Formula SAE Performance Exhaust Design

Published on November 2016 | Categories: Documents | Downloads: 138 | Comments: 0 | Views: 814
of 65
Download PDF   Embed   Report

Comments

Content

An Investigation into Formula SAE Performance
Exhaust Design and Analysis
Anthony I. McLeod1

University of New South Wales at the Australian Defence Force Academy

The Formula SAE competition demands that teams pursue synergistic designs in
creating a competitive high performance vehicle. As a result, the design of each
component fitted to a vehicle, including that of the exhaust, must be undertaken with
a sound foundation of technical understanding in conjunction with creativity and
innovation. Exhaust design is shown to make a significant contribution to engine
performance, economy and noise attenuation. Hence, this work aims to assist the
ADFA Formula SAE team of 2012 develop an understanding of current exhaust
analysis and tuning techniques such that they may be innovatively applied to the
design of a high performing exhaust system as a part of a holistic engine tuning
approach. Extensive research has been conducted into the mechanisms by which an
exhaust design may enhance engine performance and attenuate noise. In particular,
an exhaust design is understood to effect engine performance via influences upon
engine scavenging. Furthermore, the action of automotive silencers was identified to
be governed by their ability to manage the mass flow rate from the exhaust outlet as
opposed to that of acoustic theory. In addition, research has identified methods such
mechanisms may be analysed and predicted. Engine simulation software Ricardo
WAVE was used to demonstrate and analyse the performance and noise attenuation
implications of exhaust system componentry and their design parameters. A volume
restricted silencer design proposed by Professor Blair of the University of Belfast
formed the basis of further experimental and theoretical analysis of the governing
principles of silencer operation. Specifically, a derivative of this design concept was
manufactured with in-built variability to enable an experimental investigation of the
design and to also help validate data obtained using WAVE. Finally, WAVE was used
to enable a theoretical analysis which underpinned a design proposal for a high
performing silencer.

Contents
I.

II.

Introduction .............................................................................................................................................. 3
A.

Motivation .................................................................................................................................. 3

B.

Project Aims ............................................................................................................................... 3

C.

Project Methodology .................................................................................................................. 3

Part A - Literature Review........................................................................................................................ 3
A.

Exhaust design for engine scavenging performance ................................................................... 4

B.

Acoustics, Vehicle Noise and Exhaust Silencing ....................................................................... 5

C.

Exhaust Silencing ....................................................................................................................... 6

D.

Design and Modelling of Exhaust Systems ................................................................................ 9

E.

Conclusion .................................................................................................................................. 9

III. Part B –Concept Development and Investigation ................................................................................... 10
1

SBLT, School of Engineering & Information Technology, ZEIT4500
1
Final Project Report 2011, SEIT, UNSW@ADFA

A.

WAVE Model Development .................................................................................................... 10

B.

Experimental Silencer Parameter Study ................................................................................... 10

C.

WAVE Investigation – Exhaust Design Parameters for Engine Scavenging ........................... 13

IV. Part C – Preliminary Design and Design Proposal ................................................................................. 15

V.

A.

Design Requirements................................................................................................................ 15

B.

Silencing Strategy..................................................................................................................... 15

C.

Design Investigation and Definition ......................................................................................... 16

D.

Design Proposal .........................................................................Error! Bookmark not defined.

E.

Vehicle Integration of Exhaust System .................................................................................... 18

Limitations of Ricardo WAVE ............................................................................................................... 19

VI. Conclusion .............................................................................................................................................. 19
VII. Recommendations and Future Work ...................................................................................................... 19
Acknowledgements ......................................................................................................................................... 20
References ....................................................................................................................................................... 21
APPENDICES
Appendix A. Combined theoretical and experimental investigation into exhuast pipe geometry
Appendix B. Exhaust pipe optimisation using NSAGA2 and ANSYS Fluent

Nomenclature
ADFA
CFD
SPL
Q
D
̇

=
=
=
=
=
=

Australian Defence Force Academy
Computational Fluid Dynamics
sound pressure level (dB)
volume flow rate(m3/s)
pipe diameter (m)
flow velocity(m/s)

2
Final Project Report 2011, SEIT, UNSW@ADFA

A1
A2

I.

Introduction

A. Motivation
The Formula SAE® competition constitutes a variety of rules and regulations that aim to challenge design
teams whilst maintaining fairness and safety. A number of pertinent rules to this study [1] include:
 The vehicle‘s engine must be a four cycle piston engine with a maximum swept displacement of 610cc,
 An intake restrictor must be fitted with a maximum diameter of 20.0 mm for vehicles operating with
gasoline and 19.0 mm for vehicles operating with E85, and
 A vehicle‘s measured noise level must be less than 110 decibels.
It is obvious from these rules in particular that teams are challenged to form a competitive advantage via [2]
synergistic vehicle designs by applying technical knowledge innovatively as well as through the application of
advanced performance tuning techniques. In this way, teams may attain the necessary combination of power and
efficiency to be competitive throughout a series of trying auto-cross events.
The concept of ―exhaust tuning‖ has been under development for over 60 years [3]. In this time,
exhaust design has been proven to have a marked influence upon the performance and efficiency of an engine
by way of power output, specific fuel consumption, heat production and radiated noise level. It is as a
consequence of the flow on effects of such factors that the implementation of a sound understanding in the
design of an exhaust is crucial in order to obtain a high performing racing vehicle. For the benefit of
competitiveness, it is therefore important for the Formula SAE® team representing ADFA to learn to approach
the design of the exhaust in such a way that maximizes performance of the competition vehicle and ensures
compliancy.
B. Project Aims
The ADFA Formula SAE® team of 2012 has purchased a Yamaha WR450 single cylinder engine to be
integrated into a new competition vehicle. It is therefore the intent of this project to assist the team to understand
the potential benefits of exhaust tuning as well as the methods that are available in the analysis and design of an
exhaust. This project will consist of the validation of theories currently used to enhance engine output.
Furthermore this investigation will be extended to noise generating phenomenon and associated analytical and
prediction techniques. The project will employ a one-dimensional engine simulation software, Ricardo WAVE
to then undertake analysis to be validated experimentally, culminating in a final design proposal for a new high
performing silencer.
C. Project Methodology
This project utilised a series of methods to carry out an investigation into exhaust systems and their design.
Initially, extensive research constituting a literature review was undertaken to build a knowledge base requisite
of applied analysis and design. The complex trade-offs found to characterise silencer design then motivated an
experimental investigation using a promising silencer design concept that was developed and proven upon a
similar engine as the Yamaha WR450, by Professor Blair of the University of Belfast. An experimental muffler
was manufactured based upon this design concept which incorporated in-built variability to enable an
experimental parameter study of silencer attenuation. DOE methodology was utilised to conduct this
experimental study which employed an available and operable WR250 motorbike in lieu of the engine testing
rig still under development by the FSAE team. This experiment obtained insertion loss for the silencer within
the frequency domain to such that the governing principles of silencer operation could be identified. An engine
simulation model was then developed using the one-dimensional engine simulation software Ricardo WAVE.
This software then underpinned the demonstration and theoretical analysis of performance exhaust tuning and
silencer theories. Exhaust performance aspects investigated include the concept of ‗tuned length‘, the effect of
stepped pipes and diffuser components as well as the nature of ‗inertial scavenging‘ phenomena. Performance
data obtained is discussed such that these methods become yet another tool for the ADFA FSAE team to utilise
within an integrated engine tuning process. The silencer experiment was duplicated within WAVE to provide a
level of validation of the developed model. Continued silencer analysis employed the WAVE transmission loss
work bench. The conclusions drawn from these analyses then facilitated the development of a design proposal
for a high performance silencer.

II.

Part A - Literature Review

Quality design of an exhaust system requires a sound understanding of its contribution to both the overall
power output of an engine and to noise attenuation. Furthermore it is important to understand the mechanisms
that enable these contributions as well as their significance. A wide variety of sources were studied to determine
current exhaust theories, design and analysis methods as well as to better understand the restrictions imposed by
Formula SAE noise regulations.
3
Final Project Report 2011, SEIT, UNSW@ADFA

A. Exhaust design for engine scavenging performance
Performance considerations of exhaust design are a result
of the nature of the gas exchange process in a four stroke
engine. This process includes a period of valve overlap
where both the intake and exhaust valves are open
simultaneously as seen in Fig.(1). Without due regard by the
designer this period could see the induction of exhaust gases
into the cylinder as shown in Fig.(2), effectively reducing the
amount of fresh combustibles ingested and therefore overall
power.
Performance aspects of exhaust design are concerned
with minimising such residual quantities or otherwise stated
as maximising scavenging efficiency of the engine. This is
achieved, one way or another by reducing the exhaust valve
pressure during valve overlap such as to bias this exchange
Figure 1. Valve timing events showing valve
process to achieve this scavenge.
overlap
Exhaust scavenging is achieved via two methods. This is
because the exhaust phase of the four stroke cycle consists of
not only the expulsion of a high speed column of exhaust
gases but also a pressure wave. Consequently, scavenging is
achieved through techniques known as ‗wave tuning‘ and
‗inertial scavenging‘ depending on which of these
mechanisms we utilise.
The aim of wave tuning is described by Professor Blair of
the University of Belfast who states that ―the tuned exhaust
pipe harnesses the pressure wave motion of the exhaust
process to extract a greater mass of exhaust gas from the
cylinder during the exhaust stroke and initiate the induction
process during the valve overlap period.‖ This scavenging
effect is possible if a pressure wave originating from the
Figure 2.
Poorly tuned engine ingests
exhaust valve travel at the local acoustic velocity, over a exhaust gas into the cylinder during valce
tuned length such that it is reflected back to the valve face as overlap
a rarefaction wave, as seen in Fig.(3), in time to assist the gas
exchange process during valve overlap.
The phasing of the exhaust valve and the pressure wave
is dependent upon the length over which the waves travel.
Commonly known as the ‗tuned length‘, it is defined by the
length of pipe bounded by the exhaust valve and a
discontinuity in the pipe of an area ratio of 6. This being a
point that significant wave action can operate from.
In addition, scavenging may be achieved via inertial
scavenging. This is a scavenging effect achieved as a result
of the inertia of a high velocity column of gas. It functions
Figure 3. Reflection of rarefaction wave at
under the principle that a fixed volume flow rate is achieved
at a certain engine speed and for a fixed volume flow rate, exhaust pipe end which returns to the exhaust
gas velocity varies inversely with pipe diameter. There then valve
exists a pipe diameter where the scavenging effect produces a
more than proportionate amount of power than pumping
work required to achieve an effective gas velocity.
In addition to the stated performance exhaust theory there exists a number of complicating factors for the
realistic exhaust system designer. Firstly, the periodic nature of wave phenomenon in the exhaust suggests that
whilst tuning may be carried out for the benefit of power at one engine speed this will inevitably lead to poorer
performance in another [4, 6]. Furthermore, tuning of the exhaust without due regard of the interactions taking
place with other mechanisms such as similar wave action occurring in the intake, has the potential to produce
irregular shapes including troughs and peaks within the power curve. As a consequence the drivability
characteristics of the vehicle could diminish as a result of unpredictable power output behaviour.

