ONGERUBRICEERD
Lange Kleiweg 137
2288 GJ Rijswijk
Postbus 45
2280 AA Rijswijk
The Netherlands
wwwtno.nl
T +31 88 866 80 00
F +31 88 866 69 49
infodesk@tno.nl
TNO-report
TNO 2015 M11094
Damage reconstruction due to impact of highenergetic particles on Malaysia Airlines flight MH17

Date                                        August 2015
Author(s)                                   -
Copy no.                                    -
No. of copies                              -
Number of pages                      41
Number of appendices                2
Customer                                 Onderzoeksraad voor Veiligheid (DSB)
Project name                            Fragmentuitwerking
Project number                         060.15020
Classification report                  Ongerubriceerd
Classified by                             Ministerie van Defensie (NLD-MoD)
Classification date                     28 August 2015
                                              This classification will not change
Title                                        Ongerubriceerd
Report text                              Ongerubriceerd
Appendices A, B                       Ongerubriceerd

The classification designation Ongerubriceerd is equivalent to Unclassified.
This report is a translation of TNO report 2015 M11093. In case of interpretation differences
the text in TNO report 2015 M11093 is leading.
All rights reserved.
No part of this publication may be reproduced and/or published by print, photoprint,
microfilm or any other means without the previous written consent of TNO.
In case this report was drafted on instructions, the rights and obligations of contracting
parties are subject to either the General Terms and Conditions for commissions to TNO, or
the relevant agreement concluded between the contracting parties. Submitting the report for
inspection to parties who have a direct interest is permitted.
© 2015 TNO

3 / 25

1 Introduction

5 / 25

3 Visually observable damage
Four TNO employees with expertise in the field of weapon systems performances
and weapon effects on structures have recorded the visually observable damage.
Inspection of the wreckage occurred in Gilze-Rijen air base in the February 2015 –
August 2015 timeframe, involving the airplane parts present at the time, and
focussed on the forward section of the airplane.
3.1 Established state of the airplane
According to the DSB about 20-25% of the airplane was recovered (situation
February 2015). The parts are being kept on different sites on the air base. Airplane
parts belonging to the forward section of the fuselage and the left wing have been
brought together in a hanger and spread out according to the frames scheme of the
Boeing 777-200ER (see Figure 3.1). In this manner an impression is obtained of the
relative position of recovered parts to each other.

Figure 3.1: Reconstruction of the MH17 using recovered parts of the airplane (source: TNO). The
foreground shows the cockpit area and the left engine cowling is visible in the far right.
The relative height of the different parts is not is not factual.
The TNO employees gathered their impressions in view of the spread out wreckage
in the hangar. The observable damage area turned out to be incomplete because
large parts were missing. Many airplane parts have not yet been recovered
The spatial position of the recovered parts is/will be realised by another entity in a
computer environment. That work involves categorising the airplane parts by means
of identification numbers. In the following the identification numbers are mentioned
along with the airplane parts. The frames scheme offers an additional possibility to
determine locations. The frames are labelled with station (STA) numbers (see
Figure 3.2 and Figure 3.3). The difference between two STA numbers equals the
distance between two stations in inches.

7 / 25

subject to deceleration by the atmosphere, perforate the airplane structure and lose
kinetic energy with every next perforation.
3.2.1 Observability of fragment damage
Warhead fragments are being ejected in a certain preferred direction (see section
4.3), limiting perforation damage due to fragments to specific areas on the airplane.
The search is for a “ribbon” of perforations over the airplane contour.
Fragmenting warheads meant for killing air targets are, as far as currently known,
only fitted with preformed fragments. This implies that the fragment shape is fixed
and that fragments are spread out in a known fashion upon the casing of the high
explosive charge. The casing is generally made of metal and not preformed, so that
there will always be an additional (small) cohort of naturally formed fragments with a
random fragment shape. Ejection of preformed fragments results in a characteristic
holes pattern on the airplane.
3.2.2 Observed fragment damage
Figures 3.4, 3.5, en 3.6 describe observed fragment damage on the airplane parts.
Next to these parts, fragment damage has also been found in a panel from het
cockpit roof. The fragment damage is restricted to an area. No fragment damage
has been found on the recovered left panel between about STA 220 to about STA
410 (see Figure 3.2) and more remotely located panels (higher STA numbers). The
cockpit bulkhead AAHZ3064NL (see Figure 3.4) on STA 132.5 appears to be the
most forward limit of the fragment damage.

