This paper summarises the main results of an EU-funded research project, ARTEMIS (G3RD-CT-2001-00511), on noise from turbo-charged Diesel engine exhaust systems. The project started in September 2001 and ended in August 2004 and was co-ordinated by KTH. The project had 10 partners from 6 different European countries, 5 universities and 5 companies including some major truck and car manufacturers. The main objective was to develop new and improved computational tools for predicting noise from exhaust systems. New models for describing the engine as an acoustic source were developed and experimentally tested. They include a linear time-varying source model and a non-linear frequency domain model. Linear time-invariant source data was also determined both from experiments and using 1-D gas-exchange simulations. New and improved models were developed for the turbo-group including non-linear time domain models and a linear time-varying model. New models were developed and experimentally tested for sound transmission through the Diesel particulate filter included in modern Diesel engine after-treatment devices. Improved models were developed for describing perforate mufflers with high mean flow velocities. Improved experimental techniques for determination of transmission properties of duct system components were developed. Models were developed and coded for sound reflection and radiation from tailpipe openings. Full experimental validation of the Munt theory for radiation from open pipes with flow was produced. In conclusion it can be said that the project was successful and gave many useful results.
Acoustic liners are widely used to attenuate sound waves inside the aircraft jet engines. Previous research has proved that segmenting the liner and the positioning of the liner segments affect the attenuation characteristics of the liner. The combined effect of circumferentially segmented and non-locally reacting liners received little attention. The aim of this work is to investigate these effects, and to compare the properties of circumferentially segmented duct liners with those of uniform liners, in order to identify any potential benefits of circumferentially segmented liners. A new technique is proposed here; the point-matching method. Briefly, it is a straightforward numerical method based on a closed form ansatz, which fulfils the governing equations and is matched to the boundary conditions point-wise. A code, previously developed for automobile applications, is used to obtain the wave numbers of the different modes, from which the transmission loss for each mode can be calculated at the desired range of frequencies. An infinite cylindrical duct of diameter 40 cms was chosen to apply different non-locally (bulk) reacting liner configurations on. It was found that the existence of hard surfaces in a lined duct and their arrangement greatly affect the behavior of each mode and the energy distribution among them.
Both linear (frequency domain) and non-linear (time domain) prediction codes are used for the simulation of duct acoustics in exhaust systems. Each approach has its own set of advantages and disadvantages. One disadvantage of the linear method is that information about the engine as an acoustic source is needed in order to calculate the insertion loss of mufflers or the level of radiated sound. The source model used in the low frequency plane wave range is the linear time invariant 1-port model. This source characterization data is usually obtained from experimental tests where multi-load methods and especially the two-load method are most commonly used. These measurements are time consuming and expensive. However, this data can also be extracted from an existing 1-D non-linear CFD code describing the engine gas exchange process. The pressure and velocity predictions from two acoustic load cases can be used to determine the source strength and impedance at a particular location in the exhaust line. This has been done at a location downstream of the turbocharger in the exhaust system of a heavy diesel truck over a number of speeds and engine loads. This source data is then used in a linear simulation of the exhaust line to predict sound pressure levels at a free field microphone position. The predicted source data and sound output at the microphone position is validated against measured data. The results show that you can obtain reasonably accurate source data and approximate free field sound pressure level predictions using non linear simulation in a linear acoustic model of the exhaust system. This technique can be used to extend the use of linear acoustic simulations to models of the complete exhaust line with the characterized engine as a source and exhaust sound output as a result.
The most common noise reducing measure is to add sound absorbing material on the domain boundaries. The boundaries covered by the material may in sumilations be represented by the surface impedance of the material. The impedance can either be modeled or determined experimentally. The experimental determination can be done by the well known standing wave tube method or by a free field method. These free field methods enable impedance determination at any angle of incidence for bulk reacting materials, as opposed to the standing wave tube method that is restricted to normal incidence or locally reacting materials. The method prescribes a point source above the surface and measurements in two points close to the sample surface. From this, the surface impedance can be deduced through the known sound field formulation. Among other things, the impact on the accuracy of the method from the field formulation, signal conditioning and sensor type have been studied in previous work. One major concern is the finite size of the material sample, and its influence on the measurement accuracy. This has previously been investigated for highly absorbing materials and it was shown to be a low frequency problem. Therefore, we focus on the impact of the finite sample in frequencies below 2 kHz. In particular, we relate the magnitude of the impact to the properties of the tested material. Also, the influence of the mounting of the material is analyzed. The study is made through analyzing numerical simulations of the experiment for a variety of setups and materials. Theoretical discussion is provided for deeper understanding of the results. The impact of the finite sample is seen to depend on the material properties, not only the setup as previously shown. Materials with high absorption are shown to be more sensitive to these errors.
