acoustic optimization tool

SPERoN is started by selecting the AOT icon on the desktop or in the start menu. The AOT opens with the screen given in figure 2.1.

figure 2.1: Start screen

The user interface of the SPERoN software mimics the look-and-feel of a typical "explorer" type application. In the user interface we can distinguish three areas:

When no model is loaded. The model explorer and data viewer are empty. When we load a model (e.g. AOTexample.xml) these areas contain the model outline and details. The function of each of these areas will be explained in the next sections.

figure 2.2: Application with AOTexample.xml loaded

2.1.1.1. Menu bar

The menu bar contains the main menus of the application such as File, Tools and Help. From the File menu you can create a 'New' model or 'Open' and existing model. In the Tools menu with the 'Options' command lets you specify various settings for the application. The help button gives access to the help resources for the SPERoN application (such as this help document).

figure 2.3: Menu bar

2.1.1.2. Model explorer

The model explorer gives a tree view of the current model that is worked on. Under the root tree node (which has the name of the input file) we find three top-level nodes: Input parameters, Simulation runs and Results. In the Input parameters section you define the road input parameters for a simulation. In simulation runs you choose a subset of the input parameters to define the road and other parameters that comprise a simulation such as vehicle type and speed.

figure 2.4: Model explorer

2.1.1.3. Data viewer

The Data Viewer displays details about the currently selected model element from the model explorer. In the example screen shot for instance we see a texture spectrum. The data viewer changes depending on the model item that is selected in the Model explorer.

figure 2.5: Data viewer with texture spectrum

When an AOT user license has been obtained, a License Request file is needed to activate it. To generate such a license request file and activate the software, these steps should be taken:

figure 2.6: Generate a license request file

Your user data will be processed and you will be sent back a license file, f.i. LicenseData-myUsername.xml. This file is used to activate the license:

The license request window should now read your username and authentication key. You can check the connection to the SPERoN calculation server and the status of the server itself at any time through the Tools menu (see figure 2.7).

figure 2.7: Check connection to SPERoN server

By running simulations you can predict rolling noise for a certain vehicle (type) on a particular road at a certain speed. Setting up a simulation is done in the "Simulation runs" section of the model. A new simulation is added by choosing "New" in the context menu (right-click menu) of the "Simulation runs" in the model tree in the Model explorer. This opens a new simulation properties view in the Data viewer (see figure 2.9).

figure 2.9: Simulation properties view

In the simulations properties view you select the road input data and other simulation properties such as vehicle type and speed. The road data is selected in the comboboxes. In these comboboxes, you can select any road input data that is defined in the Input parameters section of the model. In addition to that, you can also directly use input from the database or read a road input data file. If you choose one of these options, the corresponding input data will also be added to the Input parameters section of the model. Options to add from the database or read a data file can also be accessed with the buttons next to the combobox (see figure 2.10).

figure 2.10: Adding input data to the simulation properties

For the vehicle type, you can choose between passenger car and truck tyre. The vehicle speed can be chosen in the range of 50 to 120 km/h. Optionally, you can add a description for the simulation run in the Description field.

When all simulation properties are defined you can start the simulation by clicking the "Run" button. The status of the running simulation will then be displayed below the "Run" button.

The context menu (right-click) of each simulation run gives additional actions that can be performed on a simulation run such as renaming, deleting, duplicating or running.

The "Results" section of the model contains prediction results of previous simulation runs. After a simulation run is started and finishes, the simulation run is removed from the "Simulation runs" section and added to the "Results" section. By clicking on the ID of a finished simulation run, you can access the output of the simulation run for viewing.

figure 2.11: Viewing simulation results in the Data viewer

The data viewer for Output has two tabs: "Noise results" and "Simulation properties". The "Simulation properties" tab lets you review the simulation settings that were used to obtain these simulation results. The "Noise results" tab shows the results for the predicted noise.