4
Final Project Report 2011, SEIT, UNSW@ADFA

Another perspective of the priorities of exhaust system design is provided by a parameter study conducted
by Sammut and Alkidas [11]. This study utilizes the engine simulation software Ricardo WAVE to quantify the
effects of and interactions between exhaust, intake and valve timing parameters. For a constant valve timing and
engine speed, Fig.(4) shows a comparison of the scavenging effect of the intake and exhaust measured in
volumetric efficiency. The data presented firstly shows that the individual contributions of the intake and
exhaust are independent as the contribution made by the exhaust is relatively constant for any intake length.
However, it is important to note that such independence should not be assumed between all parameters. All data
presented herein illustrating variation in scavenging as a function of tuned length is obtained for constant valve
timing. Variation in valve timing would inevitably change the characteristics of the overlap period and therefore
the action of exhaust scavenging. The effects of this are well documented throughout literature but assumed

Figure 4. Variation of volumetric efficiency with intake and exhaust length

constant for the purpose of this investigation of exhaust design. Secondly, data presented in Fig.(4) also
concludes that the effect of exhaust tuning is relatively small compared to the benefits of intake tuning. As a
consequence of the diminishing significance of exhaust scavenging benefits, minimizing the losses conceded to
increased pumping losses whilst achieving sufficient noise attenuation becomes of relatively high importance if
a maximum amount of power is to be derived from the engine. With the realization that there as much potential
for an exhaust system design to reduce performance as to improve it, the design of an efficient silencer becomes
crucial to the competitiveness of the vehicle. Moreover, it needs to be integrated within a system with minimal
prejudice towards efforts to attain an effective scavenge by providing low exhaust valve pressure at valve
overlap [6].
B. Acoustics, Vehicle Noise and Exhaust Silencing
Literature was consulted in order to define the problem of vehicle noise as well as to gain an appreciation of
current vehicle noise attenuation techniques such that this design issue could be effectively addressed. A noise
measurement of sound radiated from a vehicle is subject to a variety of sources including mechanical noise,
shell vibration radiated noise and duct noise where duct noise then consists of intake and exhaust tail pipe noise
[13]. Fig.(5) is provided to illustrate the prevalence of intake duct noise, being a source not considered here.
The sound pressure measured at any point in space is
relative to the radius defining the distance between the
source and the point of measurement as well as the
directivity of the source with respect to this radius vector.
Furthermore, an important consequence of the logarithmic
scale of sound measurement is that the sum of SPL from
multiple sources varies little from the maximum SPL [4] as
seen in Fig.(5). Consequently, a vehicle silencing strategy
formulated to control sound pressure at a specified location
relative to the vehicle, needs to acknowledge the most
significant source at that location in order to effectively
control the final measurement. Therefore, the following
discussions detailing exhaust tail pipe noise attenuation can
only be effective within spatial regions where this is the Figure 5. Noise level of intake and exhaust duct
dominant noise source and sound pressure contributions of noise
5
Final Project Report 2011, SEIT, UNSW@ADFA

other sources such as intake duct noise and engine
noise become negligible.
Such conclusions then underpin the next priority
in forming an efficient vehicle noise attenuation
strategy being to recognize the dominant components
of exhaust tail pipe noise. The design of an efficient
muffler should then target these most significant noise
components in order to attain the required attenuation
level whilst maintaining low resistance to flow. Tail
pipe noise component of duct noise and consists of
[8]:
1) Pulse/Engine noise which describes sound of
frequencies corresponding to harmonics of
the engine firing rate (EFR), as seen in Figure 6. Frequency analysis of radiated vehicle noise
Fig.(6). The EFR is the rate at which the showing noise corresponding to EFR frequencies
exhaust valve releases combustion gases
from the cylinder, and
2) Gas flow noise which consists of high
frequency broadband noise resulting from
pressure fluctuations inherent to the turbulent
mean gas flow in the exhaust duct.
Fig. (7) [14] indicates that the pulse noise is
dominant at low engine speeds and is superseded by
flow noise as it increases in magnitude with volume
flow rate and engine speed. A silencer design must
therefore incorporate elements that can target the
dominant noise source at the engine speed of interest.
Specifically, silencing at low engine speeds must be
concerned with discrete harmonics of the engine
firing rate while silencing at high engine speeds is
more concerned with high frequency flow noise.
Pang et al [13] shows a direct proportional
correlation between flow noise and flow velocity and
therefore exhaust pipe diameter given by Eq (1), where
the volume flow rate is a function of engine speed. The
relationship between flow noise and diameter is seen in
Fig.(8). This shows that at high engine speeds where
flow noise is dominant, a fixed volume flow tranlsates
to greater flow velocity for a smaller diameter and
therefore increased noise emissions.
.
̇

̇

̇

Figure 7. Sound pressure with engine speed showing
increasing dominance of flow noise with engine speed

(1)

The conversion of flow power to sound power is
Figure 8. Variation in flow noise with flow
identified by Wiemeler, Jauer and Brand [14] to be
velocity caused by pipe diameter
relative to an efficiency factor that is proportional to
the flow mach number. They show that a critical flow
velocity mach number of 0.25 represents a transition between flow noise generation mechanisms leading to an
icreased efficiency and increased flow noise sound pressure level (SPL).
C. Exhaust Silencing
In order to moderate exhaust tail pipe noise there exists a variety of muffler designs that are commonly
employed. The performance of a silencer may be characterized by its insertion loss defined as the difference in
measured SPL with and without the muffler fitted; its transmission loss which is defined by the difference in
SPL at the inlet and outlet of the muffler; or its effect upon the brake mean effective pressure. An efficient
silencer is defined here by a design that achieves a relatively high ratio of attenuation achieved to reduction in
engine power output. Silencers may consist of a single or a combination of standard silencing components
which include reactive, absorptive and resonator types. These components vary in the manner and efficiency
6
Final Project Report 2011, SEIT, UNSW@ADFA

with which they enable the viscous dissipation of acoustic energy. Their unique action often makes them highly
effective attenuators in discrete frequency ranges or otherwise less effective over a more general range of
frequencies. As a result hybrid silencers aim to utilise a combination of such components in order to form an
effective broadband attenuator. A summary of the acoustic theory for these common silencer types is attached in
Annex A.
Blair quotes the work of Coates [15] who shows that the sound pressure level at any point in space beyond
the termination of an exhaust system to the atmosphere, is a direct function of the instantaneous mass flow rate
from the end of the exhaust pipe, the relative distance between source and microphone and the directivity of the
pipe end. The instantaneous mass flow rate was calculated using the Eq.(2) [4].
̇

̇

(

)(

) (

)

(2)

This expression states that the radiated noise is a function of gas temperature, the discharge coefficient of the
pipe end, the outlet diameter as well as pressure wave amplitude ratio travelling in the left and rightward
direction. As a result of this direct relationship with the mass flow rate Blair states that silencing is easily
achieved given an unlimited volume able to dampen the pressure and mass flow oscillations. However, when
subject to space restrictions the design of a silencer must conform to the following empirical design guidelines:
 A silencer should have a minimum silencer-cylinder volume ratio of ten.
 If this cannot be achieved the silencer must choke the exhaust system via a restrictive muffler in order
to sufficiently damp the mass flow rate for effective noise attenuation. (However increased back
pressure will result from increased restriction, therefore a silencer with minimal choke would represent
the most efficient attenuator).
Blair [4] uses this theory to conduct a study into the effectiveness of motorcycle silencers via experimental
and numerical methods. This study tests a plenum, absorption, diffusing and side-resonant type mufflers all with
a constant silencer-cylinder volume ratio of ten. Data shown in Fig.(9) and Fig.(10) illustrates that individually
these mufflers either offer excessive reductions in the BMEP of up to 30% whilst being unable to attenuate
noise sufficiently or offer negligible effect to power and noise.

Figure 9. Torque characteristics of muffler varieties
determined by Blair

A novel ‗two-box‘ hybrid silencer design seen in
Fig.(11), comprising an absorption and diffusing
silencer component, is then verfied to result in an
average reduction in BMEP of only 7% whilst notably
attenuating noise. This is seen to be a result of the
effectiveness with which the mass flow rate at the
outlet is reduced as seen in Fig.(11). Comparisons are
shown in Fig.(13) and Fig.(14), of achieved engine
performance and measured noise emission data for this
design as well as its individual constituent components.

Figure 10. Noise characteristics of muffler varieties
determined by Blair

Figure 11. Schematic of tow-box silencer

7
Final Project Report 2011, SEIT, UNSW@ADFA

Figure 12. Exhaust outlet mass flow rate for
silencer varieties

Figure 14. Noise attenuation performance of silencer
components

Figure 13. BMEP with silencer component

Figure 15. Noise spectra at 7500 rpm with two-box
silencer

Data illustrate the potential of the volume restricted two-box silencer design as an effective and efficient
attenuator of exhaust tail pipe noise. Noise spectra in Fig.(15) shows the attenuation achieved by the two-box
silencer as well as by individual absorption and diffusing silencer components. This demonstrates the highly
non-linear interaction between the absorption and resonant/diffusing components.Acoustic theory would suggest
that the effectiveness of this particular hybrid silencer represents a combination of the attenuation of the
diffusing silencer at low frequency and the attenuation of the absorption silencer at high frequency which is to a
limited extent demonstrated within Fig.(15). However, acoustic theory is experimentally shown by Blair to be
highly ineffective in accurately predicting the achieved attenuation from a silencing element. Data in Fig.(16)
and Fig.(17) compares the experimentally obtained attenuation with that predicted by acoustic theory which

Figure 16. Noise attenuation of plenum and
diffusing type silencers

Figure 17. Noise attenuation of side-resonant silencer

8
Final Project Report 2011, SEIT, UNSW@ADFA

indeed shows poor correlation. The inaccuracy of acoustic theory is understood to stem from its specific
relevance to waves of infinitesimal amplitude which differ fundamentally from wave phenomenon experienced
in exhaust flows being waves are of finite amplitude. Methods of accurately appreciating the true nature of
exhaust waves must therefore employ an appreciation of the instantaneous mass flow rate emenating from the
outlet which is proven by Coates [15] to an accurate approach.
The design of an efficient and effective hybrid silencer is understood to be a highly complex task that may
employ acoustic theories as merely the basis of an informed estimate for attenuation in order to commence a
design process characterised primarily by experimental trial and error. Luckily, one-dimensional gas dynamic
computational programs have the capability of providing accurate estimates of the time-varying mass flow rate
from the exhuast outlet and may therefore be used to supplement the experimental silencer design process.
D. Design and Modelling of Exhaust Systems
Literature was also consulted in order to determine the capability and implementation of current numerical
techniques for the design and modelling of exhaust systems. This review identified that as unsteady gas
dynamics within an exhaust system are predominantly one–dimensional in nature, a great deal of research,
design and optimization is carried out with one-dimensional engine simulation software. These analyses also
achieve excellent agreement with experimental data providing the exhaust geometries of concern are free of
excessive curvatures or complex silencing components characterized by strong three dimensional turbulence[4].
Additional advantages of employing this simplification include the ability to simultaneously calculate the effect
of duct and silencer geometry upon engine performance and noise spectra, as well as to do so across a variety of
designs within a reasonable timeframe. Consequently, one-dimensional codes have demonstrated an ability to
enhance the efficiency of the design process [16]. Enhanced accuracy is accomplished with the use 3D CFD
and coupled 1D/3D analyses however these generally have a far greater computational cost.
E.

