Simcenter Testlab Modal Validation

2024-12-27T15:43:13.000-0500

Summary

Details

Direct YouTube link: https://youtu.be/pM4-lwlV1UQ

When performing an experimental modal analysis, Simcenter Testlab contains tools to perform checks and validate everything from the quality of the Frequency Response Functions (FRFs) collected to the natural frequency and mode shape selections.

This article gives an overview of the modal validation tools available in Simcenter Testlab:
1.   FRF Measurement Validation
1.1 Data Viewing
1.2 Driving Points
1.3 Coherence
2.   Modal Curvefitting
2.1 Band Selection
2.2 Mode Indicator Functions
2.3 Automatic Mode Selection
2.4 Complex versus Real
3.   Modal Synthesis Worksheet
3.1 Missing Modes
3.2 Error and Correlation
4.   Modal Validation Worksheet
4.1 Modal Assurance Criterion (MAC)
4.2 Complexity Check

1. FRF Measurement Validation

The “Validation” workbook is available in most structural data acquisition software (Impact Testing, MIMO FRF Testing,…). The Validation worksheet is helpful in evaluating the Frequency Response Functions (FRFs) that were measured during a modal test.

1.1 Data Viewing

Using a geometry of the test structure, the locations where measurements were acquired can be quickly seen.

In the upper left of the “Validation” workbook, under “Function Type”, select the type of data to view as shown in Figure 1.

Figure 1: The measurements selected in the upper left of the Validation can either be displayed interactively (table middle left) or all seen on the geometry (lower left) at once to find missing measurement locations (measured locations are cyan color).

In the Validation worksheet, data can be viewed:

Interactively: Using the scroll area on the middle left side.
All on the Geometry: Easily view if any measurements were missed on the test geometry.

To view the measurements interactively, highlight the measurement location of interest to view on the middle left. If desired, use the Arrow keys to move up and down to view different locations. If the “Display on Geometry” is selected in the lower left corner, the location will be highlighted on the geometry and node name shown.

To find if any measurements are missing, in the lower left corner, turn on “Display on Geometry” and select “All Measured Points” from the pulldown. All measured locations will be shown in light blue in the geometry display. Unmeasured locations will be tan colored.

1.2 Driving Points

The locations that force inputs were applied for the modal test can also be viewing on the test structure geometry.

In the lower left of the “Validation” worksheet, turn on “Display on Geometry”. Select “Reference Points” from the menu pulldown as shown in Figure 2 below.

Figure 2: The blue nodes (circled in green) are the input (or reference) force locations for the modal test.

The input or reference locations will turn to blue in the geometry display. Note that all measurement locations can be seen by right clicking in the display and choosing ”Deformed Model -> Nodes -> Markers”.

1.3 Coherence

In the upper left, choose “FRF + Coherence” from the “Function Type” pulldown and in the lower left select “Coherence values” as shown in Figure 3 below.

Figure 3: Coherence values can be visualized on geometry in the Validation worksheet. Red nodes indicate a high coherence of “1” at the wing locations. Coherence values are lower at tail (green) and fuselage (blue) which are located further away from the exciter locations.

A scale that shows the values of the coherence can be turned on in the geometry display:

Right click: Select “Coloring -> Show” under the Scale section to see the scale.
Right click: Select “Coloring -> Edit” under the Scale section to adjust the limits.

The frequency band used to calculate the coherence values is set in the lower left corner of the screen.

More information on modal data acquisition modules in these knowledge articles:

The “Validation” worksheet is used to check the measured Frequency Response Functions (FRFs). There are more validity checks that can be used when calculating modal parameters (frequency, damping, mode shapes) from the FRFs that are covered in the next section.

2. Modal Curvefitting

There are several tools and tips that can help when estimating modes (frequency, damping, mode shapes) from Frequency Response Functions (FRFs). These tools can help ensure valid modes are generated and selected.

2.1 Band Selection

In the modal curvefitter, like Polymax, the selection of the bandwidth for curvefitting is the first step. To get the best resulting stabilization diagram, place the cursors at anti-resonance locations as shown in Figure 4 below.

Figure 4: Place the double X cursor at the anti-resonances of the FRF sum in the bandwidth selection step of the Polymax modal curvefitter.

Selecting at the anti-resonance ensures that the complete data for a mode is used in the modal parameter estimation. It is not truncated.

2.2 Mode Indicator Function

While a normalized summation of all FRFs is shown in the curvefitter as a guide, another helpful function is the Mode Indicator Function (MIF). Like the sum, all FRFs are used to calculate the MIF, but singular value decomposition is also employed to give a indication of the natural frequencies of the structure.

To view a Mode Indicator Function, turn on the check box in the lower left corner of the Polymax curvefitter as shown in Figure 5.

Figure 5: A primary MIF (red) indicates the natural frequencies in the structure. The secondary MIF (green) indicates where a second mode at the same frequency is present.

