Cancel
Showing results for
Did you mean:

Simcenter Testlab Modal Analysis: Modification Prediction  Siemens Phenom

(view in My Videos)

Think design modifications can only be made on Finite Element models? Think again…!

After performing an experimental modal analysis and calculating a set of modes, each mode has a mass and stiffness matrix that can be modified.

Simcenter Testlab (formerly called LMS Test.Lab) Modal Analysis Modification Prediction can be used to:

• Add or subtract mass at a node or point
• Increase or decrease stiffness between two nodes or points
• Create tuned absorbers targeted to a frequency

After creating a group of modifications, a new set of mode shapes and modal frequencies is calculated that incorporates the changes.

Spring-Damper and Mass Modification Background

Consider a single degree of freedom system, consisting of:

• Mass (m)
• Stiffness (k)
• Damping (c)

The natural frequency (wn) is equal to the square root of the stiffness over the mass as shown in Figure 1. Figure 1: Single Degree of Freedom Mass-Spring-Damper system

A modal frequency can be increased by:

• decreasing mass
• increasing stiffness

This holds true for all structures, even more complicated ones.

A ‘Spring-damper’ modification can be used to alter the stiffness between two points as shown in Figure 2. Figure 2: Modified mass-spring system with increased stiffener

By increasing the stiffness of the spring, the modal frequency will shift higher as shown in Figure 3. Figure 3: Frequency Response Function of original system and system with increased stiffness

How would increasing stiffness address vibration issues? For example, if vibration from driving on rough road ranged from 1 to 20 Hz, increasing the first body modes beyond 20 Hz would reduce vibration experienced by the driver.

In addition to the stiffness, the damping can also be changed.

Tuned Absorber Background

A tuned absorber is a secondary mass-spring system that is added to an existing mode as shown in Figure 4. Figure 4: Tuned absorber (m2, k2) applied to mass-spring system (m1, k1)

A tuned absorber takes the original frequency of the original system and divides it into two modes.  The frequency of the first mode is lower than the original system.  The frequency of the second mode is higher than the original system as shown in Figure 5. Figure 5: Original System versus Tuned Absorber system

The mode shape of the lower frequency would have both the original system mass (m1) and the tuned absorber mass (m2) move back and forth in phase. The two masses would move back and forth out of phase in the higher frequency mode. This is illustrated in Figure 6. If the tuned absorber mass and stiffness is carefully selected, the motion on the original system can be forced to zero by the absorber. Figure 6: Original System versus Tuned Absorber system

How can a tuned absorber be used to abate a noise and vibration issue?  Consider a vehicle where the combustion frequency of engine idle excites a bending mode of the steering wheel column. A tuned absorber could be placed on the end of the steering wheel.  This would place one mode of vibration at a frequency lower than the idle which would never be excited.  The higher mode could be placed at an engine combustion frequency that the vehicle does not commonly operate at, like 30 mph.

Tuned absorbers are useful when an existing mass can be used.  For example, the airbag module at the end of the steering wheel is an existing mass that could be sprung to create a tuned absorber.  This modification would not add mass to the overall vehicle (which would have an adverse affect on fuel efficiency).  It might also be a cheaper modification than stiffening the steering column to avoid the resonant situation.

When using experimental modes for modification prediction, the following should be considered:

1. Calibration – It is important that the proper calibration was used during data acquisition. This includes both the input and response transducers.  If the accelerometer calibration was off by a factor of 100, a 2 kilogram modification could act like a 200 kilogram modification.
2. Dimensions - Proper dimensions should be used when creating the geometry. The dimensions can affect the modification prediction results.
3. Out-of-Band Modes – To have accurate modal predictions, it is advisable to have at least one mode below and one mode above the frequency of interest for the modification. It is even better if several modes above and below the frequency of interest is included. When using a modal model for a limited frequency band it is possible that important structural modifications would generate modes with a natural frequency outside the range of this frequency band. Since the original modal models are not valid at these frequencies, the predicted results will not be very reliable.
4. Proper Frequency Shift - If adding mass, modes should shift down. If they shift up instead, there is a high probability that there is a sign direction convention problem.  For example, on the input point, the +Z direction may have been substituted for the –Z direction. Check the directions again.

