaces can manipulate the sound wave direction at the target frequency using the generalized Snell’s law and phase delay [19].

In this study, a metasilencer was developed using an acoustic metasurface to reduce noise in HVAC systems. The metasilencer was designed to restrain the travel direction of the target frequency sound wave into a U-shaped configuration using a metasurface. To make the installation space of the metasilencer advantageous, the unit cell was set in the labyrinthine structure, and the length and interval of the unit cell were determined using the generalized Snell’s law and phase delay. Considering the influence of the flow noise of the HVAC system, the inlet of the unit cell has a curved structure, and the interior of the unit cell is filled with a small amount of SAM. The acoustic analysis results of the metasilencer confirmed the refractive effect of the noise reduction frequency domain and metasurface. In the speaker test, the noise reduction effect was confirmed by comparing the test results before and after the metasilencer installation, and an HVAC test was performed to confirm the noise reduction effect of the silencer, even in the case of the flow.

2. Theory

2.1. Target System

Figure 1 shows the parts of the HVAC system from a vehicle. The output voltage of the HVAC system can be expressed as a singular number, which comprises a total of six steps. The maximum voltage of the HVAC system was 13 V, and the single voltage difference of the motor was 1–2 V. Figure 1a shows the upper HVAC duct, which is divided into the discharge port center, side, front window, and side window. Figure 1b shows the heater and evaporator system. The duct attachment area of the heater and evaporator system comprises a suction port and a discharge port. Here, A–F denote the discharge ports, G is the suction port, and the upper HVAC duct is attached to A, as shown in Figure 1c. In the heater and evaporator system, except for discharge port A, the other discharge ports were noise blocked using a SAM so as not to interfere with the noise testing. Furthermore, the suction port duct was produced in such a way that the noise from the suction port did not interfere with the noise test.

Figure 2 shows the measurement results of the discharge noise when 13 V, the maximum voltage of the HVAC system, was applied. An HVAC duct was used for the noise measurement of the HVAC system. The microphone was located 500 mm away along an inclined line 45° from the center of one of the central discharge port. Other detailed measurement methods are described in Section 2.3. Figure 2a shows the 0–5000 Hz range confirmed by a color map, and Figure 2b shows Figure 2a enlarged to the 0–2000 Hz range for accurate analysis. Figure 2c shows the SPL results up to the 2000 Hz range. The acoustic measurements of the HVAC system confirmed that the range below 1100 Hz was the dominant noise range. Based on the acoustic measurement results, the silencer was set to reduce the noise in the frequency range of 800–1100 Hz. The reason for selecting 800–1100 Hz as the target frequency range is to reduce it to approximately 1 kHz, to which the human ear is more sensitive. The target frequency range was selected such that the size of the metasurface was 200 mm × 45 mm. The metasurface size constraints are discussed in detail in Section 2.2. The metasurface used herein reduces the area by approximately ±100 Hz based on the target frequency of the transmission loss analysis results. Thus, the target frequency of the metasurface was set at 900 Hz and 1000 Hz, which are at intervals of 100 Hz, in the target frequency range of 800–1100 Hz.

Figure 3 shows a 3D model with a straight pipe attached to the simple model. A simple model was designed by simplifying the shape of the HVAC duct. The simple model was manufactured using 3D printing, and the material used was acrylonitrile butadiene styrene (ABS). To test the effectiveness of the silencer, a simple model was designed, such that a straight pipe and silencer could be attached to the simple model. When only the metasilencer was installed in the simple model, the distance between the last unit cell of the metasilencer and the discharge port was narrow. Therefore, the noise reduction effect of the last unit cell may be insufficient owing to the turbulent flow generated at the outlet. To address this problem, by mounting an additional duct in front of the silencer, it is possible to prevent the noise reduction effect of the unit cell from being insufficient owing to the turbulence of the discharge port. The cross-sectional size of the duct is equal to $a$ × $b$ in Figure 3, where a and b denote the horizontal and vertical lengths of the duct, respectively. The cross-sectional size of the duct, $a$ × $b$, was set to 60 mm × 90 mm. The simple model is attached to A of the heater and evaporator system in Figure 1b, similar to an HVAC duct.