Figure 3.4: Cockpit panel left AAHZ3163NL with enlarged detail (source: DSB). The perforation
damage displays a regular pattern of larger and smaller holes. Furthermore, the skin
plating is damaged by pitting, which can be caused by the impact of many small and
hot particles, such as powder residue and molten metal. The pitting damage is found
locally; neighbouring panels are not damaged by pitting.

9 / 25

3.3.1 Observability of blast damage
Blast damage can be concealed by perforation damage, damage due to the breakup of the airplane and the ground impact. In this case, of all damage typical for blast
the dishing damage is the best visually observable.
Depression of skin panels can also be caused by bending of airplane parts during
break-up and impact of airplane parts on the ground. Several depressions have
been found that cannot be attributed to dishing with sufficient certainty.
3.3.2 Observed blast damage
In the cockpit panel on the left-hand side (Figure 3.4) the airplane skin is pressed
against the more backward region between the vertical stiffeners. This damage may
be cause by overpressure due to blast.
Figure 3.7 shows a fuselage panel to the right of the nose gear, between STA 250
and STA 330. This panel displays dishing damage. A neighbouring part of the nose
gear door (STA 184) did not have visible blast damage. However, that part of the
nose gear door is made of a honeycomb structure that is particularly resistant
against the effects of blast.

Figure 3.7: Panel AAHZ3112NL with visible dishing damage (source: TNO). Het panel is also
damaged due to break-up of the airplane and the ground impact.

The fuselage skin part AAHZ3258NL on the left-hand side between STA 230 and
STA 420 (see Figure 3.8) does not have observable blast damage. This panel
delimits the blast damage area. The lower located region of this panel is severely
deformed, probably due to break-up of the airplane and the ground impact. The
cockpit bulkhead (Figure 3.6) had also no observable blast damage.

11 / 25

Locally, holes have been observed, on a regular distance from each other, with
different hole shapes and hole areas. This fact points to a warhead with several
different specific fragment shapes.
Panel AAHZ3163NL is damaged by pitting. The pitting damage area is delimited
and can be caused by hot particles of a nearby detonating warhead.
The observed damage pattern cannot be caused by full-bore projectiles from a
medium calibre gun system, such as a gun system from a fighter aircraft or from
anti-aircraft artillery. The distance (pitch) between individual holes is too small, the
hole areas are too small and the hole area variation is too large.
Panel AAHZ3112NL and panel AAHZ3163NL display visually observable damage
that is suspected to stem from blast effects. Damage to the other inspected panels
cannot be attributed to blast with sufficient certainty. Panel AAHZ3258NL is, of the
panels without observable blast damage, the panel closest to the detonation.
Initially, blast expands spherically after a warhead detonation. However, blast flows
around an obstacle to load the backside of that obstacle as well. This makes it
possible that blast damage can occur on the right-hand side of the airplane,
following a detonation on the left-hand side. This phenomenon is illustrated in
Figure 3.10

Figure 3.10: Front view and top view with a detonating warhead to the left of the airplane (7.2 ms after detonation event) [9].
Blast (coloured lines) originating from the point of detonation also arrives at the panel on the right-hand side of the
airplane

13 / 25

Figure 4.2: Position and velocity of airplane and guided weapon (carrier of the warhead) in the
reference coordinate system. Note that the warhead position is a variable in the
terminal ballistics simulation. In other words: this figure is meant to illustrate and does
not provide the estimated position of the warhead.
Relevant velocities have been obtained from the NLR and DSB (see Table 4.1).
A possible roll angle, angle of attack and drift angle of the airplane have been
assumed to be negligible [7]. NLR has determined the probable terminal velocity of
the guided weapon by means of a fly out simulation [7].
Table 4.1: Engagement velocities at the moment of warhead detonation. Two velocities have
been considered for the terminal velocities of the guided weapon.

* The precise value is not released for publication

4.3 Warhead
Starting point for the terminal ballistics simulation is a warhead with preformed
fragments. In consultation with DSB, NLR and the Netherlands Ministry of Defence
warhead 9N314M of Surface to Air Missile (SAM) type 9M38M1 has been modelled.
4.3.1 The physical warhead
From open sources it is known that warhead type 9N314M weighs about 70 kg and
contains three different fragment shapes, denoted as bowtie, filler and square (see
Figure 4.3) [4]. To model the warhead to a sufficient level of detail supplementary
data from national sources is used.
In the course of the investigation the Russian manufacturer of SAM type 9M38M1,
JSC Almaz Antey (Алмаз-Антей), provided more insight in the functioning of the
warhead 9N314M [5]. This data is also used in the investigation.