Getting Started with Simcenter Testlab Modification Prediction

Under “Tools -> Add-in” from the main menu, select “Modification Prediction”.  If using Simcenter Testlab tokens, it requires 23 tokens total (Figure 7). Figure 7: Tools -> Add-ins -> Modification Prediction

A new worksheet called ‘Modification Prediction’ appears at the bottom as shown in Figure 8. Figure 8: Tools -> Modification Prediction on worksheet

There are two minor worksheets at the top of the ‘Modification Prediction’ worksheet:

• List Modifications – Create the set of modifications to be applied
• Predict Modes – Calculate and view new mode set with modifications

Select a set of modes for modification in the upper left corner of the ‘Modification Prediction’ worksheet. Drag and drop modal frequencies over the geometry to view the shape as shown in Figure 9 Figure 9: Select mode set in upper right, drag and drop to geometry

Viewing the mode shape aids in deciding appropriate modifications to alter the modal frequency.

Making a Modification: Spring-Damper

Choose a modification of interest.  For example, this vehicle body has a torsion mode as shown in Figure 10. Adding additional stiffness around the windshield can make the torsion mode frequency higher. Figure 10: Vehicle body torsion mode

To add a stiffener, select ‘Add Spring-Damper’ at the bottom of the screen as shown in Figure 11. Figure 11: ‘Add Spring-Damper’ button

In the ‘Add Spring Damper’ menu, define two connection points and a spring constant value. By default, with the "According to connection line" setting, the XX stiffness is added axially between the two points as shown in Figure 12. Damping can also be added. Figure 12: Select mode set in upper right

It is not necessary to type the node names into the ‘Attachment point’ menu fields.  You can simply click on the node in the geometry and it will fill into the ‘Attachment point’ automatically, as shown in Figure 13. Press the ‘Apply’ button when finished. Figure 13: Select mode set in upper right

Multiple modifications can be made.  In this case, two ‘Spring-damper’ elements were added to increase the torsion mode of the vehicle as shown in Figure 14. Figure 14: Multiple ‘Spring-Damper’ modifications

Click on the ‘Predict Modes’ tab at the top of the ‘Modification Prediction’ worksheet as shown in Figure 15. Press the ‘Calculate’ button to create a new set of modes, with the modifications applied. Figure 15: Press ‘Predict Modes’ button then the ‘Calculate’ button to calculate a new set of predicted modes

The Frequency Response Functions (FRFs) of on the points of interest are displayed.  Green is from the modified mode set, while red is from the original mode set.

Making a Modification: Tuned-Absorber

At 27.39 Hz, there is a roof pumping mode as shown in Figure 16. A tuned absorber can be applied to this mode. Figure 16: Roof pumping mode

Click on the ‘List Modifications’ minor worksheet. If desired, eliminate any previous modifications by highlighting the entire row and pressing the ‘Delete’ button as shown in Figure 17. Figure 17: Highlight complete row and press ‘Delete’ button to remove a modification

Click on the ‘Add Tuned Absorber’ button.

In the tuned absorber menu (Figure 18):

• Select an attachment point (can be done by clicking on node in geometry display)
• Enter a tuned absorber mass
• Enter a frequency to target with the absorber
• Then press ‘Tune’ button to calculate a stiffness and damping value
• Press ‘Apply’ when finished Figure 18: ‘Modify Tuned Absorber Modification’ menu

Press ‘Predict Modes’ tab to create a new set of modified modes.  Just above the ‘Calculate’ button, on the middle left, enter text in the ‘Processing data’ field to name the new data set. Press the ‘Calculate’ button to apply the tuned absorber as shown in Figure 19. Figure 19: ‘Predict modes’ and ‘Calculate’

The new modes and their frequencies are listed in the lower left of the screen.  The FRF with the red line is based on the original set of modes.  The FRF with the green line includes the tuned absorber modification.

Tuned Absorber Trivia

From a dynamics point of view, skyscrapers are equivalent to long metal beams coming out of the ground.  They have low frequency modes of vibration excited by the wind.  The top of a skyscraper can move many feet.  Many have tuned absorbers to help reduce the movement/vibration.

The Taipei 101 skyscraper contains the world's largest and heaviest tuned mass dampers, at 660 metric tons (730 short tons) as shown in Figure 20. Figure 20: Tapei 101 (left) and tuned absorber (right)

The tuned absorber is viewable by the public on an indoor observation deck at the top of the skyscraper.  It cost an estimated \$4 million to build.

Conclusion

Use Simcenter Testlab Modal Analysis Modification Prediction to simulate product design modifications on experimental modal analysis results.  The following modifications can be done:

• Add or subtract mass at a node or point
• Increase or decrease stiffness between two nodes or points
• Create tuned absorbers targeted to a frequency

After creating a new mode set, it is interesting to use the Modal Assurance Criterion (MAC) to compare the original set of modes to the modified modes as shown in Figure 21. Figure 21: Modal assurance criterion analysis of original versus modified mode set

Using the MAC analysis, it is possible to see the effects of modifications on modal frequencies and mode shapes. In the MAC table of Figure 21, red indicates modes that are 100% alike.