2.2. Metasilencer Design

Based on the acoustic measurement results, the silencer was set to reduce the noise in the frequency range of 800–1100 Hz. The metasurface used herein is characterized by a bandgap reduction of approximately ±100 Hz with respect to the target frequency. Therefore, the metasurface target frequencies were set to 900 Hz and 1000 Hz. Figure 4a shows the metasilencer duct model, where the upper and lower parts of the metasilencer reduce noise at 900 Hz and the width is set to 90 mm, while the left and right sides reduce noise at 1000 Hz and the width is set to 60 mm. In Figure 4a, $c$ is the width of the metasurface and $d$ is the height of the metasurface. The maximum width of the metasurface was set to 200 mm, and the maximum height was set to 45 mm. Figure 4b shows the shape of the unit cell structure, and the structure of the unit cell is set to a labyrinthine type. The reason for having a labyrinthine-type structure is that the target frequency can be designed at a lower height than that of the same geodesic tube, which is advantageous in terms of space constraints, where $g,$ $e,$ $h,$ and $l_n$ denote the width of the unit cell, width of the unit cell cross section, wall thickness, and length of the unit cell entrance, respectively. The width of the unit cell ($g$) was set to 21 mm and 19 mm according to the target frequency of 900 Hz and 1000 Hz, respectively; the width of the unit cell ($e$) and wall thickness ($h$) cross sections were set to 3 mm and 1 mm, respectively, regardless of the target frequency.

The metasilencer was designed by adjusting the unit cell interval ($dx$) and unit cell entrance length ($l_n$) to prevent the sound wave propagation direction of the target frequency in a U-shaped configuration. In other words, the factors for designing the acoustic metasilencer are the unit cell spacing and length of the unit cell entrance. The unit cell interval can be obtained using the generalized Snell’s law, and the length of the unit cell entrance can be obtained using the phase delay. First, the generalized Snell’s law is

$$sin{\theta }_r=sin{\theta }_i+\frac{\lambda }{2\pi }\times \frac{d\varnothing }{dx}$$

(1)

where ${\theta }_i$, ${\theta }_r$, $\lambda$, $d\varnothing /dx$, $dx$, and $d\varnothing$ denote the angle of incidence, angle of reflection, wavelength of the target frequency, phase gradient, unit cell spacing, and phase delay interval, respectively. The angle of incidence and angle of reflection can be set to refract in the desired direction, and the length of the wavelength can be obtained by determining the target frequency. In the generalized Snell’s law, the variable is the phase gradient, and setting the interval of the phase delay to a constant value in the phase gradient can induce the spacing of the unit cell. Figure 5 shows how the angles of incidence and reflection can be set, where the vertical part of the surface of the metasurface is 0°, and the direction of sound wave propagation is −90°. For example, if the incident angle and reflection angle are set to 30° and 60° on the metasurface, respectively, the incident angle is −120°, and the reflection angle is −30°. To refract the traveling direction of the sound wave in a U-shaped configuration, the incident and reflection angles were set to −90° and 0° for the upper portion, and the incident angle and reflection angle were set to 0° and 90° for the lower portion, respectively. The wavelength length of the target frequency can be obtained by $\lambda =c/f$, where 𝑐 is 343 m/s because the sound propagation medium is air. The phase delay interval is set to 30°, and the phase delay is discussed in detail in the next paragraph. Deriving the unit cell spacing according to the target frequency using the devised design dimensions yields 31.76 mm for 900 Hz and 28.58 mm for 1000 Hz.

Table 1. Design dimensions of the metasurface derived using generalized Snell’s law.

$\mathit{f}$	${\mathit{\theta }}_{\mathit{i}}$	${\mathit{\theta }}_{\mathit{r}}$	$\mathit{d}\varnothing$	$\mathit{d}\mathit{x}$
900 Hz	−90°/0°	0°/90°	30°	31.76 mm
1000 Hz	−90°/0°	0°/90°	30°	28.58 mm

shows the design dimensions of the metasurface derived using the generalized Snell’s law according to the target frequency.

Figure 6 is the model for deriving the unit cell incident pressure and transmission pressure, where $p_i$, $p_r$, $p_t$, and $l_n$ denote the incident pressure, reflection pressure, transmission pressure, and length of the unit cell entrance, respectively. The phase delay graph can be derived using incident and transmitted pressures depending on the length of the unit cell entrance. To derive this, first, knowing the rate of transmission of sound pressure that changes according to the unit cell inlet length is necessary. The transfer rate of the sound pressure formula is

$$t_p=\frac{p_t}{p_i}=\alpha +\beta i$$

(2)

where $p_i$, $p_t$, $\alpha$, and $\beta$ denote the incident pressure, transmission pressure, real value of the transmission rate of acoustic pressure, and imaginary value of the transmission rate of acoustic pressure, respectively. The incident pressure and transmission pressure values for each length of the unit cell entrance are derived using acoustic analysis, and substituted into Equation (2) to derive the real and imaginary values of the transmission rate of acoustic pressure. The phase delay formula is

