Underwater Technology 34. 2

Page 1

Vol. 34 32 No. No. 232 2017 2014 Vol.

UNDERWATER TECHNOLOGY

49

83

K Boot

Yi Wang, Menglan Duan, Huaguo Liu, Runhong Tian, Chao Peng

51

93

J Hoth, W Kowalczyk

D Cvikel, O Grøn, LO Boldreel

63

99

A Personal View... Fourth Decommissioning and wreck removal workshop

Colour reconstruction of underwater images

The use of fibre optic distributed sensing technology to detect changes in sediment overburden

Advances in deepwater structure installation technologies

Detecting the Ma’agan Mikhael B shipwreck

Book Review Extreme Life of the Sea

ISSN 1756 0543

Y Ouyang, R Hird, Md Bolton

75

Analysis of the influence of ambient conditions on the ampacity of Kevlar-armore subsea power cables

N Vedachalam, A Umpathy, GA Ramadass, MA Atmanand

www.sut.org

0A-SUT-34(2)-OFC.indd 1

24/02/17 12:19 pm


UNDERWATER TECHNOLOGY Editor Dr MDJ Sayer Scottish Association for Marine Science

Society for Underwater Technology

LJ Ayling Maris International Ltd

Underwater Technology is the peer-reviewed international journal of the Society for Underwater Technology (SUT). SUT is a multidisciplinary learned society that brings together individuals and organisations with a common interest in underwater technology, ocean science and offshore engineering. It was founded in 1966 and has members in more than 40 countries worldwide, incIuding engineers, scientists, other professionals and students working in these areas. The Society has branches in Aberdeen, London and South of England, and Newcastle in the UK, Perth and Melbourne in Australia, Rio de Janeiro in Brazil, Beijing in China, Kuala Lumpur in Malaysia, Bergen in Norway and Houston in the USA. SUT provides its members with a forum for communication through technical publications, events, branches and specialist interest groups. It also provides registration of specialist subsea engineers, student sponsorship through an Educational Support Fund and careers information. For further information please visit www.sut.org or contact: Society for Underwater Technology 1 Fetter Lane EC4A 1BR London UK e info@sut.org t +44 (0)20 3440 5535 f +44 (0)20 3440 5980

Prof DS Cronan Imperial College London

Scope and submissions

Associate Editor G Griffiths MBE Autonomous Analytics Associate Editor Subsea Engineering LJ Ayling Maris International Ltd Assistant Editor E Azzopardi SUT Editorial Advisory Board Chairman Dr MDJ Sayer Scottish Association for Marine Science

G Griffiths MBE Autonomous Analytics Prof C Kuo FRSE Strathclyde University Dr WD Loth WD Loth & Co Ltd Dr S Merry Renewable Energy Association & Focus Offshore Ltd Prof J Penrose Curtin University of Technology Prof WG Price FRS FEng Southampton University Prof MF Randolph University of Western Australia Dr R Rayner Sonardyne International Ltd Prof R Sutton Plymouth University Prof P Wadhams University of Cambridge Cover Image (top): zoonar.com/syrist Cover Image (bottom): Steve Crowther Cover design: Quarto Design/ kate@quartodesign.com

0B-SUT-34(2)-IFC.indd 1

The objectives of Underwater Technology are to inform and acquaint members of the Society for Underwater Technology with current views and new developments in the broad areas of underwater technology, ocean science and offshore engineering. SUT’s interests and the scope of Underwater Technology are interdisciplinary, covering technological aspects and applications of topics including: diving technology and physiology, environmental forces, geology/geotechnics, marine pollution, marine renewable energies, marine resources, oceanography, salvage and decommissioning, subsea systems, underwater robotics, underwater science and underwater vehicle technologies. Underwater Technology carries personal views, technical papers, technical briefings and book reviews. We invite papers and articles covering all aspects of underwater technology. Original papers on new technology, its development and applications, or covering new applications for existing technology, are particularly welcome. All papers submitted for publication are peer reviewed through the Editorial Advisory Board. Submissions should adhere to the journal’s style and layout – please see the Guidelines for Authors available at www.sut.org.uk/journal/default.htm or email elaine.azzopardi@sut.org for further information. While the journal is not ISI rated, SUT will not be charging authors for submissions.

in more than 40 countries worldwide, including over 190 Corporate Members of the Society.

Disclaimer and copyright The Society does not accept responsibility for the technical accuracy of any items published in Underwater Technology or for the opinions expressed in such items. The copyright of any paper published in the journal is retained by the author(s) unless otherwise stated. All authors are supplied with a PDF version of their papers once published. Authors are encouraged to make the PDF version of their papers free to download from their own websites.

Open Access Underwater Technology is available as Open Access. PDF versions of all published papers from Underwater Technology may be accessed via ingentaconnect at www. ingentaconnect.com/content/sut/unwt. All issues from Volume 20 (1995) onwards are available as Open Access. The Society for Underwater Technology also encourages Underwater Technology authors to make their papers available online on their personal and/or institutional websites for Open Access. Through this arrangement, the Society supports the Open Access policy not only in the UK (the Research Councils UK (RCUK) policy) but also the drive towards Open Access in other countries.

Abstracting and indexing Abstracting and indexing services covering Underwater Technology include: American Academy of Underwater Sciences (AAUS) E-Slate; Aquatic Sciences and Fisheries Abstracts (Biological Sciences and Living Resources; Ocean Technology, Policy and Non-Living Resources; and Aquatic Pollution and Environmental Policy); Compendex; EBSCO Discovery Service; Fluidex; Geobase; Marine Technology Abstracts; Oceanic Abstracts; Scopus; and WorldCat Discovery Services.

Subscription Subscription to the print version of Underwater Technology is available to non-members of the Society at the following rates per volume (single issue rates in brackets). Prices are given in GBP. Accepted methods of payment are cheque or credit card (MasterCard and Visa). Foreign cheques must be in GBP and drawn on a British bank otherwise a currency conversion surcharge is incurred. UK subscription Overseas subscription

£102.00 (£25.50 per issue) £108.00 (£27.00 per issue)

Underwater Technology is also available in electronic format via ingentaconnect as Open Access. To subscribe to the print version of the journal or for more information please email Elaine Azzopardi at elaine.azzopardi@sut.org

Publication and circulation Underwater Technology is published in March, July and November, in four issues per volume. The journal has a circulation of 2,400 copies to SUT members and subscribers

Advertising To book an advert or for more information please contact Elaine Azzopardi at elaine.azzopardi@sut.org

24/02/17 12:43 pm


www.sut.org

A Personal View...

doi:10.3723/ut.34.049 Underwater Technology, Vol. 34, No. 2, pp. 49–50, 2017

Fourth Decommissioning and Wreck Removal Workshop

T

he Decom and Wreck Removal Workshop is now a well-established addition to the Marine Alliance for Science and Technology for Scotland (MASTS) Annual Science Meeting. The workshop, jointly organised by MASTS and the Society for Underwater Technology (SUT), was held in Glasgow this year and attracted at least 80 delegates. There were 22 presentations representing the communities that have a real interest in the challenges brought by the need to decommission North Sea platforms; safely deal with shipwrecks, old and new; and remove marine renewable energy devices as they near the end of their useful lives. These three industrial sectors are very different but have much in common and, as co-organiser Karen Seath, General Manager of Decom North Sea pointed out, much to learn from each other: “The technologies and innovative ideas for offshore removals that come through the salvage community are very applicable for oil and gas, and renewables. We need to look across sectors so we can apply them. Over the last four years, we have seen conversations moving on into discussions across the sectors and that includes dialogue between academia and industry.”

The driver for the conversations is the age of the assets in the North Sea, many of which are more than 30 years old and are reaching the end of their lives. But the cost of dealing with these assets is a major concern,

with a conservative estimate from the industry of around $50bn over the next 25 years, and several speakers suggesting that a more realistic figure could be as high as $100bn. Even for an industry that is used to talking in billions, the figures are eye-wateringly high, and especially for taxpayers who will take a significant share, maybe as much as 70%, of the decommissioning costs.

Do we have to remove them, or can they simply be left in place? The key regulation that governs and imposes removal is the Convention for the Protection of the Marine Environment of the North-East Atlantic, better known as the OSPAR Convention. While it does have the power of derogation, a general feeling from the workshop was that OSPAR needs to be brought up-to-date and become more flexible. This is attributed to research-based opinions that suggest removing a platform may not be in the best interests of marine ecosystems and society, or economically sensible. Spending money wisely became a driving theme for the entire workshop after being introduced by Moya Crawford (Chair of SUT International Salvage and Decommissioning Committee) in her opening remarks. This view is not simply driven by the prodigious costs but also from the ecological and conservation sectors, which made it clear that the disturbance that would ensue from an attempt to take the seabed back to where it was before installation and

Kelvin Boot Kelvin Boot graduated from the University of Exeter with a degree in Geology with Biology. He has worked in museums, with the BBC Natural History Unit and was part of the team setting up and running the National Marine Aquarium. He is now a freelance science communicator working with the Marine Alliance for Science and Technology for Scotland, the Plymouth Marine Laboratory and other marine related organisations.

might do more harm than good. They also asserted that there is strong evidence that the jackets (legs), and mattresses (bases) of platforms are providing rich habitat for a wide range of species, enhancing rather than diminishing biodiversity in the immediate area of a platform. One session of the meeting was dedicated to looking at examples of how scientists are evaluating the value of offshore structures as habitats for marine life. The remoteness and harshness of the environment does not encourage normal biological sampling methods, however. What was apparent was the enthusiasm of the scientists and of the ‘industry’ to cooperate. While there is a presumption of removal, ideally the legislation should allow each potential

49

1-SUT279_34(2)_ut64418.indd 49

24/02/17 12:42 pm


Kelvin Boot. Fourth Decommissioning and Wreck Removal Workshop

decommission to be assessed for the most appropriate course of action. Re-use could be an option and the workshop heard some compelling arguments for re-use of topside facilities – ideas that stimulated the imaginations of delegates. For example, in the digital age even a remote platform might make useful offices or even a hotel for the more adventurous tourists or a self-contained ‘seasteading community’. Recycling was also discussed as an option. Installations comprise a wide range of materials and equipment, much of which can be re-used or re-purposed. One example was provided by Alistair Niewwenhuyse of Reflex Subsea and involved re-spooling longer lengths of flexible pipe for re-use rather than cutting it into smaller lengths for disposal. Alistair’s research and experience shows that ~50% of pipe could be re-spooled for re-use. However, as other delegates pointed out, there is still a plethora of regulation and legislation, not to mention, payable duty when bringing materials ashore, especially across jurisdictions, all of which is discouraging to innovation. There is a requirement to re-visit the available technology that was originally designed for dockside construction, as deconstruction or recycling at sea is a very different challenge. This was the crux of a presentation by Alan Edwards of Bibby Offshore, who emphasised the need for meticulous planning, surveying and data management, spiced with the need for flexibility in the face of sometimes unpredictable sea conditions and challenges such as unexpected gas releases.

The challenge of an industry, which has extensive experience of installation at sea following the construction of most of the platform being done on land, is deconstruction in a harsh and often unstable environment. Safety is as serious an issue as economic and environmental considerations. As Colin Howes of DNV GL pointed out, while the oil and gas sector has great experience in installation and operation, there is not much experience of the hazards involved in decommissioning, and we should “expect the unexpected and manage uncertainty”. The reoccurring message came through loud and clear – each project will have its own challenges and, at an early stage, experience and expertise should be identified. This could come from outside sectors such as salvage and wreck removal.

New areas, new challenges, new approaches Nobody thought of removal 30 or 40 years ago in the excitement of the new bonanza. But with new areas such as the Arctic opening up, we should be planning for future decommissioning, according to Tina Hunter of the Centre for Energy Law at the University of Aberdeen. These plans should take into consideration how building materials might be affected by very low temperatures and what the harsh conditions will mean for dismantling in situ – no two cases will be the same. Tina’s take-home message was that before you plan to build, you should plan to decommission. This might include the option to use a variation on the floating liquefied

natural gas (FLNG) production vessels, which can be removed (to a new location) rather than be dismantled – with the associated clean up, potential environmental damage and safety aspects of decommissioning a fixed platform in a harsh environment. Innovative thinking and learning from others is a key to success, and much can be learned from the salvors’ approach to such challenges. Stuart Martin gave some illustrations from Ardent Global’s approach to retrieving the Costa Concordia. His key message was: look for the best fit for each project; there is no one size fits all; and review, review, review to get the best possible experience, expertise and approach to every job. John Gillies of Shell UK echoed the need for a fresh approach to each project, suggesting that the industry moves away from the “cut and paste approach” to become more outward-looking and willing to learn from others. With discussion comes relationships and understanding, and that was very evident throughout the two-day workshop. The diversity of speakers and topics from industries, government, NGOs, the science community, lawyers and safety consultants was testament to formerly separate sectors coming together as a community to reach a common goal. There was definitely an atmosphere of wanting to work together, reflected in the avalanche of questions that followed every presentation. Conversations continued across refreshment breaks, and the feeling of cooperation and burgeoning partnerships was tangible.