Conclusion
This literature review summarises current theory and practices relating to exhaust system design,
performance analysis and optimisation. Specifically, the mechanisms by which an exhaust design contributes to
engine performance via scavenging were detailed. The nature of noise measurement, vehicle noise generation
and attenuation were described. Acoustic theory of common silencing elements was detailed, and the deficiency
of this theory in providing accurate predictions of silencer performance was addressed. The complexity of the
silencer design process was established and studies demonstrating successful design methods have been
described.
Importantly, the outcomes of this literature review enable a more informed requirement definition for a new
high performing exhaust system for the Yamaha WR450 engine. The aims of the exhaust design are summarised
as an ability to assist in engine scavenging at the desired engine speeds and also to incorporate an efficient
silencer design so as to minimise engine pumping work whilst achieving the specified noise target. To
demonstrate the implementation of the presented theories within an exhaust design, this project undertook to
varying degrees, the stages of concept development and preliminary design of the exhaust system.
Activities under the concept development stage stemmed from the manufacture of a ‗two-box‘ hybrid
silencer as well as the development of a WR450 engine simulation model using Ricardo WAVE. The conduct of
this stage consisted of an experimental investigation of achieved noise attenuation with parameter variations in
the manufactured design. Furthermore, the simultaneous conduct of this investigation with a physical engine as
well as within WAVE provided a means for comment upon the effectiveness of the design, as well as the quality
of the simulation model developed. This experimentation thereby performed both the roles of a parameter study
as well as an experimental validation of the developed WAVE model.
In addition, the concept development stage included the investigation of each of the scavenging mechanisms
detailed, using the WAVE model. (These investigations were also intended to encompass an experimental
component utilising an in-house developed engine testing rig, however this could not be accomplished due to
technical difficulties). This analysis considered all possible design parameters within a reasonable range such
that the ADFA FSAE team would be enabled to make an informed design selection as a part of an integrated
engine tuning strategy.
Finally, the preliminary design stage was conducted to provide the team with a refined design concept for a
high performance silencer based upon that which was manufactured. The proposed prototype design was
formulated using engine simulation results as well as data obtained from a silencer parameter analysis assisted
by Ricardo WAVE transmission loss work bench.
These processes thereby underpinned the aims of this project being to provide the ADFA FSAE team with
the means of designing, analysing and implementing a high performance exhaust system.

9
Final Project Report 2011, SEIT, UNSW@ADFA

III.

Part B –Concept Development and Investigation

A. WAVE Model Development
A Yamaha WR450 engine simulation model was developed to underpin the continued concept development
and analysis of the engine and associated componentry. A view of the WAVE model is provided in Fig.(18).
The model employed a host of user defined inputs available from tabulated data or otherwise from physical
measurements. Due to time and resource constraints an experimental validation could not be undertaken.
However, the validity of the model utilised within this study is supported by agreement found between the
generated engine power output prediction and data generated independently by the Cal Poly FSAE team who
managed to conduct an experimental validation of their WR450 WAVE engine model. Power curves obtained
from the WAVE model as well as by data from Cal Poly FSAE team are provided in Annex (B).

Figure 18. WR450 Ricardo WAVE model

B. Experimental Silencer Parameter Study
Having developed an engine simulation model of the Yamaha WR450, an orthogonal experiment was
designed to effect a parameter study of the ‗two-box‘ silencer design. The experiment would be implemented
within WAVE as well as upon an experimental WR250 engine using a manufactured experimental silencer. The
purpose of this experiment was to investigate the achieved attenuation of the muffler concept, in addition to the
variation in this attenuation as a function of parameter modification. Furthermore, this experiment will be used
to quantify the effectiveness of silencer design theories including that of acoustic theory and of Blair‘s mass
flow rate theory. Upon comparison of theoretical and experimental data, comment will then be made as to the
validity of the WAVE model (Only partial validation could be obtained as the WAVE model is a simulation of
the team‘s Yamaha WR450 whereas the physical experiment could only be conducted using a Yamaha WR250).
1.

Experimental Silencer Design
An experimental silencer was manufactured as per the CAD model in Fig.(19), based upon the ‗two-box‘
silencer design proposed by Blair.

Figure 19. CAD model of manufactured silencer showing initial discontinuity, absorption component and
expansion/resonator chamber

10
Final Project Report 2011, SEIT, UNSW@ADFA

In addition to the original concept, the manufactured design incorporates an initial discontinuity of area ratio
equal to six which is conformant to the findings of Blair. This was included to provide a decisive location that
may be used to define the tuned length as well as to decouple design parameters concerned with silencing from
performance aspects of the exhaust. (Blair‘s original design would instead tune from the end of the perforated
pipe. However, an industry SME advises that wave propagation through the perforated pipe enhances wave
degradation and therefore wave tuning effectiveness). This design was subject to some variation from the
original design as it was constrained by the availability of off-the-shelf (OTS) components which were preferred
in order to simplify the manufacturing process. The design may be fully disassembled as per Fig.(19), such that
parameters including the choke size, the resonator length and packing density could be varied in accordance
with the experimental intent. Table 1 shows a comparison of the non-dimensional parameters of the Blair design
and the current experimental design.
Table 1. Non-dimensional parameter of Blair silencer and mancufactured silencer

Inlet Diameter (D)
Major Diameter
Absorption Length
Resonator Length
Perforated Area
Silencer- Cylinder
Volume Ratio

Blair
46.6 mm

Experimental Design
51 mm

2.58D
8.58D
4.29D
19%
15

2.49D
9.0D
Variable up to 5.88D
25%
14.5 - 19.5

Comments
OTS component and recommended
by industry SME
OTS
OTS
Custom telescoping component
OTS

2.

Orthogonal Experiment Design
Acoustic theory predicts that the manufactured two-box silencer will achieve broadband attenuation as a
result of the combination of resonant effects of the expansion chamber and viscous dissipation of the absorption
silencer. However, as shown in experimental data obtained by Blair in Fig.(15), the operation of this silencer
does not explicitly conform to predictions underpinned by acoustic theory, nor does data show that attenuation
achieved is linear addition of the attenuation achieved by its constituent components. In contrast, Blair‘s mass
flow rate theory hypothesises that broadband attenuation achieved by this silencer is a direct consequence of the
manner with which it damps the magnitude of the mass flow rate from the exhaust outlet. Consequently, an
experiment was conducted to identify the true manner of operation of this silencer so as to provide the means for
the design of an efficient silencer for the ADFA FSAE team.
In order to attempt to validate acoustic theory, noise measurements recorded the noise spectra such that the
insertion loss of the muffler could be calculated. Hence, validation of acoustic theory could be obtained if this
data was to show agreement with transmission loss data calculated.
The experimental plan was based upon Taguchi Design of Experiment methods [30] for orthogonal
experiments. This experiment was then implemented upon a WR250 engine as well as implemented within
WAVE. Unfortunately, as the WAVE model has been developed to simulate a Yamaha WR450 engine and the
physical experiments were conducted upon a WR250 engine the comparison of these results were not able to
provide conclusive validation of the developed engine model. Instead these two sources of data were simply
used to comment on the nature of operation of the muffler as well as to confirm similarity of trending.
The chosen independent variables include the silencer parameters of resonator length, choke diameter and
packing density. Using the Taguchi L4 orthogonal array the experiment seen in Table 2 was formulated.
Table 2. Orthogonal experimental plan

Experiment
1
2
3
4

Choke Diameter (mm)
20
20
30
30

Resonator Length (mm)
100
300
100
300

3.

Packing Density (g/L)
200
100
100
200

Parameter Study Results and Discussion
Noise measurements taken at position A and position B for experimental silencer variations as well as for the
standard WR450 muffler is shown in Fig.(C1) and Fig.(C2) in Annex C. WAVE data is also provided in
Fig.(C3) which shows the predicted sound pressure at position A for each of the experimental test silencers. As
expected, comparison with experimental data at position A does not show agreement of sound pressure
magnitude due the difference in engine displacement. However, good agreement is found as to the relative
variation between designs. Measurements obtained of SPL with frequency are provided in Fig.(C4) and
11
Final Project Report 2011, SEIT, UNSW@ADFA

Fig.(C5). This data shows variation in emitted noise from the unsilenced pipe with engine speed as well as the
broadband attenuation achieved by the manufactured silencer which is seen to be equivalent to a standard
WR450 muffler. Calculated insertion loss is provided in Fig.(C6), Fig.(C7) and Fig.(C8). Data provided is
limited to a frequency of 1000 Hz as consistently high levels of attenuation are achieved for all designs at
frequencies beyond this point. Transmission loss for the each of the silencer configurations was also calculated
using WAVE for comparison, and is seen in Fig.(C9) and Fig.(C10).
The data obtained is generally supportive of the theory of silencer design concerned with the management of
the mass flow rate from the exhaust outlet as opposed to acoustic theory. For instance, within Fig.(C1), Fig.(C2)
and Fig.(C3) the most choked designs 1 & 3 record significantly lower sound pressure level recordings
compared to the less choked designs. Furthermore, both experimental and WAVE data agree that those designs
with a larger volume will attenuate noise to a greater extent. Plots in Fig.(C11), Fig.(C12), Fig.(C13) and
Fig.(C14) of outlet mass flow rate, generated with WAVE, illustrate the silencing action of the muffler
variations. These plots show data for all tested engine speeds. Prominent features include the higher peaks
recorded for the less choked designs as well as the higher steady mass flow rate recorded for the more choked
designs. This steady flow rate leading up to the peak is much more constant for choked designs, which increases
in magnitude at high engine speeds in comparison to the relatively less choked designs. This behaviour suggests
that designs utilising a 20mm orifice are likely to become aerodynamically choked leading to a rapid increase in
back pressure, but also attests to the effectiveness of a choke for the purpose of exhaust tail pipe silencing under
Blair‘s theory concerned with the outlet mass flow rate. Furthermore, fluctuations within this mass flow rate are
seen to be damped by silencers employing a larger expansion chamber volume.
However, with reference to Fig.(C1), Fig.(C2) and Fig.(C3) it could also be argued that higher attenuation is
achieved by those designs with larger expansion chambers due to an increased ability to attenuate low frequency
noise as per acoustic theory for an expansion chamber. This low frequency noise is recorded in Fig.(C4) and
Fig(F5) as a source that is relatively constant as well as relatively elevated in comparison to other regions of the
noise spectra emanating from the unsilenced pipe. To investigate this possibility, the transmission loss for each
of the silencer configurations was attained from WAVE and is shown in Fig.(C9) and Fig.(C10). (This was
conducted within the WAVE transmission loss workbench which employs the well documented two source
method to compare sound power at the inlet and outlet of the silencer). As seen in Fig.(C10), the predicted
transmission loss below 500Hz for both designs 1 and 4 is seen to be consistently up to 5dB greater than for
designs 2 and 3 which employ a smaller chamber. However, it is questionable that this extra achieved
attenuation could be the main reason for these larger designs consistently out performing corresponding designs
with an equal choke diameter and smaller chambers. This doubt is particularly pertinent as transmission loss
data predicts generally lower attenuation achieved by larger designs 1 and 4 at higher frequencies, yet
experimental data states that these larger designs record lower SPL even at high engine speeds where high
frequency flow noise becomes more dominant. Furthermore, calculated insertion loss data does not show any
significant agreement with theoretical attenuation for the silencer represented by the transmission loss data.
Subsequently, the achieved attenuation of a silencer in practice is seen to be more dependent upon the manner in
which the design manages the exhaust outlet mass flow rate to atmosphere than the attenuation predicted by
acoustic theory. Results leading to this conclusion are in line with published theory of Blair described
previously.
The value of acoustic theory is not, however, totally diminished as some agreement is found between
transmission loss data and calculated insertion loss data by way of comparative performance. Discrepancies
between these sources may also be exaggerated by inaccurate assumptions and experimental error. This would
include the assumption of nil mean flow during the transmission loss analysis and aliasing within experimental
measurements. Furthermore, the consistently high attenuation achieved at high frequency by the manufactured
design agrees with the acoustic theory of absorption silencer. (The effectiveness of an absorption silencer was
also experimentally determined by Blair and is shown in Fig.(17)). Therefore whilst acoustic theory and its
implementation may not take into account all non-idealities it may provide a good initial estimate of silencer
performance.
Further comment can be made as to the effectiveness of the manufactured design with reference to Fig.(C4)
and Fig.(C5). These show measured sound pressure with frequency at position B for engine speeds of 3000 and
7000rpm. These plots show that the test silencer offers significant attenuation averaged between 20dB and 30dB
which is also recorded for the standard WR450 muffler. The WR muffler has a silencer-cylinder volume ratio of
5 in its intended role upon a WR450 and a ratio of 7.5 for the test engine and as a result it employs a 10mm
choke in order to meet road authority noise regulations. The trade-off between silencer volume and choke is
made clear as experimental data concludes that the designed silencer of a volume ratio of 4 to 5 times greater
than the standard WR muffler, yet far less choked is able to achieve equivalent attenuation. This data thereby
emphasises the importance of exploiting available vehicle space for a silencer in order to minimise the level of
choke required and therefore minimise power losses. A comparison of these figures also illustrates the
12
Final Project Report 2011, SEIT, UNSW@ADFA

