Purpose of environmental testing

 A lot of structures suffer from damage during normal operation OR during transport

 The purpose = often to predict/avoid this kind of damage

= proof that a structure survives certain circumstances

 Ideas:
 If structure functions well during a test, it should survive normal operation
 If structure is damaged during a test, it will not survive normal operation …

 So a good test will simulate conditions which are comparable to the operational conditions
of the structure

Different types of environment tests are developed (see further)

 Environmental testing in practice?

 Measuring the response of a structure when exciting the structure such that a control
signal matches a predefined excitation level

 To do this, the drive signal needs to be constantly updated

 Control loop necessary Check difference

Feedback
with predefined
level

Control Signal

Measurements

Drive signal
Update of drive
signal
Purpose of the control loop

 To shape the drive spectrum such that the excitation of the control transducer matches
a pre-defined level (reference spectrum)
For example:
F PSD Reference RMS: 0.200 g
200.00 F PSD Low Abort RMS: 0.100 g
1.00

F PSD Low Alarm RMS: 0.142 g

F PSD UpAbort RMS: 0.399 g
F PSD UpAlarm RMS: 0.283 g
F PSD 1:+X RMS: 0.203 g

Amplitude
/Hz

Log
2
g

0.10e-6 0.00

6 200 400 600 800 1000 1200 1400 1600 2000

Hz
F PSD Reference RMS: 0.200 g
200.00 F PSD Low Abort RMS: 0.100 g
1.00

 To verify that the Ftest structure is not

PSD Low Alarm RMS:in danger
0.142 g and to shut down the test if it is
F PSD UpAbort RMS: 0.399 g
This is done
F
through the checking
PSD UpAlarm RMS: 0.283 g
of abort and alarm conditions
F PSD 1:+Y RMS: 0.198 g

plitude
/Hz

og
Types of environmental control

 Purpose is to simulate a signal which is comparable to the operation environment of the

structure

 Groups of control signals:

 Sine vibration control

 Random control
 Shock control
 Combination
 Time waveform replicator
Sine Control

 Principle:

 Excitation by a (sweeping) sine wave

 Amplitude equals a reference profile (frequency spectrum)
Sine Control

 Advantages:
 Control is very simple and straightforward
 Very precise determination of eigenfrequencies possible
 Ideal for substructures with dominant resonance mounted on basis structure
examples: - electrical components mounted on a dashboard
- parts of a satellite
If a large primary structure has a lot of
Electrical component substructures mounted on it, then the
primary structure will be a mechanical filter.
Suppose for example the dashboard of a car,
which has a random road noise excitation as
input. The dashboard will often be a
‘Sine’ input mechanical filter, causing the substructures
to be excited by a narrow-banded random
vibration that resembles a sine vibration.
Dashboard When testing the substructure, a sine
excitation will be a close to the operational
excitation

Random input
 F PSD Reference RMS: 0.700 g
200.00 F PSD Low Abort RMS: 0.351 g
1.00

Sine Control F
F
F
F
PSD Low Alarm
PSD UpAbort
PSD UpAlarm
PSD 1:+X
RMS: 0.496 g
RMS: 1.397 g
RMS: 0.989 g
RMS: 0.714 g

Amplitude
/Hz

Log
2
g
0.10e-6 0.00

6 200 400 600 800 1000 1200 1400 1600 2000

 Disadvantages: Hz
F PSD Reference RMS: 0.700 g
 200.00 1Hz per second)
Reliable tests may take lot of time (low sweep rate, F PSD Low Abort RMS: 0.351 g
1.00

F PSD Low Alarm RMS: 0.496 g

F PSD UpAbort RMS: 1.397 g
 Lot of energy at one frequency line may damage the structure (overtesting)
F PSD UpAlarm RMS: 0.989 g
F PSD 1:+Y RMS: 0.708 g

Amplitude
/Hz
 Excitation by sine wave might not be enough to determine the damage potential

Log
2
g
structures with complicated geometry and several eigenfrequencies
0.10e-6 0.00

6 200 400 600 800 1000 1200 1400 1600 2000

Hz
F PSD Reference RMS: 0.700 g
Interaction between modes may cause200.00
damage (in real life)F PSD Low Abort RMS: 0.351 g
1.00

F PSD Low Alarm RMS: 0.496 g

F PSD UpAbort RMS: 1.397 g
F PSD UpAlarm RMS: 0.989 g
F PSD 1:+Z RMS: 0.694 g

Amplitude
/Hz

Log
2
g
The mode interaction will not be generated with sine excitation (since one frequency is
excited at one time)
0.10e-6 0.00

6 200 400 600 800 1000 1200 1400 1600 2000

Hz
If the test object has complicated geometry with several
eigenfrequencies in the relevant frequency range
interaction between the vibration modes, may cause
damage. This damage will not be obtained by sine
excitation. For these case multi axial random excitation
will be necessary
Random Control

 Principle:
 Excitation by a random signal
 Frequency content of this signal equals a reference profile (power spectral density
spectrum)