In noise abatement using porous or fibrous materials, accurate determination of the surface impedance representing the absorber is decisive for simulation quality. The presence of grazing flow and non-homogeneous ambient temperature influence the reaction of the absorber and may suitably be included in a modified “effective” surface impedance. In this paper, this approach is applied to a generic case representative for the engine bay of a heavy truck, where porous shields suppress the radiated noise, e.g. during a pass-by noise test. The change in the absorption is determined numerically by solving the wave propagation through a layer of varying temperature and flow adjacent to the impedance surface for different angles of incidence. The study shows significant impact of both flow and temperature, especially for materials with low absorption. The diffuse field absorption coefficient is also derived and although the effect is less pronounced in this case, it is still important in lower frequencies and in the frequency range typical for IC engine noise. The proposed numerical method is shown to be accurate and efficient for determination of the effective impedance and moreover not limited to thin boundary layers.
Noise encapsulations are widely used in automotive industry to enclose noise sources, such as e.g. the engine or the gearbox, to reduce externally radiated noise. The sound absorption factor of the material on the inside of the noise encapsulation is obviously vital for the sound attenuation. This parameter is in most cases determined experimentally for which there are several methods. The results received from the various methods may vary as different acoustic states are examined and thus influence the choice of method. The absorption factor is crucial since it is used in specifications to material manufacturers as well as being an input parameter in modeling the performance of the noise shield e.g. during a pass-by noise test.
In this paper, two standardized measurement methods along with a third, non-standardized method, are applied to determine the properties of an absorbing material used in a commercial noise encapsulation. The methods are based on normal-, random- and oblique incident sound waves. The first and the last methods are based on measuring the acoustic impedance from which the absorption can be calculated while the random incidence method measures the absorption directly. The results retrieved from the three methods are compared and discussed in the light of the differences between them. This paper clarifies the differences and gives a practical guidance for the choice of measurement method and the use of the different absorption factors in modeling.
In vehicle applications, absorbing materials are often used to attenuate sound. In, for example, exhaust systems and on noise encapsulations, the absorber is exposed to flow. This creates a boundary layer above the absorber, which affects the impedance of the surface, and hence alters the absorption properties. In addition to this effect, the flow itself may enter the absorbent material due to high pressure and forced flow paths. An investigation of the effects that internal flow in the absorber imposes on the acoustic properties is presented. One way to describe the effect is by a change in flow resistivity. The effect is investigated for typical absorbers used in noise encapsulations for trucks. The Transfer Matrix Method is applied to calculate the resulting absorption coefficient for an absorber with changed flow resistivity due to internal flow. The possibility to model the changed properties of the absorber with internal mean flow by means of Biot theory is also explored, together with a discussion on suitable experimental methods to verify and further investigate the effects.
Accurate experimental characterization of sound absorbing materials is important to ensure good quality in simulations of larger systems and to analyze materials with unknown acoustic properties. Free field methods allow characterization of material properties at arbitrary sound incidence, which is advantageous compared to standardized methods. The errors of these methods have been studied, in particular those related to the size of the test samples. These are typically seen as oscillations about the correct value, especially at lower frequencies. The errors have been related to the source position and the sample size, but the impact of the material properties has not been investigated. In this paper, the influence of these properties on the errors are investigated through measurements and numerical simulations. The studies show a dependance on the material properties. The error results both from the pressure scattered at the sample edges and the pressure reflected at the material surface. The scattered field is shown to be stronger for materials with high flow resistivity, although the impact of this field on the result is stronger on materials with low flow resistivity. In addition, a method to reduce these errors based on an analytical formulation of the scattering is proposed. The method is applied to numerical simulations and shown to signicantly reduce the impact of the scattered field on the accuracy of the surface impedance.
