aerospace

Article
The Study of Aircraft Accidents Causes by
Computer Simulations
Paweł Szczepaniak 1, * , Grzegorz Jastrz˛ebski 1, * , Krzysztof Sibilski 1 and
Andrzej Bartosiewicz 2
1 Air Force Institute of Technology, Ksi˛ecia Bolesława 6, 01-494 Warsaw, Poland; krzysztof.sibilski@itwl.pl
2 State Commission on Aircraft Accidents Investigation, Chałubińskiego 4/6, 00-928 Warsaw, Poland;
andrzej.bartosiewicz@mi.gov.pl
* Correspondence: pawel.szczepaniak@itwl.pl (P.S.); grzegorz.jastrzebski@itwl.pl (G.J.)




Received: 23 January 2020; Accepted: 7 April 2020; Published: 10 April 2020


Abstract: Defects in an aircraft can be caused by design flaw, manufacturer flaw or wear and tear
from use. Although inspections are performed on the airplane before and after flights, accidents still
result from faulty equipment and malfunctioning components. Determining the causes of an aircraft
accident is an outcome of a very laborious and often very long investigation process. According to
the statistics, currently the human factor has the biggest share within the causal groups. Along with
the development of aviation technology came a decline in the number of accidents caused by failures
or malfunctions, though such still happen, especially considering aging aircraft. Discovering causes
and factors behind an aircraft accident is of crucial significance from the perspective of improving
aircraft operational safety. Effective prevention is the basic measure of raising the aircraft reliability
level, and the safety-related guidelines must be developed based on verified facts, reliable analysis
and logical conclusions. This article presents simulation tests carried out by finite element method
and constitutive laboratory tests leading to the explanation of the direct cause of a military aircraft
accident. Computer-based simulation methods are particularly useful when it comes to analysing the
kinematics of mechanisms and potential stress concentration points. Using computer models enables
analysing an individual element failure process, identifying their sequence and locating their primary
failure source.

Keywords: threaded connection; computer simulations; aviation safety; aircraft accidents

1. Introduction
An aircraft is an assemblage of complex and highly integrated sub-systems. It is generally
agreed that there exist certain precursors to each accident and incident. If one of these precursors is
not recognized and the underlying condition that has caused it is not corrected in time, then it can
graduate into an incident or even an accident. Aircrafts are highly engineered systems, endowed with
redundancies and fail-safe features [1].
Causal analysis of an accident or incident seeks to establish those factors that were judged to be
directly responsible in causing the event (primary causal factors) and those that contributed to the
event (secondary causal factors) by deconstructing the accident. For these causal factors, a causal
chain can usually be established for each accident or incident [2]. The advantage of causal chain
analysis is that in the case of multiple causes and multiple accidents or incidents, the common events
or elements in the chain can be identified and subjected to greatest attention. Thus, the safety system
can concentrate on those common events and maximize its responsiveness and effectiveness in cutting
down-times, and reducing or eliminating accidents. The perceived disadvantage of this approach
is that it is reactive rather than proactive. That is, the regulating authority and the industry (or the

Aerospace 2020, 7, 41; doi:10.3390/aerospace7040041 www.mdpi.com/journal/aerospace

