See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262104145

Muffler Design by Noise Transmission Loss Maximization on Narrow Band

Frequency Range

Conference Paper · May 2014

CITATIONS READS

33 466

2 authors:

Mostafa Ranjbar Milad Kermani

Cranfield University Eastern Mediterranean University
104 PUBLICATIONS 1,136 CITATIONS 7 PUBLICATIONS 114 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mostafa Ranjbar on 17 November 2016.

The user has requested enhancement of the downloaded file.

 OTEKON’14
7. Otomotiv Teknolojileri Kongresi
26 – 27 Mayis 2014, BURSA

Muffler Design by Noise Transmission Loss Maximization

Over A Narrow Band Frequency Range

Mostafa Ranjbar, Milad Kermani

Department of Mechanical Engineering, Eastern Mediterranean University,

Famagusta, North Cyprus via Mersin 10, Turkey
Email: [email protected] (M. Ranjbar); [email protected] (M. Kermani)

Abstract
Maximization of noise transmission loss in mufflers using shape modification approach is an important issue
for automobile and aerospace industries. In this paper, we focus on the noise transmission loss maximization on a
narrow band frequency of 200 Hz, one-third-octave band. This frequency range is considered as the design area for
the muffler. Various geometries, i.e., inlet and outlet radius, silencer radius and their lengths are experienced at this
regard. The influenced of geometrical dimensions of muffler parts on the amount noise transmission loss is
investigated. Numerous results from this study like the shape of inlet and outlets which can be used for designing of
more effective mufflers are reported.

Key words: Noise, Transmission Loss, Design, Narrow Band Frequency, Geometry Modification, Muffler.