Sound absorbing materials are used in many applications to reduce sound, and their soundabsorbing characteristics are most often determined experimentally since theoreticaldetermination is difficult. Sound absorption factors are used in material specifications aswell as input to numerical simulations.Several methods for experimental determination of the absorption factor exist, two of themstandardized and frequently used. It is commonly known that the absorption factorobtained by these two methods differs as different sound fields are prescribed by thestandards. However, the size of the differences has not been so well described. Due to thisdifference, the choice of method is critical in order to avoid errors in simulations andspecifications of material properties.Experimental determination of absorption factors for three commonly used absorbers wasperformed, resulting in significant differences between the two methods. Correction factorsto compensate the absorption factor determined at one acoustic state and used in anotherare given. Theory verifying the differences is also presented.
The necessity of accurate pass-by noise simulations of vehicles has increased as the requirements on noise levels is becoming stricter. Also, the design of noise reducing measures is needed early in the design process when measurements are not possible to perform. The impact of the sound absorbing materials representation on simulated pass-by noise levels from a truck is analysed in this paper. The material may be fully resolved in FEM, including bulk reaction, or represented by a surface impedance, either at normal or a specic angle of incidence. The first representation requires FEM simulations and more material data. This puts higher demands on input data, and more importantly, prevents the use of BEM simulations which signicantly would improve computational efficiency. The two latter representations may be implemented in BEM. The necessary assumption of local reaction may hold for some materials, but it is not always valid. The simulations presented in this paper show that the local reaction assumption underestimates the effect of sound absorption, giving up to 5 dB higher radiated sound power levels and pass-by noise levels up to 2 dB higher than obtained using the bulk-reacting representation. The difference is shown to depend on the material properties and the position of the source in relation to the noise shields and absorbing parts. The directivity of the radiated noise is not affected, although the regions of largest sound pressure levels are more pronounced. The choice of representation of the material is shown to be important for the simulated pass-by noise levels. To choose the level of complexity in the model, it is important to be aware of the effect this may have on the accuracy of the results in order to draw correct conclusions from the results.
The flow reversal chamber is a commonly used element in practical silencer design. To lower its fundamental eigenfrequency, it is suggested to acoustically short circuit the inlet and outlet duct. In the low frequency limit such a configuration will correspond to a Helmholtz resonator, but with a choked flow through the short circuit, the main flow will be forced through the expansion volume. For the proposed concept, the flow reversal resonator, a theoretical model is derived and presented together with transfer matrix simulations. The possible extension to a semi active device as well as the influence of mean flow on the system is investigated experimentally. Finally the concept is implemented on a truck silencer. The results indicate that the flow reversal resonator would prove an interesting complement to traditional side branch resonators. The attenuation bandwidth is broader and it can be packaged very efficiently. Mean flow effects are still an issue and should be studied further.
NOVELTY - The device has guide units (11a, 11b, 11c) that are arranged inside a pipe. The guide units divide an internal space in a curved portion (7`) of the pipe into two flow channels. The guide units comprise a set of holes running through, where the holes so narrow that sound passes through the holes between mutually adjacent flow channels but only with a certain resistance. The curved portion have curvature of such a magnitude that sound cannot pass through the flow channels without encountering a wall surface of the guide units.USE - Used for damping sound in a pipe (claimed) that supplies a gaseous medium e.g. compressed air, to a supercharged combustion engine e.g. diesel engine, that is utilized for powering a heavy vehicle e.g. truck.ADVANTAGE - The guide units divide the space in the curved portion of the pipe, where holes are so narrow that sound passes through the holes between the adjacent flow channels but only with certain resistance, thus providing a very effective damping of the sound from a turbo unit in relatively simple manner without disturbing the pattern of air flow and without auxiliary unit that would require more space.DESCRIPTION OF DRAWING(S) - The drawing shows a depiction of a sound damping device.Curved pipe portion (7`)Sound damping device (10)Guide units (11a, 11b, 11c)Flow channels (12a, 12b, 12c, 12d)
A muffler or silencer is a device used in a flow duct to prevent sound from reaching the openings of the duct and radiating as far-field sound. Reactive silencers do this by reflecting sound back towards the source while absorptive silencers attenuate sound using absorbing material. They are necessary components in the design of any exhaust or intake system for internal combustion (IC) engines. No car or truck can pass the standard noise tests required by legislation or compete on the market without them. There are three basic requirements for a modern exhaust systems; compact outer geometry, sufficient attenuation and low pressure drop.