Aerospace 2020, 7, 41 2 of 10

military operators) seek to eliminate the causal factor after the accident in order to prevent accidents
due to the same cause from happening again [2–4].
One of the commonly used disaster models is the “Swiss cheese” model proposed by Reason. In
the Swiss cheese model, an organisation’s defences against failure are modelled as a series of barriers,
represented as slices of the cheese. The holes in the cheese slices represent individual weaknesses in
individual parts of the system, and are continually varying in size and position in all slices. The system,
as a whole, produces failures when holes in all of the slices momentarily align, permitting “a trajectory
of accident opportunity”, so that a hazard passes through holes in all of the defences, leading to an
accident [5,6].
An example showing how a seemingly minor defect caused by a disaster can be the Trans World
Airlines Flight 800 (TWA 800) airliner crash. This accident is also a good example of the Swiss cheese
disaster model. TWA 800 was a Boeing 747-100 that exploded and crashed into the Atlantic Ocean
near East Moriches, New York, on 17 July 1996. The four-year National Transportation Safety Board
(NTSB) investigation concluded with the approval of the Aircraft Accident Report on 23 August 2000,
ending the most extensive, complex and costly air disaster investigation in United States history to
that time [7]. The report’s conclusion was that the probable cause of the accident was explosion of
flammable fuel vapours in the centre fuel tank. Although it could not be determined with certainty,
the likely ignition source was a short circuit. Problems with the aircraft’s wiring were found, including
evidence of arcing in the Fuel Quantity Indication System (FQIS) wiring that enters the tank. The FQIS
on Flight 800 is known to have been malfunctioning; the captain remarked on what he called “crazy”
readings from the system approximately two minutes and thirty seconds before the aircraft exploded.
As a result of the investigation, new requirements were developed for aircrafts to prevent future fuel
tank explosions. The NTCB’s final opinion is: “An explosion of the center wing fuel tank (CWT),
resulting from ignition of the flammable fuel/air mixture in the tank. The source of ignition energy for
the explosion could not be determined with certainty, but, of the sources evaluated by the investigation,
the most likely was a short circuit outside of the CWT that allowed excessive voltage to enter it through
electrical wiring associated with the fuel quantity indication system.” [5,8].
This article presents the results of finite element method (FEM) simulations and laboratory tests
leading to the detection of the “primary cause” the aircraft crash. The term primary cause, defined as
the most critical cause factor associated with a particular accident, can be deceiving and is often subject
to interpretation.
Studying the causes of an aircraft accident is usually an outcome of a very arduous and often long
investigation process. The aircraft reliability degree, apart from structural safety, is also significantly
affected by the quality of operation [9,10]. This applies to all work, both repairs as well as ongoing
maintenance. As empirical data have shown, often a seemingly insignificant shortcoming may
lead to catastrophic outcomes [9]. Based on the effects of an event, as a result of a thorough study
utilizing advanced numerical methods, it is possible to accurately diagnose the cause of the event,
which will contribute to effective prevention and possibly improve flight safety. Studies based on
advanced numerical techniques are now widely used in numerous fields of industry, the aerospace,
aviation and automotive industries in particular. In general, numerical methods are used for the
calculations in terms of machine and assembly parts at the engineering stage or when optimizing
existing structures [11–13]. The numerous research work conducted at the Air Force Institute of
Technology (AFIT) involved computer simulation research regarding the elements of an aircraft
landing gear mechanism (Figure 1a), aircraft hydraulic lines (Figure 1b), threaded connections and
an aircraft variable displacement hydraulic pump. A Computer Aid Design 3 Dimensional (CAD3D)
application, which contains a feature to study mechanism kinematics was used to recreate the geometry
of unit parts and assemblies. The study also involved using 3D object scanning techniques aimed at
recreating the complex geometry. The finite element method (FEM) analysis was used for the strength
tests. Assemblies and elements can be subjected to various numerical tests—structural strength,
flow and thermal analyses [14–17]. Nonetheless, it should be noted that numeric methods are not
 Aerospace 2020, 7, 41 3 of 10

Aerospace 2020, 7, x FOR PEER REVIEW 3 of 10

perfect, and their application for analysing technical issues is not easy and requires experience in this
infield.
this field. The entirety
The entirety of this
of this publication
publication is closely
is closely devoted
devoted to the
to the issues
issues of modelling
of modelling thethe strength
strength of a
ofthreaded
a threaded connection.
connection.

(a) (b)
Figure
Figure1.1.Examples
Examplesofof
applying
applying simulation
simulationmethods
methods inin
investigating
investigatingairairaccident
accidentcauses:
causes:(a)(a)
aircraft
aircraft
retractable
retractablelanding gear
landing gearstrut kinematics,
strut (b)
kinematics, stress
(b) concentration
stress concentrationareas and
areas and the geometry
the geometry ofofananaircraft
aircraft
hydraulic
hydraulic line.
line.

Other
Otherissues
issuesconcerning
concerning computer-based
computer-based tests involving
tests involving the
the landing
landing gear,
gear,rigid
rigidhydraulic
hydrauliclines
lines
and the pump will be addressed in further articles within this monographic cycle. This publication
and the pump will be addressed in further articles within this monographic cycle. This publication
presents
presentsdetailed
detailedresults
resultsofofthe
thenumerical
numericaland andexperimental
experimentalteststestsinvolving
involvinga ashank
shankwithwitha thread
a thread and
and
cooperating nut, which are the elements of an aircraft fuel system assembly (Figure 2). The various
cooperating nut, which are the elements of an aircraft fuel system assembly (Figure 2). The various
aspects
aspectsofofnumerical
numerical analyses
analyses ofof
threaded
threaded connection
connection areare
approached
approached inin
scientific research
scientific research [18–22].
[18–22].
Whereas
Whereasthe the authors havenot
authors have notencountered
encountered a strength
a strength analysis
analysis of a threaded
of a threaded connectionconnection with
with modelled
modelled operational wear and tear in the source literature. These studies are to confirm or contradict
operational wear and tear in the source literature. These studies are to confirm or contradict the thesis
the thesis
that, at athat, at awear
given given wear
level of level of this connection,
this connection, and underand its
under its strictly
strictly defined defined load,
load, its its load-
load-bearing
bearing capacity
capacity may bemay lost,be lost, resulting
resulting in the
in the loss loss of material
of material continuity.continuity. The conducted
The conducted simulations simulations
involved a
involved a conventionally
conventionally new thread new andthread and with wear,
with modelled modelledwhichwear,
waswhich wasbased
recreated recreated based
on the on theof
geometry
geometry
an actualofobject.
an actual object. presents
The article The article presents
results results
only for only with
a thread for aathread
modelledwithwear.
a modelled
Apart fromwear.the
Apart from the simulations, the study also involved conducting experiments, the results of which
simulations, the study also involved conducting experiments, the results of which were used to verify
were used to verify
the developed modelthe parameters
developed model parameters
and simulations and simulations conditions.
conditions.
Aerospace 2020,7,7,x41
Aerospace2020, FOR PEER REVIEW 44ofof10
10
Aerospace 2020, 7, x FOR PEER REVIEW 4 of 10