Introduction performance. Noise emission from the ventilation

Mufflers are normally used in reducing of the system, pumps, compressor, diesel engines, gas
emitted noise from ventilating systems, so controlling turbine, and power stations were investigated.
of noise radiations using a simple chamber like Ranjbar et al. presented the comprehensive study
muffler is usual [1]. Furthermore, the most important on geometry modification of mechanical structures
issue in the design process of mufflers is to develop with respect to their structural acoustic properties [9].
efficient geometries which can maximize the noise They also developed the concept of geometry
transmission loss (TL) of mufflers. In this way, it is optimization of mechanical structures to reduce the
necessary to understand the influence of geometrical radiated sound power level by various optimization
properties of muffler parts on noise TL maximization. methods [10, 11]. De lima et al. attempted to show
Wu et al. studied muffler performance by direct the application of shape and parametric optimization
mixed-body Boundary Element Method (BEM) and a techniques in the study of reactive silencer with
three-point method for evaluating TL in packed extended inlet and outlet ducts [12].Herrin et al.
silencers [2, 3]. Modification of muffler shapes by utilized the finite element method to simulate large
Genetic Algorithm (GA) and Simulated Annealing duct systems to predict the insertion loss [13].
(SA) was done by Yeh et al. [4, 6]. They discussed Acoustical topology optimization for Zwicker’s
about the graphical analysis of optimal shape designs loudness model with application to noise barriers was
and improvement of sound transmission loss for a investigated by Kook et al. [14]. They worked on a
constrained single expansion muffler. Furthermore, design method for acoustical topology optimization
they performed another survey on the shape by considering human’s subjective conception of
optimization of constrained double-chamber mufflers sound as an objective function. The main objective
with extended tubes by mathematical gradient was to minimize the loudness of a target critical band
methods [5]. They also studied the shape rate by optimizing the distribution of the reflecting
optimization of one-chamber mufflers with revers- material in a design domain. The reason of that they
flow in ducts by GA [7]. presented the topology optimized design with 1/3
Luszczynska et al. [8] did a study on the impacts octave band corresponding the critical band rates is to
of low frequency noise on human mental
demonstrate the effectiveness of the proposed radius. Also, the reverse procedure is followed for
method. silencer. It means, the experiments were done with
Another research was done is numerical analysis reducing of silencer radius from its maximum radius
of transmission loss in exhaust muffler system via for 0.0127 meter in each step until when its radius
using four poles methods and their related transfer reaches to the radius of inlet/outlet pipes.
matrixes in 2013 by Abdullah et al. [15]. In this The geometry of considered model for the
paper, we perform a study on the radiuses of the inlet muffler is shown in Fig. 1. It is a muffler with simple
and outlet tube, and also the silencer radius. The TL expansion chamber. As it is shown below, different
is being calculated on a narrow band frequency, i.e., sections of muffler contains Inlet, Ex-tube 1, Silencer
200 Hz, one-third octave band. (which it has soundproof of Polyester around the
silencer with fixed thickness, and inner interface tube
between ex-tube 1 and 2, with in-flange and out-
flange at the both sides of the silencer), Ex-tube 2,
and the outlet. The radiuses of inlet, ex-tubes,
silencer and outlet tubes are considered as r1 , r2 , r3 ,
r4 , r5, and r6 respectively. Also, r3 and r4 are outer
radiuses of in-flange and out-flange respectively. The
inner radiuses of in-flange and out-flange are r2 and
r5 in order, because they have same area as the Ex-
tube 1 and Ex-tube 2.
As more information about the considered model
Figure 1. The geometry of muffler model of muffler is that the thickness of the soundproof is
0.0127 meter [18]
A simple chamber is considred for this study.
Results
Various geometry changes on the radiuses of inlet,
Due to the page limitation in this conference, the
outlet and silencer parts were imposed to calculate
results have been tried to mention by tables and
the resulted TL. The results will produce
descriptions and figures have been omitted from
comprehensive information about the effect of each
them. In each step of the experiments, it has just tried
geometry change on the final noise TL of muffler.
to change one part of a muffler and the dimensions of
other sections have been considered fixed. The TL
Modeling
values with the original size of the muffler are 18 dB,
In this paper, we use three-point method [16] to
17.2 dB, and 15.9 dB within 170-250 Hz. The
measure the muffler TL. This method considers two
original radius of inlet is 0.0254 meter. Then, it is
points on the input and one point on the output tube
being increased by 0.0127 meter up to 0.0889 meter
of a muffler. The location of these points is
in five steps. The TL values for the first attempt when
considered from the left edge of inlet tube of muffler.
the inlet radius is increased from its original size, i.e.,
The impulse of inlet tube is done by velocity or
0.0254 meter to 0.0381 meter, over the considered
pressure, while termination of anechoic is utilized at
narrow band frequency range. The TL value for other
the outlet end [17]. Furthermore, as an analysis tool
attempts in increasing of inlet radiuses over the
for the calculation of the noise TL in muffler, the
considered narrow band frequency range are shown
MAP software [18] is used.
in table 1. By comparing the results’ values of Table
Several various methods are available by MAP
1, it is understandable that the increment of the inlet
software for the calculation TL in mufflers. As
radius leads to increase the TL value over the
indicated, the three-point method is considered for
frequency range of 170-250 Hz. In spite of existing a
the calculation of noise TL in muffler in this study.
drop is seen at attempt 5, but the TL values in that
The effect of cross-sectional radius of muffler
attempt is beyond of the original TL.
parts has been considered to be investigated in a
narrow band frequency range, namely 200 Hz one
Table 1. Original size of muffler parts and the steps
third octave band. This frequency range includes the
of changing in the inlet radiu
area between 178 to 224 Hz. This frequency range is Original Modified radius of inlet (m)
interesting for practical design applications of muffler Section
position
radius 1st 2nd 3rd 4th 5th
in the real world. (m) attempt attempt attempt attempt attempt
Inlet 0.0254 0.0381 0.0508 0.0635 0.0762 0.0889
The size of inlet tube and outlet tube is increased Ex-tube 1 0.0254 0.0254 0.0254 0.0254 0.0254 0.0254
in each step for 0.0127 meter from the original size In-flange
(inner 0.0254 0.0254 0.0254 0.0254 0.0254 0.0254
up to maximum size of silencer cross sectional radius)
 In-flange
(outer 0.0889 0.0889 0.0889 0.0889 0.0889 0.0889
As the table 4 shows us, the TL amounts have
radius) fluctuation, so it cannot be a suitable case to count
Out-flange
(inner 0.0254 0.0254 0.0254 0.0254 0.0254 0.0254 on. It probably to present fixed amounts over next
radius)
Out-flange
frequency ranges. Unfortunately, it is not an
(outer 0.0889 0.0889 0.0889 0.0889 0.0889 0.0889 appropriate matter to rely on.
radius)
Ex-tube 2 0.0254 0.0254 0.0254 0.0254 0.0254 0.0254 Since the cross sectional of Ex-tube 1 is equal to
Outlet 0.0254 0.0254 0.0254 0.0254 0.0254 0.0254 the hub’s cross section (inner radius) of the In-flange,
thus the innovation in Ex-tube 1 is implemented to
hub simultaneously.
According to this description, this section has
Table 2: TL amounts for inlet tube in considered shown you the results for the effect of changing the
frequency range radius of Ex-tube 1 and the inner radius of In-flange
Original
TL of
Modified muffler on the value of noise. Here, the original radiuses of
the
muffler
Attempt
1
Attempt
2
Attempt
3
Attempt
4
Attempt
5
both cross sections are 0.0254 meter, and 0.0127
f1-
18 14.7 21.1 31 36.5 28.5
meter is the increment which was considered for each
f2
f2-
step to reach up 0.0762 meter from central axis of
17.2 14.7 21 30.5 33 26.5
f3 muffler.
f3-
15.9 14.5 20 29 30.5 25
f4
Table 5. Original size of muffler parts and the
Now, we perform the similar study on the radius steps of changing in Ex-tube 1 and inner radius
of outlet tube. Considering the increment values for of in-flange
Modified radius of Ex-tube 1 and inner radius
the outlet radius in each attempt, the results of Section
Original
of in-flange
radius
calculated TL over the frequency range of 170-250 position
(m)
1st
attempt
2nd
attempt
3rd
attempt
4th
attempt
Hz are shown in Table 4. The original radius of outlet Inlet 0.0254 0.0254 0.0254 0.0254 0.0254
is 0.0254 meter. Then, it is being increased by 0.0127 Ex-tube 1
In-flange
0.0254 0.0381 0.0508 0.0635 0.0762

