This report was originally compiled by me
for the database of accidents. As the copyright is mine and
I was not paid to produce the report, I feel at liberty to reproduce
it. As usual, names are omitted.
the opening picture classically demonstrates how far mine-clearance
machine development has advanced in the ten years 1993-2003,
and if you want a higher resolution copy, just ask.
Accident report: NVESD's MCC in Angola
it occurred: 10:00
Date it occurred: 30th October 2003
Where it occurred: Chana Pale, Xangongo, Cunene Province
Map reference: S16° 44' 28"; E15° 02' 54"
Class: Excavation accident
Primary cause: inadequate equipment
Secondary cause: inadequate equipment
Date of main report: 31st October 2003
Organisation: [Consultancy company and Demining NGO names excised]
US Army CECOM NVESD
Mine/device: TM57 AT blast
Ground condition: grass covered, dry, agricultural
Accident notes: non injurious accident (?); mechanical detonation
Mine Accident Investigation
report of an independent investigator was made available in
November 2003. It is reproduced below, edited for anonymity.
accident involved the "Mine Clearance Cultivator", a machine
owned by the US government and under trial with [a demining
NGO] in Southern Angola.
photograph above shows the MCC alongside the T55 tank after
report was compiled by [name excised] as an independent inquiry
into the circumstances surrounding an accident in which a machine
was damaged but no one was injured. The report is written on
behalf of (and at the request of) [demining group]. The author
was within 30 kilometres of the incident site when it occurred,
engaged in an unrelated task, so was able to attend the accident
site within an hour of the accident occurring. The staff involved
were interviewed and the damage to the machine inspected with
a view to carrying out an appropriate investigation according
to the norms within Humanitarian Demining.
of incident: 30th October 2003
Time of incident: approximately 10:00 hrs
Place of incident: Chana Pale, Xangongo, Cunene Province, Angola.
Access: Road No,6(8)A from Xangongo to Ondjiva. Turn onto a
dirt road 6.3 km from Xangongo and drive 3.4 km through bush.
GPS: S16° 44' 28"; E15° 02' 54"
accident occurred at Chana Pale, in the 3rd (outer) ring of
mine belts around the city of Xangongo, Cunene Province, Angola.
ground is level in the area with scattered bush in the surroundings.
The area to be worked is clear of trees. Because of the level
ground, large areas become shallowly submerged in the rains,
and some of these only completely dry out for a short time in
the dry season. The incident occurred in one of the areas that
are shallow ponds for much of the year and over which there
is now long, dry, marsh grass. The soil was dry but friable.
people have been moving back to the area in the last year and
establishing Kimbos (brush-fenced enclosures of thatched huts).
There is a Kimbo within 200 metres of the accident site.
incident occurred in an area that had been the outer ring of
three defensive minefields around the city of Xangongo.
AT mines present are reported to have been laid mechanically
by Cuban forces in support of the government. The mines used
are reported to have been TM62B, TM57 and TM46 Anti-Tank mines.
The mine-laying machine dug holes at 3m spacing and the mines
were placed either by men or machine. There were two (sometimes
three) rows of tank mines. In some areas each mine on the outer
(enemy) side was protected with three PMD-6 AP blast mines.
Some mines in the area had been collected and removed by local
people. This was in response to a "bounty" offered for mines
across the border in Namibia. The bounty was paid by the South
African armed forces based in Namibia at that time. As a result,
erratic gaps in the mine belts occur.
No. 368 in the Database of Demining Accidents (UNMAS/GICHD)
occurred at the same place (within four metres) on 18th March
1993. The Humanitarian Demining group involved in that incident
were named [name excised]. They had deployed disarmed Soviet
T55 tanks with KMT5 roller systems mounted on the front in the
hope of carrying out rapid mechanised clearance of the area.
At that time, the area was wet and the tank was low to the ground
when a TM-57 AT mine detonated under a track (having been missed
by the rollers). The hull was breached and one occupant of the
tank was killed. The severely damaged tank was deliberately
burnt out and remains in place. The mine involved was identified
by inference because other mines found in that part of the mine
belt (located manually) were said to have been TM-57s.
clearance records were not accurately kept at that time, so
no-one is entirely certain whether the wet area around the tank
was later cleared manually (in the next dry season, or later).
