# Structural Dynamics Research Laboratory University of Cincinnati

SHAKER EXCITATION TUTORIAL

Considerations and Problems Young Engineer’s Program - IMAC 2001

Structural Dynamics Research Laborator y University of Cincinnati Cincinnati, Ohio 45221-0072 USA WWW.SDRL.UC.EDU (513) 556-2720 Copyright © 2001

Excitation Tutorial

-1-

February 5, 2001

Structural Dynamics Research Laboratory University of Cincinnati

Shaker Excitation for Experimental Modal Analysis

Typical Shaker Excitation Test Setup Schematic

Excitation Tutorial

-2-

February 5, 2001

Structural Dynamics Research Laboratory University of Cincinnati

Shaker Excitation for Experimental Modal Analysis Test Set-up for Shaker Excitation

• Physical Connections, Alignment • Instrumentation • Single Vs. Multiple Shakers • Excitation Signal Type • Digital Signal Processing • Data Quality • Post-Test Considerations • Modal Parameter Estimation

Excitation Tutorial

-3-

February 5, 2001

Structural Dynamics Research Laboratory University of Cincinnati

**Shaker Excitation for Experimental Modal Analysis Key Issues
**

• Estimate Frequency Response Functions (FRFs) Suit-

**able for Modal Parameter Estimation
**

• Minimize Digital Signal Processing Errors (Leak-

age!)

• Minimize Small Structural Nonlinearities • Multiple Reference FRF Data • Frequency Range and Resolution

Excitation Tutorial

-4-

February 5, 2001

Structural Dynamics Research Laboratory University of Cincinnati

**Shaker Excitation for Experimental Modal Analysis Basic Assumptions
**

• Linearity • Time Invariance (Stationarity, Consistency) • Observability • Reciprocity

Error Considerations

• Variance Error: Averaged value equals expected value • Bias Error: Averaged value not equal to expected value

Excitation Tutorial

-5-

February 5, 2001

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Test Object Conﬁguration
• Fixed Boundary Conditions • Free Boundary Conditions • Shock Cord • Foam Rubber • Air Suspension • Realistic Boundary Conditions • Match Impedance(s) at Boundaries • Mass Loaded Boundary Conditions
Excitation Tutorial
-6-
February 5.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Other Test Conﬁguration Considerations
• Test Fixturing • Interaction with Test Object • Test Object • Number of References • Fixed Excitation/Response Locations • Location of References (Shakers)
Excitation Tutorial
-7-
February 5.

etc. DAC. etc.Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Instrumentation
• Shaker Type (electromagnetic. hydraulic. 2001
. static + dy-
namic)
• Frequency Range • Signal Source (noise.)
Excitation Tutorial
-8-
February 5.) • Shaker Control Capability (force matching vs. mo-
tion matching)
• Speciﬁcations • Force Amplitude Range (static vs.

)
• Stiff in Direction of Excitation • Weak in Transverse Directions • No Moments or Side Loads on Force Transducer • No Moments or Side Loads on Shaker
Minimize Shaker Fixture Motion/Resonances
Excitation Tutorial
-9-
February 5. 2001
.
etc.
vacuum. etc.Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Physical Connections . screw.)
• Connect Force Transducer to Shaker with Stinger (quill.Shaker to Structure
• Mount Force Transducer on Test Object (glue.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Data Quality Issues
• Force Level (High/Low Amplitude. Shaker Connection Not Perpendicular to Test
Object
• Load Cell Not Aligned with Response Transducer at
Connection
• Low Battery Power in Transducer Signal Conditioning • Loose Cable Connections • Cables Vibrating. Bad or Intermittent Cables • Electrical and/or Radio Frequency Noise on Data • Ground Loop • 50/60 Hertz Noise • Rattles in Test Object • Unmeasured Inputs
Excitation Tutorial
-10-
February 5. Improper Frequency
Content)
• Loose Exciter Connection (Stinger) • Load Cell.

Frequency Shifts.
Rattles. Loose Stinger(s) • Monitor Force Correlation Characteristics for Multiple
Inputs
Excitation Tutorial
-11-
February 5. Amplitude Changes
• Driving Point Frequency Response Function • Cross Point Frequency Response Function • Reciprocity Check • H pq = H qp • Monitor Force Spectrum of Each Input • Equipment Failure.Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Data Quality/Consistency
• Monitor Typical Measurements • Frequency Response Functions for Noisy Data. 2001
.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-12-
February 5. 2001
.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-13-
February 5.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-14-
February 5.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-15-
February 5. 2001
.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-16-
February 5. 2001
.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-17-
February 5.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-18-
February 5. 2001
.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-19-
February 5.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-20-
February 5. 2001
.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-21-
February 5. 2001
.

Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-22-
February 5. 2001
.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-23-
February 5.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-24-
February 5.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-25-
February 5.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Test Set-Up Examples
Excitation Tutorial
-26-
February 5.

Excitation Tutorial
-27-
February 5. 2001
. Generally.Structural Dynamics Research Laboratory University of Cincinnati
Excitation Signal Considerations
The type of excitation signal used to estimate frequency response functions depends upon several factors. the excitation signal is chosen in order to minimize noise while estimating the most accurate frequency response function in the least amount of time. With the advent of the FFT. excitation signals are most often contain broadband frequency information and are limited by the requirements of the FFT (totally observed transients or periodic functions with respect to the observation window).

Structural Dynamics Research Laboratory University of Cincinnati
Classiﬁcation of Excitation Methods
• Steady State • Slow Swept Sine • Stepped Sine • Random • True Random • Periodic • Fast Sine Sweeps • Pseudo Random • Periodic Random • Transient • Burst Random • Impact • Operating
Excitation Tutorial
-28-
February 5. 2001
.

to allow nonlinear characteristics in the data to be preserved. Generally.
Distortion . excitation signals that minimize distortion are considered more favorably.This ratio is formed by taking the RMS value of the excitation signal over the observation time period (T) compared to the largest value (positive or negative) in the time period (T).
Signal to Noise Ratio . Generally.This ratio is formed by taking the RMS value of the excitation signal over the observation time period (T) over the RMS value of the noise over the same time period (T). Generally. good excitation signals have larger signal to noise ratios. 2001
. good excitation signals have larger RMS to peak ratios. since experimental modal analysis is a linear process.Structural Dynamics Research Laboratory University of Cincinnati
Excitation Signal Characteristics RMS to Peak Ratio .Distortion refers to the ability of the excitation signal. when averaged. Nonlinear characteristics must be identiﬁed by other experimental or analytical techniques.
Excitation Tutorial
-29-
February 5.

The average or ensemble refers to the total collection of contiguous time blocks that contribute to each power spectral average. 2001
. a number of nomenclature issues must be explained.Structural Dynamics Research Laboratory University of Cincinnati
Excitation Nomenclature In order to explain the way in which excitation signals are created and sent to the shaker. Capture Blocks . Burst Length . The total time of each average is equal to the sum of the number of delay blocks (N d ) plus the number of capture blocks (N c ) times the observation period (T) which is the same for all delay and capture blocks.The number of capture blocks refers to the number of contiguous blocks of time data (excitation (input) and response (output)) that are recorded or captured for each average (N c ). Average (Ensemble) . Delay Blocks .
Excitation Tutorial
-30-
February 5.Burst length is the percentage (0 to 100%) of the average or ensemble time that the excitation signal is present.The number of contiguous blocks of excitation that take place without the associated input and output data being acquired are referred to as the delay blocks (N d ). particularly in random testing.

the following ﬁgure is a schematic representation of the number of contiguous blocks of time domain data contributing to one power spectral average. therefore.
0 D 0 1 D 2
Burst Length (%) C 3 C 4 C 5 C
100
6
Number of Contiguous Time Blocks (6T)
Window Function
Total Contiguous Time Per Power Spectral Average
Excitation Tutorial
-31-
February 5. the two blocks marked "D" represent delay blocks and the four blocks marked "C" represent capture blocks. The total time for each power spectral average is.
In order to clarify the preceding terminology. 2001
.The number of power spectral averages (N avg or N a ) is the number of auto and cross spectra that are averaged together to estimate the FRF measurements.Structural Dynamics Research Laboratory University of Cincinnati
Excitation Nomenclature Power Spectral Averages . In this example. six contiguous blocks of time data (6 × T seconds of data).

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Random Excitation Methods
Typical Random Excitation Test Setup
Excitation Tutorial
-32-
February 5.

2001
.Structural Dynamics Research Laboratory University of Cincinnati
Pure Random
• Advantages: • Fair general excitation type • Fair signal to noise ratio • Fair RMS to peak ratio • Reduces distortion • Good measurement test time • Works well with Zoom • Disadvantages: • Leakage a serious problem • More averages required • Poor characterization of non-linearities • Typical DSP Window • Hanning Window
Excitation Tutorial
-33-
February 5.

Pure Random
Excitation Tutorial
-34-
February 5. 2001
.Structural Dynamics Research Laboratory University of Cincinnati
Pure Random
Power Spectrum − Pure Random
10
1
10
0
Magnitude
10
−1
10
−2
10
−3
0
5
10
15 Spectral line (bin)
20
25
30
Signal Energy Content .