$$\mathrm{phase} \mathrm{delay}={\tan }^{-1} \frac{\beta }{\alpha }$$

(3)

where $\alpha$ and $\beta$ are the real and imaginary values of sound pressure transmissibility, respectively. This is the same as in Figure 7 when the phase delay is calculated using the real and imaginary values obtained according to the length of the unit cell entrance on each metasurface and graphically derived. The length of the unit cell entrance must be selected for each phase delay interval ($d\varnothing$) to cover the entire range of −π/2 to π/2 in the phase delay graph. When the number of the unit cell is set to 6, the interval of the phase delay is 30°, and the length of the unit cell entrance corresponding to the target frequency of the metasurface was derived.

Table 2. Design dimensions of the metasurface derived using phase delay.

$\mathit{f}$	${\mathit{l}}_1$	${\mathit{l}}_2$	${\mathit{l}}_3$	${\mathit{l}}_4$	${\mathit{l}}_5$	${\mathit{l}}_6$
900 Hz	15.71 mm	12.72 mm	11.68 mm	10.97 mm	9.89 mm	5.21 mm
1000 Hz	12.00 mm	10.81 mm	10.42 mm	9.87 mm	9.50 mm	7.08 mm

shows the designed dimensions according to the target frequency of each metasurface.

Table 1 and Table 2 show that the width and height of the metasurface set to six unit cells fall within 200 mm and 45 mm, respectively. In addition, if the number of unit cells is set to 12 in the design process, the unit cell interval is calculated as a value smaller than the width of the unit cell, making the design impossible. If the number of unit cells is three, the unit cell interval can be designed to have a smaller value than the unit cell width; however, the number of unit cells is small, and the reduction effect may be reduced.

The metasurface designed using the generalized Snell’s law and phase delay has the shape shown in Figure 8a. Although the flow direction is set as shown in Figure 8a, as long as the design dimensions of the unit cell spacing ($dx$) and inlet length ($l_n$) are not changed, setting the flow in the opposite direction is acceptable. When there is a flow rate, turbulence noise is generated at the edge of the unit cell, and turbulence noise at the edge can be reduced by chamfering the edge [20]. The inlet of the unit cell was designed as a curved structure to minimize the occurrence of turbulence at the metasilencer edge. To avoid affecting the effect of the metasurface owing to the curved surface structure of the entrance of the unit cell, the maximum size of the curved surface that can be given by the size condition of the shape was set to 5 mm. In addition, considering the flow noise of the HVAC system, the inside of the unit cell was filled with a small amount of SAM, which was a surface material. Lee et al. [21] confirmed the effect of SAMs at 100 g/L and 200 g/L. A small amount (30 g/L) of SAM was filled inside the unit cell so that the noise reduction effect of the SAM would not be so large as to prevent airflow from flowing into the unit cell. Considering the effect of flow noise in Figure 8b, the edge of the unit cell inlet was designed to have a curved structure, and the inside of the unit cell was filled with a small amount of SAM.

2.3. Test Setup

Figure 9 shows a schematic of the speaker test method, where a Squadriga was used as the noise measurement device. The microphone used the PCB 378B02 model and was positioned 500 mm away from the discharge port along a line inclined at 45°. A speaker test was conducted to confirm the noise reduction effect of the metasilencer. The speaker output source used white noise, which has a nearly constant frequency spectrum over a wide frequency range.

Figure 10 shows a schematic of the HVAC test method, with the same measuring equipment, microphone models, and locations as the speaker test. In addition, except for the part where the HVAC duct or simple model was attached, the duct connection part of the heater and evaporator system was blocked with SAM, and the test proceeded. To confirm the target frequency selection and the effect of the metasilencer, HVAC tests were conducted for four cases. First, to select the target frequency for the HVAC system, HVAC tests were performed with an HVAC duct installed in the heater and evaporator system. Second, to confirm the noise reduction effect of the metasilencer when there is a flow rate, the HVAC test was conducted depending on whether the metasilencer was mounted, and a 13 V voltage of the HVAC system was applied. Third, to confirm the noise reduction effect of the metasilencer when the flow velocity was low, the HVAC test was performed with the voltage of the HVAC system set to 6 V. Finally, to confirm the effect of the metasilencer considering the flow noise, HVAC tests were conducted on the metasilencer with and without considering the flow noise.