Some of the area traversed by the T-55 was later cleared and
it was claimed that functional TM-57 mines were found. This
meant that the entire area over which the T55 roller system
was used "suspect". Prior to the MCC deployment, local people
reported that the area around the wrecked T55 tank was still
mined but the information was considered uncertain because these
people were not there at the time of the first mine clearance
attempts, having only returned to the area recently.
threat mines were believed to be TM-57 Soviet AT mines. The
pre-deployment assessment did not include an expectation of
finding PMD-6 mines in this part of the mine-belt.
the work of the MCC it uncovered and drove over a PMD-6 AP mine,
crushing it but not detonating it. Manual deminers found it
and another MUV fuze (as used in the PMD-6) in the area traversed
by the MCC. They could not find the second wooden box and believed
that it may have been broken up and driven deep into the ground.
The box has two small hinge pins of ferrous metal but these
often disintegrate with the passage of time (especially in wet
areas) so the rest of the mine can be impossible to detect with
a metal-detector. Without the fuze the mine presents, of course,
a very low threat.
photograph above shows the crushed PMD-6 recovered by the deminers.
MUV fuzes were heavily corroded but still included the "delay"
arming section at the top and the remains of a rusted arming
pin. These mines were not armed.
the PMD-6 was found (24th October), the threat assessment was
expanded to include AP mines. As there are people living in
the area and crossing the suspect land, the risk of civilian
injury was considered higher and the task's priority enhanced.
MCC in use in Angola was remotely controlled while in the suspect
area. Control was from a large armoured vehicle named the "Cougar".
The Cougar and the MCC's carrier were fitted with a total of
four cameras, with screens showing their pictures inside the
Mine Clearance Cultivator is a large "agricultural" attachment
designed to fit a heavy carrier vehicle (bulldozer). The attachment
uses shaped tines to "till" the ground and a wormscrew auger
to carry large items that are uncovered to one or other side
of the machine.
MCC is described on its designers website [ http://www.humanitarian-demining.org/demining/clearance/mcc.asp
]. See Related papers.
particular deployment had included another attachment for the
bulldozer. This was a "sifting" device intended to process "berms"
left by the MCC. Field staff pointed out that the MCC did not
leave appreciable berms and so the sifter had not been deployed.
Because it was not used (and was not intended to be used) at
the accident site, I did not examine the sifter and had no opportunity
to see it work.
16th October the MCC deployed to the accident area and it began
work on 20th October. The task-plan involved processing the
perimeters of the area beyond the flood line, starting from
the outside with a safe lane that marked out a "box". All areas
traversed by the MCC were to be cleared manually after the machine
had passed. It was intended that the safe lane should mark the
perimeter of the area, but this was varied when local people
wanted the area around the tank cleared as a priority. The perimeter
was unmarked. Marking tape was not used because the MCC needed
to be able to manoeuvre as the base line was moved forward following
manual clearance of strips processed by the MCC.
cultivated two adjacent straight-line sides of the "box" it
would work, then worked in from these "sides" towards the ruined
T55 tank. For safety, the "box" was considerably larger than
the suspect area (approximately 90 x 90 meters). The MCC made
15 passes on the East/West side, and 7 on the South/North side
leading up to the tank. On the 8th pass South to North, which
was the first on the East side of the tank, the incident occurred
as the MCC was passing the tank. [Diagram removed.]
arrived at the site at around 10:50 and found the [demining
group] deminers checking around the MCC with metal detectors.
