Development of a Radiological Characterization Submersible ROV for Use at Fukushima Daiichi

Abstract

To decommission the Fukushima nuclear power plant after the accident caused by a tsunami in 2011, characterization of the fuel debris is required. The precise location and radiological composition of the fuel debris is currently unknown and the area is submerged making it difficult to investigate the Primary Containment Vessel (PCV). An integrated system that includes both radiation detectors and sonar will allow the full localization and characterization of the fuel debris. This paper describes research completed towards the development of a complete system, on-board a low-cost, small form-factor, submersible remotely operated vehicle (ROV). A cerium bromide (CeBr3) scintillator detector for dose-rate monitoring and gamma-ray spectrometry has been integrated and validated experimentally with a 137Cs source, both in the laboratory and whilst submerged. The addition of an IMAGENEX 831L sonar has enabled technical demonstrations to take place at the National Maritime Research Institute’s facility in Japan, where the system was able to characterize the shape and size of synthetic core debris. The combination of geometrical and radiological measurements allows the real-time localization of fuel debris and isotope identification, leading to an invaluable source of information to the workers at Fukushima that will enable increased efficiency and reduce risk during the decommissioning of the site.

Full text

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
1
 Abstract - To decommission the Fukushima nuclear power
plant after the accident caused by a tsunami in 2011,
characterization of the fuel debris is required. The precise
location and radiological composition of the fuel debris is
currently unknown and the area is submerged making it
difficult to investigate the Primary Containment Vessel
(PCV). An integrated system that includes both radiation
detectors and sonar will allow the full localization and
characterization of the fuel debris. This paper describes
research completed towards the development of a complete
system, on-board a low-cost, small form-factor, submersible
remotely operated vehicle (ROV). A cerium bromide
(CeBr3) scintillator detector for dose-rate monitoring and
gamma-ray spectrometry has been integrated and validated
experimentally with a 137Cs source, both in the laboratory
and whilst submerged. The addition of an IMAGENEX
831L sonar has enabled technical demonstrations to take
place at the National Maritime Research Institute’s facility
in Japan, where the system was able to characterize the
shape and size of synthetic core debris. The combination of
geometrical and radiological measurements allows the real-
time localization of fuel debris and isotope identification,
leading to an invaluable source of information to the
workers at Fukushima that will enable increased efficiency
and reduce risk during the decommissioning of the site.
Index Terms— Fukushima Daiichi, Gamma-ray detection,
Radiation monitoring, Nuclear Decommissioning
I. INTRODUCTION
he complete characterization of the Fukushima nuclear
power plant, after the accident in 2011, is required to enable
safe and efficient decommissioning of the site. To achieve this,
many systems are being developed to remotely inspect and
characterize the internals of the PCV and surrounding
suppression chamber [1].
After the tsunami struck, the immediate area was evacuated
with the PCVs of units 1, 2 and 3 flooded for moderation and
cooling. However, this took time and several explosions
Paper submitted May 31st, 2017. This work was funded by the Engineering
and Physical Sciences Research Council (EPSRC:EP/N017749/1) and MEXT
(Japan). MJJ acknowledges the generous support of the Royal Society as a
Wolfson Research Merit Award holder.
M. Nancekievill, B. Lennox and S. Watson are with the School of Electrical
and Electronic Engineering, University of Manchester, United Kingdom, UK
e-mail: matthew.nancekievill@manchester.ac.uk
occurred causing expulsion of radioactive material and leading
to uncertainty in the precise location of the reactor fuel and
associated core materials. It is believed that molten fuel melted
through the bottom of the PCV and into the pedestal. This has
yet to be confirmed with the possibility that some of the fuel
remains in the reactor, or that it has moved sideways to the
exterior of the pedestal and PCV.
Although many portions of the site have now been
characterized, such as the grating around the PCV by Hitachi’s
shape-changing robot [2] and the flooded suppression chamber
using multiple ROVs [3], the state of the bottom of the PCV
and pedestal is still largely uncharacterized [4].