meter up to 0.0889 meter in five steps, see table 3. (inner 0.0254 0.0381 0.0508 0.0635 0.0762
radius)
In-flange
Table 3. Original size of muffler parts and the (outer
radius)
0.0889 0.0889 0.0889 0.0889 0.0889

steps of changing in the outlet radiu Out-flange

(inner 0.0254 0.0254 0.0254 0.0254 0.0254
Original Modified radius of outlet (m)
Section radius)
radius 1 st
2 nd
3 rd
4 th th
5
position Out-flange
(m) attempt attempt attempt attempt attempt (outer 0.0889 0.0889 0.0889 0.0889 0.0889
Inlet 0.0254 0.0254 0.0254 0.0254 0.0254 0.0254 radius)
Ex-tube Ex-tube 2 0.0254 0.0254 0.0254 0.0254 0.0254
0.0254 0.0254 0.0254 0.0254 0.0254 0.0254
1 Outlet 0.0254 0.0254 0.0254 0.0254 0.0254
In-
flange
0.0254 0.0254 0.0254 0.0254 0.0254 0.0254
(inner
radius) In that step of experiment, as the Table 2 has
In-
flange
0.0889 0.0889 0.0889 0.0889 0.0889 0.0889
shown us, the TL has falling trend at each phase of
(outer
radius) increasing the radius. In this part, we consider that
Out- the inlet area is fixed and tried to increase the radius
flange
(inner
0.0254 0.0254 0.0254 0.0254 0.0254 0.0254 of ex-tube 1 along with the inner radius of in-flange.
radius)
Out-
In table 7, the previous case of study is
flange
0.0889 0.0889 0.0889 0.0889 0.0889 0.0889 performed for the Ex-tube 2 radius and inner radius
(outer
radius) of out-flange. The TL results are showed in the table
Ex-tube
2
0.0254 0.0254 0.0254 0.0254 0.0254 0.0254 8.
Outlet 0.0254 0.0381 0.0508 0.0635 0.0762 0.0889
Table 7: Original size of muffler parts and the steps
of changing in Ex-tube 2 radius and inner radius of
Table 4: TL amounts for outlet in considered
out-flange
frequency ranges
Original Ex-tube 2 radius and inner radius of out-flange
Original Section
Modified muffler radius
TL of position 1st 2nd 3rd 4th
(m) attempt attempt attempt attempt
the Attempt Attempt Attempt Attempt Attempt
muffler 1 2 3 4 5 Inlet 0.0254 0.0254 0.0254 0.0254 0.0254
f1- 18 29.7 24.5 20.9 26.3 29 Ex-tube 1 0.0254 0.0254 0.0254 0.0254 0.0254
f2 In-flange 0.0254 0.0254 0.0254 0.0254
0.0254
f2- 17.2 28.9 23.9 21 26 28.8 (inner radius)
f3 In-flange
0.0889 0.0889 0.0889 0.0889 0.0889
f3- 15.9 26.9 22.9 20.9 25 27.5 (outer radius)
f4 Out-flange
0.0254 0.0381 0.0508 0.0635 0.0762
(inner radius)
Out-flange 0.0889 0.0889 0.0889 0.0889 0.0889
 (outer radius) f3-f4 15.9 12.9 10.1 8.5 7.5
Ex-tube 2 0.0254 0.0381 0.0508 0.0635 0.0762
Outlet 0.0254 0.0254 0.0254 0.0254 0.0254
Table 12 shows that the previous study is
repeated for outlet radius, Ex-tube 2 radius, and
radius of out-flange simultaneously.

