See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318792618

Structural Health Monitoring – A Design and Integration Issue

Conference Paper · December 2006

CITATIONS READS
3 655

1 author:

Christian Boller
Universität des Saarlandes
318 PUBLICATIONS   2,138 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Structural Health Monitoring of Ageing composites using acoustic methods View project

Fatigue life calculation based on NDT measurement techniques, reduction of testing efforts by intelligent testing strategies View project

All content following this page was uploaded by Christian Boller on 31 July 2017.

The user has requested enhancement of the downloaded file.

 Invited Paper to be presented at: 1st Asia-Pacific Workshop on Structural Health Monitoring,
Yokohama/Japan, December 4-6, 2006
Cover Page

Title : Structural Health Monitoring – A Design and Integration Issue

Authors : Christian Boller

Contact Address : The University of Sheffield, Dept. of Mechanical Engineering, Mappin

Street, Sheffield S1 3JD, United Kingdom

This is the Author’s Information,

Title Professor in Smart Structural Design
Affiliation The University of Sheffield, Dept. of Mechanical Engineering,
Address: Mappin Street, Sheffield S1 3JD, United Kingdom
E-mail: c.boller@sheffield.ac.uk
Web Page: www.sheffield.ac.uk/structures
Structural Health Monitoring – A Design and
Integration Issue
__________________________________________________________________________________________________________

CHRISTIAN BOLLER

ABSTRACT:

The structural health monitoring (SHM) process chain including solid state
modelling, loads monitoring, fatigue life evaluation and finally damage monitoring is
addressed in this paper. The problems arising from loads monitoring are discussed in
terms of the sensor integration point of view and the complexity one can get in easily.
The discussion goes further into the aspect of what to be defined as a sensor and what as
a sensor system respectively and the interface between the sensor system and a
complete engineering system to be monitored such as an aircraft. Finally some thoughts
are elaborated regarding the incident when SHM should be implemented into an
engineering structure. Most of the applications addressed are related to aviation.

INTRODUCTION

Much has been researched, written and published in the field of structural health
monitoring (SHM) over the past decade. Examples of the major sources of publication
include [1-9]. Most of the developments have and are currently still dealing with the
development of sensors, monitoring of loads and damage, sensor signal processing and
the SHM systems. A perspective is shown on how this SHM technology can come into
application. Aerospace [10] and civil engineering including geotechnical hazards [11] are
possibly the two areas being most addressed followed by marine, heavy machinery,
power generation, railway and automotive. Looking at SHM for aerospace applications
technologies are mainly at a laboratory stage and the question is: How can these
technologies be integrated into the complexity of an aircraft system meeting all the safety
and reliability issues related to this?

A major attraction with SHM is seen through the integration of sensors into the
engineering structure. Synonyms such as comparing the engineering structure with the
human body emerge, such as shown in Fig. 1 [12]. The synonyms can indeed lead to some
inspiration but one needs to keep in mind that sensing in biology and sensing in
engineering is of a significantly different character. While most of the sensors in biology
are of a relatively primitive nature, posses a high degree of redundancy, are reproducible
in the sense of self-healing but not replaceable and are highly networked, to just name a
few of their features, sensors in engineering are highly complex, have a low degree of
redundancy, are not reproducible but replaceable, and operate on very singular and as
such not very much networked basis. Hence if one sensor in engineering fails the
complete system can become very quickly impaired while when one sensor in biology
fails this is hardly not even registered. Consequently the integration of sensors into
structures is often not very much appreciated by operators of engineering systems such as
aircraft since it will reduce the overall reliability of the system. Integrating SHM
therefore needs a full understanding of the system’s reliability, the engineering as well as
the SHM system, such that the overall reliability of the system being possibly the major
cost driver can be optimised.

Fig. 1 An analogy of SHM according to [12]

The following chapters will try to address how SHM may have to be integrated into
the engineering system design and realisation such that SHM can turn out to be beneficial
in terms of cost and reliability.

DESIGN ISSUES

The design process of engineering structures consists of a variety of steps. It initially

requires a solid state CAD model before it moves into the assessment of structural
strength usually associated with FE modelling combined with the knowledge of the
operational loads and materials’ properties. This will then allow to know where damage is
most likely to occur which will then define where damage should be monitored.