The inability to characterize the PCV and pedestal means that
there is uncertainty in the expected dose-rates, shape and size
of physical debris, on top of or surrounding the fuel debris, as
well as to its isotopic inventory present and its distribution.
Dose-rate levels of 10 Gy.h-1 have been recorded outside the
PCV [5], with much higher dose rates anticipated within the
PCV and closer to the fuel debris. A complete characterization
will provide clarification and aid the decommissioning process.
The approach to characterization that is proposed in this
research is to use a combination of visual inspection, acoustic
mapping (2D/3D reconstructions) and radiation assessment. To
deploy the necessary payload of radiation detectors and sonar
devices within the PCV, a submersible is required that can
accommodate all of the sensors whilst retaining a small form
factor, so that it can be deployed through the relatively small
access ports that are available.
The University of Manchester has developed a low-cost
ROV called the Aqua Vehicle EXplorer for In-situ Sensing
(AVEXIS®), which was initially designed for deployment in the
Sellafield legacy ponds to aid in the characterization in this
cluttered environment [6][7]. Some initial work has also been
completed for detector integration and characterization [8].
Improvements have been made to increase the robustness and
adaptability of the ROV for integration of the required sensors
and deployment at Fukushima. A live feed video camera is
included for visual inspection and to aid navigation, with
complete manual control through a neutrally-buoyant tether,
A.R. Jones and M.J. Joyce are with the Department of Engineering,
Lancaster University, United Kingdom.
J. Katakura is with the Nagaoka University of Technology, Japan.
K. Okumura is with the Japan Atomic Energy Agency, Japan.
S. Kamada, M. Katoh, and K. Nishimura are with the National Maritime
Research Institute, Japan.
Development of a Radiological Characterization
Submersible ROV for use at Fukushima Daiichi
M. Nancekievill, A. R. Jones, M. J. Joyce, Member, IEEE, B. Lennox, Senior Member IEEE, S.
Watson, Member, IEEE, J. Katakura, K. Okumura, S. Kamada, M. Katoh and K. Nishimura
T

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
2
resistant to entanglement, with a diameter of <10mm. A sonar
system and radiation sensors have also been integrated within
the ROV.
This paper presents the experimental validation of the
resulting ROV, that contained both a CeBr3 gamma-ray detector
and IMAGENEX 831L sonar [9]. Experiments were conducted
in water tanks at Lancaster University, UK, for the testing of
the gamma-ray detector and the National Maritime Research
Institute (NMRI), Japan, for the sonar tests.
The remainder of the paper is structured as follows: Section
II gives an overview of the ROV and how the design aids the
chosen sensors for localization. Section III introduces the
integration of the CeBr3 detector and experimental results of the
gamma spectrum analysis with a small 137Cs source. Section IV
and Section V discuss the integration of the sonar used to
outline the physical dimensions and shape of the fuel debris.
Section VI discusses the results and suggests options to enable
deployment in the field and Section VII concludes the paper.
II. OVERVIEW OF THE AVEXISTM SUBMERSIBLE ROV
The AVEXIS® vehicle, shown in Fig. 1, was designed
initially for deployment in the Sellafield legacy fuel storage
ponds, to characterize underwater environments which had
restricted access [6].
The AVEXIS® vehicle consists of a central cylindrical
Perspex tube that contains all the control electronics, camera
and lighting. It has an outer diameter of 150 mm to enable
deployment through small access ports. The propulsion system
consists of 5 water pumps located at each end-cap which
provide three degrees of freedom: forwards and backwards, up
and down and rotation around the vertical axis.
The turning radius is half the length of the ROV (140 mm),
reducing the likelihood of entanglement and increasing the
range of environments within which it can be deployed.
Fig. 1. A photograph of the AVEXIS® ROV, adapted from [8].