50

1-SUT279_34(2)_ut64418.indd 50

24/02/17 12:42 pm


doi:10.3723/ut.34.051 Underwater Technology, Vol. 34, No. 2, pp. 51–61, 2017

Technical Paper

www.sut.org

Colour reconstruction of underwater images Julian Hoth* 1, 2 and Wojciech Kowalczyk1 1 University of Duisburg-Essen, Lotharstr. 1, 47057 Duisburg, Germany 2 German Aerospace Center (DLR), Institute of Communications and Navigation, Department of Nautical Systems, Kalkhorstweg 53, 17235 Neustrelitz, Germany Received 2 March 2016; Accepted 11 July 2016

Abstract Objects look very different in the underwater environment compared to their appearance in sunlight. Images with correct colouring simplify the detection of underwater objects and may permit the use of visual simultaneous localisation and mapping (SLAM) algorithms developed for land-based robots underwater. Hence, image processing is required. Current algorithms focus on the colour reconstruction of scenery at diving depth where different colours can still be distinguished, but this is not possible at greater depth. This study investigates whether machine learning can be used to transform image data. First, laboratory tests are performed using a special light source imitating underwater lighting conditions, showing that the k-nearest neighbour method and support vector machines yield excellent results. Based on these results, an experimental verification is performed under severe conditions in the murky water of a diving basin. It shows that the k-nearest neighbour method gives very good results for short distances between the object and the camera, as well as for small water depths in the red channel. For longer distances, deeper water and the other colour channels, support vector machines are the best choice for the reconstruction of the colour as seen under white light from the underwater images. Keywords: Colour reconstruction, underwater camera, unmanned underwater vehicles, marine robotics

1. Introduction Objects look very different in the underwater environment compared to their appearance in sunlight. The main reason is that the penetration of light through seawater is highly dependent on the wavelength of the light (Åhlen, 2005). Suspended particles in the water can further decrease the overall quality of underwater images (Bazeille et al., 2006; Celebi and Ertürk, 2012). * Contact author. Email address: julian.hoth@dlr.de

High quality images with correct colouring simplify the detection of underwater objects and may allow the use of visual simultaneous localisation and mapping (SLAM) algorithms developed for land-based robots underwater. Hence, image processing is required to obtain images of high quality and correct colouring. Over the last decade, significant progress has been made in this direction. Current algorithms focus on the colour reconstruction of scenery at diving depth (Iqbal et al., 2007; Celebi and Ertürk, 2012). Therefore, a significant part of sunlight is still present and different colours can still be distinguished, although they may be tainted due to light being filtered through seawater. The algorithms are often based on a simple relationship between colours under different lighting conditions. Typical image models are Beer’s law (Åhlen, 2005) and the Jaffe-McGlamery image model (Jaffe 1990; Lee et al., 2012), which are shown in Fig 1. Iqbal et al. (2007) use contrast stretching of the redgreen-blue (RGB) image, and saturation and intensity stretching of the image converted to hue intensity saturation (HIS) colour space in order to correct the colour. However, stretching requires a certain minimum amount of intensity information retained in the image. This is only true for small depths and the examples given in the paper fulfil this requirement. Other approaches require known geometric relations between the camera and the object (Lee et al., 2012) or require manual work (Chen et al., 2014). Unfortunately, at greater depth the filtering is much stronger, such that different colours are strongly tainted and can no longer be distinguished (Fig 2). Hence, simple image models are no longer applicable. To solve this, the main factor of interest for the influence on colour is the penetration depth (dp) which is a strong function of the wavelength (λ) (Fig 3).

51

2-SUT259_34(2)_ut64413.indd 51

01/03/17 2:17 pm


Hoth and Kowalczyk. Colour reconstruction of underwater images

Fig 1: Pairs of images before (left) and after correction (right) from Bazeille et al. (2006)

Fig 2: An object under normal lighting conditions (left) and under underwater lighting conditions (right)

This study investigates whether machine learning can be used to transform image data. This should work for extreme conditions as shown for example in Fig 2. To this end images of objects of different colour are obtained under underwater and regular lighting conditions.

2. Machine learning This section briefly outlines the machine learning procedures that were used in this study.

2.1. Artificial neural network Artificial neural networks (ANNs) are informationprocessing algorithms that are modelled after

the way human brains work (Aleksander and Morton, 1995). The first formulations of this method were already made in 1943 by McCulloch and Pits. Artificial neural networks consist of strongly interconnected nodes called neurons. These neurons are organised into layers: one input layer, one output layer, and one or more hidden layers inbetween. The number of nodes in the input and output layer is given due to the nature of the data being processed. In this study, there are three nodes in each of them to represent the colour channels of a 24 bit pixel. The number of nodes in the hidden layers as well as the number of hidden layers is variable (Ferreira, 1996).

52

2-SUT259_34(2)_ut64413.indd 52

01/03/17 2:17 pm


Underwater Technology Vol. 34, No. 2, 2017

Fig 3: Light penetration in ocean water and coastal waters (Carothers 2016)

Teaching neural networks is done by forming or removing connections, changing the weights and the threshold of neurons, and adding or removing neurons so that errors are minimised for a given validation set. However, in many cases the overall shape of the neural network is fixed and only the weights and thresholds are used for learning (Ferreira, 1996). A common method for training is the backpropagation algorithm, which is also used in this study. Hereby, the error derivative of the weights – i.e. the change of error depending on the change of weight – is determined by starting with the total errors at the output layer and the moving through the network towards the input layer (Rojas, 1996). This is done repeatedly until either the rate of change of the errors or the error derivatives becomes sufficiently reduced. In this study, up to 250 learning cycles are used with a learning rate of 0.1 and a momentum of 0.2 for a given neuron configuration. The number of neurons was varied between 3 and 8 per layer with the number of layer between 1 and 4.

2.2. k-nearest neighbour The k-nearest neighbour (KNN) is a non-parametric method for classification and regression. The output (classification or property value) for a given input is obtained from known input-output relations where the inputs are similar and close (i.e. ‘in the neighbourhood of’) to the sample point in question. Classification is then done by majority voting and in case of a property value the average of the outputs of the neighbouring inputs is taken (Fix and Hodges, 1951). In this study, data points are situated in a 3D space (due to three colour channels) and can have 256 different values (classifications) for each channel.

The result depends on the choice of the neighbourhood – i.e. up to what distance or how many neighbours are taken into account, and the type of distance measure has some influence (Lewicki and Hill, 2005). In this study, Euclidean with weighted averaging is used. Teaching a KNN system is usually performed by finding the size of the neighbourhood k with the lowest error by cross-validation (Lewicki and Hill, 2005).

2.3. Support vector machines Support vector machines (SMVs) are non-probabilistic linear classifiers. However, they can also be used for regression (Drucker et al., 1997). The fundamental idea is that data sets that belong to different classes are linearly separable by hyperplanes. If the data cannot be separated by linear hyperplanes it has to be mapped into a higher dimensional (embedding) space such that is becomes linearly separable (Cortes and Vapnik, 1995). Of all those hyperplanes that separate the datasets, one hyperplane has to be chosen such that the margin separating different classes is a maximum. Those data points closest to the boundary and which are required to describe the hyperplane exactly are called support vectors. Hence, teaching a support vector machine is essentially an optimisation problem. As for KNN, the support vector machine in this study starts in a 3D space and each data point is in one of 256 classes for each colour channel. Due to the likely non-linear separability the resulting machine has a much higher dimensionality. SVMs with a Gaussian Kernel are considered to be the best choice for this study. Kernel parameter (σ) is varied between 0.1 and 5.0 to obtain the spread 53

2-SUT259_34(2)_ut64413.indd 53

01/03/17 2:17 pm


Hoth and Kowalczyk. Colour reconstruction of underwater images

with the lowest error. For the optimisation, the cost of misclassification is set to a moderate value of 1.0, the convergence tolerance is 0.0010, and the width of the ε-sensitive zone is set to ε = 0.0010.

2.4. Bayesian network Bayesian networks (BN) are probabilistic graphical models based on Bayesian probability theory (Bayes 1763; Jaynes and Bretthorst, 2003). A Bayesian net describes how different states of a system represented as nodes of a graph are linked through probability, i.e. the net shows conditional interdependencies of variables via a directed acyclic graph (Ben-Gal, 2007). Bayesian networks can be learned from available data. Various procedures are available for teaching, which are separated into two categories (Cooper and Herskovits, 1992; Heckerman et al., 1995): 1 Structure learning (structure is unknown); and 2 Parameter learning (structure is known). Often, they are combined such that parameter learning is a sub-process of structure learning (score-and-search-based approach) (Friedman et al., 1997). In this study, hill climbing (Gámez et al., 2011) with an alpha parameter of 0.5 is used.

2.5. Multiple linear regression Multiple linear regression (MLR) is similar to linear regression, but instead of having one independent variable as for linear regression, the output depends on two or more independent variables. In this study, there are three separate MLRs (one for each colour channel in the reconstructed image) with three independent variables (the three colour channels of the image under underwater lighting conditions). Learning a multiple linear regression machine is done via the method of least squares. The best linear unbiased estimator is found using the Gauss-Markov theorem.

3. Methods 3.1. Laboratory test To obtain images under controlled lighting conditions, various coloured objects were illuminated by a special LED light source and regular white light. The special LED light source has a range of 450 nm to 570 nm and a mean wavelength of ¯λ = 498 nm. The mean wavelength is very close to the wavelength for maximum penetration depth in open ocean, λmax,ocean ≈ 480 nm (see also Fig 3; Smith and Baker, 1981; Mobley, 2004). Therefore, the objects looked as if they were situated in deep ocean water and illuminated by a white light from a submarine

Object

Light source z 1m

x y Camera

Fig 4: Laboratory setup for image acquisition

or underwater robot from some distance away. For the regular white light, a mercury-vapour lamp is used. The coloured objects were then photographed several times under both white light and underwater lighting conditions. The camera used had an active pixel sensor with 16.1 mpx resolution. A sketch of the laboratory setup is shown in Fig 4. The images taken were then processed as follows. First, feature-matching is done for each image set, where the same object is photographed once under white light and once under underwater lighting conditions so that the two images are exactly aligned. Therefore, every pixel in the image under underwater lighting conditions has a corresponding pixel in the image taken under white light. For feature-matching, the matchFeatures algorithm implemented in MATLAB® has been used. This method finds corresponding points of interest between pairs of images using local neighbourhoods and the Harris algorithm (MATLAB, 2016). In a second step, the images were resized such that only the coloured objects could be seen in the pictures. This was necessary as the background was not illuminated. From every image set, 10% of the pixels were chosen at random and an n × 6 matrix was built containing the colour channels (RGB) for two matching pixels in each row. The resulting matrix was then fed through statistical learning algorithms with or without pre-filters. The methods discussed in section 2 of this paper were chosen for testing. The input data for the learning methods were also varied to test the influence of different pre-processing methods on the performance of the machines. Besides feeding the raw image data (raw) directly into the statistical learning algorithms, the following methods were used for smoothing the data: • • • •

fast Fourier transform (FFT); convolution (Conv); moving average (MA); singular value decomposition (SVD).

54

2-SUT259_34(2)_ut64413.indd 54

01/03/17 2:17 pm


Underwater Technology Vol. 34, No. 2, 2017

The FFT smoother eliminates frequency noise by setting all frequencies to zero that are less than a threshold times the maximum distance of the data. The convolution smoother uses the Gaussian kernel. The size of the convolution filter was set so that the difference between the original and smoothed data is no more than 1% of the whole range of values. Moving averages was done with five consecutive entries, and the SVD smoother removes singular values that are below a chosen threshold.