 Advantage:
 More reliable test for more complex structures, with several eigenfrequencies

 Disadvantage:
 Less precise determination of eigenfrequencies …
 Not possible to bring enough energy in the structure over the whole frequency range
Shock Control

 Principle:

 Excitation by a non-stationary shock

 Shock impact equals a determined reference time signal
 Applications of Shock Control

 Examples:

 Transport: ability to survive a certain transport

 Ability to withstand a certain impact
(example: can a certain structure survive falling)
 Military: damage potential of explosions
 Safety: ability to withstand a certain nuclear explosion
Combined Mode Testing

 Principle:
 Actual vibration environment consists of:
• background random excitation
• stationary random and periodic components
 Sine components and Random Components controlled separately (Sine on
Random)

 Typical examples:
 propeller aircraft
 military jet aircraft (with or without on-board gunfire),
 helicopters
 tracked vehicles
 automobiles during acceleration or ABS braking
Time Waveform Replication

The Time Waveform Replication procedure is the current state-of-the-art solution to simultaneously replicate multiple
time traces in a set of control points. The control process is a proprietary frequency-domain Iterative Learning Control
that guarantees a fast convergence reducing the least error between the targets set and the recorded responses
Accelerated vibration testing

Dynamic tests:
Tests necessary to assess the durability of mechanical components in their entire life

Problem: The requested durability for a mechanical component is usually of thousands of hours.

It is necessary to reduce the duration of tests!

In order to reduce the testing time, some techniques have been developed to submit to mechanical
components more demanding excitations if compared to the ones they normally undergo in operating
conditions

Warning: excitations acting on components are stochastic in nature  damage accumulation is governed by
the random multiaxial fatigue

The accelerated test must replicate the same damage observable on the component at the end of its
life

C. Lalanne, Mechanical Vibration & Shock Analysis,

Vol. 5, 2009, Wiley-ISTE
 Accelerated vibration testing

Reference standards:
All the reference standards originate from military applications

• NATO
• RTCA DO160
• MIL-STD 810:
they introduce the concept of test tailoring = the accelerated test must be calibrated on the real
loadings that act on the mechanical component during usage

Implementations:
• Aerospace industry: avionics, parts of fuselage
• Automotive: common rail pipes, components subjected to strong loadings in general
• Electronics industry: printed circuit boards that operate in high vibration environment

In any case, experimental testing is performed by means of shakers (hydraulic or, more frequently,
electrodynamic, because of their wider range of operating frequencies)
 Accelerated vibration testing

The component under test can be represented as a SDOF system with base
excitation

Type of base excitation:

• Swept sine
• Random (Gaussian / not Gaussian)

In each case the excitation is monoaxial (given the wide availability of commercial shakers of this type)

For complex mechanical components standards suggest to perform sequentially the test in 3 positions of
the object, one for each Cartesian direction

Example (from MIL-STD 810 G):

• Initial swept sine to assess the natural frequencies
of the undamaged component
• Gaussian random. The standard prescribe:
o PSD spectrum of excitation (to choose as a
function of component utilization)
o Test duration
• Test repetition in the other 2 Cartesian directions
• Swept sine to evaluate possible modifications of
natural frequencies of the object
 Accelerated vibration testing

Extreme Response Spectrum (ERS) and Fatigue Damage Spectrum (FDS): parameters used to quantify
the gravity of the environment where the component works in terms of maximum strain and fatigue damage
accumulated

MIL-STD 810 recommend the usage of 2 parameters:

• Extreme Response Spectrum (ERS) It is a strong simplifying
• Fatigue Damage Spectrum (FDS) hypothesis, however it
allows the usage of the
Underlying assumptions: method already in
preliminary stages of
• Component representable as SDOF system development, when the
component under
analysis does not have
a well-defined geometry
and a FE model is not
• Stress proportional to the relative displacement of the mass available yet!
with respect of the base

• Wöhler curve representable according to the Basquin’s law

• Damage accumulation follows the Miner’s rule

Accelerated vibration testing

Extreme Response Spectrum Fatigue Damage Spectrum

ERS = maximum response acceleration of FDS = plot of the damage accumulated by

the SDOF system as a function of a the SDOF system according to Miner's rule
possible natural frequency as a function of a possible natural frequency

(useful to evaluate the maximum strain on (useful to evaluate the fatigue damage to
the system) which the system underwent)

Natural frequency of the SDOF system is varied in a range of frequencies which covers all
the resonances of interest of the mechanical component under analysis
Accelerated vibration testing

ERS
 Accelerated vibration testing
Operational data
(mechanical component in its environment)

Calculation of the Extreme Response Spectrum (ERS) &

Fatigue Damage Spectrum (FDS) for operational data

Synthesis of the PSD of the accelerated test

unknown

Calculation of ERS & FDS of the accelerating life

excitation

Are the new

N Compute a new excitation
ERS / FDS
PSD
ok? O

YES

Accelerated life testing