3. Results and Discussion

3.1. Acoustic Analysis Results

The metasilencer was designed based on the design variables calculated in Section 2.2, and acoustic analysis was performed using ANSYS 2022 R2 Acoustic software. Figure 11 shows the transmission loss graph obtained by acoustic analysis, and it can be confirmed from the graph that the noise reduction frequency range was 800–1100 Hz. The maximum transmission loss was 60 dB at 850 Hz and 62 dB at 940 Hz. Figure 12 shows the scattered field results of COMSOL and acoustic analyses of each metasurface designed to confirm the refraction effect of the metasurface. Figure 12a shows the scattered field result of the metasurface at 900 Hz, and Figure 12b shows the scattered field result of the metasurface at 1000 Hz. Consequently, there is U-shaped refraction in the scattered field, and there is a wavelength returning to the incident direction.

3.2. Test Results

The metasilencer, which confirmed the effect of the metasurface through acoustic analysis, was manufactured using 3D printing, and the material was ABS. Figure 13 shows the results of the speaker test, which was performed as described in Section 2.3, and white noise was used as the speaker output source. Here, the W/O metasilencer is the case where a straight pipe is installed instead of a metasilencer, and the W/metasilencer is the case where the metasilencer is installed. As a result, it was confirmed that the noise reduction frequency range of the metasilencer was 700–1100 Hz, and a maximum noise reduction of approximately 35 dBA was confirmed.

First, the HVAC tests were performed with a weak current, where the voltage of the HVAC system was 6 V. Figure 14 shows the HVAC test results based on the presence or absence of the metasilencer considering the flow when the voltage of the HVAC system was 6 V. Here, the W/O metasilencer and W/metasilencer were the same as the speaker test results. When the flow is weak, it can be confirmed that the noise reduction frequency range of the metasilencer is 700–1200 Hz, and a maximum noise reduction of approximately 10 dBA was confirmed.

HVAC tests were performed at 13 V, the maximum voltage for the HVAC system. Figure 15 shows the result of the HVAC test, which is the maximum voltage of the HVAC system. Here, the W/O metasilencer is a straight pipe, and the W/metasilencer 1 has a structure that considers the flow. In other words, the W/metasilencer 1 has a structure in which a curved edge is applied to the entrance of the unit cell and a small amount of SAM is filled inside the unit cell. Comparing the straight pipe and the W/metasilencer 1, when there was a flow, the noise reduction frequency range was confirmed to decrease in the range of 750–1200 Hz, and the maximum noise reduction was confirmed to be approximately 5 dBA. In addition, to confirm the effect of adding a curved edge to the entrance of the unit cell and filling a small amount of SAM inside the unit cell, an HVAC test was performed on a metasilencer without a curved edge entrance or a SAM. The result resembled the W/metasilencer 2 in Figure 15a, and when compared with the straight pipe, the noise increased over the straight pipe in the frequency range of 650–750 Hz. This revealed the noise generated by the turbulence at the end of the metasilencer. Furthermore, to confirm the effect of filling a small amount of SAM inside the unit cell, an HVAC test was performed on a metasilencer that was only given a curved edge entrance. The results were the same as for the W/metasilencer 3 in Figure 15b, and when compared with intuition, it was confirmed that the noise reduction frequency range reduced to approximately 750 to 1200 Hz, and the maximum noise reduction was approximately 3 dBA. This confirms that the W/metasilencer 1 had a greater noise reduction capability than the W/metasilencer 3.

4. Conclusions

Although the relative noise of an electric vehicle has been reduced compared with that of an internal combustion engine, a new NVH problem that has not been previously recognized is emerging. Reducing the noise of the HVAC system, which is one of the main causes of interior noise, is crucial. This study presents a metasilencer that considers the effect of flow noise to reduce the noise of an HVAC system. The main results obtained in this study are as follows:

• The target frequency of the metasilencer was selected from the acoustic measurement results of the HVAC system, and a labyrinthine-type metasilencer considering the flow noise was proposed. The metasilencer was designed using the phase delay and generalized Snell’s law.
• The inlet of the unit cell was designed as a curved structure to minimize the occurrence of turbulence at the edges of the metasilencer. In addition, a small amount of SAM was filled inside the unit cell considering the influence of the flow noise of the HVAC system.
• Acoustic analysis confirmed that the metasilencer reduced noise over a wide frequency range and that the metasurface refracts the sound propagation direction of the target frequency.
• The sound measurement results before and after installing the metasilencer in the speaker test confirmed that noise was reduced over a wide frequency range. HVAC testing confirmed the noise reduction effect of the metasilencer in the presence of flow.