3.2. Underwater experiment The second stage was to test the applicability of learning algorithms for underwater image processing under more realistic conditions. This allows a more appropriate evaluation of the method before it is implemented into a real system. Hence, an experimental verification in water was warranted. In order to be able to evaluate the results as best as possible and to automate the colour matching for the learning algorithms, the coloured objects from the laboratory test were replaced by colour patterns with known colours. Fig 5 shows one of these patterns. The patterns were such that they could easily be processed with standard pattern recognition algorithms to extract the colours from the images at the appropriate positions. One pattern consisted of 32 carefully chosen colour patches, including white, black, and several shades of grey as well as colours from the whole RGB range. The other colour patterns had random patches. The first pattern was intended for training of the machine learning algorithms. The remaining patterns were for testing the trained learning

Fig 6: Diving basin at TauchRevier Gasometer

machine on various colours that were not part of the training process. The colour patterns were first photographed in the laboratory under white light at distances of 1 m, 5 m and 8 m. This was done for various camera and zoom settings to determine the best conditions. The camera used had an active pixel sensor with 16.1 mpx resolution. The boards were then brought to Europe’s largest indoor diving basin at TauchRevier Gasometer in Duisburg, Germany. The diving basin is circular with a diameter of 45 m and a water depth of 13 m. It provides conditions that are as close to open water as possible (Fig 6), and it is therefore rated as open water for divers. Furthermore, the lighting conditions are such that greater water depths are simulated compared to the actual water depth achieved. Also, the optical properties of the water are similar to coastal water and the water is very murky. Therefore, the penetration depth of light was expected to be very low in general, with its maximum in the green colour region. The boards were taken to depths of 4 m, 8 m and 12 m and again photographed at distances of 1 m, 5 m and 8 m using the same camera that is used in the laboratory (Fig 7). Under white light, this is done for various camera and zoom settings. The resulting images were processed as for the laboratory test.

4. Results and discussion

Fig 5: One of the colour patterns used for the experimental verification

4.1. Laboratory test Table 1 shows the root-mean-square errors (RMSEs) for all combinations of the five machine learning methods and the five different input data for all colour channels of 24 bit images. Every channel has 256 possible pixel values for the laboratory test. As it is only the RMSEs, the actual deviation of a pixel colour from the expected value can differ 55

2-SUT259_34(2)_ut64413.indd 55

01/03/17 2:17 pm


Hoth and Kowalczyk. Colour reconstruction of underwater images

Board with colour pattern

dw z x y

dcb

Camera

Fig 7: Experimental setup (dcb: distance camera/board, dw: water depth) Table 1: RMSEs for different learning methods and preprocessors in laboratory setup RSME

ANN

KNN

SVM

BN

MLR

Raw FFT Conv MA SVD

57.06 56.84 54.06 62.30 56.93

30.46 9.32 10.70 80.84 30.46

12.35 11.38 11.76 48.12 11.35

60.40 60.40 60.40 62.16 60.40

43.89 43.89 43.90 43.91 43.89

significantly. The RMSEs for the red channel are slightly higher while the ones for the blue channel are slightly lower than the values given in the table. Bayesian networks and ANNs show a large RMSE independent of the input data of well above 50 px. MLRs also show a significant RMSE of around 44 px for all types of input data, although they are lower than for ANN and Bayesian networks. For KNNs and SVMs the performance depends on the input data. Both show a very large RMSEs when the image data were smoothed with moving averages. In fact, KNNs with moving averages has the largest error (almost 81 px) over all combinations, while KNNs with raw data or data decomposed with singular value decomposition shows a medium RMSE around 30 px. Much lower RMSEs (about 12 px) are observed for SVM using any type of input data except data smoothed with moving averages. The lowest RMSEs are obtained for KNN combined with convoluted data and FFT. The low performance of Bayesian networks can be expected as image data are not probabilistic. The high RMSE for MLR was also anticipated. As stated earlier, the assumption of a linear dependency between the colours as seen underwater and under sunlight does not hold beyond small depths. Smoothing the input data reduces the performance of the learning machines. Especially for KNN and SVM where the RSME is quite low, the drop in performance is significant. It can therefore be deduced that small changes of the pixel values

have a strong influence on the result. These small changes are removed by smoothing and hence the input data no longer represent the original image properly. Fig 8 shows how the approach in this study works for some sample objects when using KNN with FFT smoother in the laboratory, which is the combination with the lowest error. For every object, there is a set of three images: (a) the object under white light; (b) the same object illuminated by the special light source (see Fig 4); and (c) the reconstruction done by the algorithm using the middle image as an input. As Fig 8 shows, the colours of the object under white light are reproduced well. However, there are also some situations where the reconstruction is not optimal. Problems occur at mid left and at the top edge in Fig 8(c), where strong red colours are produced, because red light is strongly influenced by the lighting conditions (see Fig 3). It can be observed that red objects look almost the same as black and dark grey objects. Hence, regions which appear in these colours in the underwater image have to be mapped both to red and to black or dark grey. It is clear that this does not work in every situation. It can also be seen in Fig 8 that due to the strong and focused light source as well as the smooth surface of the sample objects, reflections occur in some regions of the image. The scenery in not uniformly lighted as for the object under white light. In these parts the colour reconstruction is not as good as in the rest of the image. Reflections change the relation between the colours under white light and under underwater lighting conditions.

4.2. Underwater experiment Due to the conditions onsite, it was not possible to obtain images of every colour pattern for every water depth and camera distance. In addition, not all combinations of distance and depth gave feasible results. Table 2 shows an overview of the performance.

56

2-SUT259_34(2)_ut64413.indd 56

01/03/17 2:17 pm


Underwater Technology Vol. 34, No. 2, 2017

(a)

(b)

(c)

Fig 8: (a) Object under normal lighting conditions, (b) under underwater lighting conditions and (c) reconstruction from learning machine Table 2: Performance of the underwater tests Camera distance [m]

Water depth [m]

1 4 All patterns

5 All patterns

8 All patterns except one 12 Only two patterns

All patterns except one No feasible images

8 Only one pattern No images No images

No feasible images could be obtained at a water depth of 12 m and a camera distance of 5 m. Fig 9(b) shows an image taken of pattern 2 under these conditions. As can be seen, only the measurement cord is visible in the left part of the image; the light cannot penetrate the water to the colour pattern itself. The same is true for a camera distance of 8 m and water depths of 8 m and 12 m. Hence, no images were obtained for these conditions. Fig 9 also shows some further sample images for various camera distances and water depths. These give a good indication of the visibility conditions onsite.

As there were numerous colour patches under different conditions, only the worst case (water depth 12 m, camera distance 1 m) for selected colour patches is discussed here in more detail. The observations made can also be applied to the other cases. Table 3 shows the RMSE for all combinations of the five machine learning methods and the five different input data for all colour channels of 24 bit images, i.e. every channel has 256 possible pixel values, for a water depth of 12 m and a camera distance of 1 m. The RMSEs for the red channel are 10 px to 15 px higher, while the RMSEs for the green channel and the blue channel are 5 px to 10 px lower than the values given in the table. The results show a similar structure compared to the results for the laboratory experiment. Bayesian networks and MLR show a better performance in the experiment than in the laboratory test. For ANNs, KNNs and SVMs, the errors are increased. For each learning algorithm, the errors for the different input methods are very similar except for KNNs where a strong variation between 49 px and

57

2-SUT259_34(2)_ut64413.indd 57

01/03/17 2:17 pm


Hoth and Kowalczyk. Colour reconstruction of underwater images

(a)

(b)

(c)

(d)

Fig 9: Sample images from the underwater experiment for (a) 4 m depth and 1 m distance, (b) 12 m depth and 5 m distance, (c) 8 m depth and 5 m distance and (d) 12 m depth and 1 m distance Table 3: RMSEs for different learning methods and preprocessors at 12 m water depth and 1 m camera distance RSME

ANN

KNN

SVM

BN

MLR

Raw FFT Conv MA SVD

66.52 66.55 68.03 68.35 66.15

61.13 52.69 54.08 70.41 49.98

31.04 31.73 31.91 48.41 32.81

45.50 44.78 44.52 44.05 45.50

37.19 36.85 36.86 36.89 36.85

Table 4: Learning machines used Camera distance [m] 4

Water depth [m]

8

12

71 px can be observed. Moving averages also decrease the performance of SVMs significantly. The lowest RMSEs are obtained for SVMs combined with raw data, convoluted data and FFT. Table 4 summarises the results for the different camera distances and water depths used in the experiment. It shows which learning machines give the best results separately for the three colour channels. SVMs become dominant the further away the camera is from the object and with increasing water depth. KNN is only used for small camera distances and small water depths, and is of real interest only for the red channel alone. For the green channel, KNN gives similar results to SVMs at a camera distance of 1 m and a water depth of 4 m, but in no other case. For the blue channel, only SVMs are used. It may therefore be concluded that SVMs are generally the better choice for the task of obtaining the colour of objects under white light from the underwater images, while KNN should be used only at close distances, though this is not of interest for the task of underwater navigation and object recognition.

1 R: KNN G: KNN/SVM B: SVM R: KNN G: SVM B: SMV R: SVM G: SVM B: SVM

5 R: KNN G: SVM B: SVM R: SVM G: SVM B: SVM

8 R: SVM G: SVM B: SVM

Having analysed the channels separately, it is also necessary to discuss the colours produced by the algorithms from the underwater images. Table 5 shows the results for several colour patches when using SVMs with raw data input, which is the combination with the lowest error. The left column shows colour patches as they appear in the laboratory under white light. In the middle column, the same patches are given as they look in the underwater environment in the water tank at 12 m water depth and 1 m camera distance. The right column shows the reconstruction from the algorithms. The first general observation that can be made is that there is no perceptible difference between the colour patches in the underwater environment, so there is very little information contained in the images that can be used for the reconstruction. In the first laboratory experiments, some variations between the different colours were visible (Fig 9),

58

2-SUT259_34(2)_ut64413.indd 58

01/03/17 2:17 pm


Underwater Technology Vol. 34, No. 2, 2017

Table 5: Selected results at a water depth of 12 m and a camera distance of 1 m Laboratory

Underwater

Reconstruction

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

but is not the case for the underwater condition as presented in Table 5. The second general observation is that bright colours appear darker in the reconstruction than under laboratory conditions – best seen in the red colour patches nos. 12 and 15. Therefore, there is a tendency for low pixel values to be overestimated and high pixel values to be underestimated. The high red content in the colour patches is reconstructed well. Patches nos. 1, 6, 10, 11, 12 and 15 contain a significant amount of red, and nos. 6 and no. 12 show especially good results for the red content. No. 11 shows some grey instead of orange, but this indicates problems in the combination of the channels and not in the red channel itself. In patches nos. 1 and 10, the red content is smaller and a degradation in the quality of the reconstruction can be observed immediately. This result is interesting considering that the average error in the red channel is the highest of all three colour channels. The green content in the colour patches is also well reconstructed. Patches nos. 4, 7, 8, 9 and 14 contain a significant amount of green, and the bright green patch no. 9 is excellent. The strong green patches nos. 7 and 14 show problems: except for some red, the resulting colour patches appear also green in the reconstruction even though the reconstruction is darker. In patches nos. 4 and

no. 8 the green content is slightly lower and the resulting colour is not mediocre. However, the green content can clearly be seen in the reconstruction, so the problem does not lie with the green content. For patch no. 8, the mean colour in the reconstruction actually corresponds to the colour of the patch under white light. Green light has the longest penetration depth in coastal water, and the green colour channel shows the best overall results with a lower RMSE than the other two colour channels. Hence, the results are to be expected. Regarding the blue content in the colour patches, high blue values are found in patches nos. 4, 5 and 8. Patch no. 5 is a dark blue in the laboratory with no red and green content and a small amount of blue. Patches nos. 4 and 8 are much brighter and therefore contain more blue. From these general observations, an underestimation of the blue value can be expected and is also apparent in Table 5. In patch no. 8, the blue content is higher and the reconstruction of the blue value is very good. One may observe some grey in patch no. 8, but since the green and blue values are both predicted well, the problem is more likely to be found in the red channel. A lower content of red is expected compared to the other two colours. As there is some grey, the red value is overestimated in this case. Generally, the visual results show that the reconstruction of the colours are good even in the severe optical conditions under which the images are obtained. For the deeper water depths, the results are even better and it can be expected that for clear ocean water the errors are much smaller than the ones obtained in the experiment.