6 0.Structural Dynamics Research Laboratory University of Cincinnati
Pure Random Typical Random Signal .2 −0.6 −0.2 Amplitude 0 −0. 2001
.Time Domain
Random Force 1 0.8 0.8 −1 0
0.5
2 2.5
1
1.4 −0.5 Time (Seconds)
3
3.4 0.5
4
Excitation Tutorial
-35-
February 5.

2001
.5 0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
Excitation Tutorial
-36-
February 5.5
Real Part
0
−0.5 0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
0.5 Imaginary Part
0
−0.Structural Dynamics Research Laboratory University of Cincinnati
Pure Random Typical Random Signal .Frequency Domain
Random Force 0.

2001
.Frequency Domain
10
0
Random Force
Magnitude
10
−1
10
−2
0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
200 100 0 −100 −200 0
Phase (Deg)
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
Excitation Tutorial
-37-
February 5.Structural Dynamics Research Laboratory University of Cincinnati
Pure Random Typical Random Signal .

Structural Dynamics Research Laboratory University of Cincinnati
Pure Random Typical Random Signal .Frequency Domain
10 10 Magnitude 10 10 10
2
Random Force−Averaged
1
0
−1
−2
0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
200 100 0 −100 −200 0
Phase (Deg)
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
Excitation Tutorial
-38-
February 5. 2001
.

Conventionally.
• Advantages: • Minimum leakage • Fair signal to noise ratio • Fair RMS to peak ratio • Good measurement test time • Disadvantages: • Non-linear systems generate periodic noise • Typical DSP Window • Uniform Window
Excitation Tutorial
-39-
February 5. a single time block of the pseudo random excitation signal has energy at all frequencies of the measurement.Structural Dynamics Research Laboratory University of Cincinnati
Pseudo Random A pseudo random excitation signal is a random time domain signal that is constructed from a limited sequence of random numbers. random phase spectrum at the discrete frequencies of the measurement. Therefore. the pseudo random excitation signal is constructed in the frequency domain with a uniform amplitude. 2001
.

2001
. the response will not become periodic until the startup transient has decayed to zero.Pseudo Random
The pseudo random excitation signal is applied to the exciter repetitively. At this time.Structural Dynamics Research Laboratory University of Cincinnati
Pseudo Random
Power Spectrum − Pseudo Random
10
1
10
0
Magnitude
10
−1
10
−2
10
−3
0
5
10
15 Spectral line (bin)
20
25
30
Signal Energy Content . While the excitation signal is periodic in the observation window (T). one or more averages are taken.
Excitation Tutorial
-40-
February 5.

2001
.5 Real Part 0 −0.5 −1 0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
Excitation Tutorial
-41-
February 5.Frequency Domain
Psuedo−Random Force 1 0.Structural Dynamics Research Laboratory University of Cincinnati
Pseudo Random Typical Pseudo Random Signal .5 −1 0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
1 Imaginary Part 0.5 0 −0.

Frequency Domain
10 10 Magnitude 10 10 10
2
Psuedo−Random Force
1
0
−1
−2
0
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
200 100 0 −100 −200 0
Phase (Deg)
10
20
30
40 50 60 Frequency (Hertz)
70
80
90
100
Excitation Tutorial
-42-
February 5. 2001
.Structural Dynamics Research Laboratory University of Cincinnati
Pseudo Random Typical Pseudo Random Signal .

8 −1 −1
−0.2 Imag Part 0 −0.6 0.2 −0.Frequency Domain
Psuedo−Random Force 1 0.8 0.Structural Dynamics Research Laboratory University of Cincinnati
Pseudo Random Typical Pseudo Random Signal . 2001
.4 0.5
1
Excitation Tutorial
-43-
February 5.5
0 Real Part
0.4 −0.6 −0.

the periodic random excitation signal is constructed in the frequency domain with a random amplitude. a single time block of the periodic random excitation signal does not have energy at all frequencies of the measurement.
• Advantages: • Minimum leakage • Fair signal to noise ratio • Fair RMS to peak ratio • Reduces distortion • Fair measurement test time • Disadvantages: • Slower than other periodic excitations • Special hardware needed • Typical DSP Window • Uniform Window
Excitation Tutorial
-44-
February 5.Structural Dynamics Research Laboratory University of Cincinnati
Periodic Random A periodic random excitation signal is a random time domain signal that is constructed from an unlimited sequence of random numbers. Conventionally. Therefore. random phase spectrum at the discrete frequencies of the measurement. 2001
.