Table 8: TL amounts by increasing the radius of Ex-

tube 2 and inner radius of out-flange simultaneously Table 11: Original size of muffler parts and the steps
Original TL Modified muffler of changing in outlet radius, Ex-tube 2 radius, and
of the
muffler Attempt 1 Attempt 2 Attempt 3
Attempt inner radius of out-flange
4 Modified radius of outlet radius, Ex-tube 2
Original
f1-f2 18 24 23.9 23 22.5 Section radius, and inner radius of out-flange
radius
f2-f3 17.2 23 22.9 21 21.5 position 1st 2nd 3rd 4th
(m)
f3-f4 15.9 21 20.9 20.1 19.2 attempt attempt attempt attempt
Inlet 0.0254 0.0254 0.0254 0.0254 0.0254
Ex-tube 1 0.0254 0.0254 0.0254 0.0254 0.0254
By the comparison the Table 6 and 8, it can be In-flange
0.0254 0.0254 0.0254 0.0254 0.0254
(inner radius)
understandable that TL amounts in both cases had In-flange
0.0889 0.0889 0.0889 0.0889 0.0889
downward trends, but at table 8 there is a jump in TL (outer radius)
Out-flange
amounts rather than original values. Also, TL values (inner radius)
0.0254 0.0381 0.0508 0.0635 0.0762

at table 8 are more than the amounts at table 6. So, Out-flange

(outer radius)
0.0889 0.0889 0.0889 0.0889 0.0889

the radius of Ex-tube 2 and inner radius of out-flange Ex-tube 2 0.0254 0.0381 0.0508 0.0635 0.0762
Outlet 0.0254 0.0381 0.0508 0.0635 0.0762
are more effective on maximization of TL.
In two next sections, our studies have considered
Table 12: TL amounts by increasing the outlet radius,
more details and added inlet and outlet radiuses to the
Ex-tube 2 radius, and inner radius of out-flange at the
two previous studies respectively.
simultaneously
In table, the impacts of inlet radius, Ex-tube 1
Original Modified muffler
radius, and inner radius of in-flange have showed. In TL of the
Attempt
table 9, the steps of changes have been presented. muffler Attempt 1 Attempt 2 Attempt 3
4
f1-f2 18 15.2 13 12 10.2
f2-f3 17.2 14.9 12.5 11.3 9.9
Table 9: Original size of muffler parts and the steps f3-f4 15.9 13.5 11 9.8 3.9