5. Conclusion and outlook This study tested the applicability of different learning methods for underwater image reconstruction. It showed that the KNN and SVM methods are excellent choices to perform this task in the laboratory, subject to the image data being pre-processed appropriately. It also showed that reflections resulting from a focused light source reduced the performance of the learning machines. In addition, the distinction between red objects and black or dark grey objects is very difficult and sometimes leads to inappropriate colours in the reconstructed image. During experiments in a diving basin, the KNN gave good results for short distances between the object and the camera, and for shallow water depths in the red channel. For longer distances and deeper water depths, and for the other colour channels, SVMs were the best choice for the reconstruction of

59

2-SUT259_34(2)_ut64413.indd 59

01/03/17 2:17 pm


Hoth and Kowalczyk. Colour reconstruction of underwater images

the colour under white light from the underwater images. Furthermore, the reconstruction under real conditions was much more difficult than in the laboratory due to the harsh conditions. It was shown that even under extreme conditions the reconstruction of the colours was good, although the RMSEs of the colour channels were high. It could be seen that bright colours appear darker in the reconstruction and the high colour contents were easier to reconstruct. The approach presented currently uses the original image without any additional filtering except for data smoothing. In the next stage, image pre-processing algorithms should be included to remove reflection effects and to smooth the brightness in the image. Additionally, the preprocessing after Bazeille et al. (2006) should be tested to increase the overall image quality. It may also be of interest to include information of neighbouring pixels for the calculation of the colour to remove single pixels that were not transformed correctly. Furthermore, the setup should be improved by using a distributed light source. The laboratory results show that a focused light presents addition obstacles for the algorithms. In water with a significant number of suspended particles (like in the diving basin used for the experimental verification), focused light for illumination is also not possible as the light is directly reflected by the particles. The procedure can also be used for other applications, where other ranges of wavelengths are used, to obtain the image as seen under white light from data under different lighting conditions. An example would be night sensing equipment like night-vision glasses or night vision devices in cars. Finally, the experimental setup should be implemented into a real system, e.g. underwater robots and research submarines, and tested further under various conditions. Right before deployment, the learning machines should be trained for the onsite conditions. This should be done with a suitable known colour pattern as, for instance, the one used in this study in the underwater experiment.

Acknowledgment The authors would like to thank Prof. Dr. rer. nat. Johannes Gottschling from the Chair of Mathematics for Engineers at the University of DuisburgEssen for his insights into statistical learning. Furthermore, the authors thank the TauchRevier Gasometer for allowing the use of their diving facilities and Andreas Scholz and Stefan Westermaier for spending a Sunday in cold water taking pictures for the validation of the algorithms.

References Åhlen J. (2005). Colour Correction of Underwater Images Using Spectral Data. PhD. Dissertation, Uppsala Universitet. Aleksander I and Morton H. (1995). An Introduction to Neural Computing, second edition. Boston, MA: International Thomson Computer Press, 288 pp. Bayes T. (1793). An Essay towards solving a Problem in the Doctrine of Chances. Phil. Trans. 53: 370–418. Bazeille S, Quidu I, Jaulin L and Malkasse J-P. (2006). Automatic Underwater Image Pre-Processing. In: Proceedings of Caracterisation du Milieu Marin, 16–19 October, Brest. Ben-Gal I. (2007). Bayesian Networks. In: Encyclopedia of Statistics in Quality & Reliability. Ruggeri F, Kenett R and Faltin F, (eds.). Hoboken, NJ: Wiley & Sons. Carothers K. (2016). http://oceanexplorer.noaa.gov/ explorations/04deepscope/background/deeplight/ media/diagram3.html. <last accessed 17th May 2016>. Celebi AT and Ertürk S. (2012) Visual enhancement of underwater images using Empirical Mode Decomposition. Expert Systems with Applications 39: 800–805. Chen Z, Wang H, Shen J, Li X and Xu L. (2014). Regionspecialized underwater image restoration in inhomogeneous optical environments. Optik, International Journal for Light and Electron Optics. 125/9: 2090–2098. Cooper GF and Herskovits E. (1992). A Bayesian method for the induction of probabilistic networks from data. Machine Learning 9: 309–347. Cortes C and Vapnik V. (1995). Support-Vector Networks. Machine Leaming 20: 273–297. Drucker H, Burges CJC, Kaufman L, Smola A and Vapnik V. (1997). Support Vector Regression Machines. Advances in Neural Information Processing Systems 9: 155–161. Ferreira C. (1996). Designing Neural Networks Using Gene Expression Programming. In: Abraham A, de Baets B, Koeppen M and Nickolay B. (eds). Applied Soft Computing Technologies: The Challenge of Complexity. Berlin and Heidelberg: Springer-Verlag. 517–536. Fix E and Hodges JL. (1951). Discriminatory analysis, nonparametric discrimination: Consistency properties. USAF School of Aviation Medicine Technical Report 4. Friedman N, Geiger D and Goldszmidt M. (1997). Bayesian network classifiers. Machine Learning 29: 131–163. Gámez JA, Mateo JL and Puerta JM. (2011). Learning Bayesian networks by hill climbing: efficient methods based on progressive restriction of the neighborhood. Data Mining and Knowledge Discovery 22: 106–148. Heckermann D, Geiger D and Chickering DM. (1995). Learning Bayesian networks: The combination of knowledge and statistical data. Machine Learning 20: 197–243. Iqbal K, Salam RA, Osman A and Talib AZ. (2007). Underwater Image Enhancement Using an Integrated Colour Model. International Journal of Computer Science 34: 529–534. Jaffe JS. (1990). Computer modeling and the design of optimal underwater imaging systems. IEEE Journal of Oceanic Engineering 15: 101–111. Jaynes ET and Bretthorst GL. (2003). Probability Theory: The Logic of Science: Principles and Elementary Applications. Cambridge: Cambridge University Press, 753 pp. Lee D, Kim G, Kim D, Myung H and Choi H-T. (2012). Vision-based object detection and tracking for autonomous navigation of underwater robots. Ocean Engineering 48: 59–68. Lewicki P and Hill T. (2005). Statistics: Methods and Applications. Tulsa: StatSoft, Inc.

60

2-SUT259_34(2)_ut64413.indd 60

01/03/17 2:17 pm


Underwater Technology Vol. 34, No. 2, 2017

MathWorks Inc. (2016). MATLAB http://de.mathworks. com/help/vision/ref/matchfeatures.html. <last accessed on 17th May 2016>. McCulloch W and Pitts W. (1943). A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biophysics 5: 115–133.

Mobley CD. (2004). Light and Water: Radiative Transfer in Natural Waters. Office of Naval Research.Rojas R. (1996). Neural Networks: A Systematic Introduction. Berlin and Heidelberg: Springer-Verlag. 502 pp. Smith RC and Baker KS. (1981), Optical properties of the clearest natural water (200–800 nm). Applied Optics 20: 177–184.

61

2-SUT259_34(2)_ut64413.indd 61

01/03/17 2:17 pm


4 – 6 April 2017 Southampton UK

The hands-on ocean technology exhibition and training forum

REGISTER FOR *FREE

International exhibition showcasing 340+ exhibitors from across the globe

Cutting edge conference presenting the latest developments in the industry

Hands-on demonstrations held in 180+ workshop sessions

Industry associated meetings held by leading organisations Ocean Business @OceanBusiness #oceanbiz @OceanBusinessShow OceanBusinessShow

* The exhibition, workshops, associated meetings and Ocean Careers are FREE to attend. There is a fee to attend the conference. Organised by:

In partnership with:

Ocean Careers event providing recruitment and careers advice

Networking opportunities at the welcome drinks, wine trail and gala dinner

oceanbusiness.com

Hosted by:

+44 (0)1453 836363

info@oceanbusiness.com


doi:10.3723/ut.34.063 Underwater Technology, Vol. 34, No. 2, pp. 63–74, 2017

Technical Paper

www.sut.org

The use of fibre optic distributed sensing technology to detect changes in sediment overburden Y Ouyang1, R Hird2* and MD Bolton2 1 21 Thornton Court, Girton, Cambridge CB3 0NS, UK 2 Schofield Centre, University of Cambridge, Cambridge CB3 0EL, UK Received 10 May 2016; Accepted 15 August 2016

Abstract Fibre optic cables can be used as sensors to monitor changes in temperature and strain through the analysis of backscattered light. This can be linked to changes in the ambient conditions surrounding the cable. Active distributed temperature sensing relies on an external heat source relative to the fibre optic cable to measure the properties of, and changes in, the surrounding medium. An experiment was conducted using fibre optic sensing technology to monitor changes in sediment overburden. Fibre optic cables were buried in a channel containing saturated sand and water with an external heat source. The depth of overburden sediment above the cables was reduced, while the associated temperature response along the cable was monitored. This paper explains the characteristics of heat transfer from an active heat source to the surrounding soil medium providing a means to translate from the temperature measurement to the associated overburden thickness. The techniques used here are intended to be applicable to measurements of seabed scour above buried power cables.

for preventive maintenance and asset management strategies. The principle of fibre optic monitoring is based on the detection of backscattered light to quantify changes in temperature and strain from OFS. Backscattered light is composed of three main spectral components: Rayleigh, Raman and Brillouin, of which Raman and Brillouin are currently used to report temperature and/or strain along the optical fibre. Raman spectra differ from Brillouin spectra by having a fixed frequency but variable intensity. In comparison, Brillouin interaction exhibits a frequency shift through the scattering process. The correlation of the frequency shift is linearly dependant on the refractive properties of the optical fibre, and thus the strain and temperature changes within. Unlike Raman scattering, Brillouin scattering can propagate over long distances without significant attenuation, allowing for analysis of longer cable lengths where necessary.

Keywords: Distributed temperature sensing, buried power cables, seabed scour, optic fibre sensor

1. Introduction Fibre optic cables are often included in the power cable assembly to transmit telecommunications (Fig 1). When used as an optic fibre sensor (OFS), they can measure the conductors’ thermal and structural integrity during operation by monitoring optical signal attenuation, which acts as an early warning system to detect changes along the power cable. This additional application has gained pace over recent years, almost to a point where it can be considered a prerequisite in subsea cable installations * Contact author. Email address: rh500@cantab.net

Fig 1: Typical power cable assembly (adapted from DNV GL, 2014)

63

3-SUT265_34(2)_ut64415.indd 63

24/02/17 12:44 pm


Ouyang et al. The use of fibre optic distributed sensing technology to detect changes in sediment overburden

Applications of OFS used for cable condition monitoring have been documented by Hara et al. (1999) for monitoring temperature distribution in a 500 kV DC submarine power cable, and by Zhao et al. (2014) for the real-time monitoring of strain in a three-phase 110 kV subsea cable. Svoma et al. (2009) discussed the benefits of integrated fibre optic arrays within power cables to measure thermal runaway created by seabed sediment mobility, which can alter the thermal conductivity of the medium surrounding the cable, leading to hotspots. Continuous and real-time monitoring of temperature can also reveal cold spots caused by removal of sediment that once surrounded the cable, possibly leading to free spans if unchecked. Fractures in the cable from seismic activity or anchor drag can also be detected (Omnisens, 2015). Ultimately, remote monitoring along cable arrays can optimise future power cable design (Fromme et al., 2011) and cable installation procedures by measuring thermal efficiency in known sediment types. A distributed temperature sensing (DTS) system can be used to measure spatial and temporal variability in the subsurface at a high resolution along cables over tens of kilometres in length. Two forms of measurement are currently recognised: active and passive sensing. Passive sensing can be used to monitor the ambient temperature of the subsurface when an OFS is buried beneath the surface or placed on top of the seabed. Active sensing relies on an external heat source that may vary depending on changes to internal and external properties of the surrounding medium. An example would be power cables where electrical resistance within the conductor gives rise to internal heat. The amount of heat generated by an electrical cable depends on the load being carried as well as its construction and insulation thermal rating. A maximum temperature reaching 90 °C is considered for a subsea cable using polyethylene insulation (Pilgrim et al., 2013) for example. DTS systems have been successfully demonstrated in various applications, such as monitoring terrestrial pipelines (Inaudi and Glisic, 2010), changes in soil moisture content (Steele-Dunne et al., 2010) and in hydrological temperature-depth profiling (Arnon et al., 2014). This paper, however, focuses on sensing the changes in sediment overburden around a simulated power cable buried in an aquatic environment using Brillouin optical time domain analysis (BOTDA). Seabed sediment mobility challenges the effectiveness of the subsea cabling industry, especially in shallow water where strong tidal currents are located and where seabed materials are often noncohesive. Around the UK strong seabed currents,

accentuated during tidal surges and storms, are commonplace. Offshore territories with names like ‘The Wash’ and ‘Race Bank’ conjure images synonymous with transient shifting seabed bathymetry. Some of these environments have largescale offshore wind farm developments where inter-array and interconnector seabed cables need to be buried to ensure optimum thermal efficiency and to prevent damage caused by third-party activities. Seabed sediment transport can be prevented by mitigating against scour or infilling scour pits with sandbags and rock fill. Related cold spots in cables can be detected by occasional DTS monitoring, but the gradual thinning of sediment cover above a cable in an area of high sediment transport should be detectable as a gradual reduction in local temperature. Using active DTS, Zhao et al. (2012) experimentally demonstrated how scour can be detected around subsea pipelines. But limitations include the need for the OFS with an active heat source to be installed to the side of the pipeline, which would add to cost as well as create operational issues. Power cables, in comparison, generate heat during operation owing to electrical resistance and can provide the means for active monitoring of changes in thermal gradient. Extending the work of Zhao et al. (2012), but on a smaller scale and with a power cable rather than a pipeline in mind, the sensitivity of the DTS using Brillouin scatter to measure temperature variation due to changes in sediment thickness is demonstrated in this study. A long channel containing sand with a covering of water forms the basis of the model presented in this paper. A heat tape is embedded in the sand together with an OFS, where the heat tape mimics the power cable in generating heat. Sediment cover is manually reduced and the temperature change monitored showing the sensitivity of the monitoring system to measure real time changes in overburden.