The available payload weight of the AVEXIS® is
approximately 1.5 kg, which is sufficient to accommodate
several sensors. Initially, a gamma-ray detector and sonar were
integrated within the vehicle, with enough contingency for the
later addition of a neutron sensor if required.
An adaptable communication protocol is required to enable
high-definition visual inspection, with no delay in image
processing, control of the ROV’s actuators or outputting of
measurements from the detector payloads. It was also desirable
to make the tether as small as possible to reduce the possibility
of entanglement and to reduce drag effects on the ROV. This
led to the use of a Ethernet-over-power set-up, also known as
powerline technology, similar to systems used in OpenROV
vehicles [10]. Using this technology, transfer speeds of >200
Mbps can be achieved through a local area network created via
a laptop, a powerline adaptor on the surface and a powerline
adaptor in the ROV. The same two wires were used to power
the vehicle as shown in Fig. 2. All detectors attached to the
ROV are powered via separate voltage regulators from the main
power supply on the surface.
A Raspberry Pi™ Zero is used for the main control of the
ROV and a Raspberry Pi™ Camera is used to return a live video
feed with LED strips on the sides of the ROV to aid in visual
inspections in poorly lit conditions.
Irradiation tests conducted on some of the constituent
components indicated that the TID tolerance of the device is
greater than 300 Gy(Si). With a target operational lifetime of 10
hours, this is considered sufficient, assuming that the dose rate
will be approximately 10 Gy.h-1.
Fig. 2. A schematic diagram of the ROV communication protocol and
power through the 2-wire tether.
III. CERIUM BROMIDE DETECTOR IMPLEMENTATION
A. Detector overview
The gamma-ray detector chosen for deployment within the
ROV to characterise Fukushima Daiichi is a CeBr3 inorganic
scintillator. CeBr3 is a promising, relatively-new scintillation
material that has been suggested as one of the best alternatives
to replace the more established thallium-doped sodium iodide
(NaI(Tl)) [11]–[13]. CeBr3 has the advantage of having a very
low intrinsic background radioactivity, unlike other alternatives
such as LaBr3 ([11], [14]) and its energy resolution surpasses
that of NaI(Tl) [13]. It also has good pulse linearity [12], [15],
desirable timing properties [15]–[17]. Further advantages of
the CeBr3 detector that contributed to its selection as the
detection material are: its high detection efficiency, making it
ideal for in-situ monitoring [14], as well as its radiation
robustness to gamma radiation [18].
The detector consists of a 10 mm diameter crystal coupled
with a photomultiplier, connected to an integrated, high-voltage
supply unit. This high voltage unit allows operation of the
detector via a 5 V voltage regulator within the ROV. The
detector unit was manufactured by Scionix [19] in the
Netherlands and supplied by JCS Ltd., UK. A schematic
diagram of it is shown in Fig. 3.
B. CeBr3 Gamma detector integration
During initial integration of the CeBr3 gamma-ray detector
with the ROV, an RG178 coaxial cable was connected to the
signal output from the detector through the ROV. This resulted

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
3
in two tethers from the ROV, however, the RG178 coaxial cable
has a diameter of 2 mm and can be wrapped around the ROV
tether. Future iterations of the system will incorporate the
coaxial cable within the tether.
As an added layer of protection, the detector was placed in a
water-tight aluminum housing inside the ROV to mitigate any
unforeseen leaks. The detector can be seen with and without the
casing in Fig. 4.
Fig. 3. A dimensional schematic elevation of the CeBr3 inorganic
scintillator detector, model VS-0087-50, manufactured by Scionix,
Netherlands [19].
Fig. 4. A photograph of the CeBr3 inorganic scintillation detector,
without and with aluminum casing, left and right respectively, with a
UK new penny of diameter 20 mm included far left for scale [8].
