Time Domain Transfer Path Analysis: Listening to Paths!
After creating a Transfer Path Analysis (TPA) model that characterizes the noise and vibration of a structure, it is possible to listen to it!
Why listen to a TPA model? Listening interactively to different paths can help gain additional insight versus traditional assessments of sound. For example:
2D Graph versus Sound: Just because a potential modification shows a large drop in decibel level in a two dimensional graph (for example, amplitude vs rpm), it does not mean that it is better sounding. Listening to the sound before and after modification helps assess the changes with more insight.
Transient sounds: Transient sounds, for example injector tick, are not well represented in a frequency domain model. Having the time history contribution allows metrics like time varying loudness to be used to assess the sound. Time varying loudness cannot not be calculated from a frequency based curve.
This article explains the steps for listening to a Transfer Path model: 1. Background 2. Getting Started 3. Creating Time Histories 4. Navigator Attributes 5. Listening to Path Contributions
To create time histories from a Transfer Path model, a frequency domain model is used in conjunction with the measured time histories. The measured input time histories are applied to the model in the frequency domain to calculate output time histories of the various path contributions (Figure 1).
Figure 1: Measured input time histories are applied to the model in the frequency domain to calculate output time histories.
This frequency domain Transfer Path Model is typically composed of:
Frequency Response Functions (FRFs): FRF functions characterize the TPA structure. Force is the input (or denominator) of the FRF function, and the output (numerator) is usually either acceleration or sound pressure.
Operating Data: Orders, spectrums, or other functions taken from the operating condition of interest. These might be directly measured forces or indirect measurements (acceleration or otherwise) used to derive the force.
The operating time histories of the model inputs are also needed. The time histories must have the following characteristics:
Acquired Simultaneously: All channels must have been acquired at the same time to have consistent phasing in the time domain. They cannot be acquired in batches and patched together.
Consistent Sampling Rate: All channels used in the time domain model must have the same sampling rate. Having channels in the same recording with different sampling rates will not work with the Simcenter Testlab software.
The frequency range of the time history will either be limited by the frequency range of the FRF measurements or the operating data measurements, whichever is lower. Generally speaking, it is easier to measure to a higher frequency on the operating data than the FRF measurement. Some further guidelines:
The FRFs are often acquired using an impact hammer, and it is difficult to measure to a high frequency range depending on the hammer tip used. See the knowledge article "Selecting a Modal Impact hammer tip".
If the TPA model is composed of orders, it might be prudent to include additional orders (higher and lower) so that the frequency content of the model covers a suitable range for sound playback.
The steps for doing a time domain transfer path listening in Simcenter Testlab are covered next.
2. Getting Started
In Simcenter Testlab Transfer Path Analysis, select the input time history and add it to the Input Basket as shown in Figure 2.
Figure 2: Right click on the input time history throughput file and select “Add to Input Basket”.
From Simcenter Testlab main menu, turn on the following add-ins (Tools -> Add-ins): Time Domain TPA (35 tokens) and TPA Synthesis (62 tokens) as shown in Figure 3.
Figure 3: From Simcenter Testlab, choose “Tools -> Add-ins” and turn on “Time Domain TPA” and “TPA Synthesis”.
Both add-ins do not need to be on at the same time. The “Time Domain TPA” add-in is used to create path contribution time histories. The “TPA Synthesis” add-in is used for interactive listening after the path contribution time histories have been calculated. Both options rely on Transfer Path Analysis (98 tokens) or OPAX (73 tokens) being activated and on.
3. Creating Time Histories
The first add-in to be used is “Time Domain TPA”. After turning on the add-in, a “Time Domain” subsheet is added to the TPA Results workbook (Figure 4).
Figure 4: Time Domain sub sheet (lower right) is added to TPA Results workbook.
In the Time Domain subsheet (Figure 5), do the following:
Read Input Basket: Click on the read input basket button. The throughput file should appear just below the button, and the red column of PointId information (lower left) should turn green with the letter X in each box.
Cell Selection: Click on any of the cells with the green background to view the time histories. Use the arrow key to move up and down the cells to view the time data.
Figure 5: The “TPA Results” workbook can be used to visual the input time histories.
Next, the time history data acquired to build the original model (direct forces, indirect accelerations, etc) will be transformed into path contribution time histories.
In the “TPA Results” workbook (Figure 6):
Select the desired frequency range of the path contribution time histories (use the default if nothing specific is desired)
Press the “Calculate” button.
Figure 6: After pressing the “Calculate” button, path contribution time histories are created.
The results can be viewed in the Navigator worksheet.
4. Navigator Attributes
The path contribution time histories will be found in a new analysis that ends with the extension “TTPA” (assuming the defaults are used).
When viewing the time data, there is a mixture of the originally acquired data and the newly calculated contribution time histories in the same file.
To tell the difference, some attribute columns can be added in the Navigator (Figure 7) to help differentiate the time histories. Right click at the top of the columns and choose “Select Columns” to add the attributes.
Figure 7: Right click at the top of columns and choose “Select Columns”.
In the “Select Columns” menu (Figure 8), do the following:
Set the Origin to Testlab, and the Type to Block.
Enter “tpa” in the Quick Find area.
Choose the attributes “TPA Path” and “TPA Result Type” and use the arrow to put them in the selected field.
Figure 8: Put the attributes “TPA Path” and “TPA Result Type” in the Selected area.
Press OK when finished.
Possible attribute values include (Figure 9):
Total Contribution: The calculated total time history at the target location (Point ID field). This can be used for comparison to the original measured time history at the target location.
Partial Contribution: Many of the Point ID fields or target location names are the same. Partial Contribution indicates a path contribution from a location (TPA Path) to the target location. For example, the contribution of a left engine mount (TPA Path) to the driver ear (Point ID).
Load-MI: To calculate the contribution of a path, the force time history must be calculated. This is indicated by the LOAD-MI label.
Figure 9: Attributes on the time histories include Total Contribution, Partial Contribution, and Load-MI.
With the time histories calculated, they can be listened to interactively.
5. Listening to Path Contributions
With the “TPA Synthesis” add-in set to ON (Figure 10):
Click on the Definition subsheet located at the top of the screen.
If desired, select a smaller time segment for listening. Use the Double X cursor and press the “Apply segment selection” button.
Figure 10: In the Definition subsheet of the “TPA Synthesis” workbook, a partial time segment can be selected for playback.
With the time segment of interest selected, switch to the Replay subsheet, and do the following (Figure 11):
Choose the Replay subsheet at the top of screen
Pick the Transfer Path model for listening. The top row is the originally measured data, and the model is contained in the rows below. Any models created with “TPA Component Editing” will also be listed. While replay is in session, click on the different models to hear the differences.
Press the Replay button to start listening.
Use the check boxes on the right side to turn paths on and off while listening. Paths cannot be turned on and off on the original measured data, they can only be turned on and off on Transfer Path models.
Figure 11: Use the Replay subsheet of “TPA Synthesis” workbook interactively listen to the path contributions.
Instead of turning on and off individual paths, if the groupset feature of Simcenter Testlab was used, entire sets of paths can be turned on and off at the same time.