2. Experiment setup The test setup consisted of ten 1 m long sections of a concrete U-channel (referred to here as ‘channel’) that had been locked together with a silicone sealant placed between each joint. The dimensions of the channel cross-section are presented in Fig 2(a). A vertical wall comprising 12 mm thick waterproofed plywood was clamped at each end of the channel using corner clamps at the top and a gravity block at the base. Sealant was applied to seal gaps between the channel and the wood panel ends. At the lower corner of the wood ~50 mm above the base of the channel, an 18 mm hole had

64

3-SUT265_34(2)_ut64415.indd 64

24/02/17 12:44 pm


Underwater Technology Vol. 34, No. 2, 2017

Fig 2: Sketch of the experimental setup: (a) dimensions of the channel; (b) placement of heat tape and fibre optical cables within the sand; and (c) plan view showing position of PT-100 thermometers (all dimensions in millimetres unless stated)

been drilled to accommodate the heating tape. The self-regulating heating tape was 12 mm wide and 10 m long, supplied by Heat TraceTM, UK as 75FSS2-A rated at 75 W per metre (Failsafe Super, 230 V metal sheathed heating cable). Selfregulating heat tapes are cables whose local resistance increases when the temperature in that location rises beyond target, thereby reducing the local heat output while the remainder of the cable generates heat as before. The aim is to supply heat only where it is needed to maintain the source temperature. The cable was supplied with a termination kit consisting of a silicone shoe and a socket that isolates the 240 V power input cables with an earth cable and bracket that connected to the outer cable. The tape supply wires were connected to a terminal block and then to a 16 A, 14 m long cable. The cable connected to a socket containing a 16 amp type C residual current circuit breaker (RCBO) with a 30 mA trip to protect the heating tape during operation. To protect the exposed connection against surface water, the entire terminal block was immersed in ACC silicone adhesive (RS 448-0286). The cable passed through the wooden end panels on the channel and the annulus was back filled with high temperature resistant sealant. The potting compound for the terminal blocks was also applied on the cable after drying, where it entered and exited the wood panel enclosure and ensured

a waterproof seal. The heat tape was supported using wooden blocks during initial setup of cables and sensors in the channel. Fibre optic cables (Mayflex Excel OS2 4C 9/125), specific to measuring temperature by removing the influence of strain, were laid at the base of the channel directly below the heating cable and 100 mm to the side (Fig 2b). These were fixed to the base of the channel using duct tape. The cable was one continuous 100 m length and was placed in the 10 m long channel with 4 m loops exposed outside the channel at each end. This was so that the cable could be doubled back along the channel at the same elevation or placed at a different elevation. The significant length of these loops was necessary to distinguish and identify these cable sections during analysis. As a separate and independent point of temperature reference, sheathed, four-wire, ceramic platinum resistance thermometers (PT-100) were placed at two locations along the channel (Fig 2c) and attached to the OFS closest to the heat tape (2-1) using a cable tie. The PT-100 (supplied by Labfacility, UK) were monitored using a PicoÂŽ Technology PT-104 data logger connected to a computer. Two loggers were used with eight PT-100 sensors. After placement of 20 m length of cable with PT-100 sensors at the base of the channel, the channel was flooded with water before placing the

65

3-SUT265_34(2)_ut64415.indd 65

24/02/17 12:44 pm


Ouyang et al. The use of fibre optic distributed sensing technology to detect changes in sediment overburden

brown medium-grained sand. The sand was then gradually added across the channel to remove air and thus achieve uniform density. The particle size distribution of the sand in Fig 3 indicates the distribution of particles as shown by the coefficient of uniformity (Cu) and coefficient of concavity (Cc), which are derived from the particle size (D) passing at 60%, 30% and 10%. The sand was placed up to the level of the heating cable so that it rested on the sand surface. A length of OFS was then attached to the heating cable with cable ties and a PT-100 sensor was attached to the cable at 2.4 m and 7.0 m along the channel. Two 10 m sections of OFS were then placed at the same depth as the heating tape and

Fig 3: Particle size distribution of sand used in the channel

at a horizontal separation of 100 mm from it. Sand was then added and another duel cable length placed at a higher elevation above the heating tape, as shown in Fig 2b. A cross-section of the cable arrangement at each elevation is presented in Fig 2c. A single length of cable was then placed above the final layer of sand to measure conditions at the simulated seabed. The cable positions were identified as 1, 2 and 3 representing a reduction in elevation towards the seabed, with ‘-1’ referring to vertical alignment with the heating tape and ‘-2’ as 100 mm offset from it. Thus cable reference 1-1 represents the OFS at the base of the channel directly below the heat tape and ‘3-2’ the OFS located above the heat tape offset by 100 mm separation. The OFS along the heat tape is referenced as ‘2-1’. The OFS along the seabed is referenced as ‘4’. The depths of all the OFSs were measured incrementally along their length. Variations in depth up to 20 mm off the intended cable elevation are due to localised variations during manual placement of the sand. During activation of the heat tape, the surrounding soil temperature would increase and begin to affect the temperature of the water above. It was therefore necessary to recirculate the water using small submersible pumps located at each end of the channel. The water at the surface was also cooled by adding blocks of ice to closer simulate typical water temperatures around the UK at seabed depth. In Fig 4, (a) shows the setup of the

Fig 4: View from the start of the channel (0 m) towards the opposite end (at 10 m): (a) during initial setup; (b) during placement of sand around cables; and (c) final setup with retaining partitions where sediment removal later takes place

66

3-SUT265_34(2)_ut64415.indd 66

24/02/17 12:44 pm


Underwater Technology Vol. 34, No. 2, 2017

Fig 5: Measurement reported by DITEST STA-R at the beginning and the end of the test period

experiment prior to introduction of sand, (b) at a midpoint during laying the OFSs, and (c) the final setup with vertical partitions marking intended sediment removal.

3. The setup of fibre optic sensing system Following the cable laying and subsequent burial, OFS was connected to a Brillouin-based fibre optic distributed temperature and strain analyser known as DITEST STA-R (manufactured by Omnisens™ Ltd, Switzerland). The measurements reported by the DITEST analyser can achieve a minimum spatial resolution of 0.5 m, with temperature accuracy of 0.2 °C, and the system setup requires access to both ends of the OFS cable. The analyser sends two counter-propagating waves that can be coupled through a non-parametric non-linear process, where the energy transfer from one wave (referred to as ‘pump’) feeds into the other (called a ‘probe’). The sensing process is highly accurate and efficient in identifying the position-dependent frequency information, by pulsing one of the optical waves and observing the local coupling on the counterpropagated wave. It allows each measurement to be completed within 7 mins with the finest setting mode. The measurements collected by DITEST STA-R report data in terms of Brillouin frequency shift which has a linear relationship with temperature, such that 1 MHz = 1 ºC increase. Each cable section shown in Fig 2b is highlighted in Fig 5, including 4 m OFS ‘loops’ that occur at each end of the channel. The largest and smallest shifts in frequency are associated with the cable 2-1 and cable 4, respectively.

4. Baseline checks to establish boundary conditions The initial reference temperature conditions had to be established first, so that later removal of sediment above the cable at specific locations could be correctly interpreted based on the changes in temperature. To understand the temperature distribution generated by the heat tape when it is activated, the change in temperature (ΔT) can be obtained by taking the difference between two measurements collected by the BOTDA, as shown in Fig 5. Fig 6 shows the temperature development throughout the testing period based on the initial temperature around 18 °C reported by PT-100 sensor. The temperature changes rapidly when the heating tape is switched on, where 50% of temperature is generated within the first 30 mins and the temperature then increases slowly towards eventual stabilisation. The heat tape was switched off at 320 mins and shows the rapid initial loss of temperature followed by a gradual temperature reduction leading to eventual ambient temperature condition. Two heating phases were conducted over the course of two separate days and are shown in Fig 7

Fig 6: Change in temperature over a full heating cycle as recorded by BOTDA

67

3-SUT265_34(2)_ut64415.indd 67

24/02/17 12:44 pm


Ouyang et al. The use of fibre optic distributed sensing technology to detect changes in sediment overburden

Fig 7: Temperature comparison between two heating phases on day 1 and day 2 at selected points along the channel adjacent to the heating tape

for specific locations along the channel. This comparison serves to demonstrate the reliability of the heating tape system to generate a consistent selfregulating heat source, which potentially highlights any major temperature changes in the experiment between the two days. The results in Fig 7 show that the change in temperature in day 2 was slightly less than day 1. This would be consistent with less temperature variability because of an improved soil packing around the cable (Woodside and Messmer, 1961). Some sediment consolidation would be expected over a 24 hr period. Similarly shaped plots indicate that the cable is generating the same temperature output over the same time step. The OFS 4 section was laid on the simulated seabed at the interface between the sand and the water

to measure water temperature. Water temperature recorded by OFS 4 varied because of partial embedment of the cable in some areas, which detected conduction heat flux from the sand as well as water temperature (Meininger and Selker, 2015). As a result, the readings from cable 4 were not used, and point reference readings were monitored instead using the PT-100 sensors. The water temperature through the heating phases was 17 °C with no appreciable increase caused by heat convection from the sand. This value is comparable with average summer values for the mid-North Sea water temperature (ignoring depth related variability). Variations in temperature between days 1 and 2 were also compared to understand the correlation in temperature from heat source with its adjacent cables.

68

3-SUT265_34(2)_ut64415.indd 68

24/02/17 12:44 pm


Underwater Technology Vol. 34, No. 2, 2017

Fig 8: Temperature comparison between the heating tape at OFS 2-1 and adjacent OFSs 1-1(a), 1-2(b), 2-2(c), 3-1(d) and 3-2(e) at position 1.5 m in the channel

Fig 8 shows the temperature comparison between the cables at the heating tape (2-1) and an adjacent cable, both at a point 1.5 m along the channel. Although the data show some variations during early stages of heating, there is similar temperature distribution from the heat tape across the channel on both days. Soil consolidation as identified in Fig 8 at 1.5 m along the channel may be regarded as insignificant.

5. Simulation of sediment removal The excavation of sediment was undertaken at two zones along the channel once an optimum temperature was established along the heat tape of ~24 °C. The water temperature was cooled using ice to an

average temperature of 12 °C. One excavation zone was located between 3 m and 4.5 m (zone 1) and the other between 6.5 m and 8.5 m (zone 2). Each zone comprised three stages of reduction in overburden sediment thickness above the cables. Sediment was retained using vertical partitions and then removed by hand to avoid damaging the fibre optic cables. A cross-section along the channel in Fig 9 shows sediment depth at each stage together with the position of OFSs. The excavation at the base was undulating owing to slumping of the sand beneath the vertical retaining partitions. No cables were exposed in excavation stages 1 and 2. In the final excavation (stage 3), cables were exposed (3-1 and 3-2) in both zones. The change recorded by 1-1 and 1-2 OFS

69

3-SUT265_34(2)_ut64415.indd 69

24/02/17 12:44 pm


Ouyang et al. The use of fibre optic distributed sensing technology to detect changes in sediment overburden

Fig 9: Position and excavation stage of sediment along channel

Fig 10: Change in temperature at specific points along the channel in response to sediment overburden excavation at (a) 3.5 m and (b) 7.0 m, with reference to (c) undisturbed sediment at 1.5 m. Arrows and numbers refer to commencement and stage of excavation, respectively

located beneath the heating tape are not discussed further in this paper as they form part of a wider investigation into changes in thermal gradient. Fig 10 presents the change in soil temperature at three separate positions on the OFS within the channel 300 mins after the start of the test with three stages of sediment removal. The temperature changes are shown by cable positions 2-1, 2-2, 3-1 and 3-2. Figs 10a,b refer to the positions where sediment removal took place. Fig 10c shows the temperature change where no sediment was removed,

though arrows show the time when sediment was removed at the two excavation zones. The changes shown after excavation of sediment are discussed in the following paragraphs. Excavation stage 1: there is a small negative change at 3.5 m and 7 m associated with cable locations 3-1 and 3-2; otherwise a small increase in temperature was detected between 0.3 °C and 0.5 °C. In the undisturbed soil, there is an increase in temperature at 1.5 m between excavation stages 1 and 2 of 0.8 °C.