of changing in inlet radius, Ex-tube 1 radius and

inner radius of in-flange As the tables 10 and 12 show us, increasing the
Original
Modified inlet radius, Ex-tube 1 radius and inner radiuses of these parameters such inlet/outlet, Ex-
Section radius of in-flange
position
radius
(m)
1st 2nd 3rd 4th tubes, and inner radiuses of in/out- flanges make to
Inlet 0.0254
attempt
0.0381
attempt
0.0508
attempt
0.0635
attempt
0.0762
decrease the TL amounts. These experiments have
Ex-tube 1 0.0254 0.0381 0.0508 0.0635 0.0762 been continued up to 0.0127 meter left to the
In-flange
(inner radius)
0.0254
0.0381 0.0508 0.0635 0.0762
soundproof of silencer over the 170 to 250 Hz.
In-flange
0.0889 0.0889 0.0889 0.0889 0.0889 Finally, the survey has been done on the outer
(outer radius)
Out-flange
0.0254
0.0254 0.0254 0.0254 0.0254 radiuses of in/out-flanges. This is done because of
(inner radius)
Out-flange
observing the impacts of changing the inner shapes of
(outer radius)
0.0889 0.0889 0.0889 0.0889 0.0889
silencer by creating changes on cross-sections.
Ex-tube 2 0.0254 0.0254 0.0254 0.0254 0.0254
Outlet 0.0254 0.0254 0.0254 0.0254 0.0254
Table 13: Original size of muffler parts and the steps
In this section of trial, inlet area is considered of changing in outer radius of out-flange
along with extrude and inner radius of internal flange Section
Original
radius
Modified outer radius of in-flange
1st 2nd 3rd 4th
for increasing the radius. The table 10 has shown us position
(m) attempt attempt attempt attempt
that the TLs amounts have downward trend. Inlet 0.0254 0.0254 0.0254 0.0254 0.0254
Ex-tube 1 0.0254 0.0254 0.0254 0.0254 0.0254
In-flange
0.0254 0.0254 0.0254 0.0254 0.0254
Table 10: TL amounts by increasing the inlet radius, (inner radius)
In-flange
Ex-tube 1 radius and inner radius of in-flange (outer radius)
0.0889 0.0889 0.0889 0.0889 0.0889