The aim of this chapter is to discuss acoustical design and analysis of IC-engine exhaust and intake systems. The specific problems of modern intake systems made from plastic material with non-rigid walls are not discussed because not much has been published on this subject. The theory and techniques presented can be used also for other applications such as compressors and pumps and to some extent also for air-conditioning and ventilation systems.
A folded side-branch resonator has been designed and tested on a heavy truck. The design procedure is based on numerical simulations of the complete exhaust line and cold flow measurements of resonator inlet resistance and end-correction. A reduction in exhaust noise of as much as 7dB relative to the standard exhaust system shows the potential of the suggested technique to handle low frequency problems without increased volume or backpressure.
This paper presents an analytical wave decomposition model for predicting the transmission loss a cylindrical silencer with both annular and baffled micro-perforated screens. Numerical simulation shows the fundamental characteristics as well as the potential to achieve large attenuation using micro-perforations. The numerical model is verified by measurements using the 2-microphone technique and shown to be a useful tool in practical design. Clear from the analysis is the sensitivity of the micro-perforated silencer to changes in both porosity and overall layout.
The invention relates to a silencer device (10) for connection to a piston compressor of a motor vehicle, which device comprises - a housing (11) with an inlet (12) intended for connection to a suction line or to an air outlet of a piston compressor, and an outlet (13) intended for connection to an air inlet of a piston compressor or to a compressed air line, - a throughflow duct (20) extending through the housing from the inlet to the outlet to allow air to flow from the inlet to the outlet via this throughflow duct (20), and - a quarter-wave resonator (30) situated in the housing and comprising an elongate resonator duct (31) which is curved in its longitudinal direction, has at its one end an inlet aperture (32) situated in a wall of the throughflow duct, and has a cross- sectional area larger than or equal to that of the throughflow duct, the other end (33); of the resonator duct being closed. The invention relates also to a motor vehicle provided with such a silencer device.
This paper presents a derivation of the 2-port matrix for a folded quarter-wave side-branch resonator including higher order modes but neglecting flow interaction effects. The model is restricted to coaxial geometries and two-folds. The derivation is based on the mode matching technique and is verified by measurements done on prototypes. A notable result from these experiments is the effects of a slit-like leakage close to a rigid wall. A parametrical study finally investigates the influence of various lengths and area ratios of the resonator resulting in a set of design rules.
The classic Herschel-Quincke tube is a parallel connection of two ducts yielding multiple noise attenuation maxima via destructive interference. This problem has been discussed to different degrees by a number of authors over the years. This study returns to the basics of the system for the purpose of furthering the understanding of the conditions necessary for noise attenuation and especially their sensitivity to mean flow. First, the transmission loss for an N-duct system with mean flow and arbitrary conditions of state in the different ducts is derived. Next, the two types of conditions yielding the attenuation maxima are studied. In addition to a discussion of the underlying physics, generic expressions for frequencies at which maximum attenuation occur are presented. Experiments without mean flow generally show good agreement with theory based on straight duct elements. However, more detailed models may be required for accurate simulations in the presence of mean flow. A simple model compensating for the losses associated with bends is shown to improve the results significantly for the geometry studied.
The inclusion of flow-acoustic interaction effects in linear acoustic multiport models has been studied. It is shown, using a T-junction as illustration example, that as long the acoustic system is linear the required information is included in a scattering matrix obtained by experimental or numerical studies. Assuming small Mach numbers and low frequencies-as in most automotive silencer applications-the scattering matrix for the T-junction can be approximated using quasi-steady models. Models are derived that holds for all possible configurations of grazing and bias flow in the T-junctions. The derived models are then used to predict the performance of a novel silencer concept, where a resonator is formed by acoustically short-circuiting the inlet and outlet ducts of a flow reversal chamber. The agreement between experiments and simulations is excellent, justifying the use of the quasi-steady modeling approach.