15 / 25

Table 4.2 provides a summary of the most important performance differences
between the designs. Design II uses the specified fragment properties and
detonation position according to Almaz Antey [5]. The corresponding ejection
angles and fragment velocities of design II have been determined with the Split-X
software v.5.3.1.0. Design III adopts all warhead performances according to Almaz
Antey without any adaptations.
Table 4.2: Performances of three warhead 9N314M designs. Design I is based upon national
sources, design III is based upon Almaz Antey information [5]. Design II uses the
geometric design according to Almaz Antey, but the corresponding ejection angles
and fragment velocities are calculated by TNO.

* This property is not released for publication.
** Zero degrees in longitudinal direction pointing forward.

TNO rates design II as being the most realistic for the purpose of this investigation
because of the physical basis of the design. The main difference with design III is
the smaller angular range for the fragment ejection. Note that the warhead model
only contains preformed fragments. Other fragments that occur with the break-up of
the SAM are not included in the model.
4.3.3 Warhead modelling in Split-X
The Split-X software used in this investigation is commercially available. The
supplier Numerics reports on its website that Split-X it is based on analytical
procedures and engineering approximations calibrated using experimental results
[10]. Split-X is validated (i.e. model results correspond with test results) for the
fragment ejection angles and corresponding fragment velocities of warheads with
natural fragments as well as preformed fragments.
With the start of every study, TNO assesses the suitability of software for producing
sufficiently truthful results. Split-X has been validated by TNO for several different
ammunition types and warheads. On this basis Split-X has been found suitable for
use in this study. TNO uses Split-X and related models regularly for study
assignments from the Netherlands Ministry of Defence and other sponsors.
However, details on the TNO-internal validation process cannot be released due to
classification guidelines.
4.3.4 Simulation of warhead fragment trajectories
The fragments move in a straight line from the point of launch. The point of launch
is a position on the warhead with a given launch angle and velocity. The terminal
ballistics simulation calculates trajectories for each individual fragment. The finite

17 / 25

5 Damage matching
In this chapter the observed damage pattern and the simulated damage pattern are
compared. The terminal ballistics simulation makes it possible to vary position and
orientation of the warhead relative to the airplane until both fragment damages
match with each other as good as possible.
5.1 Variation of warhead position and orientation
Figure 5.1 provides an overview of the simulation strategy. The simulation provides
an impact pattern from the fragments on the airplane contour. Starting point for the
terminal ballistics simulation is the warhead position (X, Y, Z) and orientation
(azimuth, elevation) according to the initial assessment by NLR [7].

Figure 5.1: Variation of the warhead position and orientation. Starting point for the simulations is
the initial position and orientation according to NLR. The blue dotted line is an aid to
determine the absolute distance between the detonation point and the airplane
contour.

The terminal ballistics simulation is carried out as follows for each of the three
warhead designs:
1. Determination of the fragment impact pattern with the position and
orientation according to the initial assessment by NLR.
2. Variation of warhead position along the SAM direction of flight;
determination of the influence on the fragment damage pattern.
3. Variation of warhead position perpendicular to the SAM direction of flight;
determination of the influence on the fragment damage pattern.
4. Adjustment of the original warhead position (matching).
5. Adjustment of the orientation of the adjusted warhead position (matching).

19 / 25

5.3 Comparison of observed and simulated fragment damage
The matching makes use of the observed boundary of the fragment damage area,
the matching conditions derived for this area, and the observed distance (pitch)
between individual fragments (see Chapter 2).
5.3.1 Fragment distribution within the damage area
The simulated ejection of individual fragments requires some clarification. The
figures 4.3 and 4.4 show a sudden transition in the thickness of the outer geometry
of the warhead. Due to the applied simulation model, a sudden transition occurs in
the fragment velocities over the airplane outer geometry. The consequence for the
fragment damage area is shown in Figure 5.3.

Figure 5.3: Simulated fragment impact on the airplane contour (on left the impact pattern and on right the fragment
trajectories). The impact from bowtie, filler and square fragments is colourised red, blue en yellow, respectively.
The transition between the areas with and without square fragments corresponds with a leap in the fragment
velocities, resulting in a strip without impacts.