(a) (b) (c)

(a) (b) (c)
Figure 2. A computer solid model of the studied fuel system assembly: (a) cross-section, (b) view
Figure
Figure
without 2. 2.AAcomputer
housing, computer solidmodel
solid
(c) simplified modeloffor
model of the
the studied
studied
strength fuel
fuel system
system
analysis. assembly:(a)(a)
assembly: cross-section,(b)(b)
cross-section, view
view
without
without housing,
housing, (c)(c) simplified
simplified model
model forfor strength
strength analysis.
analysis.
2.2.Simulations
Simulationsand andExperimental
ExperimentalTests Tests
2. Simulations and Experimental Tests
The
Thefirst
firststage
stageofofthethestudy
studyinvolved
involvedmapping
mappingaaspecific
specificthread
threadwithin
withinaasolid
solidmodel,
model,which
whichwaswas
The
conventionallyfirst stage
new, of
with the
thestudy involved
parameters setmapping
out in the a specific
literature. thread
A within
seven-pitch a solid model,
M14x1.5 which
thread waswas
conventionally new, with the parameters set out in the literature. A seven-pitch M14x1.5 thread was
conventionally
subjected new, with the parameters set out in the literature. A seven-pitch M14x1.5 thread was
subjectedtotothethetests.
tests.This
Thismodel
modelwas wasthen
thenmodified
modifiedinintermstermsofofthethepitches
pitcheson onthetheshank
shank(Figure
(Figure3).
3).
The subjected to theinvolved
tests. This model was then modified in termspitch,
of thebased
pitches on the shank (Figure 3).
The modification involved modelling the wear of shank thread pitch, based on the measurementsof
modification modelling the wear of shank thread on the measurements of
an The modification involved modelling the wearmodelof shank thread pitch, based on the measurements of
anactually
actuallyworn
worn thread.
thread. The Thedeveloped
developedCAD CAD modelwas was subjected
subjected to to discretisation.
discretisation. The Thecreated
created
an actually
numerical worn
modeldoes thread.
doesnotnot The developed
reflect thread CAD model was
pitch microdamage subjected to
and the discretisation.
contact The created
numerical model reflect thread pitch microdamage and the contact surfaces are surfaces are
“numerically”
numerical model
“numerically” smooth, does
which not
is a reflect
sort of athread pitch microdamage
simplification. A set of and the contact
“non-penetration” contact surfaces
assembly are
smooth, which is a sort of a simplification. A set of “non-penetration” contact assembly functions
“numerically”
functions smooth, which is a the sortjoint.
of a simplification. aAdiscrete
set of “non-penetration” contact assembly
were usedwere used
for modelling for modelling
the joint. Developing Developing
a discrete microstructure microstructure
in the formin of the form of a
a microscopic
functions
microscopic were used
numerical for
modelmodelling
is not an the
easyjoint.
task, Developing
but in some a discrete
cases can microstructure
be of crucial in the
importance.formAnof a
numerical model is not an easy task, but in some cases can be of crucial importance. An example of
microscopic
example of suchnumerical
an modelinisreality
approach not an is
easy task, butby
presented inthe
some cases can
authors of be of
the crucialinvolving
analysis importance. theAn
such an approach in reality is presented by the authors of the analysis involving the numerical tests of
example
numerical of such an
tests ofelements approach
braking [18,21]. in reality
system elements [18,21].is presented by the authors of the analysis involving the
braking system
numerical tests of braking system elements [18,21].

(a) (b)
(a) (b)
Figure
Figure3.3.View
Viewofofa athread
threadcontour
contourwith
withmodelled
modelledshank
shankwear:
wear:(a)
(a)all
allpitches,
pitches,(b)
(b)two
twopitches
pitchesmagnified.
magnified.
Figure 3. View of a thread contour with modelled shank wear: (a) all pitches, (b) two pitches magnified.
Thematerial
The materialused
usedtotoconstruct
constructthe
theshank
shankofofthe
thefilter
filterwith
withthe
thestudied
studiedthread
threadisisthe
theTi8Al1Mo1V
Ti8Al1Mo1V
titanium
titanium The material
alloy
alloy with used
withthe
the to construct
short-term
short-term the shank
strength
strength of937
of ofMPa,
937 the filter
MPa, yield
yield with the
point
point ofofstudied
910MPa
910 MPathread is the Ti8Al1Mo1V
andYoung’s
and Young’s modulus
modulus
titanium alloy with the short-term strength of 937 MPa, yield point of 910 MPa and Young’s modulus
Aerospace 2020, 7, 41 5 of 10
Aerospace 2020, 7, x FOR PEER REVIEW 5 of 10