C. Initial Characterisation
To determine the impact that the waterproof housing has on
the sensitivity of the CeBr3 detector, a set of experiments were
undertaken in the laboratory. Two separate sources were used;
a 330 kBq 137Cs, and a 50 kBq 22Na source. The signal from the
detector was passed through and processed by a single-channel,
mixed-field analyser (MFA) manufactured by Hybrid
Instruments Ltd., UK [20]. Comparative Multichannel analyser
(MCA) plots from these tests are shown in Fig. 5 and Fig. 6.
Fig. 5 and Fig. 6 show that the photopeaks are consistent with
one another between the tests conducted with and without the
aluminum casings. However, as expected, the relative intensity
of the peaks does vary. This is due to the increased distance and
shielding effects of the aluminium casing that is located
between the detector and source.
A summary of the change in observed throughput of counts
during the 10-minute experiments, in and out of the casing, can
be seen in Table I.
Fig. 5. Energy spectra in terms of counts as a function of channel
number for the CeBr3 detector exposed to a 137Cs source for 10
minutes. The top spectrum shows results taken without the casing, the
bottom spectrum that for the detector within casing, as labelled,
adapted from [8].
Fig. 6. As for Fig. 5 but for the 22Na source. The top graph shows
results for the detector without casing, and the bottom graph represents
those for the detector within casing as labelled, adapted from [8].
The results in Table I, indicate that the number of counts
were reduced, by a magnitude of approximately 11-14, when
Channel
Channel
Count
Channel
Channel
Count
Count
Count (103)

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
4
the casing was used. The difference in magnitude of the
attenuation between the two separate sources was assumed to
be due to the contrasting energy spectra of these sources. This
is expected, with the number of counts for the lower-energy
spectrum of 22Na, dropping in magnitude by 14 times in
comparison to 11 times for the 137Cs spectrum.
TABLE I
COMPARISON OF NUMBER OF COUNTS WITH AND WITHOUT
THE ALUMINUM CASING, ADAPTED FROM [8].
10 minutes
Pulses per second
No Casing
Casing
No
Casing
Casing
137Cs
source
216291
±465
174552
±418
360
±19
291 ±17
22Na
Source
19705
±140
12485
±112
33 ±6
21 ±5
By conducting these experiments, an understanding of the
magnitude of count reduction was found to inform the correct
measurement of gamma radiation dose-rate, whilst also
verifying the expected spectra that would be observed.
D. Integrated ROV Experiments
Having verified the operation of the detector within the
aluminum casing, a set of experiments was undertaken to
compare the effect of mounting the detector inside and outside
of the ROV. These tests were conducted on a laboratory
benchtop as shown in the photograph in Fig. 7, where the
AVEXIS® is shown next to two lead bricks which were used as
a precaution to shield the source from the user. In this test, the
detector was powered via the main circuitry in the ROV and the
signal was outputted through the RG178 coaxial cable tether.
Fig. 7. A photograph of the benchtop test arrangement of the CeBr3
gamma-ray detector integrated within the ROV, adapted from [8].
Fig. 8 shows the response of the detector to a 10-minute
exposure to a 330 kBq 137Cs source, with the detector located
outside the ROV (top) and then inside the ROV (bottom). It
can be seen that there are fewer counts when the detector is
mounted inside the ROV compared to when it was located
outside the ROV. This is due to the increased distance to the
source and also, to a lesser extent, the attenuation of the Perspex
tube of 3 mm thickness. The results suggest that despite this
reduction in counts, the detector is still capable of isolating the
spectrum of a relatively weak 137Cs source, whilst operating
from within the ROV.
E. Submerged experiments
A set of experiments was conducted to verify the operation
of the CeBr3 detector placed within the ROV whilst it was
submerged. This was to determine the ability of the system to
locate radioactive sources underwater, which could then be
applied to the decommissioning of Fukushima Daiichi.
Fig. 8. Spectra of photon counts as a function of channel number for
the CeBr3 detector exposed to a 137Cs source for 10 minutes. Top,
integrated within the ROV, bottom outside the ROV, adapted from [8].