The noise predictions with SPERoN produce data on a number of receiver positions. The standard CPX and SPB positions are available by default as receiver positions. For a certain receiver position, the total noise levels and noise spectra are displayed. Besides the total level and spectrum, the contribution of each of the noise generation mechanisms such as vibration, airflow, cavity noise and residual (e.g. vehicle) noise is displayed both as a spectrum and as a noise level.

The noise levels and spectra that are predicted are the SPB and CPX level for the tyre group that was selected for the simulation. They do not necessarily correspond to the standard indices for SPB and CPX (either survey or investigatory) methods.

Additional operations with simulation results can be accessed from the context-menu (by right-clicking on the simulation run ID). The results of a simulation can be removed (menu item "Delete"), renamed (menu item "Delete"), exported to a CSV or XML file (menu item "Export") or copied to the clipboard for further processing (menu item "Copy"). The graphs can also be copied to the clipboard by using the context menu on the graph and choosing Copy to clipboard.

Road texture is generally measured with laser profilometers. A profilometer measures the height of the surface as function of the position. This done by moving a laser distance sensor over the road surface with a controlled displacement. The horizontal movement and the distance between laser and road surface are recorded while the laser scanner moves over the road surface. This results in a 2D road texture profile. To add third dimension to the measurement setup, the scanning procedure is repeated on parallel measurement lines. The ISO standard 13474-1/2/3, “Characterization of pavement texture by using surface profiles” describes the requirements for the measurement equipment and data processing into more detail.

figure 3.1: 3D laser profilometer

For accurate predictions with the SPERoN model, there are some minimum requirements on the measurement data that have to be fulfilled. The required input for the SPERoN model are at least 6 parallel texture profile lines measured with a sampling interval of 1 mm or smaller and a distance of 10 mm between the profile lines. This requirement is necessary to feed the model enough information about the transverse variation in texture. The length of each profile is at least 2 m. This is necessary to allow a full revolution of a (passenger car) tyre on the road surface. For truck tyres, it is advisable to use longer texture profiles.

When the requirements on number of profiles and/or profile length are not met, then one could resort to preprocessing to convert the measurement data in the required format. For example, if a single texture profile for a large length is available, one could cut the profile in smaller pieces and assume these shorter pieces as being parallel profiles. The application does not restrict this kind of processing, but it is disadvised by the authors. You should be cautious with the results obtained in this way as there is no way of checking the validity of this approach without a thorough analysis.

Input of texture data from user supplied data is possible in two formats: the XML format and the CSV format (e.g. exported by Excel and Matlab). All data in all SPERoN model input files is assumed to be in SI units (meter, kilogram, second).

The XML file format is the same as is used for the input files of the AOT software. An example of an XML input file is given in example 3.1. The tags in the XML file are self-explanatory. The <profileData> contains the actual profile data as a long value array with the values separated by whitespace.

<?xml version="1.0"?>
	<texture>
	 <userTexture ID="Testfield10_B_west">
	  <userDescription>
	   <datetime>2006-11-06</datetime>
	    <location>Kloosterzande field 10, west track, pos B</location>
	    <material>25 mm PAC 4/8 - 65 mm PAC 11/16</material>
	    <method>2.5D</method>
	    <info>M+P measurement for AOT project, 3m texture beam</info>
	   </userDescription>
	   <textureProfile>
	    <textureSettings>
	     <measurementLength>2.9000</measurementLength>
	     <longitudinalSpacing>2.0000e-004</longitudinalSpacing>
	     <numParallelTraces>20</numParallelTraces>
	     <parallelSpacing>0.0100</parallelSpacing>
	    </textureSettings>
	    <profileData>
            0.012763 0.012832 0.012763 ... ...
	    </profileData>
	   </textureProfile>
	  </userTexture>
	 </texture>

The CSV file format assumes that the measured texture profiles are stored column wise in a comma separated values file. The expected separator is the semi-colon (;). The first column of the input file should be the longitudinal position of profile. Note that the first line in the input file is considered a header line and is therefore ignored. An example of a CSV input file is given in example 3.2. Since there is no way to deduct the distance between the traces from the input file, you will be asked in a dialog for this value.