ofof120
120GPa.
GPa.TheThe modelled
modelled nut is made
nut is madeof of30H2N2M
30H2N2Malloy alloysteel.
steel.
TheThe calculations
calculations adopted
adopted a short-term
a short-term
strength of 723 MPa, yield point of 620 MPa and Young’s modulus
strength of 723 MPa, yield point of 620 MPa and Young’s modulus of 210 GPa [23]. of 210 GPa [23].
Thediscrete
The discretemodel
model (Figure
(Figure 4a)
4a) has
hasgreen
greenarrows
arrowsmarking
markingthe fastening
the fastening points
points(restraint) andand
(restraint) red red
arrows marking the load. The construction
arrows marking the load. The construction of a spatial
spatial discrete model was based on solid elements in
model was based on solid elements
in the shape of tetrahedrons, with a maximum size of approx. 1.6 mm.
the shape of tetrahedrons, with a maximum size of approx. 1.6 mm. Grid
Grid refinementswithin
refinements withinthe
thearea
ofarea of interest
interest were developed
were developed usingusing grid controls,
grid controls, four Jacobian
four Jacobian points,
points, withwith the minimum
the minimum element
element size of
size of approx.
approx. 0.06 mm. 0.06
The mm. The
total total number
number of nodes of was
nodes was 1,230,607,
1,230,607, whereas whereas
the totalthenumber
total number of
of elements
elements was 825,131. The share of elements with a shape factor below 3
was 825,131. The share of elements with a shape factor below 3 was around 95%. There were nowas around 95%. There were
no identified
identified distorted
distorted elements.
elements. The numerical
The numerical tests tests
utilizedutilized theoptimization
the “h” “h” optimization method method and
and resulted
inresulted in the development
the development of a convergence
of a convergence graph at
graph aimed aimed at increasing
increasing the accuracy
the accuracy of theofobtained
the obtained
results.
results. The optimization criterion was the resultant stress value, as per the Huber–Mises
The optimization criterion was the resultant stress value, as per the Huber–Mises hypothesis, and the hypothesis,
and the deformation
deformation energy
energy error error
was was specifically
specifically relative.relative. The post-optimization
The post-optimization energyenergy
errorerror
waswasbelow
below 6%, which demonstrates satisfactory compliance of the results according
6%, which demonstrates satisfactory compliance of the results according to the data. The general to the data. The
general dependence on Huber–Mises reduced stress σred [23–27]:
dependence on Huber–Mises reduced stress σred [23–27]:
2 2
�σx −σσy−� σ +2�σ y σ− σzσ� 2++(σ − σσxx))2 + 3�τ
(σzz −
2
s
(1)

σred σ= =� + − 2 2 τ2 xz 2 �
2 ++ τyzτ2yz+ τ+
x y y z 
red 22 + 3 τxyxy xz (1)

where: x(y,z)—normal
where:σσx(y,z) —normalstress towards
stress towardsthethe
x (y,x z)
(y,axis of the
z) axis ofcoordinate system,
the coordinate τxy(yz,xz)—shearing
system, stress
τxy(yz,xz) —shearing
within the xy (yz, xz) plane cross-section.
stress within the xy (yz, xz) plane cross-section.

(a) (b)
Figure4.4.Solid
Figure Solidmodel
modelgrid
grid
forfor
thethe studied
studied fuel
fuel system
system assembly:
assembly: (a) (a) complete
complete discrete
discrete model,
model, (b) (b)
within
within
the the studied
studied thread-refinement.
thread-refinement.