Two tests were conducted at Lancaster University’s wave
tank. The same 330 kBq 137Cs source used in the previous
experiments was placed against the internal surface of one of
the sides of the wave tank and the ROV was then placed in the
water as close as possible to this source. After this initial
verification stage, the ROV was positioned 50 cm away from
the side of the tank. The tank wall was made of approximately
10 mm thick glass. The experimental set-up is shown in Fig. 9.
To remain consistent with the previous experiments, two 10-
minute exposures were taken at each location with the spectra
given in Fig. 10.
It can be seen in Fig. 10 that there is a reduced count in
comparison to previous tests due to the increased attenuation
Channel
Channel
Count
Count

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
5
caused by the distance and water separating the source and the
detector. However, the same structure to the spectra can be seen
with the peaks observed as expected. The spectra observed were
less defined for the submerged test at 50 cm from the tank side
but was still recognisable as a 137Cs source.
Fig. 9. A photograph of the set-up for the submerged ROV gamma-ray
detector experiments. The 137Cs source was surrounded by lead bricks
and placed up against the side of the tank, with the AVEXIS® parallel
to the source whilst submerged, adapted from [8].
Fig. 10. Spectra of photon counts as a function of channel number for
the CeBr3 detector exposed to a 137Cs source for 10 minutes, whilst
integrated within the ROV and submerged underwater. Top, with the
ROV pressed against the side of the tank. Bottom, with the ROV
approximately 0.5 m away, adapted from [8].
This experiment suggests that the detector is capable of
characterising radioactive isotopes through water. A source
with increased activity, by definition, would be detected more
quickly and would require shorter exposures to return
comparable spectra. However, due to health and safety
requirements it is more difficult to replicate this in the
laboratory. However, these tests verify the working principle of
the CeBr3 gamma-ray detector whilst integrated within the ROV
and submerged.
As the data are outputted in real-time, the successful spectral
analysis and dose-rate monitoring of the radioactive fuel could
enable specific nuclides to be localised within a variety of
radioactive material.
As the AVEXIS® is adaptable to accommodate the sensors as
required by the operator, it is feasible to replace the CeBr3
gamma-ray detector by other detectors, such as a single crystal
diamond (made by chemical vapor deposition) neutron detector
with a 6Li convertor foil. This could enable the mapping of
thermal neutron flux, complementing the gamma-ray detection
functionality.
IV. INTEGRATION OF SONAR DATA TO ENABLE LOCALISATION
OF RADIOACTIVE DEBRIS
In a cluttered environment such as inside the PCVs at
Fukushima Daiichi, knowledge of the physical layout is
required to ensure the safe operation of the ROV. Geometric
data will also enable the fusion of outputted radiation detector
data with a physical map of the reactor, further improving the
assessment of the nuclear waste.
The integration of the sonar was supported through
collaborators from NMRI, Nagaoka University of Technology
and the Japan Atomic Energy Agency. The objective was to
characterize the physical layout of the unknown areas beneath
the PCV in the pedestal.
The sonar used was the IMAGENEX 831L [9], which was
chosen due to its design for use in enclosed pipe systems. This
requires that the device is small and relatively light, allowing
for integration onto an ROV to detect fuel debris. Table II
shows the sonar specifications. This table shows that the
communication protocol is Ethernet TCP/IP, therefore, it is
possible to use the same powerline technology previously
discussed to gather outputted data, removing the need for an
additional tether: The sonar was powered from the ROV via the
two-wire tether to the surface and the data were output across
the same two wires. This layout can be seen in Fig. 11.
To validate the operation of the sonar, submerged tests
were undertaken at a test pond operated by Forth Ltd.,
Maryport, UK. The pond is approximately 8 m wide, 25 m long
and 6 m deep, which is representative of a Sellafield legacy
pond, whilst the depth is representative of the PCVs at
Fukushima Daiichi.
Count
Count
Channel
Channel

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
6
TABLE II
SPECIFICATIONS OF THE IMAGENEX 831L SONAR.