	position;	trace 1;	trace 2;	...	trace 6
	0.0002;	-8.11E-04;	 6.32E-04;	...	-9.38E-04
	0.0004;	-7.99E-04;	 6.27E-04;	...	-6.55E-04
	0.0006;	-9.91E-04;	 5.62E-04;	...	-1.58E-03
	0.0008;	-5.35E-04;	 5.04E-04;	...	-1.94E-03
	0.0010;	 2.48E-04;	 5.40E-04;	...	-2.30E-03
	0.0012;	 3.39E-04;	 4.56E-04;	...	-2.66E-03
	0.0014;	 3.68E-04;	 3.48E-04;	...	-2.66E-03
	0.0016;	 3.21E-04;	 8.53E-05;	...	-2.19E-03
	0.0018;	 3.34E-04;	-4.70E-05;	...	-2.04E-03
	…
	0.9996;	 4.65E-04;	 4.49E-04;	...	-5.97E-04
	0.9998;	 1.64E-04;	 4.58E-04;	...	-4.84E-04
	1.0000;	-1.84E-04;	 4.47E-04;	...	-4.38E-04

The measurement of absorption involves the determination of the difference between the energy levels of two related sound fields: the incident and reflected sound fields. For this type of measurements, complex methods are required to achieve acceptable accuracy levels. In general, two absorption systems are used: the extended surface method (ESM) and the sealed tube method. Each of these systems has its specific operation application. The ESM best suited for road surfaces with high absorption. The sealed tube method performs best at low absorption. Both methods are described in ISO standards: ESM in ISO 13472-1 and the sealed tube in ISO/DIS 13472-2.

The result of the measurement is the acoustic absorption coefficient as a function of frequency. Generally this quantity is represented as a third octave spectrum. However, the accuracy of the SPERoN propagation model improves when a spectrum with a higher frequency resolution (e.g. 12th octave band or small-band FFT spectrum) is used as input.

In the current version of the SPERoN model, only absorption data is used. Therefore, you do not need to supply the complex acoustic impedance spectrum. The spectrum of the acoustic absorption coefficient is sufficient for the propagation model. The spectrum can be read in from a CSV file with two colums: a column with frequency values and and a column with values for the absorption coefficient at each frequency. An example of a CSV input file is given in example 3.3.

	frequency;	absorption
		250;		0.069
		315;		0.089
		400;		0.071
		500;		0.084
		630;		0.055
		800;		0.126
		1000;		0.023
		1250;		0.024
		1600;		0.037
		2000;		0.006
		2500;		0.194
		3150;		0.115

Besides entering absorption data using measurement data, you can also define the absorption data by using a constitutive model for the road absorption. In the AOT, two models for road absorption are implemented: the Hamet model and the Attenborough model. A new model can be entered by right-clicking on the "Acoustic impedance" input node in the AOT model view and selecting "New model..." from the context menu.

In the data viewer panel for the acoustic impedance model, a form is displayed showing the model parameters in the bottom section and a graph of the current absorption curve derived from these parameters at the top. Using the buttons below, road surface layers can be added and removed, thus allowing for porous surfaces with multiple, different layers. The constitutive parameters for each road surface layer are:

figure 3.2: Data viewer containing an absorption model for a road surface

The measurement procedure for the determination of the air flow resistance of road surfaces on-site is method developed in the EU project ITARI. The method was chosen to be as close as possible to the procedure described in the standard DIN EN 29053 - “Acoustics; materials for acoustical applications; determination of airflow resistance (1993-05)”. The latter procedure describes the measurement of airflow resistance in laboratory conditions.

