App Note 12 www.vibetech.

com 2/8/2021

MEscope Application Note 12

Calculating ODS-FRFs from Bridge Impact Data
The steps in this Application Note can be carried out using any MEscope package that includes the VES-3600 Advanced
Signal Processing option. Without this option, the steps can still be carried out using the AppNote12 project, but the results
cannot be saved. These steps might also require MEscope software with a more recent release date.

APP NOTE 12 PROJECT FILE

• To retrieve the Project for this App Note, click here to download AppNote12.zip
This Project file contains numbered Hotkeys & Scripts for carrying out the steps of this App Note.
• Hold down the Ctrl key and click on a Hotkey to display its Script window
INTRODUCTION
In this Application Note, test data acquired during an impact test of the bridge pony truss shown below is processed in
MEscope to yield a set of Operating Deflection Shape Frequency Response Functions (ODS-FRFs). The ODS-FRFs are
then used to display the Operating Deflection Shapes (ODS’s) of the bridge truss in animation on a 3D model of the pony
truss.
Modal testing is ideally done under controlled excitation and time-invariant (stationary) conditions, using one or more
sources of excitation.
• To calculate Frequency Response Functions (FRFs), the excitation forces and their corresponding responses must
be simultaneously acquired
Excitation force is typically measured with a load cell; and vibration response is typically measured with an accelerometer,
although velocity or displacement sensors can also be used.
• In an output-only or operational modal test, the excitation forces are not measured

Model of a Bridge Pony Truss.

Page 1 of 8
App Note 12 www.vibetech.com 2/8/2021

For this App Note, a pony truss of a bridge was excited with a drop-weight impactor. The impact points are labeled on the
model below.
• An impactor was necessary to excite the lowest frequency modes of the bridge, which were below 15 Hz
• However, the impactor created peak forces that were far greater than any load cell could withstand
Since the impact force was not measured, FRF measurements could not be calculated. When the excitation force is not ac-
quired, a Transmissibility measurement is typically calculated instead. A Transmissibility (TRN) is defined as,
𝐃𝐅𝐓 (𝐫𝐨𝐯𝐢𝐧𝐠 𝐫𝐞𝐬𝐩𝐨𝐧𝐬𝐞)
𝐓𝐑𝐍 =
𝐃𝐅𝐓 (𝐫𝐞𝐟𝐞𝐫𝐞𝐧𝐜𝐞 𝐫𝐞𝐬𝐩𝐨𝐧𝐬𝐞)
• The DFT is the Discrete Fourier Transform
• A DFT is calculated from a uniformly-sampled time waveform using the FFT algorithm
• The reference response sensor must remain fixed throughout an output-only test
REQUIREMENTS FOR AN ODS
An Operating Deflection Shape (ODS) is any motion at two or more DOFs (points & directions) on a machine or structure.
• An ODS defines the relative motion between two or more DOFs on a machine or structure
▪ A frequency-based ODS is defined at a specific frequency
▪ A time-based ODS is defined at a specific moment in time
• Each shape component of an ODS must have the correct magnitude & phase relative to all other shape compo-
nents
To ensure that all measured responses have the correct relative magnitudes & phases, one of the following acquisition meth-
ods must be employed.
• Simultaneously acquire all responses
• Simultaneously acquire some Roving responses and a fixed Reference response
▪ A set of simultaneously acquired data is called a Measurement Set
MULTIPLE MEASUREMENT SETS
Most field vibration measurements are made with equipment that can only acquire a few channels of data at a time. There-
fore, multiple Measurement Sets of data are required.
• Each Measurement Set consists of one or more different Roving responses and the same Reference response
• To preserve the relative phase among all responses, the same Reference response (or responses) must be acquired
with all Measurement Sets
STATIONARY STRUCTURE
During a field test, the total time required to obtain all measurements can be substantial, and the machine or structure could
undergo physical changes from the beginning to the end of the test.
If the physical properties of the structure or machine change during a test, its vibration characteristics could also change. For
example, outdoor temperature changes due to sunshine can cause changes in material stiffness. Damping characteristics can
change during a prolonged test. Unmeasured excitation force levels can also change during a test.
• A structure is said to be stationary if the Auto spectrum of one of its responses does not change from the beginning
to the end of the test
When overlaid on one another, if two or more Auto spectra calculated for the same DOF of the test article are essentially the
same, the structure is said to be stationary.
• If the amplitude of an Auto spectrum for the same DOF changes during a test, the structure is still considered to be
stationary, but re-scaling is required to define ODS’s from the data