They were removing a large amount of debris from the explosion,
most of which was on the surface. A grassfire around the explosion
had either burned out or been extinguished.
reported to the Site Manager and explained that I had been asked
by [the Demining group] to make an investigation. The Site Manager
did everything possible to assist me, providing an interpreter
and helping me to ask questions and understand answers. All
[demining group] personnel were open and frank at all times.
personnel interviewed were [four names excised].
team had deployed at around 08:00 and the MCC machine was brought
into the suspect area from its place of safe-keeping nearby
at around 08:30. It was manually driven to the edge of the safe
lane and thereafter operated remotely from the Cougar armoured
command vehicle about 120 metres away (as far away as possible
while still allowing a camera view through trees). The threat
of MCC damage from concealed roots (left after burning trees
down to make charcoal) meant that progress was cautious. The
MCC's tines had already been damaged by contact with concealed
roots of dead trees at the site. The MCC made one pass South
to North on the West side of the T55, passing as close as possible
to it. It was driven (RC) back to the South side and lined up
for a pass on the East side of the tank. Manual deminers checked
the area it had just traversed. Then the MCC made its first
that pass on the East side of the T55 (and as its cultivator
attachment began to pass alongside the T55 turret) there was
[demining group] hand-held radio communication was not possible
because the use of these radios interfered with the video signal
from the cameras mounted on the MCC. The images from these cameras
allowed the machine to be controlled with precision by operators
who did not have direct line of sight themselves (an external
camera on the command vehicle did have direct line of sight).
the detonation, the signal from the MCC's video cameras was
lost. The machine stopped moving but its engine was running.
[The Site Manager] left the command vehicle using the door on
the safe side and warned the rest of the team to stay behind
the vehicle while he went to investigate. He went to the T55
crossing the cleared area and stood on the T55 tank to assess
the damage. A grass fire was spreading away from the accident
site but was not providing an immediate threat to the MCC. The
MCC motor was still running. He was able to pass from the tank
to the MCC in order to switch the motor off.
I arrived at the site, the MCC was standing alongside and slightly
behind the T55 tank. The distance between them was about 40
centimetres at its closest point.
found that the device had detonated on the left-hand side (facing
forwards) of the MCC's cultivating attachment. It had apparently
detonated above ground level and on top of the auger to the
rear of the attachment, exploding as it passed to the side of
the machine. Damage to that side of the auger and entire cultivating
attachment was severe, with large parts separated. This means
that the mine detonated on the T55 side of the MCC, and there
was some evidence of sooting and light fragmentation damage
to the side of the T55 tank. The tank was in poor condition
and some damage may have been old. There was no evidence that
either the tank or the MCC were physically lifted or moved sideways
by the detonation.
rear window of the MCC cab had sprung inward (unbroken) and
the control panels come loose. There was also some damage to
the cab structure, with one door jammed and another twisted.
The glass inside the camera on that side was broken, although
the cover (presumably polycarbonate) was not.
left side of the cultivator attachment was wrecked as shown
below (this picture was taken after the machine had been withdrawn
from the suspect area).
examined the machine thoroughly before it was moved (driven
manually) backwards along its own tracks out of the suspect
area. For this manoeuvre, all personnel except the driver, [the
Site Manager], withdrew 100 metres behind the armoured command
vehicle (Cougar). A deminer then checked where the MCC had been
standing and found further debris from the detonation. None
of that debris was from the device involved in the accident.
There was no crater, but there was evidence of a point of initiation
with "teeth" broken from the auger driven into the ground -
indicating that the device was above those teeth when it detonated.
Damage to the auger's drive-train on that side was catastrophic
(all parts separated, including parts of the triple-width chain)
which confirms where the detonation occurred.
picture below shows the auger damage.
sooting on the cultivator backplate supported the inference
about the place of detonation. I checked the ground at the point
of initiation but could find no evidence of the device involved.