Many of our engineering structures today including aircraft and civil engineering
buildings have been designed decades back in the past and are due to operate for decades
still further in the future. One of the most remarkable examples in aviation in that regard
is the US Air Force B-52 Stratofortress bomber. Would a component have to be replaced
on such an old aircraft availability of a CAD or generally digital model would be most
unlikely. Luckily this problem can be overcome these days as well as a follow-on
structural assessment to be made which allows even a structure from the past to be
adequately designed for future operations. The following paragraphs will describe this
process along a structure being much simpler than an aircraft structure and are expected
to demonstrate where the complexity of SHM integration has to be seen.

Solid State Model Generation

The engineering structure to be considered here is the frame of a mountain bike.
Amazingly many of the mountain bike frame manufacturers design their frames mainly
out of experience and do not possess any FE modelling as such. Fig. 2 shows how a CAD
and finally FE model was generated using a laser scanning device. The laser produces a
light stripe and scans the structure which is recorded by a camera and converted into a
digital point cloud to be stored in a data base. The point cloud model is then subsequently
converted into a polygon mesh, then structured on the basis of NURBS patches and
refined as a CAD model that can then further be converted into an FE model.

Polygon mesh NURBS patches CAD model FE model

Fig. 2 Laser scanning of a structure and conversion into a CAD and FE model

Stress-Strain Analysis
To analyse the strength of a structure requires the knowledge of the loads to be
applied. If those are not known to a sufficient degree they need to be determined. Finding
the locations where this would be best done requires the FE model and some assumptions
to be made with regard to the different loads to be applied. Fig. 3 shows the potential
loads to be applied on a mountain bike frame. They look to be quite limited however a
deeper analysis shows in what complexity one will be in. Fig. 4 shows the result for the
pedal load case only and allows the locations of stress and strain concentration
(highlighted by circles) to be determined at which a stain gauge for monitoring the pedal
load sequence would be placed best. This exercise can now be performed subsequently
for all the different loads applied on the bicycle frame individually and further
superimposed with potential load combinations occurring in service. This will then allow
a strategy to be developed on how to monitor the load sequence best being applied to the
structure.

x
Saddle y Steering

FX
MZ MY

Rear Front
Pedals

Fig. 3 Typical loads applied to a bicycle frame

1000 N
Saddle
moment
3 450 Saddle
45°

Nm force
1700 N
(via rigid 2
body)
Restrained
(direct BC)

2400 N
(via rigid
(via rigid- body)
body
Restrained
1
Pedal

Fig. 4 Stress distribution on mountain bike frame obtained for the pedal load case

The limited number of 11 load cases and their combinations already shows that finding
the optimal solution for a limited number of load monitoring sensors is by far a trivial
problem. This among others makes it understandable why a complex aircraft such as the
Eurofighter Typhoon has currently no more than 16 locations for loads to be monitored
which makes it already a sufficiently complex problem. Strains being monitored can
further be used for validation of the FE model and thus for further improvement of the
assessment procedure.

Fatigue and Damage Assessment

The ability of monitoring a load sequence appropriately in accordance to the true
operational conditions allows damage and thus the initiation of damage to be determined
more precisely. Load sequences monitored as a time sequence can now be assessed in
accordance to a rainflow cycle counting procedure which allows peaks and troughs of the
time signal to be allocated to cycles characterised in terms of their mean stress and stress
amplitude. Each cycle is then accumulated into a matrix describing mean stress over
stress amplitude which is characteristic for the load spectrum applied. This matrix can
then be again applied to describe a deterministic randomised in-service loading sequence
of similar characteristic to the one having been monitored which in combination with a
fatigue life evaluation post-processor allows damage accumulated in each of the elements
of the FE-model to be estimated and thus also to predict the incident when damage is due
to occur. Fig. 5 shows an example of such a component indicating the intensity of damage
accumulated. This damage distribution map does indicate which locations are most likely
to fracture and where the implementation of a damage monitoring system may be useful.