Periodic Random
The periodic random excitation signal is applied to the exciter repetitively. 2001
.Structural Dynamics Research Laboratory University of Cincinnati
Periodic Random
Power Spectrum − Periodic Random
10
1
10
0
Magnitude
10
−1
10
−2
10
−3
0
5
10
15 Spectral line (bin)
20
25
30
Signal Energy Content .
Excitation Tutorial
-45-
February 5. the response will not become periodic until the startup transient has decayed to zero. the ﬁrst average of data is taken. while the excitation signal is periodic in the observation window (T). After sufﬁcient time has elapsed to allow for both the input and output to become periodic. This process is repeated until sufﬁcient averages have been taken.

Structural Dynamics Research Laboratory University of Cincinnati
Burst Random
• Advantages: • Good general excitation • Minimum leakage • Fair signal to noise ratio • Fair RMS to peak ratio • Reduces distortion • Good measurement test time • Disadvantages: • Special hardware needed • Voltage feedback excitation ampliﬁer • Typical DSP Window • Uniform or Exponential Window
Excitation Tutorial
-46-
February 5. 2001
.

Structural Dynamics Research Laboratory University of Cincinnati
Burst Random
Power Spectrum − Burst Random
10
0
10
−1
Magnitude
10
−2
10
−3
0
5
10
15 Spectral line (bin)
20
25
30
Signal Energy Content . 2001
.Burst Random
Excitation Tutorial
-47-
February 5.

2001
. Current Feedback: Current feedback refers to the types of ampliﬁers in the exciter system that attempt to match the current supplied to the shaker to the excitation signal. the exciter will attempt to prevent the armature from moving. This effectively means that the acceleration of the armature will follow the excitation signal. no attempt is made to exactly match the excitation signal. if a zero voltage signal is sent to the exciter system. if a zero voltage signal is sent to the exciter system. Therefore. This effectively means that the displacement of the armature will follow the excitation signal. this means that the displacement. preventing any force to be applied by the exciter system.
Excitation Tutorial
-48-
February 5. the exciter will allow the armature to move. Typically. Note that this is normally an open loop control process. Therefore. particularly electromagnetic. Voltage Feedback: Voltage feedback refers to the types of ampliﬁers in the exciter system that attempt to match the voltage supplied to the shaker to the excitation signal.Structural Dynamics Research Laboratory University of Cincinnati
Burst Random Exciter Systems: Exciter systems. velocity or acceleration of the armature of the shaker will attempt to match the excitation signal. attempt to match the excitation signal to some physical characteristic of the exciter.

5
-1
-1.5
500
1000
1500
2000
Signal from load cell (Voltage Feedback)
1.5
0
-0. 2001
.Structural Dynamics Research Laboratory University of Cincinnati
Burst Random Signal to shaker
1.5
-1
-1.5
1
0.5
1
0.5
500
1000
1500
2000
Signal from accelerometer
6 x 10
-3
4
2
0
-2
-4
-6
500
1000
1500
2000
Excitation Tutorial
-49-
February 5.5
0
-0.

burst random and periodic random together with cyclic averaging.Burst Pseudo Random
Power Spectrum − Burst Periodic Random
10
1
10
0
Magnitude
10
−1
10
−2
10
−3
0
5
10
15 Spectral line (bin)
20
25
30
Signal Energy Content . 2001
.Structural Dynamics Research Laboratory University of Cincinnati
Hybrid Random Excitation Methods Several random excitation methods have recently been demonstrated that are hybrid methods involving combinations of burst random and pseudo random.Burst Periodic Random
Excitation Tutorial
-50-
February 5.
Power Spectrum − Burst Pseudo Random
10
1
10
0
Magnitude
10
−1
10
−2
10
−3
0
5
10
15 Spectral line (bin)
20
25
30
Signal Energy Content .

Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Summary of Excitation Signal Choices
Excitation Signal Characteristics Steady State Sine Minimize Leakage Signal-to-Noise Ratio RMS-to-Peak Ratio Test Measurement Time Controlled Frequency Content Controlled Amplitude Content Removes Distortion Characterize Nonlinearity * Special Hardware Required No Very High High Very Long Yes Yes No Yes Yes * No Yes No Fair Good Fair Ver y Short Yes * Yes * No No Yes No Yes * No Yes * Yes * No Yes No No Yes No No Fair Fair High Fair Low Ver y Shor t No Fair Very Shor t Yes * No No Fair Yes Fair Yes Fair Yes High Yes Low Yes Fair Pure Random Pseudo Random Random Periodic Chirp Impact Burst Random
Excitation Tutorial
-51-
February 5. 2001
.

Structural Dynamics Research Laboratory University of Cincinnati
Shaker Excitation for Experimental Modal Analysis Summary/Conclusions/Discussion
Excitation Tutorial
-52-
February 5. 2001
.