The ground at that place had been passed by the MCC's tines,
and was disrupted to a depth of 8-10cm. There was scrap metal
contamination alongside the T55 tank and the tank itself made
a fast metal-detector search difficult. I did not spend hours
looking for evidence of the mine involved because I had no reason
to expect to find any evidence.
asked about the MCC's past performance and was told that it
had largely been deployed in safe areas as a data-gathering
and testing exercise. During this time, dummy mines provided
by its owners had sometimes been placed in the ground at a shallow
depth to help assess its ability to unearth mines. It had unearthed
some of these concrete dummies, but it had also missed at least
one and run over it with its tracks causing damage. (See photograph
below.) The relatively hard soil in the dry season caused regular
tine-damage, breaking shear bolts and sometimes welds in the
cultivator attachment. This meant that the tines could not be
deployed to dig deeply, even in the softer ground of the accident
this year, prior to [the Site Manager's] employment, the MCC
had been moved to its first genuinely suspect area. As it drove
into the area cultivating a strip that would become a safe-lane
after manual clearance, it detonated an AT mine on top of the
auger and on the left side of the attachment. The mine was probably
in the open inside the attachment and damage was far less extensive
than in this accident, but the machine had to be withdrawn for
extensive repair and the replacement of parts. On that occasion,
the rear window of the cab shattered. Remote control of the
machine was also lost on initiation, and the machine stopped
advancing (presumably as it is designed to do).
deployment at the accident site, the MCC has missed and crushed
with its tracks at least one PMD-6 mine and probably two. One
was found during manual demining as follow-up. A second intact
fuze was found.
MCC struck an explosive device as it passed close to a ruined
T55 with roller-attachment that had also struck an explosive
device and been severely damaged in 1993. The MCC was being
operated safely and while paying due cognisance to all the available
information about threats and local concerns.
fact that local people believed the area was dangerous may be
of greater significance than is obvious. Frequently, local people
will move unknown or suspect devices out of their way, putting
them in a place that everyone knows should be avoided (such
as the obvious wreck of the T55 tank). People may have placed
ERW items close to the tank. The fact that they knew that a
deminer died inside the tank may have meant that it would be
disrespectful for them to enter the tank itself, so placing
the items close by might be expected. The items moved to the
vicinity of the tank could have been far more extensive than
mines. However, I found no evidence of penetrative or fragmentation
damage that implied that the device involved was anything other
than a large blast mine.
explosive device involved in this incident is most likely (by
inference) to have been a TM57 with 6.34 kg of TNT. The extent
of the damage implies that it may have been a TM-62B with a
main charge of 7.5 kg TNT.
mine appeared to have initiated as it left the auger and so
below or beyond the left-side upper support arm of the cultivator
and alongside the auger-drive mechanism. This means that the
detonation was confined on one side by the cultivator and on
the other by the wrecked T55 tank. The T55 is embedded in the
ground and has not moved (crusted earth around the wheels and
rollers was unbroken). Reflection from the tank may have had
the effect of intensifying the blast damage to the cultivator
and auger mechanism.
cultivating attachment is so severely damaged that it may not
be economic to repair it.
While I have not carried out an extensive engineering assessment
in a workshop, repair requirements to the "bulldozer" carrier
seem to be limited to cab-repair and the replacement of one
major bolt-on ball-joint, where the left arm of the "C" frame
is attached. The electronics involved in the RC and Video systems
will also require repair/replacement.
line with normal HD accident investigations, "recommendations"
are included in the hope of providing guidance to avoid any
repetition of the accident and to increase safety.
machine had only worked on a total of a few hundred square metres
of suspect ground when this accident occurred. Over that few
hundred square metres, its performance had been demonstrably
poor (with two known missed and two detonated devices). The
missed AP mines mean that it must always be followed by full
manual clearance, and the devices it leaves behind may well
be damaged (as with the PMD-6 found at this site), so the follow-up
requires increased caution in response to an elevated risk assessment.