Thenumerical
The numerical strength
strength tests
tests using
usingthethehypothesis
hypothesisononthe thehighest
highest shape
shapedeformation
deformation energy are are
energy
mainlyutilized
mainly utilized to research
to research and calculate
and calculate elements elements
within within therange
the elastic elastic range [13,27,28].
[13,27,28]. The
The phenomenon
ofphenomenon of material
material failure failureisunder
under stress much stress
more is much more
complex thancomplex than just theofappearance
just the appearance of the
the first permanent
first permanent
deformations deformations
[27]. [27]. is
Parting failure Parting
caused failure is caused
by stress by stress
resulting resulting inof
in overcoming overcoming
the cohesion of the
forces
cohesion forces between material particles. In the case of calculating material failure
between material particles. In the case of calculating material failure conditions, it is recommended conditions, it is
recommended to apply Mohr’s hypothesis. According to the Mises model, the
to apply Mohr’s hypothesis. According to the Mises model, the limit of elasticity and the yield limit of elasticity and
the yield
point point
are of the are
same of the sameIt value.
value. It is assumed
is assumed that reaching
that reaching this value,
this value, hencehence the beginning
the beginning of
of plastic
plastic deformation, equals the loss of thread’s load-bearing capacity [27]. Therefore, strength
deformation, equals the loss of thread’s load-bearing capacity [27]. Therefore, strength hypotheses for
hypotheses for ductile materials are very often called the yield conditions. Given the auxiliary nature
ductile materials are very often called the yield conditions. Given the auxiliary nature of the numerical
of the numerical calculations in terms of solving engineering issues, the authors believe that applying
calculations in terms of solving engineering issues, the authors believe that applying a universal
a universal reduced stress model is sufficient. Other authors have also used the model as per the
reduced stress model is sufficient. Other authors have also used the model as per the Huber–Mises
Huber–Mises hypothesis within their numerical studies involving a threaded connection.
hypothesis within their numerical studies involving a threaded connection.
The highest stress values (Figures 5a,b and 6a,b) were obtained as expected, at the interface point
betweenhighest
The stress values
the elements of the(Figure 5a,b and
modelled, worn Figure 6a,b) connection.
threaded were obtained Theas simulation
expected, atconditions
the interface
point between the elements of the modelled, worn threaded connection. The simulation conditions
 Aerospace
Aerospace 2020,2020, 7, x FOR
7, x FOR PEER
PEER REVIEW
REVIEW 66ofof1010
Aerospace 2020, 7, 41 6 of 10
assumed
assumed loading
loading thethe filter
filter cover
cover element
element witha astatic
with staticpressure
pressureof
of 8.5
8.5 MPa.
MPa. Under
Under such
such load,
load,very
very
large
large areas
areas of of
the the modelled
modelled thread,
thread, thethe first
first three
three pitches
pitches experience
experience significantly
significantly exceeded
exceeded yield
yieldpoint
point
assumed loading the filter cover element with a static pressure of 8.5 MPa. Under such load, very large
value.
value.
areas of the modelled thread, the first three pitches experience significantly exceeded yield point value.

(a) (b)
Figure 5. The stress(a)charts for the studied shank thread with simulated wear and (b)experimental failure
load: 5.
(a)The
graph cross-section withstudied
a grid (first 2 are the most loaded pitches), (b) separated area with
Figure
Figure 5. The stress charts
stress forfor
charts thethe
studied shank shank thread
threadwith simulated
with wear
simulated wearandandexperimental failure
experimental failure
a clearly exceeded material yield point.
load:
load:(a)(a)
graph
graphcross-section with
cross-section witha agrid
grid(first
(first2 2are
arethe
themost
mostloaded
loadedpitches),
pitches),(b)
(b)separated
separatedarea
area with
with a
a clearly
clearly exceeded
exceeded material
material yield
yield point.
point.

(a) (b)
Figure6.6. The
Figure The outcomes
outcomes of ofthe
theoptimization
optimization algorithm as per
algorithm method
as per h: (a)h:convergence
method graph,graph,
(a) convergence (b)
(b)reduced
reducedstress
stress (a)
distribution, cross-section,
distribution, magnification,
cross-section, withwith
magnification, a grid for the
a grid forfirst (b) thread
the thread
first pitch.pitch.

Figure 6. The outcomes of the optimization algorithm as per method h: (a) convergence graph, (b)
Sucha asituation
Such situationmay
mayconfirm
confirm the
the load
load capacity
capacityloss
lossofofaathread
threadwith
withsuchsuchmaterial
material properties,
properties,
reduced stress distribution, cross-section, magnification, with a grid for the first thread pitch.
geometry,
geometry, diagram
diagram and andloadload value.
value. TheThe results
results of numerical
of numerical teststests
enabledenabled also
also to to determine
determine the
the features
features
of the of the
experimental experimental
test bench. test
The bench. The experimental tests were conducted for two threaded
Such a situation may confirm theexperimental
load capacity tests
losswere
of aconducted
thread with forsuch
two material
threadedproperties,
connections
connections
with a similar with a similar wear
operational operational
level. wear
It has level.
been It developed
has been developed
test test methodology,
methodology, using ausing a
specially
geometry, diagram and load value. The results of numerical tests enabled also to determine the
specially
designed designed
andexperimentaland constructed
constructed test bench test bench (diagram—Figure
(diagram—Figure 7). 7).
Thewere The test
test liquid, liquid,
withfor with
a similara similar
pressure
features of the
pressure value, resulted in test
the bench.
loss of The
the experimental
threaded tests
connection’s load conducted
capacity and, two threaded
ultimately, in the
value, resulted
connections with ina the loss operational
similar of the threaded wearconnection’s load capacity and,test
ultimately, in the “shearing”
“shearing” of thread pitches and the loss oflevel. It has
tightness been
within developed
the studied assembly.methodology, using a
The experimental
of thread
specially pitches and
designed and the loss of tightness within the studied assembly. The experimental test results
constructed
test results are shown in Figure 8. test bench (diagram—Figure 7). The test liquid, with a similar
are shown in Figure 8.
pressure value, resulted in the loss of the threaded connection’s load capacity and, ultimately, in the
“shearing” of thread pitches and the loss of tightness within the studied assembly. The experimental
test results are shown in Figure 8.
Aerospace 2020,7,7,x41
Aerospace2020, FOR PEER REVIEW 7 7ofof1010

Figure 7. Diagram of a test bench for static loading of an aircraft fuel filter assembly: 1—filter, 2—NC
filter body, 3—pressure sensors, 4—one-way valve, 5—manual hydraulic pump, 6—analogue signal
translator, 7—computer.