Fig. 5 Damage accumulated on a complex component using FE-Fatigue

(Courtesy of nCode)

Damage Monitoring and Sensors Patterns

The knowledge of a damage to occur as described in the paragraph before and shown
in Fig. 5 is not sufficient to simply bond some transducers that will act as sensors and
possibly actuators respectively. What is also required is the knowledge along where
damage (i.e. a crack) would propagate and what size of crack would have to be reliably
recognised. This may be again associated with an increasing number of sensors which
would be in contradiction to the reduced number of sensors and actuators requested by the
engineering system operator such as an airline. Here some true current deficits of SHM
become apparent. One of them is the lack of appropriate tools that allow the SHM process
to be simulated. First approaches are provided in terms of using genetic algorithms and
other optimisation procedures. However this is still far from being a holistic approach and
tool allowing the monitoring process to be described and the number of sensors to be
determined. Such tools would however significantly help to configure the appropriate
sensor/actuator pattern.

Associated to that discussion is the discussion of what is a sensor by itself. Fig. 6

shows an example where a set of 12 transducers (sensors/actuators) are used in the form
of the Smart Layer from Acellent Technologies [13]. These transducers interface with the
Smart Suitcase which generates actuation data, records sensor data and may further
process the sensor data in accordance to requirements being set. In the classical sense this
would mean 12 additional transducers for monitoring the component considered, which
may not be appreciated by an engineering system operator and may become further of an
object for transducer optimisation and minimisation.

Data
Amplifier
Acquisition

Switch
Processor
Box

Waveform
Amplifier
Generator
Smart Layer®
with 12 transducers Smart Suitcase™

Smart Sensing System with 1 input and output

Fig. 6 Logic an interfaces of a sensing system for SHM

Indeed the number of transducers becomes irrelevant when the complete signal
generation and processing unit becomes a part of the sensor system, in practical terms
when the Smart Layer and Smart Suitcase from Acellent Technologies becomes a unity
that can be considered as the smart sensing system highlighted in Fig. 5. This system now
only has a single interface between the remaining engineering system, such as an aircraft,
for which the requirements in terms of input and output as well as reliability can be
clearly defined and controlled. Hence whatever is inside the smart sensing system is not
relevant as long as the system meets the requirements set at the interface to the
engineering structure or system. Problems of reliability and redundancy are therefore
transferred to the SHM system supplier similar to the approach applied for any avionic or
engine system in aviation. Virtually the biological analogy shown in Fig. 1 would be
indeed applicable under these conditions provided the sensing system would meet the
requirements accordingly.
INTEGRATION ISSUES

There are good reasons as to when SHM should be integrated into an engineering
structure. For aerospace applications those can be well identified along the structure
provided by the ATA Maintenance Steering Group MSG-3 document [14] of which an
overview logic prepared by Boeing is shown in Fig. 7 [15].

Principal Structural Elements

Fatigue Environmental Accidental & Discrete

Damage Damage Damage

Damage Corrosion Typical Discrete

Safe Life
Tolerance Inspection Accidental Source

Types of Integrated Corrosion Types of Types of

Inspection Prevention Programme Inspections Inspection

DTR Evaluations EDR Evaluations ADR Evaluations

Approved Initial Inspection and Maintenance Programme

SHM relevant

Fig. 7 ATA MSG-3 maintenance process logic breakdown [15]

Following that logic damage can occur either due to fatigue, environmental effects or
accidental/discrete damage. Fatigue and environmental damage are usually effects which
have been accounted for during design and which are less predictable at the conception of
an aircraft. Most of their impacts are gathered from experience such as analysing the
leaders of an aircraft fleet ageing over time. Hence it is possibly not relevant and even
advisable to implement SHM for fatigue and environmental monitoring at the very
beginning of the life of an aircraft since SHM would be without function over at least half
of an aircraft’s life.

However there are also good reasons for implementing SHM at the very beginning of
an aircraft’s life and this is when damages due to accidents have to be monitored. Of
course not each type of accident can be predicted but there are different locations along an
aircraft which are more prone to accidental damage than others and which are shown in
Fig. 8. These include doors where access and delivery ground vehicles may collide under
uncontrolled conditions or similar situations that may occur with respect to engine pylons
in terms of ground vehicles circulating under the wings. Another more environmentally
related damage is uncontrolled water spillage around galleys and lavatories. SHM
systems are also very much advisable to be implemented from day one in an aircraft with
respect to loads monitoring. This is nowadays becoming more and more standard with
modern fighter airplanes [16] where load spectra can vary significantly. However this is
also very much discussed these days in the context with commercial aircraft and here
specifically in the context of hard landing monitoring devices [15]. Landing gears are
possibly the most critical interface between the aircraft and the ground impact and are
currently one of the major components for aircraft structural failure. So far judgement of
the landing impact is left with the crew only but a more objective judgement would be
beneficial in the light of the various landing gear failures that happened in the past.