increasing prevalence of high frequency flow noise at high engine speed, a trend which is also shown by Honda
et al in Fig.(7).
A comparison of all SPL spectra data obtained experimentally is compared with corresponding data attained
from the WAVE model in Fig.(C15), Fig.(C16), Fig.(C17) and Fig.(C18). Obvious discrepancy should be
expected as the data sources are generated by different engines. Despite this, however, fair agreement is found
highlighting the capability of the developed WAVE engine model.
Finally, noise measurement data obtained from position C is given in Fig.(C19) and Fig.(C20). This data
shows no clear indication of any significant resonances that are absent from unsilenced data.
4.

Conclusion
The base silencer design as proposed by Blair is here shown to have high potential as an effective silencer.
With further design refinement and testing, the efficiency of this concept may also be appreciated. The
expansion chamber has shown enhanced sensitivity to the non-idealities present within an exhaust silencing
application. As a result, more significant correlation is found between the volume of this component than with
its length as per acoustic theory. However, acoustic theory was demonstrated to generate predictions of
absorption silencer performance with relatively high accuracy.
This experiment established the priority for silencer design as to control the mass flow rate from the exhaust
outlet. Acoustic theory was determined to have limited effectiveness in predicting silencer performance. The
usefulness of acoustic theory within silencer concept development is recognised.
Fair agreement was found with WAVE data obtained despite variation in engine displacement used to
generate the data sets. This agreement was represented mainly by similar trending. As per theory detailed in the
literature review, the value of a one-dimensional software for concept development is shown to be founded in its
ability to quickly describe unsteady gas flow throughout an engine.
C. WAVE Investigation – Exhaust Design Parameters for Engine Scavenging
A literature review identified that exhaust performance considerations are resultant of the nature of the gas
exchange process in a four stroke engine. The following investigation was undertaken using WAVE to
demonstrate the potential of the identified exhaust tuning strategies. This included the variation in ‗wave tuning‘
and ‗inertial scavenging‘ with exhaust pipe geometry. Specifically, this study was concerned with identifying
the extent of variation in engine performance possibly accomplished via the design of an exhaust pipe for a
single cylinder engine assuming a constant intake length and valve timing. The purpose of this investigation is
therefore to inform the ADFA FSAE team of the methods commonly incorporated within the design of an
exhaust system, that aim to enhance or merely shape an engine‘s performance characteristic. The following
findings should therefore act as a tool to be used in conjunction with many other powertrain parameters to
obtain a desired engine performance target.
1.

Exhaust Wave Tuning
As stated by Professor Blair of the University of Belfast ―the tuned exhaust pipe harnesses the pressure wave
motion of the exhaust process to extract a greater mass of exhaust gas from the cylinder during the exhaust
stroke and initiate the induction process during the valve overlap period.‖ This scavenging effect is possible if a
pressure wave originating from the exhaust valve travels over a tuned length such that it is reflected back to the
valve face as a rarefaction wave in time to assist the gas exchange process during valve overlap. The coincident
phasing of valve overlap and the arrival of pressure waves, seen in Fig.(D1), is dependent upon the length over
which the waves travel known as the ‗tuned length‘. Resultant valve mass flows and pressure differentials are
provided in Fig.(D2) and Fig.(D3).
As per Fig.(D4), Fig.(D5) and Fig.(D6), this scavenging effect is characteristic of a certain engine speed
where the correct phasing occurs. These figures shows the variation of residual gas fraction with tuned length
and the resulting effect upon torque and power output of the engine as a result of an increased delivery of
combustibles. Here relatively small variation in residual gas fraction is seen to have a marked effect upon
torque. Furthermore, when this effect is achieved at high engine speeds a highly significant influence is
exercised over the shape of the power curve. This data therefore demonstrates the significance of the exhaust
wave tuning and resultant scavenging effect achieved. Finally, Fig.(D7) is provided to demonstrate the possible
effect of interaction between intake and exhaust tuning. As seen, the scavenging ratio (a measure of the quantity
of fresh combustible mixture ingested to the engine) is greater than unity at 3000rpm for a 1450 mm tuned
length. This suggests that at this point wave tuning is interacting with other tuning effects to achieve a greater
scavenge than could be achieved alone. Such interactions are also important when defining the design target for
exhaust tuned length such that these interactions are used to their full potential.
To summarise, a contour plot of residual gas fraction with tuned length and engine speed is provide provided
in Fig.(D8). The region representing the most effective scavenging effect is seen in blue. An inverse relationship
is seen to exist between exhaust length and tuned engine speed. Since pressure waves travel within the exhaust
13
Final Project Report 2011, SEIT, UNSW@ADFA

system at the acoustic speed which is a function of temperature, the hyperbolic nature of this relationship is
therefore present due to the asymptotic nature of heat transfer from the exhaust pipe.
Stepped Pipe Tuning
As a part of an investigation into exhaust
wave tuning techniques, the industry practice
of utilising stepped pipes was considered. A
schematic of a stepped exhaust pipe is shown
in Fig.(20). A stepped pipe offers extra degrees
of freedom in wave tuning practices as there
exists more discontinuities able to create
rarefaction wave reflections. Furthermore this
characteristic can also lead to varied heat
transfer properties. To illustrate, Fig.(D9) is
provided. This plot is the product of a 1500mm Figure 20. Simplified schematic of stepped pipe
pipe with a fixed expansion at 500mm and
another expansion whose location in the pipe varies between 510mm to 1490 mm. Wave action from the pipe
end at a tuned length of 1500mm is tuning at 4000 rpm and secondly at 7500rpm shown by two regions of
relatively low residual gas fraction. In addition, wave action from the first expansion occurs at 500mm
enhancing the region of low residual at 7500 rpm. Of note, data shows a noticeable change within each of these
regions as a result of the location of the second expansion. To illustrate Fig.(D10) is provided which shows
variation in exhaust gas temperature (in blue) as well as acoustic velocity (in green) with position for two
stepped pipes with location of the steps indicated. Fig.(D10) specifies that these shifts in the tuning behaviour of
the pipe can be attributed to the effect stepping behaviour of the pipe has on heat transfer, average gas
temperature and the average acoustic speed which are indicated to vary slightly. In addition to this effect,
Fig.(D11) and Fig.(D12) show that for these same two stepped pipe designs, the intermediate expansions are in
fact also able to reflect rarefaction waves for the purpose of scavenging, despite being of a lesser magnitude.
Specifically, these figures show that for two different stepped pipes (parameters indicated in plot caption), a
variation in the arrival of the first smaller wave is recorded, giving rise to corresponding change in the recorded
valve mass flow rate purely as a consequence of the unique positioning of the intermediate discontinuity.
Diffusers
The most effective exhaust component design for wave
scavenging is that of the diffuser as seen in the schematic
provided in Fig.(21). This type of exhaust component is
known to be able to tune over a wider range of engine
speeds offering superior scavenging and engine
performance. For comparison however, a generally
accepted rule-of-thumb states that a third of the length of
the diffuser is used in the calculation of the total effective
tuned length.
To demonstrate the action of the diffuser, Fig.(D13)
shows residual quantity achieved for a diffuser 400mm Figure 21. Simplified schematic of diffuser
long and with a taper angle of 6.34 degrees attached to a
variable length of pipe. What is instantly noticeable is the greater dominance of the scavenging region compared
to that of the straight pipe. Again Fig.(D14), shows residual quantity for a larger diffuser of 600mm in length
with a taper angle of 6.65 deg. Once more this shows a vastly greater scavenging ability than that shown by a
single straight pipe. Finally, Fig.(D15), is showing residual gas fraction for a diffuser larger still, of 900mm in
length with a taper angle of 6.65 degrees, however no significant benefit is seen to be gained by the extra length.
Through the course of this study a diffuser was manufactured in order to obtain experimental validation of
this theory and to promote the diffuser as an innovative technique to enhance overall engine power. However, as
a result of the continuing inoperability of the team‘s engine testing rig this validation could not be undertaken.
2.