For this reason, full manual follow-up is likely to take as
long (or longer) than it would take to clear the area manually
in the first place. The detonated AT mines have done extensive
and costly damage to the machine and imply that its design (intended
to move mines to one side) is flawed.
threat mines in the area are the TM57 and TM62B AT mines and
the PMD-6 AP mine. The ground is not electromagnetic, although
there is fragmentation in the immediate vicinity of the T55
tank. The ground is relatively soft and the area is clear of
significant undergrowth. It also has easy access. These facts
mean that the area could have been cleared manually with relative
ease and with no unusual costs.
cost of the MCC operation is not known, but can be presumed
to be high given the cost of the machine and its control vehicle,
their deployment, their running costs and their staffing by
dedicated support personnel. It has worked slowly (very slowly
when compared to designer's claims - see Related papers) and
has required frequent running repairs. My estimate of the repair
costs involved in both accidents (including the cost of sending
staff from the USA to oversee repairs) far exceeds the cost
of a ten-man manual clearance team for several months. This
leads me to the opinion that the MCC is an uneconomic tool for
use in support of wide-area Humanitarian Demining at a site
recommend that there would be little or no value in repairing/replacing
the damaged cultivating attachment. However, it is possible
that another attachment might make the machine more useful despite
its gross weight (the bulldozer is so heavy that it cannot be
driven over road bridges in the region).
the MCC is to be repaired, I recommend that the repair should
include the installation of a means of removing the vehicle
from a suspect area after a blast without having to drive it
manually. This may be achieved by fitting a properly armoured
cab (so protecting the RC systems) and duplicating controls
so that a secondary system can be switched on after a blast
- one that is unaffected by the magnetic surge associated with
the first large blast. Fitting an armoured cab would also reduce
the risk involved in withdrawing the machine from a suspect
area manually if this is required.
The suspect area
recommend that the ruined T55 tank should be moved from the
area, if possible, so allowing easier access to the ground around
and beneath it and its roller attachment.
entire area around the T55 tank to a distance of 40 metres each
side North and South, 25 metres West and 70 metres East should
be cleared/recleared using proven methods involving the removal
of all metal fragments. Preliminary checks close to the T55
imply that there is not extensive fragmentation over a wide
area, so the use of good metal-detectors should be viable in
all except a few square metres. These should be excavated with
a minimum 15 cm of topsoil removed. This clearance should be
done as a priority in order to rebuild local confidence in [the
demining group] and its operations.
following description of the MCC was found on a US government
website (December 2003).
The Mine Clearing Cultivator (MCC) is a large tine array designed
to uncover buried AT mines from the soil without destroying
its ability to support agriculture. The tine array lifts the
mines to the surface without creating a berm or spoil. A hydraulically
powered auger casts the mines to both sides of the vehicle for
naturalisation. The MCC is a kit designed for installation onto
150KW (200HP)-class dozers. The kit contains a tine array, auger
assembly, hydraulic power unit, linkage assembly and an ultrasonic
depth control system for the tine. The current system is mounted
onto a Liebherr 742B dozer. The kit can be integrated onto a
Caterpillar D7R dozer and other tractors as well.
MCC Assembly: 7,420 kg (16,324 lbs.)
Hydraulic power unit: 1,889 kg (4,156 lbs.)
Liebherr 742B (Dozer): 22,636 kg (49,800 lbs.)
Clearance Width: 3.66 m (144 in.)
Clearance Depth: 20-38 cm (8-15 in.) depending on vegetation
Tine Array: 23 tines spaced @ 14.3 cm (5-5/8 in.)
Auger Operation: 25 rpm
Hydraulic Power Unit Output: 132 kw (177 hp)
"Clearance rate of 4000 square meters/hour in light-to-moderate
vegetated and denuded areas to a depth of 25 cm (10 in.).
mines as deep as 38 cm (15 in.) in light soils.
Tine arrangement survived live mine blasts equivalent to 6 kg
(13.4 lbs.). Damage was field repairable with parts in the support
date, no fused AT mines have been initiated during operations.
" Small mines pass through the auger assembly.
" Remote control operation tends to slow clearance rate.
" Terrain limited.
" Total system weight exceeds 32 metric tons.
" In desert sand, the mines tend to stall in auger.
statement that "no fused mines have been initiated during operations"
was not true even before the latest accident. No mention of
using another "ground sifting" machine after the MCC was made
at the website when the above information was downloaded. In
discussion with the head of the establishment that has designed
the MCC, the researcher was told that the MCC's detonation of
mines was a "success" because there were two fewer mines around
as a result - regardless of the exorbitant costs involved.