External noise

Train Noise Expert allows the calculation of external noise levels at various receiver locations. Predictions can be made as static ‘standstill’ or under train pass-by conditions.

External noise calculation

External noise calculations are performed for each situation by selecting in the menu ‘Project’ > ‘Run external noise calculation’. This will first create a new results folder in the project folder, and then prepare the calculation input files for each situation. The calculation procedure will then be called for each situation in turn.

The progress of the calculation is shown in the run log. Please note that complex models could take 5-10 minutes per situation to calculate.

Once the calculation has completed, the results will be presented in the results view.

A calculation is assumed to be ‘standstill’ if the train speed is set to a value of 0.0 km/h. For speeds greater than 0.0 km/h, the ‘pass-by’ calculation is performed.

Note, the “time increment” parameter (see info for the situations model object) is an important aspect of the calculation. This affects how many time steps will be performed in the calculation, which influences the time taken for the calculation. In addition, in the results, the Lmax parameter is the highest sound pressure level recorded over all of the time steps, and therefore this can be affected by time increment setting.

Fit group to measurements

When measured sound pressure level data is available for individual sources, or groups of sources, this can be used to refine the model data.

First ensure that the source is assigned to a group. Also ensure that a receiver has been created that represents the location of the sound pressure level measurement. Then select Project > ‘Fit group to measurement…’

In the window that appears, select the situation for the train, track, operating speed and environment that are representative of the measurement conditions. Also select the group that you would like to fit to the measurement data, and the appropriate receiver.

There are two fitting modes that may be selected; the overall L_Aeq,Tp value, or the ¹⁄₃ octave spectrum L_eq,Tp value.

For the overall L_Aeq,Tp, specify the value and click ‘OK’. The calculation routine will calculate the sound pressure level at the selected receiver for the specified group and conditions. It will then determine the difference between the calculated and specified L_Aeq,Tp values, and prepare to apply the correction to all sources in the specified group. This correction is applied by means of adjusting the insertion gain parameter for each source in the group. All frequencies will be adjusted by the same amount. (n.b. A dialog will display to request confirmation before applying this correction.)

For the spectrum L_eq,Tp, specify the values for each ¹⁄₃ octave frequency band and click ‘OK’. Note that the frequency bands may be populated by pasting csv data (20-20,000 Hz, e.g. from a spreadsheet row) and clicking the ‘set’ button. The calculation routine will calculate the sound pressure level at the selected receiver for the specified group and conditions. It will then determine the difference between the calculated and specified L_eq,Tp values in each frequency band, and prepare to apply the correction to all sources in the specified group. This correction is applied by means of adjusting the insertion gain parameter for each source in the group. The amount of adjustment will be frequency dependent. (n.b. A dialog will display to request confirmation before applying this correction.)

Export for noise mapping

Noise map protocols require source emission data for specific heights, and Train Noise Expert offers the capability of exporting detailed source models to data that is compatible with strategic noise mapping software. This is currently a feature in Beta (see Application preferences to enable Beta options), and requires the licence for external noise calculations.

To export data for noise mapping, first ensure that results from an external noise calculation are loaded. Then please select Project > ‘External noise’ > ‘Export for noise mapping…’. The following settings dialog will appear.

First the specific situation is selected by the user, and then a set of line source transfer function (abbreviated as ‘line Xfer’ in the user interface) calculations are performed. For this, line sources are assigned a reference sound power level and are assumed to be the full length of the train, and are positioned at the centre of the track, at heights of 0, 0.5, 2, 3, 4 and 5 metres above the railhead. Corresponding sound pressure levels are calculated for all receivers in the situation.

Once the line source transfer functions have been determined, the user is able to select one of the noise mapping protocols, an optimised receiver, and to calculate/export the contributions from each source height accordingly. During this procedure, the contributions of each source in the model are apportioned (on an energetical basis) to the closest defined heights for the appropriate noise map protocol. These are as follows:

• HARMONOISE/IMAGINE: 0.0, 0.5, 2.0, 3.0, 4.0 metres above railhead.
• Dutch RMR: 0.0, 0.5, 2.0, 4.0, 5.0 metres above railhead.
• CNOSSOS-EU: 0.5, 4.0 metres above railhead.
• TNPM: 0.0, 0.5, 2.0, 4.0, 5.0 metres above railhead.

Train Noise Expert calculates the noise contributions from each of the source height categories, and computes the corresponding sound power level per metre for each source height using the line source transfer functions.

In the case of Dutch RMR, a +30 dB correction is applied to account for the fact that Dutch RMR calculations typically assume emission values in terms of sound power level per kilometer.

If “Export rolling terms” was selected (CNOSSOS-Rail or HARMONOISE/IMAGINE only), then rolling noise terms are exported in terms of roughness, contact filter, and wheel/track transfer functions (i.e. sound power per unit roughness).

If “Include rolling noise” is selected, then the rolling sound powers are included in the derived line sources. If “Export rolling terms” is selected, then usually this should not be selected, in order to prevent duplication of rolling sources in your noise map model.

Data is exported to .CSV files in the corresponding results folder of your project.

Currently there is no standard procedure for easily importing this type of data into noise mapping software. We are liaising with providers to encourage a standard protocol to be established in the future, but for now it is hoped that data from the .CSV files will be able to be pasted in to the appropriate fields in the noise mapping softwares’ interfaces. Please let us know about how you get on with this feature, so we can refine/improve it in future versions.