There is another important issue with aircraft which drives SHM to be implemented
into aircraft at the very beginning and not being related to MSG-3 which is the extension
of the damage tolerance design principle in the way reported by Schmidt et al. [17]. The
implementation of damage monitoring systems into structures, such as frames and
stringers of a fuselage structure, being badly accessible for traditional manpower related
inspection methods, could allow those components to be continuously monitored. This
would permit the assumptions for damage tolerance design to be improved such that
either inspection intervals could be extended or allowable stresses to be increased, the
latter resulting in lower structural weight.

Doors

Galleys &
Pylons
Lavatories Landing Gear

Fig. 8 Locations on aircraft for SHM to be implemented from the beginning

There is a further major issue which is related to the integration of SHM into the
aircraft system which is the interface between the SHM system and the data management
of the aircraft itself. Since automated monitoring is already done to a larger extent and for
a longer period on engines and avionics there already exist standards and protocols
related to data management at the interfaces between these systems and the aircraft. It will
therefore be less than likely that these standards and protocols will be changed once SHM
will be implemented. Hence the interface described in Fig. 6 is likely to be well defined
from the aircraft’s side. A view on those types of SHM integration issues into an existing
aircraft platform is provided in Fig. 9 for a commercial aircraft [15] and in Fig. 10 for a
military fighter airplane [18]. It should be kept in mind that the logic shown in Fig. 10 is
already well in place due to the operational loads monitoring system already existing
onboard the Eurofighter Typhoon [16]. Fig. 9 underlines the similarities of protocols and
interfaces for engines and flight controls once SHM will be implemented.
BENEFITS

The rewards of implementing SHM have already been mentioned to some degree
before. They range from saving weight through enhanced use of damage tolerance design
to avoidance of gradual damage progressing due to accidental damage resulting in initial
barely visible damage. These benefits have so far not been quantified but they can be
considered to become significant.

Air/ground communication Wireless ground link Maintenance access terminal Data storage Flight crew displays

Airplane health manager Flight crew and dispatch alerting

Cabin core Avionics CCS Structures Structures Flight controls Propulsion

health health health health health manager health and utilities
manager manager manager manager sensors manager health manager
SHM-003

Fig. 9 SHM integration into the health management of a commercial airplane [15]

Downloaded A/C Data Evaluation

RIU
RIU
[Reconfigurable Interface]
[Reconfigurable Interface]

ASH
ASH
Aircraft System Health
Aircraft System Health

EHM
EHM
Engine Health Monitoring PMDS/ MDP
Engine Health Monitoring
SHM
SHM
Arising
Structural Health Monitoring
Structural Health Monitoring Off Aircraft Data
SPS(HM)
SPS(HM) Arising
Secondary Power System
Secondary Power System
Health Monitoring Arising
Health Monitoring
ASH ExCS SPS NDI

SHM EHM LOG AIS

(2*)
EFLogSW
EFLogSW System Services
Arising
Eurofighter Logistics Software
Arising Eurofighter Logistics Software
Packages
Consumption Data Packages

Maintenance Work Order

Maintenance Work Order

Interface
Interface Industry
to Industry [Server] Industry
to Industry [Server]

Fig. 10 Health management integration of the Eurofighter Typhoon [18]

Another aspect where SHM can contribute is aircraft operability. Many of the
components to be inspected are located in areas highly hidden, where up to 99% of the
inspection effort can be related to disassembly and re-assembly of components only. This
effort could be avoided if an SHM system would be in place. This becomes specifically
relevant if the component to be monitored is along the critical path of a maintenance
process which becomes increasingly likely with the change from the flight hours based
clearly defined letter checks to the more usage based and operator defined MSG-3
maintenance procedure where damage critical components can appear along the critical
path of a maintenance process on a random basis. Keeping in mind that every hour of
aircraft maintenance saved allows to increase an aircraft’s operability and that each hour
of aircraft non-operation will cost the operator easily a remarkable four digit sum in
US$ it becomes obvious that an SHM system may easily pay off in a relatively short
period of time.

CONCLUSIONS

Designing engineering structures such as aircraft on the basis of SHM requires some
care. All structural components requiring inspection have to be differentiated into two
groups.