The obtained value of the threaded connection destructive force is 84 and 86 kN (Figure 8) for
shank No. 1 and 2, respectively, which corresponds to the value of the hydrostatic pressure inside
the assembly housing at a level of 8.2–8.3 MPa. Shank elongation confirmed by experimental tests,
relative to7.7.
Figure
Figure the tensileofof
Diagram
Diagram force
a test(increase
bench forof
a test bench internal
loadingpressure
static
for static within
loading of an aircraft
of an aircraft
fuelthe fuel
filter
fuel filter assembly)
assembly:
assembly: is of
1—filter,
1—filter, linear and
2—NC
2—NC
filter
flexible body, 3—pressure
character, in sensors,
accordance 4—one-way
with Hooke’s valve,
law 5—manual
[27]—R 2 hydraulic
coefficient ofpump, 6—analogue
approximately 0.99signal
and 1.00.
filter body, 3—pressure sensors, 4—one-way valve, 5—manual hydraulic pump, 6—analogue signal
translator, 7—computer.
translator, 7—computer.

The obtained value of the threaded connection destructive force is 84 and 86 kN (Figure 8) for
shank No. 1 and 2, respectively, which corresponds to the value of the hydrostatic pressure inside
the assembly housing at a level of 8.2–8.3 MPa. Shank elongation confirmed by experimental tests,
Shank tensile force Fr [kN]

relative to the tensile force (increase of internal pressure within the fuel assembly) is of linear and
flexible character, in accordance with Hooke’s law [27]—R2 coefficient of approximately 0.99 and 1.00.
Shank tensile force Fr [kN]

Shank elongation Δl [mm]

Shank elongation Δl [mm]

Figure 8. The results of experimental tests involving rupturing threads of two shanks.

Figure 8.
The obtained The results
value of theof experimental
threaded tests involving
connection rupturing
destructive forcethreads of two
is 84 and 86shanks.
kN (Figure 8) for
shank No. 1 and 2, respectively, which corresponds to the value of the hydrostatic pressure inside
the assembly housing at a level of 8.2–8.3 MPa. Shank elongation confirmed by experimental tests,
relative to the tensile force (increase of internal pressure within the fuel assembly) is of linear and
flexible character, in accordance with Hooke’s law [27]—R2 coefficient of approximately 0.99 and 1.00.
With the elongation value within the elastic range at a level of 3.4 mm, there is a sudden loss of
thread load-bearing capacity, pitch rupturing and a simultaneous loss of assembly tightness.
Figure 8. The results of experimental tests involving rupturing threads of two shanks.
Aerospace 2020, 7, 41 8 of 10

3. Conclusions
Many aircrafts operated in the military reach a state referred to as aging at continuous operational
service. That is why military aircrafts are in constant use for many decades, maintaining satisfactory
operational capabilities. Aging of an aircraft is not the same as it becoming obsolete. The time when an
aircraft reaches the aging state is usually much more difficult to determine. It is important to distinguish
between the characteristics of the structure of a young aircraft and an aging aircraft. Sustainment of an
aircraft is the act of keeping it operational (i.e., airworthy). Maintenance of an aircraft (that is, the work
done by mechanics in keeping it airworthy) is one aspect of sustainment. However, sustainment also
includes the engineering analyses and tests needed to determine an adequate maintenance plan for the
aircraft. Sustainment is life management. One task of sustainment is the determination of structural
inspections based on damage tolerance principles. These inspections protect against failure from
defects that could be in the structure because of manufacturing or from operational service [29].
In the absence of support from the aircraft manufacturer and the lack of original technical
documentation, it was necessary to reproduce those documentations by reverse engineering methods.
Such works was carried out at the Air Force Institute of Technology. These works allowed the
maintenance of continuous operation of post-Soviet combat aircraft in the Polish Air Force.
Unlike commercial aircraft structures, military aircraft airframe structures are rather fatigue
resistant. Therefore, in the case of combat aircraft, disasters caused by fatigue of the airframe structures
are rare. More frequent causes of combat aircraft disasters are engine damage or malfunctions of
seemingly insignificant aircraft subsystems and damages of equipment elements. This article was
devoted to the study of a fighter aircraft crash. The use of the finite element method allowed the
detection of the cause of damage in the fuel system, which was the direct cause of the accident.
The first step of this case study was developing of the geometrical model of the threaded
connection. This geometrical model was based on the measurements obtained in respect of an actual
object. The formulated boundary conditions (load, fastening) allowed to conduct numerical calculations
for a geometrical model with recreated operational wear. The results of numerical tests confronted
with environmental studies can be used as a base to evaluate the actual load-bearing capacity of a
threaded connection for a given level of operational wear. At the same time, the experimental test
results enabled evaluating and validating the developed numerical method.
Based on the results of the numerical tests, it is possible to estimate a feasible load-bearing capacity
of actual, operated threaded connections within a studied fuel system assembly. Using a computer
simulation method enabled to verify the validity of the thesis in the analysed case. The Huber–Mises
reduced stress values σred in the interfacing areas of the shank thread pitches and an aircraft fuel
assembly nut, with modelled operational wear, obtained in the course of the simulation tests are
consistent with the experimental results. With the assumed internal fuel load of 8.5 MPa over significant
areas of the studied thread, the reduced stresses σred exceed the yield point value. Such a state of
the stress is a piece of evidence of lost load-bearing capacity of the threaded connection, which was
confirmed by the results of experimental tests.
There are works devoted to forensic engineering of the causes of aviation accidents focused
on reconstruction of flight trajectories [30,31]. Simulation tests of flight dynamics in pre-disaster
time intervals are particularly important in the absence of sufficient information registered by Flight
Data Recorder (FDR) [6,32]. In cases of full registration of flight parameters by FDR, simulation
reconstruction of flight parameters is generally not necessary. However, questions arise regarding the
impact and damage of aircraft components to the occurrence of a disaster.
This article’s objective was to assist the Air Force in making difficult yet necessary choices
regarding its aging fleets in order to fulfil national security objectives at lowest cost. We have shown
the usefulness of computer methods in the process of reaching to indicate the direct cause of an aircraft
crash. It is an example of the use of computer simulation methods in forensic engineering. A simulation
study of the cause of a fighter aircraft crash was carried out. The use of the finite element method
allowed the detection of the cause of the fuel system failure, which was the direct cause of the accident.
Aerospace 2020, 7, 41 9 of 10