Inertial Scavenging
Inertial scavenging describes a scavenging effect that is enabled by the inertia of a high velocity column of
exhaust gas escaping from the cylinder. As seen in the simplified schematic in Fig.(22), the interaction between
this column of gas and the gas exhange process during valve overlap may see a build up of pressure energy at
the exhaust valve that acts to assist in engine breathing. This mechanism functions under the principle that a
fixed volume flow rate is achieved at a certain engine speed and for a fixed volume flow rate through a pipe, gas
14
Final Project Report 2011, SEIT, UNSW@ADFA

velocity varies inversely with pipe diameter. Consequently,
there then exists a pipe diameter where the scavenging
effect produces a more than proportionate amount of power
than pumping work required to achieve an effective gas
velocity. Under these circumstances an increase in engine
power will be realised.
To demonstrate Fig.(D16) is provided, which depicts the
variation of scavenging effect measured in residual gas
fraction with differing pipe diameters for a constant tuned Figure 22. Simplified schematic of inertial
scavenging
length of 1200mm and for a silenced exhaust system. A
constant diameter 35 mm pipe is seen to scavenge well between 4000 and 5000 rpm. The addition of a 42mm
step sees a shift in this scavenging effect to 5500rpm and similarly if this step is increased to 48mm in diameter,
a further increase in the scavenged engine speed is observed. A pipe consisting of multiple steps and therefore
including pipe diameters of 35mm, 42mm and 48mm achieves a scavenging effect which almost represents the
addition of scavenging effects for both single step pipes. A consideration of the corrresponding pumping torque
data in Fig.(D17), may be used to explain the transfer of energy from the piston, to pressure energy within the
exhaust gas and then to its kinetic energy which eventually promises a scavenging effect capable of making
proportionately more power than that lost in pumping. A comparison of trending in residual gas fraction and
pumping torque for the constant 35mm pipe shows that within the range of engine speed of 4000rpm to
5000rpm, where most significant scavenging is achieved, pumping torque data is recorded as the greatest of all
data sets. This increase in pumping torque acknowledges the transfer of energy from the piston to exhaust gas
which in turn assists in cylinder scavenge. A similar phenomenon is observed for all other data sets such that a
relative increase in pumping torque corresponds to the range of engine speeds where inertial scavenging is
achieved. Valve mass flow rate data provided in Fig.(D18) and Fig.(D19), illuminates this theory further. For
data at 4000rpm, the highest mass flow rates are attributed to the 35mm pipe. However, for data at 7500 rpm the
stepped pipes are seen to record the highest valve mass flow rates. Finally, the subsequent torque and power
curves are provided in Fig.(D20) and Fig.(D21).
Contours of residual gas fraction for exhaust tuned length and pipe diameter were generated to explore this
concept further. Three plots are provided in Fig.(D22), Fig.(D23) and Fig.(D24), one each for the engine speeds
of 6000rpm, 7000rpm and 8000rpm. Data is summarised in Table (3).
Table 3. Pipe parameters for optimal scavenging

Pipe Diameter (mm)
Tuned Length (mm)

6000rpm
724
31

7000rpm
559
31

8000rpm
586
34

These plots draw attention to the minimum residaul gas fraction which is achieved with a specific
combination of pipe length and diameter. Comparison of data at 6000rpm and 7000rpm suggests that wave
tuning is the dominant mechanism at these engine speeds. This can be concluded as a constant diameter has
been maintained, suggesting no significant variation in inertial scavenging achieved between these two engine
speeds. Furthermore, a longer pipe is seen to contribute to scavenging at lower engine speed which is in line
with previously discussed wave tuning concepts. However, a comparison of data at 7000rpm and 8000rpm
shows that inertial scavenging is seen to gain significane again due to devaition from this logic. Now at
8000rpm, peak scavenging is achieved with a longer tuned length than at 7000rpm, which is contrary to wave
tuning trends, as well as a larger diameter. This larger diameter can then be understood to underpin a required
flow velocity such that inertial scavenging is maximised.
The reciprocal exchange of energy from exhaust gas inertia to pressure energy which works to assist in
pumping is seen in Fig.(D25) which shows variation in pumping torque with the length of constant 35mm
diameter pipe. Here, at low engine speeds longer pipes offering back pressure record higher relative pumping
losses. However, at higher engine speeds the maintenance of flow enery within longer pipes works to assist the
piston during the exhaust stroke leading to a reduction in pumping losses relative to shorter pipes.

IV.

Part C – Preliminary Design and Design Proposal

A. Design Requirements
As per the design of any engineering product, the design of a performance exhaust system must be
conducted relative to specified requirements. Therefore before implementing knowledge gathered as a part of
the concept development stage, customer requirements were explicitly stated. For a performance exhaust for the
WR450 engine, they are stated in order:
15
Final Project Report 2011, SEIT, UNSW@ADFA

1.
2.

3.
4.
5.

Adequate insertion loss—in accordance with FSAE regulations the noise measurement taken near the
exhaust must not exceed 110dB;
Back pressure minimal—to maximize vehicle competitiveness throughout the competition the
implemented silencer should introduce minimal engine power losses by way of back pressure and be
integrated within an exhaust system that provides a desired torque and power output characteristic.
Size—the silencer must be able to be easily integrated within the vehicle;
Cost—low cost desirable;
Durability—high durability desirable.

B. Silencing Strategy
In accordance with findings of the literature review, the formulation of a preliminary design proposal was to
be carried out relative to the identified ‗noise problem‘. In this case ‗noise problem‘ was established from an
estimate of the SPL spectra, measured under FSAE conditions, emanating from a 1000mm stepped pipe, which
was obtained from WAVE. This data is provided in Fig.(E1) in Annex I. It shows the predicted sound pressure
relative to a ceiling of 108dB. This target was calculated in accordance with theory detailed in Annex B. This
calculation recognises that the intake is the second most dominant source of noise upon a vehicle, as well as that
any other sources of noise varying from the maximum of more than 15dB offers a negligible addition to the total
SPL measurement. A pessimistic estimation of the intake noise is taken as 100dB, and therefore with the
addition of a 108dB exhaust noise contribution, a total measurement of 109dB would be achieved which is in
accordance with noise level design requirement.
Predicted noise spectra for the WR450 at its test engine speed of 7000rpm shows that the noise measured
from the outlet of the stepped pipe constitutes a number of significant contributions from a range of frequencies
corresponding to flow noise as well as to the engine firing rate. Therefore the proposed silencer is required to
offer broadband attenuation of up to 20 dB to satisfy noise the level design requirement.
C. Design Investigation and Definition
The definition of a final design proposal comprised a process of systematic analysis and selection, of silencer
components. Each of the significant silencer components including the absorption silencer, expansion chamber
and the choke were investigated individually such that the final silencer assemblage would represent the option
best able to satisfy the design requirements.
In line with findings of experiments conducted, the analysis of these individual components was conducted
relative to their governing principles. Since good correlation was found between acoustic theory and achieved
attenuation for an absorption silencer, the analysis of this component was based upon the predicted acoustic
transmission loss, which ignores the non-idealities of an exhaust silencing application. In contrast, the analysis
of the choke and expansion chamber volume was conducted upon the developed WAVE engine model such that
variation in outlet mass flow rate and consequent attenuation could be appreciated.
In accordance with findings of the WAVE enabled investigation into exhaust scavenging mechanisms, the
final silencer design proposal will be implemented upon a 1000mm stepped pipe as this design offers the highest
level of inertial scavenging for a fixed tuned length. Tuning of this integrated system will then be undertaken in
order to demonstrate the process of exhaust tuning relative to a performance target. In this case, the performance
target will be the unsilenced performance trend such that the attainment of this target will help demonstrate the
efficiency of the silencer design proposed by way of minimal back pressure.
1.

Absorption Silencer Component
Acoustic theory of an absorption silencer was seen to hold true during experiments. Therefore an analysis of
transmission loss is recognised to represent a reasonable prediction of achieved performance if not merely
relative performance. Fig.(E2) shows variation in attenuation with increasing diameter assuming the current
51mm diameter perforated pipe is used and holding the length of the component and packing density as
constant. Attenuation is seen to increase linearly with diameter. This result agrees with acoustic theory such that
the increased depth of sound absorbing material increases the viscous dissipation of sound energy via interaction
with particle oscillations. Fig.(E3) shows variation in attenuation with length of the absorption silencer
component. Attenuation is seen to asymptote such that large increases in length are required for only a fractional
increase in attenuation. Acknowledging that the absorption silencer component will constitute the majority of
the weight of the overall silencer, the attained data was analysed in terms of attenuation achieved per unit mass.
Fig.(E4) shows that for a specified increase in attenuation, an increase in diameter represents a more efficient
means than increasing the length. As a result, the proposed silencer design should incorporate as large a
diameter as possible that still remains conformant to the size constraints specified in the design requirements.
Fig.(E5) shows variation in attenuation with sound absorbing material density. This plot shows the
convergence of attenuation to an asymptote. Consequently, this data suggests that a density greater than 150g/L
offers a negligible increase in attenuation for the extra weight. Finally, Fig.(E6) shows variation in attenuation
16
Final Project Report 2011, SEIT, UNSW@ADFA

with perforated area of the inner pipe used. Negligible variation is illustrated therefore the selection of
perforated tube used will be constrained the requirement for durability of the silencer, as too high a perforated
area will allow violent exhaust gas flow to degrade or remove sound absorbing material.
2.

Choke Diameter
The choke was proven to practice significant control over emitted noise within experiments conducted which
is demonstrated further in Fig.(E7). This shows the ability of the choke to scale the emitted noise levels via
practicing direct control over the outlet mass flow rate. Fig.(E8) is also provided in order to emphasise the
silencing action of the choke. This data is generated within the WAVE Transmission Loss Workbench, which
conducts a comparison of sound power at the inlet and outlet of a silencer under nil mean flow conditions. As a
result, the choking of an expansion chamber element is seen to have minimal effect in the absence of mean flow.
In accordance with the design requirement for engine power losses, Fig.(E9) and Fig.(E10) are provided.
These plots show that for a choke diameter greater than 26mm a minimal effect upon engine scavenging,
measured in total residual quantity, and brake torque is predicted. Meanwhile a choke of 26mm also achieves an
increase in overall attenuation of up to 5dB making this a highly efficient component for silencing.
3.

Expansion Chamber
Having acknowledged the trends in attenuation achieved via absorpion parameters as well as choke, the
volume of the expansion chamber was varied to gain a similar appreciation. Variation in predicted outlet mass
flow rate with expansion chamber length with fixed diameter is provided in Fig.(E11). The maximum mass flow
rate recorded is seen to decrease consistently until a length of 150 mm is reached. A negligible change in peak
mass flow rate is found beyond this length and instead a phase shift is noted. Similar behaviour is seen in
Fig.(E12) which illustrates the resultant SPL measurement taken under SAE noise test conditions. A consistent
reduction in SPL is recorded up to a length of 150mm at which point the negligible change in the magnitude of
the mass flow rate results in no further reduction in SPL.
B. Design Proposal
The conducted parameter study was used to inform the formulation of a final design concept. Data generated
justified the selection of silencer parameters such that concepts could be verified using the developed engine
model. So as to minimise size and weight of the silencer, a conservative choke diameter of 30mm was selected
to exercise meaningful control over the outlet mass flow rate without deliberately increasing engine pumping
losses. The diameter of the silencer was identified to represent the most significant factor per unit mass, for
increasing absorption attenuation and silencer volume. As a result, in order to minimise weight of the silencer
and with a consideration of space constraints relevant to the current vehicle‘s side pod arrangement, a diameter
of 175mm was selected. Again to minimise weight, a minimal length was sought for the absorption silencer
component. Data suggested that lengths beyond 300 mm were subject to diminishing returns in terms of
attenuation and so this was selected as the final absorption length. By doing so the requirement for extra
chamber volume would also be minimised. The packing density was selected to be 150g/L as acquired data
suggested diminishing returns beyond this value. Finally, expansion chamber length was increased until a
satisfactorily low outlet mass flow rate was obtained giving a prediction of under 108 dB as per design goal.
The ability of WAVE to accurately calculate the instantaneous mass flow rate from the exhaust outlet, being
the definitive measure of silencer performance, underpins confidence within this design proposal. The
characteristics of the proposed silencer design are provided in Table (4).
Table 4. Design Proposal

Component
Resonator Chamber

Parameters
Length: 150 mm
Diameter: 175 mm

Attenuation Characteristics
Dampen outlet mass flow rate thereby
providing broadband attenuation.