Frequency
2.25 MHz
Transducer
Profiling type, fluid
compensated
Transducer Beam Width
1.4o conical
Range Resolution
1/250 of full scale range
Min. Detectable Range
50 mm
Max. Operating Depth
1,000 m
Max. Cable Length
Standard: 100 m CAT5e
Interface
Standard: 10 Mbps Ethernet
TCP/IP
Connector
Subconn MCBH-8M-SS
Power Supply
20-32 Vdc at less than 5 W
Dimensions
61 mm dia. X 343 mm
length
Weight: In Air
In Water
1.2 kg
0.4 kg
Materials
6061-T6 Aluminium &
Polyurethane
Finish
Hard Anodize
Fig. 12 shows the output from the sonar. A feature consistent
with the corner of the tank can be seen, with the AVEXIS®
approximately 0.5 m from the tank wall and 0.8 m from the
bottom. The two small shapes at the bottom of the tank are
drainage hoses with a diameter of 10 cm. This shows the
suitability of the system for characterizing debris at the bottom
of the PCVs in Fukushima, as broken piping or cylindrical fuel
rods are likely debris that will require localization and
characterization.
From these tests, it was determined that the sonar data could
be retrieved accurately allowing an underwater environment to
be characterized in a 2D plane. To complete a more thorough
and representative set of tests to verify the suitability of the
sonar to characterize expected radioactive fuel debris, a set of
experiments were conducted in Japan.
Fig. 11. A photograph of the sonar unit integrated with the ROV.
Fig. 12 An example output scan from the sonar unit on-board the
ROV in submerged tests with a grid scale of 0.4 m per division,
indicating features consistent with the environment, as labelled.
V. FULLY INTEGRATED EXPERIMENTS
The ROV was taken to a facility operated by Japan’s NMRI
in Tokyo. This was to determine if both the CeBr3 gamma-ray
detector and sonar could be integrated within the ROV at the
same time and used to characterize synthetic debris.
Radioactive materials could not be placed in the tank and
therefore, success for the gamma-ray detector was determined
by whether it was able to collect background radiation
measurements in real-time whilst the ROV was characterising
the physical shape and size of the debris. The tests illustrated
that the detector was able to do this.
To test the sonar’s capabilities, synthetic debris was placed
at the bottom of the tank. This was representative of what might
be expected at Fukushima and is shown in Fig. 13. A layer of
sand was placed over this debris to represent sediment that has
built up in the PCVs over the course of the last 6 years. Fig. 14
shows the results of a scan of the simulated debris that was
made during the tests. This scan shows a dome representing the
debris with the sharp edges of the mock, broken fuel assemblies
and other debris, and provides evidence that the sonar could be
used to characterize the debris in the Fukushima Daiichi power
plant.
Fig. 13. A photograph of the synthetic debris of the physical form
expected to be found at the bottom of the PVCs at Fukushima Daiichi.

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
7
Fig. 14. The output sonar image from tests at the NMRI to characterize
synthetic debris.
VI. DISCUSSION
The ability to operate a CeBr3 gamma-ray detector, remotely,
whilst submerged on an ROV has been validated
experimentally. Whilst submerged, the CeBr3 gamma-ray
detector was able to detect the number of counts over a set time
period at different pre-determined positions, leading to an
approximate source location. Using the approximate location
of any source, the ROV could then be positioned so that the
detector is exposed to the higher counts required to carry out
the more detailed spectral analysis required to identify
constituent isotopes. This two-stage analysis could also aid in
reducing the radiation damage to the ROV, as proximity to the
source would not be reduced (and thus exposure to the radiation
not exacerbated) until necessary.