• For these modes with a MAC of 100%, the frequencies were changed by the modifications, but the mode shape was not.
• The orange color indicates modes that are 80% alike. In this case, we can see that the second mode of the original structure at 27.4 Hz was shifted upward to 30.6 Hz, and the shape was significantly changed.

Questions?  Email  peter.schaldenbrand@siemens.com or contact Siemens PLM GTAC Support.

Other Articles:

• ,
• Structural Dynamics Testing Creator

Hi Friends!

It's possible to use this option in OMA cases?. I'm acquiring output measuremetns without input measurements (in a chasis on the bus, only excited for the main egine).

Greetings

Alvaro.  Siemens Phenom

Hello Alvaro,

Operational Modal shapes and Operating Deflection shapes cannot be used in Modification Prediction.   Since OMA does not measure or use the forces these modes are unscaled and modal mass is zero / undefined.

Kevin Enthusiast

Hello,

" The modes don't have driving points, therefore this type of scaling can't be done."

What does it mean?

I proceed "modification prediction"  with the results of experimental modal testing with two shaker.

Regards,

Kim  Siemens Phenom

Hello Kim,

If you don't have a driving point, ie, one location with the same Point ID with Force and Acceleration the modes cannot be scaled.   After your mode(s) is / are calculated, right click on one to get its properties.  Look at the Mode Details.  Modal Mass and Modal Stiffness are likely zero and if so, then some tools like Modification Prediction and Modal Scaling cannot be done.  Also some mode complexity checks on the Validation worksheet like mass sensitiity and MOV cannot be done.

By having a driving point with acceleration and force at one location that can be used to determine mass and thus scale the mode.  Since the equations of motion are mass*acceleration + damping*velocity + stiffness*displacement = force we can calculate the mass at that location if we have the rest measured or estimated through modal analysis.

Without the driving point we still get frequency, damping and the mode shape but no modal mass, no modal stiffness and no scaling.  Siemens Phenom

I was reviewing this article and wanted to explain a few things:

1)  How to check the mass sensitivty of the modes or as PJS states "Proper Frequency Shift - If adding mass, modes should shift down. "

2)  How to check the linearity or complexity of the modes.

3)  How to determine an appropriate masss for the tuned absorber.

(1)  On the Modal Validation worksheet press the Complexity button to calculate the Compact Mode Complexity which will list one complexity value per mode.  Make sure the Modal Validation worksheet is showing the Table by selecting the Table / Geometry view.

See how Mass sensitivity is mostly negative for all the modes and for each reference.  This means if I add a mass to the structure then the frequency of the mode will shift downward as expected.  The MOV shows what percentage of the responses will have a negative shift in frequency when the mass is added.  If Mass sensitivity and MOV shows the opposite of what is expected the reference channel direction is probably incorrect and should be edited and the mode shapes recalculated.

(2) Note here that the MPC (Modal Phase Collinearity) is close to 100% indicating we have a very linear mode shape or one without a lot of complexity.  If my modes were not very linear I would investigate as to why before proceeding with Modification Prediction.   A low MPC index indicates a rather complex mode, due either to local damping elements in the tested structure or to errors in the data measurement or to mistakes in the analysis procedure. (3)  PJS shows using 6 kg for the tuned absorber which has the following effect: That seems effective but a little bit overkill to add 6 kg.  What I would suggest is scaling the modes to determine what I call the effective mass of the mode and use 10% of that for the mass of the tuned absorber.  To do that:

On the Modal Validation worksheet, select to Scale to Unity Length, provide a new processing name and press the Scale button. When scaled to unity legth I consider the modal mass to be the effective mass of the mode.   Then right click on the mode of interest (27.397 Hz here) and get its Properties.  On the Details tab note that the Modal Mass is 6.8 kg.  This means for the tuned absorber we should use 0.68 kg and not 6 kg!

For more information on modal scaling, see FAQ432 "How is the scaling of the modes in LMS Test.Lab? " installed with the Test.Lab rev17 or Testlab rev 18 documentation or find it on the Global Technical Assistance center (GTAC) under the Solution Center. Let's try that and see the effect.    It's still effective but with much less mass, better damped amplitude and the neighboring mode is not effected much at all.  Contributors
Latest articles