There are several types of MIF functions available. The “Complex MIF” has peaks that correspond to natural frequencies. Other MIFs “dip” or bottom out at natural frequencies.
Under the “Advanced” button, it is possible to have as many MIF functions as references used in the test. There is a “primary” MIF, a “secondary” MIF, etc. These MIFs do not correspond to the physical input locations. Instead, they indicate if there are two modes at the same frequency. Mathematically, without the additional references, the additional MIFs cannot be calculated.

Having multiple modes at the same frequency Is common in symmetric structures (for example a tube, disk, etc). Not only will there be a peak in the secondary MIF that corresponds to a peak in the primary MIF, but there will also be two columns of “s” in the stabilization diagram.

Both the sum and the MIF functions are used as guides to the location of modes when using the stabilization diagram which is covered in the next section. The modal selections in the stabilization diagram are calculated from all of the individual FRFs, not from the sum or MIF.

2.3 Automatic Mode Selection Tool

After selecting the band and choosing a function (sum or MIF) as a guide, the next step is to select the modes of the structure. This is done in the stabilization diagram.

Under “Tools -> Add-ins” there is an add-in called “Automatic Modal Parameter Selection” which can assist in the selection of modes from the stabilization diagram (Figure 6).

Figure 6: Add-in menu of Simcenter Testlab

Turn the add-in on adds a button called “Select Poles” below the stabilization diagram (Figure 7).

Figure 7: With the "Automatic Modal Parameter Selection" add-in turned on, modes can be automatically selected from the stabilization diagram by choosing the "Select Poles" button at the bottom of the screen.

Choose “Select Poles”. The modes that were automatically selected will be shown in a list on the left side of the screen. They will have a “AMPS” next to their frequency and damping.

The modes should be treated as suggestions. The user should still evaluate the selections further. Also, the automatic modal selection missed a mode, the user can select modes from the stabilization diagram.

More information on modal curvefitting in the knowledge articles:

2.4 Complex versus Real Shapes

In the upper left of the mode shape calculation, “Real” or “Complex” can be selected as shown in Figure 8.

Figure 8: Complex or Real can be selected in the upper left of the modal curvefitter “Shapes” step.

The “Complex” versus “Real’ setting affects the mode shapes that are calculated during the modal curvefitting process. An example of the a shape calculated with the "Complex" setting versus the "Real" setting for the same modal frequency is shown in Figure 9 below.

Figure 9: The phasing between different locations is not perfectly in and out of phase in a complex shape (left) versus a real mode shape (right).

When “Real” is selected, the shape that the modal curvefitter fits based on the FRF data is forced to be either in or out of phase at every measurement location. This is done even if the underlying FRF data does not support this assumption.

When should "Real" or "Complex" be selected?:

Simple metal structures mode shapes are typically all in or out of phase. These are called "normal modes" and correspond to simulation results from Nastran Solution 103, etc. Using the "Complex" or "Real" setting will not change the shape (or they will be very similar). In this case, for best correlation to simulation results use the "Real" setting to cleanup any small complex behavior.
Some structures can exhibit non-linear or complex behavior. For example, loosely bolted joints or rubber couplings may cause complex behavior. In this case, the true mode shape might be distorted by using "Real". "Complex" should be used instead.
Sometimes the underlying FRF data may not be well acquired. If a normal mode shape is expected, setting "Real" will ensure a normal mode shape. Ideally, the FRF data would be re-acquired in this case before forcing the shape to be normal using the "Real" setting.

More knowledge articles on complex mode behavior in the articles:

Simcenter Testlab: Multi-Run Modal

3. Modal Synthesis

The “Modal Synthesis” worksheet is a standard part of the modal analysis add-in of Simcenter Testlab. It is used to compare a predicted Frequency Response Function (FRF) based on the modes selected versus the measured Frequency Response Function at the same location (Figure 10).

Figure 10: The “Modal Synthesis” worksheet compares the measured FRF (red) with a predicted FRF (green) based on modal curvefit selections (upper left). One measurement location is shown at a time which is controlled by the highlighting a row in the FRF table on the left side. In this case, four displays are shown because four reference inputs were used in the test.

In the “Modal Synthesis” worksheet, the modes selected in the stabilization diagram are listed on the upper left side. The synthesized FRFs (green) are shown against the original measured FRF (red). The measurement location is selected via the “FRF table” on the middle left of the screen.

3.1 Missing Modes

One use of “Modal Synthesis” worksheet is to see if any modes were missed during the selection process in the stabilization diagram. In Figure 11 below, some modes in the 5000 Hz range are missing:

Figure 11: Modes are missing in the 5000 Hz range. The measured FRF (red) contains peaks that are not in the synthesized FRF (green).

There are peaks in the measured FRF (red) that are not present in the synthesized FRF (green). In this case, the user should return the stabilization diagram and select appropriate modes in the frequency range where they are missing.