With a source of greater activity, a reduced time of exposure
would be required to identify a given radioisotopes to the same
extent. In this research, a duration of 10 minutes was chosen to
confirm the operational abilities of the system with a relatively
small source. It is anticipated that isotopic identification will
be possible at a much quicker rate, dependent on source activity
and ROV distance. In the deployment at Fukushima Daiichi,
the level of activity is not expected to be a problem but, rather,
the robustness of the system under high levels of exposure will
need to be better understood and the ability to separate
important, constituent isotopes from more prominent, and more
widely-dispersed nuclides.
The integration of a sonar system capable of outputting the
physical size and dimension of the internal state of Fukushima
Daiichi, further aids in the process of localisation and
characterisation of the fuel debris. It could indicate if a source
of gamma-ray radiation is likely to be buried beneath other
physical debris, and also be used to indicate its location relative
to the walls of the PCV. This is critical data for engineers to
decommission the site safely and efficiently.
It may also be possible to return a 3D image of the debris
within the reactor using the sonar data and a relative positioning
system. This would combine each 2D plane of sonar data,
transforming it into a 3D model of the reactor. The data received
from the CeBr3 gamma detector can then be overlaid onto this
image in real-time. This would be a valuable source of
information for decommissioning engineers.
The ROV has also been designed with adaptability and
modularity. The connections and power circuit can also be used
to integrate a neutron detector, specifically such as a scCVD
diamond detector. This could be used to provide data about the
fuel debris within the Fukushima Daiichi pedestal, which is
distinct from 137Cs because caesium is understood to be
dispersed widely throughout the water covering the core, the
suppression chamber and surrounding facility; thus it is not
associated uniquely with the debris. The data received from this
neutron detector could also be overlaid on a representative 3D
map leading to a greater understanding of the fuel debris
situation in Fukushima Daiichi. An example scCVD diamond
detector to be integrated in future is shown in Fig. 15.
Fig. 15. A photograph of a scCVD diamond neutron detector with a
UK new penny included for scale.
Before this device is installed it will also be necessary to test
it in the presence of a significant source of alpha radioactivity
and neutrons to be consistent with the Fukushima Daiichi PCV
environments.
VII. CONCLUSION
Integration of a CeBr3 gamma-ray detector into an underwater
submersible compatible with the characterizing needs of the
inside of the Fukushima Daiichi nuclear power plant has been
described. This has been validated experimentally through
bench-top and submerged testing, which highlighted the
reduced counts seen due to attenuation by the ROV and water,
whilst the spectra are still recognizable. Further
characterization of the capabilities of this device will be
possible with a higher-activity gamma radiation source and
multiple submerged sources at specialist facilities.
Integration and experimental validation of a working sonar
has been completed to determine the physical location and size
of debris inside Fukushima Daiichi. Tests conducted both in the
UK and Japan outlined the capability of the sonar to physically
characterize submerged material consistent with this
requirement.
Improvements to both the submersible and gamma-ray
spectrometry functions will be conducted with the aim of
overlaying positional data of gamma sources in relation to the
internal dimensions of the PCV. This will aid the effort to
decommission the nuclear power plant and lead to targeted
removal of fuel debris, potentially reducing risk to workers and
increasing efficiency of decommissioning.

This work is licensed under a Creative Commons Attribution 3.0 License. For more information, see http://creativecommons.org/licenses/by/3.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TNS.2018.2858563, IEEE
Transactions on Nuclear Science
8
REFERENCES
[1] K. Oikawa, “R & D on Robots for the Decommissioning of
Fukushima Daiichi NPS International Research Institute for Nuclear
Decommissioning,” Feb. 2016.
[2] Tokyo Electric Power Company, “Development of a technology to
investigate inside the Reactor Primary Containment Vessel (PCV),”
2015.
[3] IRID, “Robots working inside the buildings at Fukushima Daiichi
NPS(Part III),” 2016. [Online]. Available:
http://irid.or.jp/en/research/gengorov_trydiver/. [Accessed: 16-Apr-
2017].
[4] IAEA, “Events and highlights on the progress related to recovery
operations at Fukushima Daiichi Nuclear Power Station,” 2017.