Absorption
Component

Length: 300 mm
Diameter: 175mm
Packing Density: 150 g/L
Diameter: 30 mm

Target high frequency flow noise

Choke
Silencer-Cylinder
Volume Ratio

Scale outlet mass flow rate thereby
providing broadband attenuation.

24

17
Final Project Report 2011, SEIT, UNSW@ADFA

Predicted SPL measurement for the proposed silencer configuration is provided in Fig.(I13). Unsilenced
SPL data is also provided on this figure to demonstrate the broadband attenuation characteristics of the design.
Data sets are also seen to diverge slightly at higher engine speeds where the increasing dominance of flow noise
is experienced. This deviation attests to the effectiveness of the absorption silencer component at attenuating
this high frequency flow noise. Performance variation as a result of the implementation of the proposed silencer
design is shown in Fig.(I14). As per design requirements, the efficiency of the proposed silencer design is
defined as its ability to achieve a required level of attenuation with minimal effect upon performance. As such,
performance with and without the silencer fitted to a 1000mm stepped pipe is given. The addition of the silencer
was seen to cause a significant change in the mean exhaust temperature over the tuned length of the exhaust
system leading to a reduction in the effective tuned length of the system. The resultant change in wave tuning is
illustrated in Fig.(I15) which shows a time plot of the inward travelling waves at the exhaust valve. The increase
in exhaust temperature causes the inward wave to arrive earlier than the wave generated by an unsilenced
exhaust system causing a reduction in scavenging effectiveness at this engine speed. In order to investigate a
strategy to retrieve peak torque and power, contours provided in Fig.(I16), Fig.(I17) and Fig.(I18) of residual
quantity, brake torque and power with exhaust length in addition to a 500mm header pipe, were generated.
These plots show that an increase of exhaust length by 200mm is able to achieve similar performance obtained
from the unsilenced system without becoming subject to unsteady behaviour seen to exist for longer lengths.
Therefore with an extension of the exhaust length to 1200mm, Figs.(I14), Fig.(I19) and Fig.(I20) show that a
negligible reduction in overall power and torque is achieved for the silenced engine. This data also shows a
number of other performance variations resultant of the addition of the silencer. As per Fig.(I15), whilst the
extended exhaust pipe has attained the correct phasing of pressure waves with valve overlap, the intensity of this
wave is seen to be degraded by the extra distance of pipe travelled. The other significant consequence of silencer
addition, has been a marked increase in scavenging and torque at high engine speeds. With reference to
Fig.(I19), Fig.(I20) and Fig.(I21), a large amount of inertial scavenging has been achieved beyond 7000rpm
leading to a significant reduction in pumping losses as well as residual gas fraction.
In conclusion, the exhaust system design process has demonstrated the effectiveness and efficiency of the
proposed silencer as an integrated component within a performance exhaust system. Specifically, the proposed
silencer design is demonstrated to satisfy design requirements in terms of noise attenuation targets and back
pressure. Furthermore this process has demonstrated the tuning effect of silencer addition to an exhaust system
and described a strategy to attain a required engine performance characteristic of the silenced exhaust system.
By no means does this design process identify this particular tuning strategy as the best but instead merely sets
out to demonstrate the implementation of exhaust tuning theories discussed.
D. Vehicle Integration of Exhaust System
The process of vehicle integration of an exhaust system provides numerous constraints to the design
parameters of an exhaust system. In particular, due to limited number of mounting positions of a silencer, tuned
length could be understood to be constrained to discrete values. However, this would merely require the
innovative combination of exhaust techniques discussed such that a desired tuning effect is achieved. For
example, if silencer integration dictates that the exhaust length must differ from that which directly offers the
desired wave tuning effect a number of alternate strategies are available. These may include:
 the manipulation of pipe diameter so as to target this performance characteristic with inertial
scavenging,
 the use of exhaust pipe insulating wraps or coatings to tailor the mean exhaust gas temperature and
therefore the effective tuned length, or
 the implementation of an appropriately sized discontinuity to define the tuned length prior to the
silencer,
 the use of stepped pipes to target a performance characteristic with a combination of partial wave
reflections and inertial scavenging.
Whilst considering concerns of exhaust integration, it is important to explicitly state that by minimising pipe
bends within the exhaust will act to minimise overall back pressure and engine pumping losses. Consequently,
supplementary reports provided in Appendix A and Appendix B, demonstrates the performance effect of pipe
bends on heat transfer and engine performance as well as the optimisation of pipe geometry for pressure loss
using ANSYS Fluent and the optimisation routine NSGA2.
Finally, it is important to consider the achieved exhaust flow Mach number during the selection of exhaust
pipe parameters. This is in accordance with findings of Wiemeler, Jauer and Brand [14] who direct correlation
between Mach number and flow noise generation efficiency.

18
Final Project Report 2011, SEIT, UNSW@ADFA

V.

Limitations of Ricardo WAVE

As a one-dimensional software WAVE is limited in the accuracy with which it can represent any part of an
engine which inherently consists of three dimensional flow behaviour to some extent. As a result spurious data
can sometimes be obtained from simulations conducted, seen within some of the provided plots as seemingly
random spikes. However, the majority of data is conformant to a definite trend which is for the most part where
the value in this software is derived.
Of particular note, by utilising a one-dimensional simplification the model relies highly upon the quality of
its inputs rather than directly resolving aspects of engine operation. As a result, there exists a central
requirement for the validation of the model whilst it is under development. For example, a flow noise efficiency
factor is user defined input as a part of the process of acoustic acquisitions within the post processing program
WAVE-post. Standard values for this input, as with many inputs, were used which may need to be verified
within the process of model development.

VI.

Conclusion

The aim of this work is to underpin all future development by the ADFA FSAE team, of a high performance
exhaust system for a single cylinder engine. In accordance with findings presented, the presented silencer design
will be able to provide the required noise management capability without prejudice to engine performance.
Performance exhaust tuning techniques have been discussed and demonstrated such that this design should be
able assist in shaping the torque and power output characteristic of the engine for the benefit of vehicle
competitiveness. Moreover, the importance of conducting of exhaust tuning as a part of an integrated engine
performance tuning process has been identified. The presented design and analysis methodology has provided a
meaningful demonstration of the silencer operation and design. Finally this demonstration has culminated in the
proposal of an efficient silencer design.

VII.

Recommendations and Future Work

The complexity of operation and analysis of the exhaust provides a wide scope. This research has attempted
to provide the basic foundations of exhaust and silencer design and analysis however in doing so, depth of
research has been sacrificed for breadth. As a result continued work should hope to explore more specialised
techniques of exhaust design and analysis to extend upon the basic concepts presented herein. Some topics of
interest would stem from tuning interactions assumed constant within this study. For instance, the simultaneous
tuning of the exhaust and intake as well as valve timing is documented as a significant method of engine
performance optimisation. In addition, tuning methods identified herein suggests potential for innovative
integrated designs that combine the use of inertial scavenging and wave tuning in a synergistic manner that may
be worthy of investigation. For a single cylinder engine there is limited further work that could be undertaken in
the way of exhaust tuning via pipe design. However a hypothesis was formed during the course of this study that
specialised exhaust components such as Helmholtz chambers have been used within industry to not only provide
a means of targeted silencing but also to enhance wave tuning effects via wave interaction. The implementation
of such a component in this way would be expected to enhance performance as well as to justify a lighter
silencer and therefore it seems worthy to recommend an investigation into the feasibility of the idea.
Furthermore, commercial products such as those in
Fig.(39) incorporate components that are unexplained by
this study but may offer extra performance benefit and
therefore may represent another opportunity for further
work.
It was found during the course of this research, that a
range of studies into silencer design used other forms of
silencer concepts as the basis of a design optimisation. In
particular a text by Munjal entitled ‗Acoustics of ducts
and mufflers with application to exhaust and ventilation
system design‘ was used by a number of studies who
were concerned with the implementation of reactive
silencers. Ideally, future work conducted by the ADFA
FSAE team would be able to identify whether a reactive
silencer concept would be able to surpass the current Figure39. Commercial exhaust system with novel
proposed design in terms of attenuation, back pressure components labelled ‘Powerbomb’ and ‘Megabomb’
and weight.
Finally, much literature is available as to the implementation of coupled 1D/3D analyses within this topic
area. WAVE openly admits to enhanced inaccuracy when dealing with complex components and the
development of this capability within any aspect of exhaust design would represent a powerful design tool.
19
Final Project Report 2011, SEIT, UNSW@ADFA

Acknowledgements
The author would like to gratefully thank a variety of important individuals that helped throughout the course of
this study. Thanks goes to thesis supervisor, Dr Warren Smith for providing much needed guidance during the
course of what always seemed to be a grossly under defined problem. To Mr Alan Fien, for your willingness to
offer your vast technical insight. To Mrs Marion Burgess for your patience and understanding despite the
tribulations of this project. Much thanks goes to SEIT workshop staff particularly Doug Collier and Marcos De
Almeida for their assistance throughout the design and manufacturing process. Thanks to members of the FSAE
team and fellow engineers whose support was invaluable and who at times managed to make this project an
enjoyable process. Finally, to my girlfriend who showed amazing patience over the course of a very long year of
work as well as offering much needed support over the course of this degree. Thanks to my family who are well
deserving of official recognition of all their support over the many years.

20
Final Project Report 2011, SEIT, UNSW@ADFA

References

[1]

SAE, "FSAE Inspection Sheet," ed: SAE, 2011.

[2]

SAE-Australiasia. (2011, 07 May). Competition Overview [Internet Web Page]. Available:
http://www.saea.com.au/formula-sae-a/competition-overview

[3]

G. P. Blair, "Design and Simulation of Engines: A Centruy of Progress," SAE International1999.

[4]

G. P. Blair. (1999). Design and Simulationof Four-Stroke Engines.

[5]

J. Robinson. (1994). Motorcycle tuning

[6]

Smith and Morrison, Scientific Design of Exhaust & Intake Systems, 2009.

[7]

A. G. Bell. (2001). Four Stroke Performance Tuning.

[8]

D. Winterbone and. R. Pearson, Design Techniques for Engine Manifolds - Wave action methods for IC engines.
London and Bury St Edmunds, UK: Professional Engineering Publlishing Limited, 1999.

[9]

M. Ashe, G.Blair, G.Chatfield, D.Mackey, "Exhaust Tuning on Four-Stroke Engine: Experimentation and
Simulation," The Queen's Univeristy of Belfast; OPTIMUM Power Technology2001.

[10]

Yunquig Li, Jincheng Wang and Peng He, "Study on the exhaust system parameters of a small gasoline engine,"
Beihang University 2008.

[11]

G. Sammut and. A. Alkidas, "Relative Contributions of Intake and Exhaust - Tuning on SI Engine Breathing - A
Computational Study," Oakland University 2007.

[12]

J.D. Irwin and E.R. Graf, Industrial Noise and Vibration Control. New Jersey: Prentice-Hall 1979.

[13]

J. Pang et al. "Flow Excited Noise Analysis of Exhaust," Ford Motor Company; Gates Coporation2005.

[14]

A. Jauer, J. Brand and D. Wiemeler, "Flow Noise Level Prediction Methods of Exhaust System Tailpipe Noise,"
Tenneco, Germany 2008.