3.2 Correlation and Error

Each predicted FRF is compared to the measured FRF with a correlation and error percentage. The correlation and error percentage are displayed above each FRF as shown in Figure 12.

Figure 12: Correlation and error percentage is displayed above the measured FRF (red) and predicated FRF (green).

The percentages mean:

Correlation: How close the shape of the predicted and actual FRFs match. This number should be as close to 100% as possible.
Error: The error percentage is the difference in amplitude between the predicted and measured FRFs. It should be as close to 0% as possible.

The synthesized/predicted FRF for different measurement locations can be viewed by highlighting the corresponding row in the FRF table. Once a row is highlighted, the arrow keys can also be used to change the location being viewed (Figure 13).

Figure 13: Highlight a row corresponding to a measurement location in the “FRF table” to view

A table of correlation and error percentages for all measurement locations can be generated. First turn on “Generate correlation-error list” under “Synthesize parameters” in the FRF table as shown in Figure 14.

Figure 14: Under “Synthesize parameters” in the FRF table turn on “Generate correlation-error table”

To generate the table of the correlation and error percentage for all measurement locations, press the down arrow in the lower left corner of the FRF table as shown in Figure 15.

Figure 15: Press the down arrow to generate a list of correlation and error percentages for all measurement locations.

The software will step thru each location one at a time. To see the table, it must finish stepping through all the locations. The “Speed:” control can be used to make the process faster.

Once the correlation and error table appears on screen, it can be written to Excel via copy/paste. Click on the “Output name” in the upper left of the table to highlight all fields and use “Ctrl-C” and “Ctrl-P” to copy to Excel.

4. Modal Validation Worksheet

The “Modal Validation” worksheet contains further checks that can be performed on the modal selections from the stabilization diagram.

4.1 Modal Assurance Criterion (MAC)

The Modal Assurance Criterion (MAC) calculation indicates the similarity of the mode shapes that were selected via the stabilization diagram. Mode shapes that are similar will have a value close to 100%, mode shapes that are difference and unique will have close to 0%.

Press the “Auto-MAC” button on the left side of the “Modal Validation” worksheet to perform a MAC calculation within a mode set as shown in Figure 16.

Figure 16: Modal Assurance Criterion matric (top) in the “Modal Validation” worksheet.

In the display area, choose “Matrix/Geometry” in the upper right. A MAC matrix is shown containing the values of the MAC for every mode in the set:

Diagonal: Along the diagonal, the values are 100% and shown as tall bars in red. These are the mode shapes compared to themselves which are always 100% alike.
Off-Diagonal: Modes compared to other modes are on the off diagonal. Normally these values should be close to zero (blue) if every mode shape is unique.

A high off diagonal (orange/yellow) indicates that some modes have similar shapes at different frequencies. This should not be possible.

If the off-diagonals are high, this might indicate:

Some modes were selected twice. Solution would be to remove them from the modal results.
Not enough measurement locations (nodes) were measured to uniquely identify the shapes. Solution is to measure mode locations on the structure.

Note that clicking on a bar in the MAC matrix will display both modes in the left-right display below.

More information on Modal Assurance Criterion in the knowledge article:
• Modal Assurance Criterion (MAC)

4.2 Complexity Check

Many times, normal or real modes are expected. The amount of complexity in the shapes can be accessed using the “Complexity” button on middle left side (next to the Auto-Mac button) of the Modal Validation worksheet. Using the "Table" view, the complexity check results look as shown in Figure 17:

Figure 17: Complexity check results in Modal Validation worksheet.

List of metrics:

Mode Over Complexity Value (MOV): The MOV index should be high (near 100 %) for a high quality mode.
Mass Sensitivity: If mass is increased, resonance frequency should decrease. “-” sign. If a reference has one or two modes have “+” sign then perhaps the mode is not well excited by that reference. If there are lot of “+” signs, then perhaps the reference direction is not correct and needs to be reversed.
Mode Participation (MP): High value (near 100%) means the mode is well excited by the input location. Modal Phase Collinearity (MPC): An indicator called the Modal Phase Collinearity (MPC) is an indicator that checks the degree of complexity of a mode. The closer to 100%, the less complex the mode.
Mean Phase Deviation (MPD): Another indicator for the complexity of unscaled mode shape vectors is the Mean Phase Deviation (MPD). This index the phase scatter of a mode shape. The MPD value should be low (near 0 degrees) for real normal modes.
Scatter: The last column lists the phase scatter for each mode as being either "high", "low" or "?". Scatter is a composite metric based on MPC and MPD. "Low" scatter is usually preferred. The ratings "high", "low", etc are determined by Table 1.

Table 1: The Scatter rating is a composite metric based on the MPC and MPD values.

Hope this article is a helpful guide for modal validation! Questions? Email peter.schaldenbrand@siemens.com.

Related Links:

KB Article ID# KB000156298_EN_US

Summary Details

Simcenter Testlab Modal Validation

Summary

Details

KB Article ID# KB000156298_EN_US

Contents

Associated Components