[5] Tokyo Electric Power Company, “Development of a technology to
investigate inside the Reactor Primary Containment Vessel ( PCV
),” Tokyo, Apr. 2015.
[6] A. Griffiths, A. Dikarev, P. Green, B. Lennox, X. Poteau, and S.
Watson, “AVEXIS-Aqua Vehicle Explorer for In-Situ Sensing,”
IEEE Robotics and Automation Letters, vol. 3766, pp. 282–287,
Jan-2016.
[7] A. R. Jones et al., “On the Design of a Remotely-Deployed
Detection System for Reactor Assessment at Fukushima Daiichi,” in
IEEE Nuclear Science Symposium 2016, 2016.
[8] M. Nancekievill et al., “A Remote-operated System to Map
Radiation Dose in the Fukushima Daiichi Primary Containment
Vessel,” EPJ Web Conf., vol. 170, 2018.
[9] IMAGENEX, “IMAGENEX Model 831L Digital Pipe Profiling
Sonar,” 2011.
[10] OpenROV, “OpenROV 2.8 Kit,” 2017. [Online]. Available:
https://www.openrov.com/products/openrov28/. [Accessed: 14-Apr-
2017].
[11] E. García-Toraño, B. Caro, V. Peyrés, and M. Mejuto,
“Characterization of a CeBr3 detector and application to the
measurement of some materials from steelworks,” Nucl. Instruments
Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc.
Equip., vol. 837, pp. 63–68, 2016.
[12] K. S. Shah et al., “CeBr3 Scintillators for Gamma-Ray
Spectroscopy,” IEEE Trans. Nucl. Sci., vol. 52, no. 6, pp. 3157–
3159, 2005.
[13] P. Guss, M. Reed, D. Yuan, A. Reed, and S. Mukhopadhyay,
“CeBr3 as a room-temperature, high-resolution gamma-ray
detector,” Nucl. Instruments Methods Phys. Res. Sect. A Accel.
Spectrometers, Detect. Assoc. Equip., vol. 608, no. 2, pp. 297–304,
2009.
[14] V. Peyres, T. Crespo, M. Mejuto, and E. García-Toraño,
“Measurement of NORM samples with CeBr3detectors,” Appl.
Radiat. Isot., vol. 126, no. February, pp. 307–310, 2017.
[15] R. Billnert, S. Oberstedt, E. Andreotti, M. Hult, G. Marissens, and
A. Oberstedt, “New information on the characteristics of 1 in.×1 in.
cerium bromide scintillation detectors,” Nucl. Instruments Methods
Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip., vol.
647, no. 1, pp. 94–99, 2011.
[16] S. S. Alam et al., “VECC array for Nuclear fast Timing and angUlar
corRElation studies (VENTURE),” Nucl. Instruments Methods
Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip., vol.
874, no. May, pp. 103–112, 2017.
[17] N. D’Olympia, S. Lakshmi, P. Chowdhury, E. G. Jackson, J. Glodo,
and K. Shah, “Sub-nanosecond nuclear half-life and time-of-flight
measurements with CeBr3,” Nucl. Instruments Methods Phys. Res.
Sect. A Accel. Spectrometers, Detect. Assoc. Equip., vol. 728, pp.
31–35, 2013.
[18] W. Drozdowski, P. Dorenbos, a J. J. Bos, G. Bizarri, A. Owens, and
F. G. a Quarati, “CeBr3 Scintillator Development for Possible Use
in Space Missions,” IEEE Trans. Nucl. Sci., vol. 55, no. 3, pp.
1391–1396, 2008.
[19] SCIONIX, “SCIONIX,” 2017. [Online]. Available:
http://scionix.nl/. [Accessed: 02-Jun-2017].
[20] Hybrid Instruments Ltd., “Single Channel Mixed-Field Analyser,”
2016. [Online]. Available:
http://hybridinstruments.com/products/MFAX1.3.html. [Accessed:
15-Jun-2017].