Author Contributions: Conceptualization, P.S., A.B. and K.S.; methodology, P.S. and G.J.; software, P.S. and G.J.;
validation, P.S., G.J. and A.B.; formal analysis, P.S. and K.S.; investigation, P.S., G.J., and A.B.; resources, P.S.; data
curation, P.S., G.J. and A.B.; writing—original draft preparation, P.S. and K.S.; writing—review and editing, K.S.;
supervision, P.S. and K.S.; funding acquisition, P.S. All authors have read and agreed to the published version of
the manuscript.
Funding: This research was internally funded by the Air Force Institute of Technology.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. MacLean, L.; Richman, A.; Hudak, M. Failure Rates for Aging Aircraft. Safety 2018, 4, 7. [CrossRef]
2. Sampath, S.G. Safety and Service Difficulty Reporting. In Aging Aircraft Fleets: Structural and Other Subsystem
Aspects; RTO Lecture Series 218, RTO-EN-015; Defense Technical Information Center: Fort Belvoir, VA,
USA, 2000.
3. Masys, A. Aviation accident aetiology: Catastrophe theory perspective. Disaster Prev. Manag. 2004, 13, 33–38.
[CrossRef]
4. Wood, R.H.; Sweginnis, R.W. Aircraft Accident Investigation, 2nd ed.; Endeavor Books: Casper, WY, USA, 2006.
5. Bibel, G.D. Beyond the Black Box: The Forensics of Airplane Crashes; The John Hopkins University Press:
Baltimore, MA, USA, 2008; ISBN1 0801886317. ISBN2 9780801886317.
6. Maryniak, J.; Lasek, M.; Sibilski, K.; Żyluk, A. Some Thoughts on Aircraft Accidents Reconstruction and
Simulation: A Forensic Engineering Case Studies (AIAA 2008-190 CP). In Proceedings of the 46th AIAA
Aerospace Sciences Meeting and Exhibit, Reno, NV, USA, 7–10 January 2008. [CrossRef]
7. Shepard, J.F. Learning from a tragedy: Explosions and flight 800. Eng. Sci. 1998, 61, 19–29, No 2.
8. Anon. In-Flight Breakup Over the Atlantic Ocean Trans World Airlines Flight 800 Boeing 747-131, N93119 near
East Moriches, New York, July 17, 1996; Report No NTSB/AAR-00/03; National Transportation Safety Board:
Washington, DC, USA, 2000.
9. Klich, E. Flight Safety; Łukasiewicz Institute for Sustainable Technologies Press: Radom, Poland, 2011;
ISSN 978-83-7789-024-0. (In Polish)
10. Lasek, M.; Sibilski, K.; Żurek, J. Aviation accidents aetiology from catastrophe theory point of view, ICAS
2010 proceedings Grant I. (ed.) CD-ROM. In Proceedings of the 27th Congress of the International Council of
the Aeronautical Sciences, Nice, France, 19–24 September 2010; ICAS 2010-707TP. ISBN 978-0-9565333-0-2.
Available online: http://www.icas.org/ICAS_ARCHIVE/ICAS2010/ABSTRACTS/707.HTM (accessed on
7 April 2020).
11. Balac, M.; Grbovic, A.; Petrovic, A.; Popovic, V. FEM analysis of pressure vessel with an investigation of
crack growth on cylindrical surface. Maint. Reliab. 2018, 20, 378–386. [CrossRef]
12. Chodurski, M.; D˛ebski, H.; Samborski, S.; Teter, A. Numerical strength analysis of a supporting frame for a
universal head of a palletizing robot. Maint. Reliab. 2015, 17, 374–378. [CrossRef]
13. D˛ebski, H.; Koszałka, G.; Ferdynus, M. The application of FEM in analysing the structure of a supporting
frame in a trailer with variable operating parameters. Maint. Reliab. 2012, 2, 107–113. Available online:
http://www.ein.org.pl/2012-02-03 (accessed on 7 April 2020).
14. Baranowski, P.; Damaziak, K.; Małachowski, J. Brake systems studies using numerical methods. Maint. Reliab.
2013, 4, 337–342. Available online: http://ein.org.pl/2013-04-07 (accessed on 7 April 2020).
15. Lisowski, E.; Panek, M. CFD modeling of vane pump vanes operation. Maint. Reliab. 2004, 2, 36–41. Available
online: http://www.ein.org.pl/2004-02-08 (accessed on 7 April 2020).
16. Lonkwic, P.; Różyło, P.; D˛ebski, H. Numerical and experimental tests of a progressive gripper structure using
the finite element method. Maint. Reliab. 2015, 4, 544–550. [CrossRef]
17. López-Campos, J.A.; Segade, A.; Casarejos, E.; Fernández, J.R.; Vilán, J.A.; Izquierdo, P. Finite Element Study
of a Threaded Fastening: The Case of Surgical Screws in Bon. Symmetry 2018, 10, 335. [CrossRef]
18. Yu, Q.; Zhou, H.; Wang, L. Finite element analysis of relationship between tightening torque and initial load
of bolted connections. Adv. Mech. Eng. 2015, 7, 1–8. [CrossRef]
19. Kopecki, T.; Zacharzewski, J. Fatigue life and stress state analysis of cracked thin-walled plate under cycles
axial tension. Maint. Reliab. 2006, 3, 19–26. Available online: https://www.ein.org.pl/ein/2006-03-0 (accessed
on 7 April 2020).
Aerospace 2020, 7, 41 10 of 10