[15]

S. W. Coates, "The Prediction of Exhaust Noise Characteristics of Internal Combustion Engines ", The Queen's
University of Belfast, 1974.

[16]

Muthukumar Yadav, Kiran, Tandon and Raju, "Optimized Design of Silencer - An Integrated Approach," The
Automotive Research Association of India, Pune, India 2007.

[17]

Silvestri, Morel, Goerg and Jebasinski, "Modeling of Engine Exhaust Acoustics," Gamma Technologies, BMW
AG, J. Eberspacher, GmbH & Co.1999.

[18]

Wrtz and Mazzoni, "Application of WAVE in Motorcylce Prototyping," Ducati Motor S.p.A,, Bologna, Italy.

[19]

Honda et al, "Honda, Kodama, Wakabayashi,Nakayama, Morimoto and Ueda," Kokushikan University, Japan
2005.

[20]

Rose, Marshland and Law, "Optimisation of the Gas-Exchange System of Combustion Engines by Genetic
Algorithm," in 4th International Conference on Autonomous Robots and Agents, Wellington, New Zealand, 2009.

[21]

Massey, Williamson and Chuter, "Modelling Exhaust Systems Using One-Dimensional Methods," Flowmaster
(UK) Ltd. ; ArvinMeritor 2002.

[22]

Montenegro and Onorati, "A Coupled 1D-multiD Nonlinear Simulation of I.C. Engine Silencers with Perforates
and Sound Absorbing Material," Politecnico di Milano 2009.

21
Final Project Report 2011, SEIT, UNSW@ADFA

[23]

Montenegro and Onorati, "Modeling of Silencer for I.C. Engine Intake and Exhaust Systems by Means of an
Integrated 1D-multiD Approach," Dipartimento di Energetica - Politecnico di Milano2008.

[24]

Zhang and Romzek, "Computational Fluid Dynamics Applications in Vehicel Exhaust System," Eberspaecher
North America, Inc.2008.

[25]

J. Middleberg, T. Barber, S. Leong, E. Leonardi and K. Byrne, "Determining the Acoustic Performanec of a
Simple Reactive Muffler using Computational Fluid Dynamics," presented at the The Eight Western Pacific
Acoustics Conference, Melbourne, Australia, 2003.

[26]

Shah, Kuppili, Hatti and Thombare, "A Practical Approach towards Muffler Design, Developement and Prototype
Validation," 2010.

[27]

S. Sen, "Predction of Flow and Acoustical Performance of an Automotive Exhaust System using 3D CFD," TATA
Technologies Ltd.2011.

[28]

J. Caradonna, "Advanced Computational Aero-Acoustic Simulation of Complex Automotive Exhaust Systems,"
Faurecia Emissions Control Technologies2011.

[29]

Lu Lirong, Jin Xiaoxiong, Peng Wei and He Wei, "Application of Flow Field Simulation Technique to the Study
of Exhaust Noise of Car," presented at the IEEE Vehicle Power and Propulsion Conference, Harbin, China, 2008.

[30]

W. Y. Fowlkes a. C. M. Creveling, Engineering Methods for Robust Product Design- Using Taguchi Methods in
Technology and Product Developement. Reading, Massachusetts: Addison Wesley, 1954.

[31]

Bureau of Meteorology. (05 Nov 11). Sound Attenuation Calculator. Available:
http://www.csgnetwork.com/atmossndabsorbcalc.html

[32]

Bureau of Meteorology. Humiditiy Calculator. Available: http://www.bom.gov.au/lam/humiditycalc.shtml

[33]

T. J. Schultz, "Acoustical Uses for Perforated Metals: Principles and Applications," I. P. Association, Ed., ed:
Industrial Perforates Association Inc, 1986.

[34]

N. Huff, "Materials for Absorptive Silencer Systems," Owens Cornering Automotive Solutions2001.

22
Final Project Report 2011, SEIT, UNSW@ADFA

Annex A
Summary of Acoustic Theory for Automotive Silencers

1.

Absorptive/Side-Resonant Silencer
The absorptive and side-resonant silencers operate under the principles established for the use of perforated
metals in acoustic treatments. These principals differentiate between the design parameters of the perforated
materials utilised within the design, which in turn specify if the action of the silencer to be through the resonant
action of the perforated material or via viscous dissipation within sound absorbing material placed behind.
Parameters such as the Transparency Index [33] in Eq.(A1) or otherwise Blair‘s empirical relations [4] in
Eq.(A2) and Eq.(A3) may be used to distinguish between these types of silencer which are concerned with
perforation pattern of the material used. However, since the transparency index measure is only capable of
distinguishing between these variations of silencer beyond 10 kHz, being a frequency fairly well beyond the
significant spectrum present in an exhaust, it is not predominantly used for this purpose within this application.
Aonversely, Blair‘s relations were developed specifically for automotive silencers and are therefore much more
relevant.
(A1)
( )
( )
( )

( )
(

)

(A2)

(A3)

An absorption silencer utilises perforated material that shows negligible preference to the transmission of
any region of the frequency spectrum through the material and into the side chamber, otherwise known as the
‗transparency approach‘ [33]. By permitting acoustic wave energy within the side chamber it is made to reflect
from the outer shell and constructively interfere with sound waves entering the chamber. This interference then
establishes standing waves characterised by increased amplitude of particle oscillation, within the region
between the perforated pipe and the silencer housing. As seen in Fig.(A1), sound absorbing material fills this
region where wave superposition is predicted to occur.
Consequently, viscous dissipation of sound is achieved
as particle kinetic energy is converted to thermal energy
via interaction with the sound absorbing material. As per
Fig.(A2), correlation is shown between the radial
distance between the silencer shell and perforated pipe
and the largest wavelength capable of superposition
within the thickness of the sound absorbing material.
Consequently, this figure shows that for an increase in
thickness of the sound absorbing material, a significant
increase in attenuation is achieved for noise of longer
wavelength.
The Transparency Index can however be informative
through the evaluation of the Access Factor, which
represents a measure of the perforated metal‘s ability to
obstruct the entry of acoustic waves and has the effect of
scaling the absorption factor of the silencer[33] (The
absorption factor is defined as the transmission loss
Figure A1. Schematic of absorption silencer
expressed as a fraction of the incident sound energy).
23
Final Project Report 2011, SEIT, UNSW@ADFA

Annex A
Summary of Acoustic Theory for Automotive Silencers

Figure A2. Variation in attenuation with
frequency for thickness of absorption silencer

Figure A3. Access Factor vs frequency and
Transmission Index of perforated sheet metal

This is therefore a measure of the degradation in the
ability of the silencer to attenuate acoustic energy as a
result of perforation parameters.
As per Fig.(A3),
perforated metals with a transparency index of less than
6500 begin to have a noticeable effect on the attenuation
achieved at frequencies below 2000 Hz (frequencies up to
this point are considered significant sources of vehicle
noise).
Literature provides additional design consideration
relevant to the manufacture and implementation of an
absorption silencer are proposed which include the
following:
 It is recommended that the perforated pipe be
manufactured with stabbed holes rather than blind holes as
seen in Fig.(A4). This has the effect of increasing the
discharge coefficient for flow into the side chamber from
the central pipe and reducing turbulent eddies produced by
gas flowing over the sharp edges of blind holes. [4]
 While taking the transparency approach, it is also
important to consider that an excessive perforated area of
a tube may enable violent exhaust flow through the
silencer to degrade the sound absorbing material and even
attempt to rip it from the side chamber. It is therefore
recommended to use perforates with holes of diameter
between 2 and 3.5mm. [4]. (The addition of a layer of
stainless steel wool is also recommended by industry
SMEs).
 An absorptive silencer is most effective at
attenuating high frequency noise. Therefore this
component is recommended to be one of the last within
the exhaust system such that turbulent flow preceding the
absorption silencer has limited opportunity to build up this
high frequency component.
 The choice of sound absorbing material as well
as the density of the packing will lead to variation in the
achieved transmission loss as per Fig.(A5) [34]. This data
reiterates that the maximum wavelength absorbed
increases with thickness of sound absorbing material. In
addition, an increase in the material density from 100g/L
to 200g/L is seen to accompany a reduction in absorption
achieved. This highlights that a density too high will
restrict the entry of acoustic waves into the side chamber,
while a density too low is also acknowledged to become
less effective in achieving viscous dissipation of acoustic
energy.
In contrast a side-resonant silencer, whilst sharing a
similar form as the absorption silencer, does not utilise
packing material and instead provides attenuation over a
relatively narrow band of frequencies as per Fig.(A6). This
is achieved through the resonance of the side cavity at its
natural frequency. The design of this type of silencing
component is governed by Eq.(A4) to Eq.(A7). The design
variables, seen in Fig.(A7), are shown to dictate the
resonant frequency of the component as well as the
attenuation achieved at the resonant frequency [4].

Figure A4. Schematic of perforated pipe
showing standard blind holes and stabbed
holes

24
Final Project Report 2011, SEIT, UNSW@ADFA

Annex A
Summary of Acoustic Theory for Automotive Silencers

Figure A6. Attenuation predicted for a side-resonant
silencer

Figure A5. Variation in attenuation with density of sound
absorbing material

Figure A7. Design parameters of side-resonant silencer
element



(A4)
(A5)

(


)

(A6)



(A7)
hole conductivity

25
Final Project Report 2011, SEIT, UNSW@ADFA

Annex A
Summary of Acoustic Theory for Automotive Silencers
2.

Diffusing Silencer/Expansion Chamber
The expansion chamber as seen in Fig.(A8), is designed to absorb acoustic wavelengths equivalent to the
natural frequency of the chamber. The transmission loss of an expansion chamber is given by Eq.(A8) to
Eq.(A11).
[

(

)

(

)]

(

)

(

)

(

)

(

)

(A8)

(A9)
(A10)
(A11)

As per Fig.(A9) [12], the attenuation is seen to be periodic with frequency. In addition, the maximum
attenuation is seen to increase with area ratio of the chamber (m). These relations have been experimentally
determined to have value up to a frequency of 1500 Hz. In practical terms, Fig.(A9) suggests that a longer
chamber will offer increased attenuation at lower frequency, hence why this component is employed within
silencers to address low frequency noise corresponding to the engine firing rate.

Figure A8. Schematic of diffusing silencer/ expansion chamber

Figure A9. Plot of theoretical attenuation of expansion chamber with design parameters

26
Final Project Report 2011, SEIT, UNSW@ADFA

Annex A
Summary of Acoustic Theory for Automotive Silencers

3. Hershel-Quincke Tube
As seen in Fig.(A10), the Hershel-Quincke tube is a device in which sound waves from a common source
travel through two tubes of different lengths and recombine, producing reinforcement or cancellation of sound
depending on the difference in path length. Unfortunately, no further description can be provided as this concept
is poorly documented with regards to acoustic attenuation.

Figure A10. Example of a Hershel Quincke tube silencing component

4.

Helmholtz Resonance Silencer
A Helmholtz chamber is seen in Fig.(A11) attached to the header
pipe of an aftermarket Akrapovic exhaust system. It is an acoustic
filter element that operates under the principles of a spring-mass
system where the equivalent mass and spring force components are
defined by the structural parameters of the chamber seen in Fig.(A12).
The Helmholtz silencer is classified as a band-stop filter which offers
attenuation at a specified frequency defined by Eq.(A12) and
Eq.(A13).