70

3-SUT265_34(2)_ut64415.indd 70

24/02/17 12:45 pm


Underwater Technology Vol. 34, No. 2, 2017

The removal of up to 40 mm of sediment allows an ingress of water, which cools the cables closest to the surface. The small reduction in temperature due to removal of overburden also effects the cables at 2-1 and 2-2, though the overall reduction is masked by the continual increase generated by the heat tape. This linear increase in temperature from the heat tape is shown at 1.5 m where no sediment disturbance takes place. Excavation stage 2: change in temperature in zone 1 occurs after 21 mins and in zone 2 after 7 mins. The change affects all cable locations but especially at 3-1 and 3-2, where the temperature reduction is about 3.4 °C. At 1.5 m along the OFSs in the channel, there is an overall increase in temperature of 0.4 °C. As the overburden further reduces (Fig 9), there is 20 mm cover above the cable 3-1 and cable 3-2. The delay in temperature response between zones 1 and 2 was detected and is due to the excavations not being performed simultaneously, as zone 1 was excavated first. Excavation of the sediment also took longer to perform in zone 2 because the zone is wider. The reduction in temperature in cables

2-1 and 2-2 is of the order of 1.4 °C before the start of excavation stage 3. Excavation stage 3: there is an immediate change in temperature in zones 1 and 2 affecting all cables but especially cable 2-1, which reduces by 6 °C (at 3.5 m) and 10 °C (at 7 m). At 1.5 m, there is an overall increase in temperature of 0.4 °C. The exposure of cables 3-1 and 3-2 to the surface water results in an immediate change in temperature. This drop in temperature is larger in cable 3-1 because it is closer to the heat tape and had a higher temperature before exposure. Cables 3-2 and 2-2 show similar temperature change magnitudes during excavation stage 3. However, the exposure of cable sections from cable 2-1 and 3-1 demonstrates the sensitivity of the fibre optic sensor by instantly detecting the change in soil/water environment. The change in temperature at 2-1 in zone 1 was 5 °C, whereas in zone 2 the change was 9 °C due to exposure to the surface water. The temperature distribution because of sediment removal can also be shown as a function of channel length. Fig 11 shows the change in temperature

Fig 11: Change in temperature along the channel, following (a) excavation stage 2 and (b) excavation stage 3

71

3-SUT265_34(2)_ut64415.indd 71

24/02/17 12:45 pm


Ouyang et al. The use of fibre optic distributed sensing technology to detect changes in sediment overburden

along the channel at the end of excavation stages 2 and 3. The most significant changes in temperature are associated with the cable attached to the heat tape, whereas OFS 3-2 and 2-2 shows noticeable changes with smaller magnitude due to their physical positions being further away from the heat source. As expected, the change in temperature is more accentuated after the third excavation (shown in Fig 11b) than after the second excavation (shown in Fig 11a). In the undisturbed sections between the excavations zones, temperature fluctuations also occur. This is likely related to the disturbance of the water column and creation of some vortex during sediment excavation, which will cause some sediment erosion. In addition, behind the retained sections some slumping of the sand occurred. The disturbances illustrate the sensitivity of the DTS system for measuring temperature change. The change in temperature along the fibre optic sensor is presented as a heat map in Figs 12 and 13. The contour map represents horizontal slices at four cable locations along OFSs 2-1, 2-2, 3-1 and 3-2, and shows the temperature development in the

channel throughout the entire testing period. The temperature scale has been adjusted between plots because of the magnitude of temperature proximal to the heat tape. The time when each excavation commences is shown on the left side of each map as arrows. There are several elongate hot spots shown in the contour plots, notably in cable 2-1 at 6.5 m and around cable 3-1 at 8.7 m. This is likely to be associated with higher density soil around the cable or the relative installation position of fibre optic sensor to the heating tape. The OFS at position 2-1 was only directly attached to the heating tape at the cable tie positions leading to thermal discrepancies along the channel. The fibre optic sensing system has detected very subtle changes in temperature variation to the order of 0.3 °C. The positions of sediment removal are clearly shown as reductions in temperature or cold spots.

6. Discussion With the heat generated from the heat tape increasing linearly during the experiment, the change in

Fig 12: Temperature contour map showing horizontal slices at cable positions: (a) OFS 2-1; and (b) OFS 2-2

72

3-SUT265_34(2)_ut64415.indd 72

24/02/17 12:45 pm


Underwater Technology Vol. 34, No. 2, 2017

Fig 13: Temperature contour map showing horizontal slices at cable positions: (a) OFS 3-1; and (b) OFS 3-2

sediment thickness can be observed as both a stabilisation and/or reduction in temperature by OFSs. Despite a difference in temperature between heat tape and surface water, a minimum of 10 mm sediment removal could be detected from the temperature in the soil medium. The effective detection of sediment disturbance between excavation zones also demonstrates the sensitivity of DTS system to capture temperature change throughout the test. The thermal capacity of the soil depends on the grainsize and density, as well as grain type of the material. Sand was used in this experiment for ease of setup and elimination of air during placement above the fibre optic and heat tape. The use of a fine sand/silt mix would change the thermal conductivity and may show a significantly higher temperature difference between cable and seabed. The sensitivity of the DTS to monitor changes in sediment thickness would therefore increase. Changes in the sand packing density and consolidation along the channel would also occur during burial

of the power cable beneath the seabed. In granular sediments, changes in consolidation will be rapid. However, as shown in the rapid placement of sand around cables in this experiment, the variation and incremental increase in density around the cable over time can be determined by indirect measurement of thermal changes. The use of self-regulating heat tape as a source of heat at roughly constant temperature was convenient in the context of the experiment aimed at showing the capability of the optical fibre to detect local changes of soil cover via changes of temperature. In a field application with a real power cable, it will be necessary to calibrate the OFS within the cable after burial and prior to use, as well as at intervals during minimum and maximum current load, in order to offset the corresponding changes in heat generated by electrical resistance. Such further research and development can now be justified. As shown in this experiment, active DTS has the capacity to act as a monitoring system to measure

73

3-SUT265_34(2)_ut64415.indd 73

24/02/17 12:45 pm


Ouyang et al. The use of fibre optic distributed sensing technology to detect changes in sediment overburden

changes in the sediment thickness. Brillouin backscattered distributed sensing has the potential to monitor cables up to at least 60 km from one commercially available analyser channel. Such techniques could reduce the reliance on, or enhance the interpretation of data from, routine seabed surveys to locate scour pits. This would be especially beneficial during months when treacherous sea-state conditions limit the deployment of geophysical arrays and when storm surges may amplify seabed erosion due to strong currents or even bed liquefaction. Active fibre optic monitoring could also aid in monitoring scour mitigation during rock dumping activities to ensure successful remediation.

Acknowledgments The authors are grateful for the assistance of the technicians at the Schofield Centre and to Heat Trace™, UK for the generous donation of the heat tape for use in this experiment. The authors are also grateful to Mr Patrick Martin for helpful comments during drafting of this manuscript. The first and second authors acknowledge the financial support of Hird GeoWorks Ltd.

References Arnon A, Lensky NG and Selker JS. (2014). High‐resolution temperature sensing in the Dead Sea using fiber optics. Water Resources Research 50: 1756–1772. DNV GL. (2014). DNV-RP-J301: Subsea power cables in shallow water renewable energy applications. DNV Recommended Practice. <last accessed March 2016>. Fromme M, Christiansen W, Kjaer SV and Hill W. (2011). Distributed temperature monitoring of long distance

submarine cables. In: Proceedings of the SPIE 21st International Conference on Optical Fibre Sensors (OFS21), 15–19 May, Ottawa, Canada. Hara T, Terashima K, Takashima H, Suzuki H, Nakura Y, Makino Y, Yamamoto S and Nakamura Y. (1999). Development of long range optical fiber sensors for composite submarine power cable maintenance. IEEE Transactions on Power Delivery 14: 23–30. Inaudi D and Glisic B. (2010). Long-range pipeline monitoring by distributed fiber optic sensing. Journal of pressure vessel technology 132: 011701. Meininger TO and Selker JS. (2015). Bed conduction impact on fiber optic distributed temperature sensing water temperature measurements. Geoscientific Instrumentation, Methods and Data Systems 4: 19–22. Omnisens. Case study: Monitoring the temperature of 120 km Malta – Italy Interconnector. http://www.omnisens.com/ ditest/3431-power-cables.php. <last accessed March 2016>. Pilgrim J, Catmull S, Chippendale R, Tyreman R and Lewin P. (2013). Offshore wind farm export cable current rating optimisation. In: Proceedings of EWEA Offshore Conference, 19 – 21 November, Frankfurt, Germany. Steele‐Dunne SC, Rutten MM, Krzeminska DM, Hausner M, Tyler SW, Selker J, Bogaard TA and van de Giesen NC. (2010). Feasibility of soil moisture estimation using passive distributed temperature sensing. Water Resources Research. 46: 1–12. Svoma R, Smith C and Conway C. (2009). Integrated condition monitoring for subsea power cable systems. In: Proceedings of 20th International Conference on Electricity Distribution Part 1, 8–11 June, Prague, Czech Republic. Woodside W and Messmer JH. (1961). Thermal Conductivity of Porous Media. I. Unconsolidated Sands. Journal of Applied Physics 32: 1688–1699. Zhao XF, Li L, Ba Q and Ou JP. (2012). Scour monitoring system of subsea pipeline using distributed Brillouin optical sensors based on active thermometry. Optics & Laser Technology 44: 2125–2129. Zhao L, Li Y, Xu Z, Yang Z and Lü A. (2014). On-line monitoring system of 110 kV submarine cable based on BOTDR. Sensors and Actuators A: Physical 216: 28–35.

74

3-SUT265_34(2)_ut64415.indd 74

24/02/17 12:45 pm


doi:10.3723/ut.34.075 Underwater Technology, Vol. 34, No. 2, pp. 75–82, 2017

Technical Paper

www.sut.org

Analysis of the influence of ambient conditions on the ampacity of Kevlar-armoured subsea power cables N Vedachalam, A Umapathy, GA Ramadass and MA Atmanand National Institute of Ocean Technology, Ministry of Earth Sciences, Chennai, India Received 18 May; Accepted 27 August 2016

Abstract This paper presents the electro-thermal modelling and simulation, done on a standard Kevlar-armored subsea electrooptic cable based on the finite element analysis (FEA) approach, to determine the cable ampacity when operated in air, water and buried in the seabed under the relevant ambient environmental conditions. Compared to the ampacity in air at 30 ºC, the ampacities in air, water with no flow condition and seabed buried conditions with a thermal conductivity of 0.8 W/m-k could be 1.37 times, 1.57 times and 1.5 times, respectively, at an ambient temperature of 5 ºC; and the same could be 1.13 and 1.36 and 1.31 times, respectively, at 20 ºC. Under the influence of convective ambient flow fields, the ampacities could be increased only up to a maximum of 9% and 3%, with air and water flows of 20 m/s and 5 m/s, respectively. The results serve as a guideline for determining the cable ampacities for various environmental applications, based on the available cable current rating under specified ambient operating conditions. Keywords: Ampacity, cables, finite element analysis, kevlar armour

1. Introduction A reliable power cable is the lynchpin for subsea systems used in oil and gas installations, offshore wind energy farms interconnecting multiple turbines to power aggregating stations, tidal energy farms involving multiple marine current generators, long step-out power transmission, mineral mining systems, and other intervention systems (Nidhi et al., 2015; Vedachalam 2013; Vedachalam and Andreasen 2013; Vedachalam et al., 2013). A kevlar-armoured cable provides multiple advantages including higher strength-to-weight ratio, excellent corrosion resistance, easier handling, reliability, low armour losses and life expectancy compared to steel armoured types. In light of these benefits, the use of Kevlar-armoured subsea cables is on the rise (Shilpa et al., 2015; Vedachalam et al., 2015).