20. Kopey, B.; Kopey, V.; Lyskanych, M. Tool-joint thread modeling by finite element method. Drill. Oil Gas
2005, 22, 201–204. Available online: http://journals.bg.agh.edu.pl/WIERTNICTWO/index.php?vol=2005-01
(accessed on 7 April 2020).
21. Sun, Y.J.; Liao, R.D. The Effect of Helix on the Nonlinear Analysis of Threaded Connection. Adv. Mater. Res.
2010, 148–149, 1741–1744. [CrossRef]
22. Wittenberghe, J.V.; De Baets, P.; De Waele, W. Modelling of preloaded threaded pipe connections.
In Proceedings of the 8th National Congress on Theoretical and Applied Mechanics, Brussels, Belgium, 28–29
May 2009; Available online: https://lib.ugent.be/catalog/pug01:673974 (accessed on 7 April 2020).
23. Dacko, M.; Borkowski, W.; Dobrociński, S.; Niezgoda, T.; Wieczorek, M. Finite Element Method in Structural
Mechanics; Arkady: Warszawa, Poland, 1994. (In Polish)
24. Gawecki, A. Mechanics of bar Structures and Materials; Poznań University of Technology Press: Poznań,
Poland, 2003.
25. Loska, A. Remarks regarding the modelling of operating processes using scenario-based techniques. Maint.
Reliab. 2012, 2, 92–98.
26. Markowski, T.; Mucha, J.; Witkowski, W. FEM analysis of C-fram rigidity in a device for manufacturing
clinching joints. Maint. Reliab. 2013, 1, 51–57.
27. Brzoska, Z. Strength of Materials; PWN: Warsaw, Poland, 1972. (In Polish)
28. D˛ebski, H. Experimental and numerical studies involving post-critical behaviour of composite columns with
an omega cross-section. Maint. Reliab. 2013, 2, 106–110. Available online: http://www.ein.org.pl/2013-02-03
(accessed on 7 April 2020).
29. Underwood, P.; Waterson, P. Systemic accident analysis: Examining the gap between research and practice.
Accid. Anal. Prev. 2013, 55, 154–164. [CrossRef] [PubMed]
30. Calkins, D.E. Aircraft accident flight patch simulation and animation. J. Aircr. 1994, 31, 2. [CrossRef]
31. Dietenberger, M.A.; Haines, P.A.; Luers, J.K. Reconstruction of Pan Am New Orleans accident. J. Aircr. 1985,
22, 8. [CrossRef]
32. Sibilski, K. Some Thoughts on Mathematical Models for Aircraft Accidents Simulation. In Aviation Safety;
Soekha, H., Ed.; VSP Publishing Company: Utrecht, The Netherlands, 1997.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).