(A12)
(A13)

The acoustic power transmission coefficient is then defined by
Eq.(A14) and shown in Fig.(A13) for a specified chamber.
(

(






) )

Figure A11. Helmholtz chamber

(A14)

Figure A12. Schematic of Helmholtz
chamber

27
Final Project Report 2011, SEIT, UNSW@ADFA

Annex A
Summary of Acoustic Theory for Automotive Silencers

Figure A13. Acoustic power transmission coefficient for an example Helmholtz chamber

28
Final Project Report 2011, SEIT, UNSW@ADFA

Annex B
WAVE model validation data -WR450 Power curves

WAVE Model Power Curve
40

35

30

Engine Power(hp)

25

20

15

10

5

0
1000

2000

3000

4000

5000
6000
RPM

7000

8000

9000

Figure B1. WR450 power curve generated with developed WAVE model

Figure B2. WR450 power curves provided by Cal Poly FSAE team

29
Final Project Report 2011, SEIT, UNSW@ADFA

10000

Annex C
Muffler Parameter Study - Experimental Data

SPL Magnitude at Position A with Engine Speed for Parameter
Variations and Commercial WR450 Muffler
120

SPL (dB)

115

Design 1

110

Design 2
Design 3
Design 4

105

WR450 Muffler

100
0

2000

4000

6000

8000

10000

Engine Speed (rpm)
Figure C1

SPL Magnitude at Position B with Engine Speed for Parameter
Variations, Unsilenced and Commercial WR450 Muffler
125
120

SPL (dB)

115
Design 1

110

Design 2
Design 3

105

Design 4
WR

100

Unsilenced

95
90
0

2000

4000

6000

8000

10000

Engine Speed (rpm)
Figure C2

30
Final Project Report 2011, SEIT, UNSW@ADFA

Annex C
Muffler Parameter Study - Experimental Data

Figure C3

31
Final Project Report 2011, SEIT, UNSW@ADFA

Annex C
Muffler Parameter Study - Experimental Data

SPL with Frequency at 3000 rpm at position B
110
100
90
SPL (dB)

Design 1
80

Design 2
Design 3

70

Design 4
60

WR Muffler
Unsilenced

50
40
0

200

400

600

800

1000

1200

1400

1600

1800

2000

Frequency (Hz)
Figure F4

SPL with Frequency at 7000 rpm at position B
120
110
100
Design 1

SPL (dB)

90

Design 2

80

Design 3

70

Design 4
WR Muffler

60

Unsilenced

50
40
0

200

400

600

800

1000

1200

1400

1600

Frequency (Hz)

Figure F5

32
Final Project Report 2011, SEIT, UNSW@ADFA

1800

2000

Annex C
Muffler Parameter Study - Experimental Data

Insertion Loss at 3000 rpm
40

SPL (dB)

30
20

Design 1

10

Design 2
Design 3

0
-10

0

100

200

300

-20

400

500

600

700

800

900

1000

Design 4

Frequency (Hz)
Figure F6

Insertion Loss at 4000 rpm
50

SPL (dB)

40
30

Design 1

20

Design 2
Design 3

10

Design 4

0
-10

0

100

200

300

400

500

600

700

800

900

1000

Frequency (Hz)
Figure F7

SPL (dB)

Insertion Loss at 7000 rpm
40
35
30
25
20
15
10
5
0

Design 1
Design 2
Design 3
Design 4

0

100

200

300

400

500

600

700

800

Frequency (Hz)

Figure F8

33
Final Project Report 2011, SEIT, UNSW@ADFA

900

1000

Annex C
Muffler Parameter Study - Experimental Data

Variation in transmission loss with silencer design at 3000 rpm
90
Design 1
Design 2
Design 3
Design 4

80

70

Transmission Loss (dB)

60

50
X: 2620
Y: 40.65

40

30

20

10

0

0

500

1000

1500
Figure F7
Frequency (Hz)

2000

2500

3000

Figure F9

Variation in transmission loss with silencer design at 3000 rpm
45
Design 1
Design 2
Design 3
Design 4

40

Transmission Loss (dB)

35

30

25

20

15

10

5

0

100

200

300

400
Frequency (Hz)

500

Figure F10

34
Final Project Report 2011, SEIT, UNSW@ADFA

600

700

800

Annex C
Muffler Parameter Study - Experimental Data

Figure C11

Figure C12

35
Final Project Report 2011, SEIT, UNSW@ADFA

Annex C
Muffler Parameter Study - Experimental Data

Figure C13

Figure C14

36
Final Project Report 2011, SEIT, UNSW@ADFA

Annex C
Muffler Parameter Study - Experimental Data
Design 1: 4000 rpm

Design 3: 4000 rpm

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

100

SPL (dB)

SPL (dB)

100

80

60

40

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

80

60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

600

Design 2: 4000 rpm

1800

2000

100

SPL (dB)

SPL (dB)

1600

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

100

80

60

40

1400

Design 4: 4000 rpm

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

800
1000 1200
Frequency (Hz)

80

60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

2000

Figure F15

Design 1: 3000 rpm

Design 3: 3000 rpm

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

100

SPL (dB)

SPL (dB)

100

80

80

60

40

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

600

Design 2: 3000 rpm

1800

2000

100

SPL (dB)

SPL (dB)

1600

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

100

80

60

40

1400

Design 4: 3000 rpm

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

800
1000 1200
Frequency (Hz)

80

60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

600

Figure F16

37
Final Project Report 2011, SEIT, UNSW@ADFA

800
1000 1200
Frequency (Hz)

1400

1600

1800

2000

Annex C
Muffler Parameter Study - Experimental Data

Design 1: 7000 rpm

Design 2: 7000 rpm

150

150
WAVE:SPL at muffler outlet
Theoretical resonator chamber transmission loss
Experimental data

WAVE:SPL at muffler outlet
Theoretical resonator chamber transmission loss
Experimental data

SPL (dB)

100

SPL (dB)

100

50

0

50

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

0

2000

0

200

400

600

Design 3: 7000 rpm

800
1000 1200
Frequency (Hz)

1400

1600

1800

2000

Design 4: 7000 rpm

150

150
WAVE:SPL at muffler outlet
Theoretical resonator chamber transmission loss
Experimental data

WAVE:SPL at muffler outlet
Theoretical resonator chamber transmission loss
Experimental data

SPL (dB)

100

SPL (dB)

100

50

0

50

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

0

2000

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

2000

Figure F17

Design 1: 8000 rpm

Design 2: 8000 rpm

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

100

SPL (dB)

SPL (dB)

100

80

80

60

40

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

600

Design 3: 8000 rpm

1800

2000

100

SPL (dB)

SPL (dB)

1600

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

100

80

60

40

1400

Design 4: 8000 rpm

WAVE:SPL at muffler outlet
Mean SPL measurement
Experimental data

120

800
1000 1200
Frequency (Hz)

80

60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

600

Figure F18

38
Final Project Report 2011, SEIT, UNSW@ADFA

800
1000 1200
Frequency (Hz)

1400

1600

1800

2000

Annex C
Muffler Parameter Study - Experimental Data

SPL with Frequency at 3000 rpm at position C
95
90
85
80

Series1

75

Series2

70

Series3
Series4

65

Series5

60
55
50
0

100

200

300

400

500

600

700

800

900

1000

Figure C19

SPL with Frequency at 7000 rpm at position C
100
95
90
85
Series1

80

Series2
75

Series3

70

Series4

65

Series5

60
55
50
0

100

200

300

400

500

600

700

800

Figure C20

39
Final Project Report 2011, SEIT, UNSW@ADFA

900

1000

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D1

Figure D2

40
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D3

Figure D4

41
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D5

Figure D6

42
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D7

Figure D8

43
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D9

Spatial Temperature and Speed of Sound in stepped pipe- Design 1

Spatial Temperature and Speed of Sound in stepped pipe- Design 2

1100

1100
Pipe step 1
Pipe step 2

1000

1000

900

900

Speed of Sound (m/s) / Temperature (K)

Speed of Sound (m/s) / Temperature (K)

Pipe step 1
Pipe step 2

800

700

600

500

400

800

700

600

500

0

0.5

1

1.5

400

0

0.5

Position (m)

1
Position (m)

Figure D10

44
Final Project Report 2011, SEIT, UNSW@ADFA

1.5

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D11

Figure D12

45
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D13

46
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D14

47
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D15

48
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D16

Figure D17

49
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D18

Figure D19

50
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D20

Figure D21

51
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D22

Figure D23

52
Final Project Report 2011, SEIT, UNSW@ADFA

Annex D
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure D24

Figure D25

53
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

SPL at muffler outlet: 3000 rpm

SPL at muffler outlet: 4000 rpm

140

140
SPL prediction at outlet
max tolerated SPL: 108 dB

120

100

SPL (dB)

SPL (dB)

120

80
60

40

SPL prediction at outlet
max tolerated SPL: 108 dB

100

80
60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

SPL at muffler outlet: 7000 rpm

1400

1600

1800

2000

140
SPL prediction at outlet
max tolerated SPL: 108 dB

120

SPL prediction at outlet
max tolerated SPL: 108 dB

120

100

SPL (dB)

SPL (dB)

800
1000 1200
Frequency (Hz)

SPL at muffler outlet: 8000 rpm

140

80
60

40

600

100

80
60

0

200

400

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

40

2000

0

200

400

Figure E1

54
Final Project Report 2011, SEIT, UNSW@ADFA

600

800
1000 1200
Frequency (Hz)

1400

1600

1800

2000

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E2

Figure E3

55
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Comparison of Transmission Loss to Weight of Absorption
Silencer
Ratio increase in Transmission Loss

4
3.5
3
2.5
2

Diameter

1.5

Length

1
0.5
0
0.00E+00 2.00E+00 4.00E+00 6.00E+00 8.00E+00 1.00E+01 1.20E+01
Ratio increase in weight
Figure E4

Figure E5

56
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E6

Figure E7

57
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Transmission loss with choke diameter

Transmission loss with choke diameter

30

30
20mm
30mm
no choke

20
15
10
5
0
-5

20mm
30mm
no choke

25

Frequency (Hz)

Frequency (Hz)

25

20
15
10
5
0

0

200

400

600

800
1000 1200 1400
Transmission Loss (dB)

1600

1800

-5

2000

0

200

400

Transmission loss with choke diameter
30
20mm
30mm
no choke

Frequency (Hz)

25
20
15
10
5
0
-5

0

200

400

600

800
1000 1200 1400
Transmission Loss (dB)

1600

1800

2000

Figure E8

58
Final Project Report 2011, SEIT, UNSW@ADFA

600

800
1000 1200 1400
Transmission Loss (dB)

1600

1800

2000

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E9

Figure E10

59
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E11

Figure E12

60
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E13

Figure E14

61
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E15

Figure E16

62
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E17

Figure E18

63
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E19

Figure E20

64
Final Project Report 2011, SEIT, UNSW@ADFA

Annex E
WAVE Parameter Study – Exhaust Design for Scavenging Performance

Figure E21

65
Final Project Report 2011, SEIT, UNSW@ADFA

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close