Modern Kevlar-armoured umbilicals used for complex offshore developments comprise a combination of power conductors, optical fibres and hydraulic passes (Nexans, 2007). Failure of these cables leads to huge economic losses and requires costly interventions. Accelerated ageing and failure of cables are mainly due to the failure of the power conductor insulation when they are operated at temperatures higher than the design ratings (Holyk et al., 2014; Karahan and Kalenderli, 2011). Hence their current ratings/ampacities should be selected based on both the current carrying requirements and the environmental conditions in which they are expected to operate. Environmental conditions mainly include temperature, flow field and the thermal characteristics of the surrounding medium. The power cables are generally specified for ampacity at specific ambient conditions, including temperature and surrounding fluid medium. When they are used under different environmental conditions, their ampacity needs to be determined for the specific environment conditions (Marshall and Fuhrmann, 2015). Utilising a cable with reduced ampacity leads to increased conductor temperature, which is detrimental for the cable insulation, while an oversized cable leads to underutilisation, increased cost and underwater handling difficulties. With increasing water depths, ambient temperature decreases and hence the cables specified for shallow-water applications can handle higher currents when operated in deeper waters owing to increased conductive heat transfer (Ramesh et al., 2013). In tidal locations, higher water velocity fields increase the convective heat transfer, and hence, the cable could handle higher currents (Pham and Martin, 2009). In the case of offshore wind farms, where multiple turbines are connected to a central grid, the

75

4-SUT266_34(2)_ut64416.indd 75

01/03/17 3:57 pm


Vedachalam et al. Analysis of the influence of ambient conditions on the ampacity of Kevlar-armoured subsea power cables

submerged and air exposed sections experience different ambient conditions (Pilgrim et al., 2013). Seabed buried cables have higher conductive heat transfer performance due to the low ambient temperature and higher thermal conductivity of the moisture saturated backfill (Waite et al., 2009). Thus selecting a suitable cable taking into consideration the current carrying requirement and the specific operating environmental conditions is a challenge. The next section of this paper describes the limitations in conventional standards in determining the ampacity of present generation Kevlararmoured subsea cables, which are expected to operate in a wide range of ambient conditions, and the present day numerical modelling capabilities in ampacity determination. Section 3 presents the finite element analysis (FEA) done on an industry standard Kevlar-armoured subsea power cable, and section 4 shows the FEA results of the ampacity when the cable is operated under various environmental conditions.

2. Need for finite element analysis Dissipation of the heat generated due to current flow in power cables is based on conduction, convection and radiation mechanisms. Heat conduction is based on Fourier’s law, which states that the heat flux is proportional to the ratio of temperature over the surrounding space. Convection is based on Newton’s law, which states that the flow of heat is proportional to the temperature difference. Radiation is based on Stefan-Boltzmann law, which states that the quantity of heat radiated is proportional to the difference in temperature to the power of four. In the case of buried cables, the dissipated heat undergoes diffusion, which gets transferred to the surrounding region in a slow, space limited decaying fashion through conduction and convention processes (Maximov et al., 2016). Safe ampacity is a function of the internal and external cable system components that comprise the thermal circuit of the cable and its boundaries. Cable manufacturers normally specify the ampacity of the cable for a specified ambient temperature and surrounding medium for which the cable is designed for continuous operation. When the cable is used at environments other than the specified, the ampacity corresponding to the environment needs to be identified for safe and optimum utilisation of the cable. The standards for computing the ampacity of the cables other than the specified temperature does not consider the changes in ambient medium and the surrounding fluid flow fields.

Hence analytical approaches or FEA methods should be used to determine the environment specific ampacity (International Electrotechnical Commission (IEC), 2003). The traditional analytical method of computing temperature distribution and cable ampacity based on multiple approximations was first carried out by Neher and McGranth (N-M) in 1957 based on the method developed by Pashkis and Baker (1942), which was later adopted as the IEC60287 standards (2006). The N-M method of ampacity calculations is based on the thermal-electrical analogy and is realised with simplified boundary conditions and step functions. The assumptions simplify the mathematical formulation from complex 3D to 1D, and further to algebraic equations. It could be applicable only in homogeneous ambient conditions and for simple cable geometries (Baazzim et al., 2014; Marshall and Fuhrmann, 2015). Further, such approximations lead to inaccuracies in ampacity calculations and forces application engineers to use unnecessarily large safety factors. Taking into consideration the limitations in adopting the published standards and analytical models, IEC TR 62095-2003 recommends the use of elaborate FEA models for precision ampacity determination for modern special purpose power cables. The models would be based on the complex set of differential equations to solve the complex 3D heat transfer problem of arbitrary and variable boundary conditions (IEC, 2003). FEA based cable ampacity determinations under transient and steady state conditions are being done using numerical tools such as COMSOL multi physics, ANSYS Fluent, CYMCAP, USA mp and ETAP (Anders et al., 2010; Shackleton et al., 2007; Teja and Rajagopal, 2014; Vahidi and Mahmoudi, 2012; Vaucheret et al., 2005). The tools utilise techniques based on finite element, finite volume, boundary element, overset grid and perturbed methods. These techniques analyse the electrothermal behaviour of power cables used in underground locations, trenches, open top trays and covered trays. Kevlar-armoured cables are recent developments in the subsea power transmission industry, and FEA on these cables are seldom available. In spite of the excellent mechanical and corrosion resistance properties, Kevlar’s low thermal conductivity demands the need for carrying out FEA to understand the electrothermal behaviour of the cable and determine the ampacity under various environmental conditions (other than the specified ambient medium and fluid fields) (Ventura and Martelli, 2009).

76

4-SUT266_34(2)_ut64416.indd 76

01/03/17 3:57 pm


Underwater Technology Vol. 34, No. 2, 2017

3. FEA methodology and validation The thermal field around the cable medium is governed by the differential equation:

heat transfer rate (q) crosses the boundary is specified as: k

∇⋅( ∇ ) = −Q + c

∂T ∂t

(1)

where K is the thermal conductivity, T is the temperature at any point, Q is the heat generation rate per unit area, c is the specific heat capacity and t is the time. In the steady-state thermal analysis of 2D media, equation 1 reduces to: ∂ ∂T ∂ ∂T kx + ky +Q = 0 ∂x ∂x ∂y ∂y

1 ∂T kx 2 ∫∫ ∂x

2

+ ky

∂T ∂y

2

dx dy

−∫∫ QT Q dx dy

(3)

The boundary condition on the surface of the cable system (usually air, water or soil) may be represented as a convection boundary, where the derivative of the temperature on this boundary in a direction normal to the boundary is proportional to the temperature difference with respect to the ambient temperature. The condition where the

2

1 ∂T I= ∫∫ k 2 ∂x

∂T +k ∂y

2

1 + ∫ Γ1 h(T 2 − 2TaT )d Γ1 2

dx dy − ∫ ∫ QT Q dx dy ∫ 2 qTd Γ2

(5)

The FEA is performed by the discretisation of the entire domain of analysis into smaller subdomains shown in equation 6, followed by selecting a suitable interpolation function that provides an approximation of an unknown solution with the sub-domain (Al-Saud et al., 2008; Maximov et al., 2016): Te

3

∑N T j =1

I

(4)

Thus, the boundary condition including both the conductive and convective heat transfers is represented as:

(2)

In any homogenous medium of a given thermal conductivity and heat generation rate, equation 2 can be solved for the temperature at any point in the cable region, subjected to specified boundary conditions. The FEA is based on the solution to equation 2, namely that T(x, y) is that which minimises the function given by:

∂T = h( h T −Ta ) ∂n

e j

e j

(6)

where Nje are the linear interpolation function of x and y. The minimisation of equation 5 is performed over each individual finite element defined by the associated nodes. Based on the described FEA principles and the recommendations of IEC TR 62095, FEA is performed for the Kevlar-armoured subsea standard umbilical used in a deepwater remotely operated vehicles (ROVs; Vedachalam et al., 2015). The Kevlar-armoured cable, with the specifications shown in Table 1 and the constituent material properties shown in Table 2, is modelled and simulated

Table 1: Specifications of the power cable (Vedachalam et al., 2015) Detail Power conductors Optical conductors Mechanical member Overall

Description 3 cores of 6 mm2 copper conductors with polyethylene insulation rated for 6600 V, 12 A load, 70 °C continuous operation, resistance of 3.2 Ω/km at 30 °C, 100 nF/km capacitance 6 numbers of single mode fibres inside steel tube of 5 mm diameter Two layers of aramid yarns capable of providing safe working load of 11 T and nominal breaking load of 63 T 600 bar pressure rated cable with overall diameter of 37.5 mm insulated with polyethylene and with a minimum dynamic bending diameter of 1.5 m

Table 2: Properties of the cable system components Type

Thermal in W/m.C

Polypropylene insulation Thermoplastic elastomer insulation Kevlar armour Copper conductor

0.22 0.4 0.04 386

conductivity

Specific heat capacity in J/kg.C

Mass density in Kg/m3

1900 1350 340 383.1

905 1380 1440 8954

77

4-SUT266_34(2)_ut64416.indd 77

01/03/17 3:57 pm


Vedachalam et al. Analysis of the influence of ambient conditions on the ampacity of Kevlar-armoured subsea power cables

Fig 1 shows the simulation results, which found that the steady-state temperature rise in the conductor is 70.41 ºC. The results match the manufacturer’s specifications with an accuracy of 0.6%, which validates the accuracy of the model. The model is further experimentally validated by measuring the change in the electric resistance of the power-carrying copper conductors in a known length of the same cable, after circulating 30 A DC under an ambient of 34.6 ºC in a closed chamber until a steady state temperature is reached. The resistances are measured every hour after energising the cable. The steady-state temperature is found to comply with the model results for the same environmental conditions with an accuracy of 0.4%. The validated model is extended for further analysis. Fig 2 shows the simulation results for the transient temperature rise in the power conductors.

Fig 1: Electrothermal model at 30 A at 30 ºC

for temperature rise under various amplitudes of current using Infolytica MagNet v7 and ThermNet v7 FEA software v7.4.1 (Ramesh et al., 2015). The MagNet v7 software is used to model the cable components and the current flow in the electrical conductors, and the coupled ThermNet v7 software is used for simulating the transient and steady-state temperature in the cable components and the surrounding medium. To validate the accuracy of the model, the cable is powered with a direct current of 30 A, and the ambient air temperature around the cable is set at 30 ºC with zero air velocity. This condition exactly conforms to the cable manufacturer’s specifications on ampacity, ambient medium and temperature for which a peak temperature rise of 70 ºC is expected in the power conductors.

4. Performance under various environments The validated cable model is used for analysing the ampacity under ambient conditions, including: an air medium with varying temperature and flow field; a water medium with varying temperature and flow field; and a buried cable with varying soil temperatures and thermal conductivities. The relevance of the analysis of the influence of these variable environmental conditions in specific offshore systems is shown in Fig 3.