Page 2 of 8
App Note 12 www.vibetech.com 2/8/2021

HOW THE BRIDGE DATA WAS ACQUIRED

The bridge was excited with an impactor that consisted of a sliding weight (like a bar bell weight) on a pole mounted to a
base-plate. The weight had handles on two sides so that two people could lift it. When the weight was lifted and then re-
leased, it slid down the pole and impacted the bridge.
The impactor was located at a fixed point near the middle of the bridge span, as shown in the first figure above. This location
was chosen to excite as many resonances of the bridge as possible.
• A uni-axial accelerometer was mounted near the impactor to acquire the Reference response during each impact
• A Roving tri-axial accelerometer was used to acquire the 3D response at each of the numbered Point on the bridge
trusses, as shown in the first figure above
During each impact of the bridge, a 4-channel analyzer was used to simultaneously acquire three tri-axial Roving accelera-
tion signals and the Reference acceleration signal. The bridge was impacted three times for each Measurement Set, and tri-
spectrum averaging was used in the analyzer to reduce extraneous noise.
TRANSMISSIBILITY
Transmissibility is the traditional way of making a frequency response measurement when the excitation force (or forces)
cannot be measured.
• An FRF is defined as the ratio of the DFT of a response divided by the DFT of a force that caused the response
• A Transmissibility is calculated in the same way as an FRF, but the spectrum of a fixed Reference response is sub-
stituted for the spectrum of the excitation force in the denominator of the Transmissibility
• A Transmissibility is defined as the ratio of the DFT of a Roving response divided by the DFT of a fixed Refer-
ence response
DIFFICULTY WITH TRANSMISSIBILITY’S
• Assumption 1: If the spectrum of the un-measured force (or forces) is assumed to be “a relatively flat spectrum”,
then a peak in the Auto spectrum of each response is caused by the excitation of a resonance
• The difficulty with a Transmissibility is that a resonant frequency is displayed as a “flat spot”, not as a peak
• A ““flat spot”” is caused by the division of one response DFT by another where both contain the same resonance
curve
A set of ODS-FRFs is created by multiplying a set of Transmissibility’s by a reference Auto Spectrum, thus creating meas-
urement functions with a peak at each resonance.
• An ODS-FRF has a peak at each structural resonance, so a frequency-based ODS can be displayed and extracted
from a set of ODS-FRFs at each resonant frequency

Page 3 of 8
App Note 12 www.vibetech.com 2/8/2021

STEP 1 - AUTO SPECTRUM VERSUS TRANSMISSIBILITY

• Press Hotkey 1 Auto Spectra Versus TRNs
The Auto spectra for all the Roving DOFs of the bridge are overlaid on the left, and the Transmissibility between each Rov-
ing DOF and the Reference DOF are overlaid on the right.
• A resonance peak in shown at 6 Hz in the Auto spectra and a ““flat spot”” is shown at 6 Hz in the Transmissibil-
ity’s.

Auto spectrum & Transmissibility at the Same Frequency.

THREE WAYS TO CALCULATE AN ODS-FRF
In MEscope, three different types of data can be used to calculate an ODS-FRF,
1. Roving & Reference response time waveforms
2. Transmissibility between a Roving & Reference response, and a Reference Auto spectrum
3. Auto spectrum of a Roving response and a Cross spectrum between the Roving & Reference response
When Transform | ODS-FRFs is executed in a Data Block window, the following dialog box opens

Page 4 of 8
App Note 12 www.vibetech.com 2/8/2021

• If Roving & Reference time waveforms are provided, an ODS-FRF is calculated by first calculating a Transmissi-
bility between the Roving & Reference response, and then multiplying the Transmissibility by the Reference Auto
spectrum
• If a Transmissibility and a Reference Auto spectrum are provided, an ODS-FRF is calculated by multiplying the
Transmissibility by the Reference Auto spectrum
• If Roving Auto & Cross spectra are provided, an ODS-FRF is calculated by replacing the magnitude of the Cross
spectrum with the Auto spectrum
In this App Note Auto & Cross spectra are used to calculate ODS-FRFs.
When multiple Measurement Sets of data are acquired, each ODS-FRF contains the correct magnitude of a response (the
Auto spectrum of a Roving response) together with the correct phase relative to the Reference response for each Measure-
ment Set.
ODS-FRFs IN DISPLACEMENT, VELOCITY, OR ACCELERATION UNITS
Another advantage of an ODS-FRF is that it can be integrated from acceleration to velocity units or doubly integrated to
displacement units.
• Frequency-based ODS’s extracted from ODS-FRFs provide the true deflection of a structure or machine in engi-
neering units
STEP 2 - CALCULATING ODS-FRFs
• Press Hotkey 2 Calculate ODS-FRFs
The BLK: Unscaled ODS-FRFs window will open with 60 ODS-FRFs overlaid in it. Three ODS-FRFs represent the accel-
eration in the X, Y & Z directions at each of the 20 Points on the bridge.
BLK: Unscaled ODS-FRFs is shown on the left below, BLK: Pony Truss Auto Spectra on the upper right, and BLK:
Pony Truss Cross Spectra on the lower right.
BLK: Pony Truss Auto Spectra contains both the Roving Auto spectra and the Reference Auto spectra calculated for the
time waveforms acquired from the bridge pony truss. BLK: Pony Truss Cross Spectra contains the Cross spectra between
each roving DOF and the Reference DOF.