simultaneously Out-flange
(inner radius)
0.0254 0.0254 0.0254 0.0254 0.0254
Original Modified muffler Out-flange
0.0889 0.0762 0.0635 0.0508 0.0381
TL of the (outer radius)
muffler Attempt 1 Attempt 2 Attempt 3 Attempt 4 Ex-tube 2 0.0254 0.0254 0.0254 0.0254 0.0254
f1-f2 18 14.9 12.9 11 9.5 Outlet 0.0254 0.0254 0.0254 0.0254 0.0254
f2-f3 17.2 14 12 9.9 8.5
Table 14: TL amounts while the outer radius of out- [1] Magrab, E.B., Environmental Noise Control; John
flange is changed from 0.0762-0.0381 inches Wiley & Sons, New York, 1975.
Original Modified muffler [2] Wu, T. W., and Wan, G. C., 1996, “Muffler
TL of the
muffler
Attempt Attempt Attempt Attempt Performance Studies Using a Direct Mixed-Body
1 2 3 4
f1-f2 18 14.6 12 9.8 8.7 Boundary Element Method and a Three-Point
f2-f3 17.2 13.9 11 8.8 7.7 Method for Evaluating Transmission Loss,” ASME
f3-f4 15.9 12.7 9 6.7 4.9
Transactions, J. Vib. And Acoust., 118, pp. 479-484.
As the table 14 shows us, it is understandable [3] Wu, T. W., Cheng, C. Y. R., and Zhang, P., 2002,
that the reduction of outer radius of out-flange which “A Direct Mixed-Body Boundary Element Method
is related to outtake of silencer causes to decrease the for Packed Silencers,” J. Acoust. Soc. Am., 111, pp.
TL value over the frequency range of 170-250 Hz. 2566-2572.
Furthermore, by comparison these two last cases, [4] Yeh, L.J., Chiu, M.C. and Lai, G.J., “Computer
it is crystal clear that the influences of outer radius of Aided Design on Single Expansion Muffler under
out-flange is more striking than the impacts of outer Space Constraints”, Proc. Nineteenth C.S.M.E.
radius of in-flange. Because the TLs over the National Conf., Taiwan, ROC, pp. 625-633.
reductions of radiuses decreased dramatically in out- [5] Yeh, L.J., Chang, Y.C., Chiu, M.C., Lai, G.J. and
flange than in-flange. Her, M.G., 2003, “Shape Optimization of
Constrained Double-chamber Muffler with Extended
Tubes by Mathematical Gradient Methods”,
Conclusion
In this paper, the effect of geometry modification Proceedings of the 3rd Conference on S.M.E.,
of muffler parts on the noise TL over a narrow band Taiwan, ROC, pp. 519-526.
frequency range, is investigated and reported. The [6] Yeh, L.J., Chang, Y.C., Chiu, M.C., 2005, “Shape
computations are performed by the educational Optimal Design on Double-Chamber Mufflers Using
version of MAP software. Simulated Annealing and a Genetic Algorithm”,
The most significant thing that has to be Turkish Journal Engineering Environmental Science,
considered is about existing symmetry in geometry of 29, pp. 207-224
the shape, and if the focus was on changing one [7] Chui, M. C., 2010, “Shape Optimization of One-
section of the shape, the other parts were in original Chamber Mufflers with Revers-Flow Ducts Using a
size. Genetic Algorithm”, Journal of Marine Science and
The results of this study show that outlet area has Technology, Vol. 18, No. 1, pp. 12-23
the fluctuation outcomes and cannot be a matter to [8] Luszczynska, M. P., Dudarewicz, A.,
make decision based on it for optimal muffler shape. Waszkowska, M., Szymczak, W., Kowalska, M. S.,
Totally, TL amounts have been increased 2005, “The Impact of Low Frequency Noise on
according to the below actions: Human Mental Performance”, International Journal
1. By increasing the inlet radius of Occupational Medicine and Environmental Health,
2. By decreasing the radius of Ex-tube 1 18(2), pp. 185-198
and inner radius of in-flange [9] Ranjbar, M., Hardtke, H.-J., Fritze, D., and
3. By decreasing the radius of Ex-tube 2 Marburg, St., 2010, “Finding the Best Design within
and inner radius of out-flange Limited Time: A Comparative Case Study on
4. By decreasing the inlet radius, Ex-tube Methods for Optimization in Structural Acoustics,”
1 and inner radius of in-flange Journal of Computational Acoustics, 18(2), pp. 149-
5. By decreasing the outlet radius, Ex- 164.
tube 2 and inner radius of out-flange [10] Ranjbar, M., Marburg. St., Hardtke, H.-J., 2012,
6. By decreasing the outer radius of in- “Structural-Acoustic Optimization of a Rectangular
flange Plate: A Tabu Search Approach”, Journal of Finite
7. By increasing the outer radius of out- Elements in Analysis and Design, 50, pp. 142-146.
flange [11] Ranjbar, M., Marburg. St., Hardtke, H.-J., 2013,
“Vibroacoustic Optimization of Mechanical
Structures: A Controlled Random Search Approach,”
Acknowledgment
The evaluation version of MAP software is used Advanced Material Research, 622-623, pp. 158-161.
in this study. MAP is developed by T. W. Wu from [12] De Lima, K. F., Lenzi, A., Barbieri, R., 2011,
the University of Kentucky, USA. “The study of reactive silencer by shape and
parametric optimization techniques”, ELSEVIER
Journal, Applied Acoustics, Vol. 72, pp. 142-150.
References
[13] Herrin, D. W., Ramalingam, S., Cui, Z., Liu, J.,
2012, “Predicting insertion loss of large duct systems
 above the plane wave cutoff frequency”, ELSEVIER
Journal, Applied Acoustics, Vol. 73, pp. 37-42
[14] Kook, J., Koo, K., Hyun, J., Jensen, J. S., Wang,
S., 2012, “Acoustical topology optimization for
Zwicker’s loudness model – Application to noise
barriers”, ELSEVIER Journal, Computer Methods in
Applied Mechanics and Engineering, Vol. 237-240,
pp. 130-151.
[15] Abdullah, H., Abu, A., Muhamad, P., Sahekhaini,
A., Quen, L. K., 2013, “On Theoretical of
Transmission Loss in Exhaust Muffler System”,
Advanced Materials Research Vol. 647, pp. 848-853.
[16] Wu, T. W., and Wan, G. C., 1996, “Muffler
Performance Studies Using a Direct Mixed-Body
Boundary Element Method and a Three-Point
Method for Evaluating Transmission Loss,” ASME
Transactions, J. Vib. And Acoust., 118, pp. 479-484.
[17] Cui, Z., and Huang, Y., 2012, “Boundary
Element Analysis of Muffler Transmission Loss with
LS-DYNA,” 12th International LS-DYNA® Users
Conference.
[18] Wu, T. W., 2012, “MAP V0.90 User’s Guide,”
University of Kentucky, USA.

View publication stats