In reality the fragment velocity leap will be less sudden, so that the two strips with
fragment impacts will merge into each other. For the best matching result the pitch
between individual fragments in the wide strip with impacts has been used.
The regular holes pattern, typical for a warhead with preformed fragments is also
visible in the simulation. See Figure 5.4 with a side view (X-Z plane in the reference
coordinate system) of the airplane. When one imagines the warhead as a (wooden)
barrel, the preformed fragments are ordered along the circumference of the barrel,
as well as along the length of the barrel. The ordering along the length would then
follow the “staves” of the imaginary barrel.

21 / 25

Figure 5.6 provides an illustration of the matching perpendicular to the assumed
SAM velocity vector. Moving a half metre towards the airplane matches best with
the observed pitch between individual fragments (see Figure 3.9).

Figure 5.6: Matching perpendicular to the assumed SAM velocity vector. The red impacts
represent the initial situation. The relative distance between fragments decreases
when the warhead detonates closer (blue impacts) and increases when it detonates
further away (green impacts).

The results of the applied matching procedure for the different warhead designs are
available in Appendix A. The best match is provided in Table 5.1 and Figure 5.7.
Note that the match is a combination of position and orientation; if one would
change the position, the orientation changes as well. The corresponding smallest
distance between the geometric centre of the warhead and the cockpit is about 3 m.

Table 5.1: “Best match” warhead position (geometric centre) and orientation in the reference
coordinate system

23 / 25

non-compliancy with the set conditions (see Section 5.2). Figure 5.8 illustrates the
differences between the best fitting 70 kg and best fitting 40 kg warhead. The 70 kg
match is better.

Figure 5.8: Red: fragment impacts for “best match” warhead design II (70 kg 9N314M). Blue:
fragment impacts for “best match” warhead design C (hypothetical 40 kg). Design C
results in a less fitting match.

Design C is extreme, in the sense that the angular range of the fragment ejection is
made as large as physically possible. Only with an extreme angular range it proves
possible to remotely approximate the observed damage pattern.
The damage pattern of a lighter warhead closer to the airplane does not resemble
the damage pattern of a heavier warhead further away from the airplane. Therefore,
TNO judges the hypothesis that a lighter warhead can cause the observed damage
as being improbable.

25 / 25

7 References
[1] Dutch Safety Board, Rapport van eerste bevindingen – Crash van Malaysia
Airlines Boeing 777-200 vlucht MH17, Hrabove, Oekraïne – 17 juli 2014, The
Hague, September 2014.
[2] Deursen, J.R. van, Springpunt en schadebeeld (Vertrouwelijk), TNO
quotation number 919486, quotation letter 15 EBP/032, Rijswijk,
29 January 2015.
[3] Assignment agreement between DSB and TNO, inzake het uitvoeren van de
opdracht met betrekking tot het opstellen van een advies in verband met
MH17, TNO-reference SP50749, 16 February 2015.
[4] Internet, wwwwhathappenedtoflightmh17.com, website visited March 2015.
[5] JSC Almaz Antey, Materials – according to the characteristics of the rockets
BUK, requested by experts NLR (English translation), Reference nr.
01-09/548K, 29 July 2015.
[6] http://www.turbosquid.com/3d-models/3d- … el/816230, website visited March 2015.
[7] NLR, flight data, e-mail message, 28 Januari 2015, National Aerospace
Laboratory (NLR).
[8] JSC Almaz Antey, Findings of expert assessments on MH17 accident –
Modulation results (English translation), presentation, 6 August 2015.
[9] TNO, Numerical simulation of blast loading on Malaysia Airlines flight MH17
due to a warhead detonation, TNO report 2015 M10626, Rijswijk, May 2015.
[10] Numerics, SPLIT-X - Fragmentation Warhead Expert System, folder,
download from website www. numerics-gmbh.de, accessed April 2015.

Appendix A | 2 / 9

A.1.1 Design I, terminal velocity ~600 m/s
The simulation result is shown in Figure A.1. The most favourable warhead position
and orientation fulfils the match conditions. Figure A.2 shows an alternative
warhead position and orientation resulting in a reasonable match. This shows that
there is some ‘room to manoeuvre’ (solution space) when establishing the match.

Figure A.1: Impact pattern (above) and corresponding fragment trajectories (below) of warhead design I,
velocity ~600 m/s, position [-0.4 m, -3.5 m, 3.7 m], orientation [-17°, 7°] and number of hits 769.