The first group includes those components where SHM is worth to be implemented
from the very beginning. These are the components that may be prone to any sort of
overloading or accidental damage as well as those where the information of their integrity
can help to enhance the idea of damage tolerance and thus allow weight to be saved.
Another type of sensor that contributes to SHM and which is worth to be implemented at
the very beginning of an aircraft’s life is a loads monitoring sensor. Much care is required
in finding the appropriate locations and combinations for these sensors such that the
information generated can be well used for damage accumulation and damage initiation
prognosis which will define when and where damage monitoring will have to be done.

The second group consists of those components which are prone to fatigue and
environmental damage. These damages are not due to be expected before half life of the
aircraft and as such the SHM system may not be advisable to be installed before.
Furthermore there may be a variation in which components may have to be equipped
depending on the past operational conditions of the aircraft.

The number of transducers (sensors/actuators) used for damage monitoring is not

relevant when the transducers are combined with the signal generation and processing
unit to a complete monitoring system and the monitoring system itself only possesses a
single interface to the aircraft maintenance system. In that case the complete monitoring
system can be considered as one single sensor and it is the reliability of the information
transferred from the system through its single interface which contributes to the overall
reliability of the aircraft. Since other monitoring systems for engines, flight control and
avionics systems already exist it will be more than likely that the protocols used for those
systems will also be applied for an SHM system hopefully operating in the not too far
future.
 REFERENCES

1. Chang F.-K. and A. Güemes (Ed.s): Proceedings of the Internat. Workshop of Structural Health
Monitoring, DEStec Publ. Inc. Lancaster/PA, USA (1997, 1999, 2001, 2003, 2005)
2. Balageas D. (Ed.) (2002), Boller C. and W.J. Staszewski (Ed.s) (2004), Güemes A. (Ed.) (2006):
Proceedings of the European Workshop on Structural Health Monitoring, DEStec Publ. Inc.
Lancaster/PA, USA
3. Mita A. (Ed): Proceedings of the Asia-Pacific Workshop of Structural Health Monitoring
Yokohama/Japan, (2006)
4. Smart Structures and Materials & Nondestructive Evaluation and Health Monitoring Symposium, San
Diego/CA, USA, SPIE Bellingham/WA, USA, since 1994
5. Journal of Structural Control and Health Monitoring, (Ed.: L. Faravelli), J. Wiley and Sons
6. Structural Health Monitoring – An International Journal, (Ed.: F.-K. Chang), SAGE Publ.
7. Smart Materials and Structures, (Ed.s: V.K. Varadan and R.O. Claus), Inst. of Physics Publ.
8. Intelligent Material Systems and Structures, (Ed.: D. Inman), SAGE Publ.
9. Smart Structures and Systems, (Ed. C-K Choi, J-M Ko, B.F. Spencer Jr. and F. Casciati) Techno-Press
Ltd.
10. Staszewski W.J., C. Boller and G.R. Tomlinson: 2003: Health Monitoring of Aerospace Structures; J
Wiley & Sons
11. Wenzel H. and D. Pichler, 2005: Ambient Vibration Monitoring; J. Wiley & Sons
12. Speckmann H. 2006: Structural Health Monitoring (SHM): One Technology of the Airbus
Intelligent-Airframe philosophy; Proc. of the Asia-Pacific Workshop of Structural Health Monitoring
Yokohama/Japan (Ed.: A. Mita)
13. www.acellent.com
14. ATA MSG-3, Operator/Manufacturer Scheduled Maintenance Development; Rev. 2005.1; Air
Transport Association of America
15. Akdeniz A., 2005: Structural Health Management Technology Implementation on Commercial
Airplanes; Boeing White Paper
16. Hunt S.G. and I.G. Hebden, 2001; Validation of the Eurofighter Typhoon structural health and
usage monitoring system; Smart Mater. Struct. 10, pp. 497 - 503
17. Schmidt H.-J. and B. Schmidt-Brandecker, 2001: Structure Design and Maintenance Benefits from
Health Monitoring Systems; in: Structural Health Monitoring (Ed. F.-K. Chang), CRC Press, pp. 80 –
101
18. Boller C. and M. Buderath, 2006: Fatigue in Aerostructures – Where Structural Health Monitoring
can Contribute to a Complex Subject; manuscript in press for publication in Phil. Transactions of the
Royal Society

View publication stats