Overlaid ODS-FRFs on the Left and Auto & Cross Spectra on the Right.
• Drag the Blue Splitter bar to the left to display the M# properties in each Data Block window

Page 5 of 8
App Note 12 www.vibetech.com 2/8/2021

DIFFERENT FORCE LEVELS

Transmissibility is a ratio of the DFTs of two responses. The excitation force level can change throughout a test, which
commonly occurs with impact testing.
• Multiple Measurement Sets of Transmissibility’s can be calculated from data acquired with as few as two simulta-
neous acquisition channels
• The Transmissibility in each Measurement Set is the ratio of a different Roving response divided by the same
Reference response
• Assumption 2: If a varying excitation force level affects all responses in a linear manner, then the variations in
response levels are “canceled out” of a Transmissibility measurement
RESCALING ODS-FRFs
A Transmissibility automatically accounts for the effects of a varying force levels between multiple Measurement Sets, but
an ODS-FRF does not.
• If ODS-FRFs are calculated from multiple Measurement Sets of data, they must be corrected to account for chang-
es in the excitation force level between Measurement Sets
To correct the magnitudes of every ODS-FRF in a Measurement Set (i), each ODS-FRF is multiplied by a scale factor (SFi)
defined by,
Sets

G k
SFi = k =1

Sets  Gi
Sets = number of Measurement Sets
G i = the averaged value of the reference response Auto Spectrum for Measurement Set(i)
• The scale factor (SFi) corrects each ODS-FRF magnitude according to the average level of all the reference re-
sponse Auto spectra
STEP 3 - OVERLAYING THE REFERENCE AUTO SPECTRA
• Press Hotkey 3 Overlay the Reference Auto Spectra
When Hotkey 3 is pressed, the Reference Auto spectra from 20 Measurement Sets are overlaid.
• It is clear from the figure below that the impact force level was different for each of the 20 Measurement Sets.

20 Reference Auto Spectra Overlaid.

Therefore, after ODS-FRFs are calculated from the 20 Measurement Sets of Auto & Cross spectra, they must be re-scaled
before meaningful frequency-based ODS’s can be displayed from them.

Page 6 of 8
App Note 12 www.vibetech.com 2/8/2021

STEP 4 - RE-SCALING THE ODS-FRFs

• Press Hotkey 4 Re-Scale the ODS-FRFs
The ODS-FRFs must be re-scaled to compensate for differences in the impact force levels applied when each of the Meas-
urement Sets was acquired. This was observed by overlaying the Reference Auto spectra in Step 2. These same Reference
Auto spectra will be used to re-scale the ODS-FRFs.
• A Band cursor was displayed from 4 to 18 Hz so that the Average Reference spectrum was only calculated over
that frequency span

Un-Scaled and Re-Scaled ODS-FRFs

When Hotkey 4 is pressed, the BLK: Scaled ODS-FRFs window will open on the right with 60 scaled ODS-FRFs overlaid
in it. These re-scaled ODS-FRFs are used in the next step to display the frequency-based ODS at each resonance peak on a
model of the bridge truss.

Page 7 of 8
App Note 12 www.vibetech.com 2/8/2021

STEP 5 - DISPLAY FREQUENCY-BASED ODS’s

• Press Hotkey 5 Display Frequency-Based ODS
In order the display ODS’s in animation on the bridge model, an M# Link is created linking each M# in the BLK: Scaled
ODS-FRFs window to a DOF of the bridge model in STR: Pony Bridge. This is done with a Script command.
Animation will begin from the Line cursor position in BLK: Scaled ODS-FRFs. The animated shape is also being normal-
ized to remove some of the random phase from the ODS. The Complexity Plot on the lower right shows the complex ODS
and the normalized ODS data side-by-side.
• Drag the cursor to another resonance peak to display its frequency-based ODS

ODS of the Bridge Truss at 6 Hz Dominated by the First Bending Mode Shape.
STEP 6 - REVIEW STEPS
To reviews the steps of this App Note,
• Press Hotkey 6 Review Steps
CONCLUSIONS
Multiple Measurements Sets of impact response-only data were post-processing in MEscope to calculate a set of ODS-FRFs
from which ODS’s were displayed in animation.
• To display a valid ODS from a set of ODS-FRFs, they must have correct magnitudes & phases relative to one an-
other
This is guaranteed if all response channels of data are simultaneously acquired. Unfortunately, most structural testing is
done by acquired a few channels of data at a time in multiple Measurement Sets.
When the Reference response Auto spectra from 20 Measurement Sets were overlaid, their levels were different, indicating
that the impact force levels were different between Measurement Sets. These differences among Measurement Sets were cor-
rected by re-scaling each Measurement Set of ODS-FRFs using an average Reference response from all 20 Measurement
Sets.
After the ODS-FRFs were re-scaled, realistic ODS’s were observed by displaying ODS’s at the resonance peaks in the ODS-
FRFs, including a dominant first bending mode shape of the bridge truss at 6 Hz,

Page 8 of 8