The airflow resistance is measured with a special device in which airflow and pressure drop in the contact area of the tyre/road interface is simulated. A schematic view of the measurement setup is given in figure 3.3. The measurement principle is as follows. A definite constant air flow q is pressed through a definite area A of the road surface. The differential pressure Δp = p_meas - p_atm necessary to maintain the constant air flow q through this area is measured as a function of q. The air flow resistance Ξ of the surface is calculated from air flow rate q and differential pressure Δp by Ξ = Δp/q. Here Ξ is still a function of the area A so we define the specific air flow resistance R_s ^* = Ξ·A = Δp/u, which is independent of the size of the area A. u is the velocity of the air flow through the surface area A. In the evaluations given here the specific flow resistance R_s ^* is calculated for a reference velocity of u = 0.0125 m/s. As this specific air flow resistance does not match exactly the definition of the specific air flow resistance R_s in standard DIN EN 29053 the denotation R_s ^* is used here for the specific air flow resistance of road surfaces determined according the measurement procedure described above.

figure 3.3: Schematic overview of the setup for airflow resistance measurements

The flow resistance is a single value quantity which does not depend on the frequency. The unit for this quantity is Ns·m^-3.

The mechanical properties of the surface are assumed can be explained with a basic mass-spring-damper model where the spring represents the stiffness k_s and a damping c_v . The amplitude of the mechanical impedance for this system is given in figure 3.4. This typical graph shows the frequency dependence of the impedance: the low frequency and high frequency regions are governed by stiffness and mass, respectively, whereas damping dominates the mid-frequency region around the resonance frequency f_res

figure 3.4: Basic principle for mechanical impedance

Mechanical impedance can be measured in various ways. Since the mechanical impedance is a rather new road input parameter, no standards exist that describe how this quantity should be measured. In the development of the SPERoN model, we designed a measurement device to measure the driving point mechanical impedance on road samples in the laboratory. In figure 3.5 we see an overview of the measurement setup. We used a shaker to excite the road surface sample and used a so-called impedance head to simultaneously measure the input force and driving point vibration.

The shaker is placed upside down in springs, so to avoid any unwanted vibrations of the measured surface. The impedance head is connected rigidly to the shaker and excited with a sinusoidal force using a frequency sweep. Since both the input (force) and the output (acceleration) are measured with the impdance head, the results are independent of the exact force amplitude used and of the configurations of the system above the impedance head.

figure 3.5: Measurement setup for mechanical impedance measurements

With respect to the measurement setup, care should be taken to some aspects that influence the measurement results:

figure 3.6: Impedance head and measurement foot

Mechanical impedance data can be selected from the measurement database or the user can import his own data. An XML file format (see example 3.4) can be used to import the user data, which allows information on the measurement location, material, etc. to be included, or a CSV file (see example 3.5) can be used that contains only the complex impedance data.

	<?xml version="1.0"?>
	 <mechanicalimpedance>
	  <userMechanicalimpedance ID="Testfield_36">
	   <userDescription>
	    <datetime>2008-11-10T00:00:00</datetime>
	    <material>15 mm porous Regupol 6510 G</material>
	    <method>M+P measurement on surface sample</method>
	    <location>Kloosterzande field 36</location>
	    <info>M+P measurement for AOT project</info>
	   </userDescription>
	   <complexSpectrum>
	    <cPt><f>100</f><ZRe>-482300.48</ZRe><ZIm>-65079.30</ZIm></cPt>
	    <cPt><f>125</f><ZRe>-480536.20</ZRe><ZIm>-77301.49</ZIm></cPt>
	    <cPt><f>160</f><ZRe>-479587.85</ZRe><ZIm>-89961.08</ZIm></cPt>
	    <cPt><f>200</f><ZRe>-487571.68</ZRe><ZIm>-95420.20</ZIm></cPt>
	    ... ...
	   </complexSpectrum>
	  </userMechanicalimpedance>
	 </mechanicalimpedance>
          

	freq;	Zreal;			Zimag
	100;	-482.3e3;	-65.1e3
	125;	-480.5e3;	-77.3e3
	160;	-479.6e3;	-90.0e3
	200;	-487.6e3;	-95.4e3
	... ...
	4000;	9734.0e3;	-1797.1e3