70.4 °C

11

10

10.5

9.5

9

8.5

8

7.5

7

6.5

6

5.5

5

4

4.5

3

3.5

2

2.5

1.5

1

0.5

Average temperature, R (temperature)

0

Temperature °C

R 71 69 67 65 63 61 59 57 55 53 51 49 47 45 43 41 39 37 35 33 31 29

Time (hr)

Fig 2: Transient with air 30 ºC, 30 A and 0 m/s air velocity

78

4-SUT266_34(2)_ut64416.indd 78

01/03/17 3:57 pm


Underwater Technology Vol. 34, No. 2, 2017

1.5

Air flow velocity

Marine current turbines – subsea grids

Wind turbines – aggregating stations

1.46

0 m/s 1 m/s 3 m/s 5 m/s 10 m/s 20 m/s

1.4

% of ampacty at 30 °C

Air temperature

1.36 1.3

1.37 1.24

1.25

1.2 1.1

1.10

1.13

1.0 10

5

Water temperature

Oil an gas installations

20

30

40

0.95

0.9

Air temperature in °C

Fig 4: FEA results for operation under various conditions in air

Soil thermal conductivity

Long step out power transmission

Intervention systems (ROV)

Fig 3: Influence of environmental conditions in offshore systems

1.6

% of ampacty at 30 °C

Water flow velocity

1.57 1.51

1.5

1.47 1.4

1.40 0. m/s 0.2 m/s

1.3

1.36

0.5 m/s 1.27

1 m/s 5 m/s 1.2 5

10

20

30

Water temperature in °C

4.1. Air medium Portions of subsea cables used for interfacing offshore wind turbines with power aggregating stations, as well as unreleased portion of cables used for subsea intervention systems such as ROV, need to be operated in an air medium involving variable ambient air temperature and air flow velocity. When the cable passes through various environments, the section of the power cable which is exposed to higher environmental stresses determines the ampacity of the cable. FEA is performed with the validated model operating in an air medium with a range of ambient temperature of 5 ºC to 40 ºC and air velocities ranging from 0 m/s to 20 m/s, and the results are plotted in Fig 4. It shows that the ampacities of the cable at 5 ºC ambient temperatures and air velocities of 0 m/s and 20 m/s are 1.37 and 1.46 times that at 30 ºC, respectively. At 30 ºC, the air velocity of 20 m/s increases the ampacity to 1.1 times that at zero air velocity conditions. 4.2. Water medium Subsea cables, used for interconnections in marine current renewable energy systems, hydrocarbon recovery installations, long step-out power transmissions and intervention systems, are exposed to variable ambient temperature and water flow velocities dependent on the location they are installed or operated in. When the depth of the installation

Fig 5: FEA results of operation in various conditions in water

increases, the ambient water temperature decreases and is about 2 ºC to 4 ºC at 1000 m water depths (Ramesh et al., 2013). In tidal energy farms, the cables experience high and variable water flows. As the farms are located in less than 100 m water depths close to the shore, the ambient temperatures are in the range of 20 ºC – 28 ºC (Pham and Martin, 2009). FEA is performed on the validated model with water as the operating medium, ambient water temperature ranging from 5 ºC to 30 ºC and water velocities of 0 m/s to 5 m/s. The results are plotted in Fig 5, which shows that the ampacities of the cable at 5 ºC ambient temperatures and water velocities of 0 m/s and 5 m/s are 1.57 and 1.61 times that at 30 ºC, respectively. At 30 ºC, the water velocity of 5 m/s increases the ampacity by 2.5% compared to that at zero air velocity conditions. The simulated results showing the steady-state temperature of the cable and the water medium at a conductor steady-state temperature of 70 ºC under an ambient temperature of 10 ºC and zero water velocity conditions are given in Fig 6.

4.3. Soil medium Subsea cables mainly used for long step-out power delivery to enhanced hydrocarbon recovery systems and cables used for cross-country power transmission

79

4-SUT266_34(2)_ut64416.indd 79

01/03/17 3:57 pm


Vedachalam et al. Analysis of the influence of ambient conditions on the ampacity of Kevlar-armoured subsea power cables

Fig 6: Simulation output at 10 ºC and still water ambience

1.5

0.8 W/m-k

1.50

1.0 W/m-k

Fig 8: Simulation output at 10 ºC and 0.8 W/m-k thermal conductivity

1.2 W/m-k

1.43 1.41 1.4

1.33

30

1.31 1.3 1.21 1.19

1.2 5

10

20

30

Soil temperature in °C

Fig 7: FEA results of operation in various conditions in soil

are usually buried. They use special water jet based cable trenching equipment to prevent damage due to falling objects. The ampacity of the cable varies with the depth of burial, which is a function of the soil temperature and associated thermal properties. The heat transfer properties in a buried soil medium mainly depend on the soil ambient temperature and its thermal conductivity, both of which are normally identified through geotechnical methods prior to design and laying of the power cable. Ambient soil temperatures can vary from 2 ºC to 28 ºC, depending on the location and the backfill with soil thermal conductivities in the range of 0.8 W/m-k–1.2 W/m-k (Waite et al., 2009). FEA is performed with the validated model in the range of temperature and thermal conductivities. The results are plotted in Fig 7, which shows that the ampacities of the cable at 5 ºC ambient temperatures and soil thermal conductivities of 0.8 W/m-k and 1.2 W/m-k are 1.5 and 1.52 times that at 30 ºC, respectively. The simulation results showing the steady-state temperature of the cable and the soil medium at a conductor steady-state temperature of 70 ºC under an ambient temperature of 10 ºC and thermal conductivity of 0.8 W/m-k are illustrated in Fig 8.

% increase in ampacity over 30 ºC air ambience

% of ampacty at 30 ºC

1.52

Air

Water

25 20 15 10 5 0 0

5

10

15

20

Ambient fluid velocity in m/s

Fig 9: Influence of fluid flow in cable ampacity compared to 30 ºC in air

The obtained results were analysed to understand the performances under convective heat transfer conditions due to fluid flow around the cable, and the results are plotted in Fig 9. It can be seen that the ampacities could be increased only up to a maximum of 9% and 3% for air and water flows of 20 m/s and 5 m/s, respectively.

5. Summary and discussion The FEA performed on the standard Kevlararmoured subsea power cable to determine the cable ampacity when operated in air, water and buried in the seabed under the relevant ambient environmental conditions reveals the following findings: • The ambient temperature plays a significant role in the ampacity determination compared to the surrounding fluid velocity.

80

4-SUT266_34(2)_ut64416.indd 80

01/03/17 3:57 pm


Underwater Technology Vol. 34, No. 2, 2017

Table 3: Derived coefficients for ampacity determination Environmental condition Coefficient a

b

–0.0146 3 –0.0143 3 –0.0149 0 –0.0146 1 –0.0147 9 –0.0145 7

1.447 1.461 1.497 1.513 1.527 1.536

–0.0132 2 –0.0132 2 –0.0132 2 –0.0132 2

1.631 1.643 1.651 1.669

–0.0131 5 –0.0131 5 –0.0131 5

1.564 1.574 1.584

Air medium 0 m/s 1 m/s 2 m/s 5 m/s 10 m/s 20 m/s Water medium 0 m/s 0.5 m/s 1 m/s 5 m/s Soil medium 0.8 W/m-k 1.0 W/m-k 1.2 W/m-k

• Compared to the ampacity in air at 30 ºC, the ampacities in air, water with no flow and seabed buried conditions with a thermal conductivity of 0.8 W/m-k could be 1.37 times, 1.57 times and 1.5 times, respectively, that at an ambient temperature of 5 ºC; and the same could be 1.13 and 1.36 and 1.31 times, respectively, that at 20 ºC. • Under ambient flow fields, the ampacities could be increased only up to a maximum of 9% and 3%, with air and water flows of 20 m/s and 5 m/s, respectively. To ensure a wider use of the identified results, the curves are represented in the form of equations, using the curve fitting tool box of MATLAB software. Y=aX+b

(7)

where X is the ambient temperature and Y is the ampacity factor. The values of coefficients a and b for the analysed cases are listed in Table 3.

Acknowledgment The authors gratefully acknowledge the support extended by the Ministry of Earth Sciences, Government of India, in funding this research.

References Al-Saud MS, El-Kady MA and Findlay RD. (2008). A new approach to underground cable performance assessment. Electric Power Systems Research 78: 907–918. Anders GJ, Coates M and Chaaban M. (2010). Ampacity calculations for cables in shallow troughs. IEEE Transactions on Power Delivery 25: 2064–2072. Baazzim MS, Al-Saud MS and El-Kady MA. (2014). Comparison of finite-element and IEC methods for cable thermal

analysis under various operating environments. International Journal of Electrical, Computer, Energetic, Electronic and Communication Engineering 8: 484–489. Holyk C, Leiss HD, Grondel S, Kanbach H and Loos F. (2014). Simulation and measurement of the steady-state temperature in multi-core cables. Electric Power Systems Research 116: 54–66. International Electrotechnical Commission (IEC). (2003). Electric cables – Calculations for current ratings – Finite element method. Technical report TR 62095: 2003. International Electrotechnical Commission, 69pp. IEC (2006). Electric Cables - Calculation of The Current Rating. International Standard 60287 Part 1-1. International Electrotechnical Commission. Karahan M and Kalenderli O. (2011). Coupled electrical and thermal analysis of power cables using finite element method. In: Vikhrenko VS. (ed.) Heat transfer—engineering applications. Croatia: IntechOpen, 205–230. Marshall JS and Fuhrmann AP. (2015). Effect of rainfall transients on thermal and moisture exposure of underground electric cables. International Journal of Heat and Mass Transfer 80: 660–672. Maximov S, Venegas V, Guardado JL, Moreno EL and Lopez R. (2016). Analysis of underground cable ampacity considering non-uniform soil temperature distributions. Electric Power Systems Research 132: 22–29. Nexans. (2007). Nexans customised offshore cablessubsea communication and control. Available at: http:// imistorage.blob.core.windows.net/imidocs/0215-28p007%20 customised%20offshore%20cables.pdf. <last accessed 12 December 2016>. Nidhi V, Rajesh S, Aarthi AP, Ramesh NR, Vedachalam N, Ramadass GA and Atmanand MA. (2015). Estimation of reliability of underwater polymetallic nodule mining machine. Marine Technology Society Journal 49: 131–147. Pashkis V and Baker H. (1942). A method for determining the steady state heat transfer by means of electrical analogy. ASME transactions 104: 105–110. Pham CT and Martin VA. (2009). Tidal current turbine demonstration farm in Paimpol-Brehat (Brittany): tidal characterisation and energy yield evaluation with telemac. In: Proceedings of the 8th European wave and tidal energy conference, 7–10 September, Uppsala, Sweden, 181–188. Pilgrim J, Catmull S, Chippendale R, Tyreman R and Lewin P. (2013). Offshore wind farm export cable current rating optimisation. In: Proceedings of EWEA Offshore Conference, 19–21 November, Frankfurt, Germany, 10pp. Raja SN, Basu S, Limaye AM, Anderson TJ, Hyland CM, Lin L, Alivisatos AP and Ritchie RO. (2015). Strain-dependant dynamic mechanical properties of Kevlar to failure: structural correlations and comparison to other polymers. Materials Today Communications 2: 33–37. Ramesh S, Ravichandran M, Ramadass GA and Atmanand MA. (2013). Dissolved oxygen as a tracer for intermediate water mixing characteristics in the Indian Ocean. Current Science 105: 1724–1729. Ramesh R, Umapathy A, Babu SM, Vedhachalam N, Venkatesan K, Harikrishnan G, Subramanium AN, Ramadass GA and Atmanand MA. (2015). Heat dissipation studies on sub- sea cables wound on winches. In: Proceedings of the IEEE Underwater Technology Conference, 23–25 February, Chennai, India. Shackleton D, Abib L and Balena R. (2007). Electrical and thermal design of umbilical cable. In: Proceedings of the 7th International Conference of Power Insulated Cables, Jicable 2007, 24–27 June, Paris, France. 6pp.

81

4-SUT266_34(2)_ut64416.indd 81

01/03/17 3:57 pm


Vedachalam et al. Analysis of the influence of ambient conditions on the ampacity of Kevlar-armoured subsea power cables

Teja AD and Rajagopal K. (2014). Thermal analysis by conduction, convection and radiation in a power cable. IOSR Journal of Electrical and Electronics Engineering 9: 51–56. Vahidi B and Mahmoudi A. (2012). Determination of the ampacity of buried cable in non- homogenous environmental condition by 3D computation. Journal of Electrical Engineering & Technology 7: 384–388. Vaucheret P, Hartlein RA and Black WZ. (20005). Ampacity derating factors for cables buried in short segments of conduit. IEEE Transactions on Power Delivery 20: 560–565. Vedachalam N. (2013). Review of challenges in reliable electric power delivery to remote deep water enhanced oil recovery systems. Applied Ocean Research 43: 53–67. Vedachalam N and Andreasen HB. (2013). Challenges in realizing reliable subsea electric power grid for tidal energy farms. Marine Technology Society Journal 47: 80–93.

Vedachalam N, Ramesh R, Muthukumaran D, Aarthi A, Subramanian A, Ramadass GA and Atmanand MA. (2013). Reliability-centered development of deep water ROV ROSUB 6000. Marine Technology Society Journal 47: 55–71. Vedachalam N, Umapathy A, Ramesh R, Babu SM, Muthukumaran D, Subramanian A, Harikrishnan G, Ramadass GA and Atmanand MA. (2015). Ampacity derating analysis of winch-wound power cables: a study based on deep-water ROV umbilical. IEEE Journal of Oceanic Engineering 41: 462–470. Ventura G and Martelli V. (2009). Thermal conductivity of Kevlar 49 between 7 and 290 K. Cryogenics 49: 735–737. Waite WF, Santamaria JC, Cortes DD, Dugan B, Espinoza DN, Germaine J, Jang J, Jung JW, Kneafsey TJ, Shin H, Soga K, Winters WJ and Yun T-S. (2009). Physical properties of hydrate bearing sediments. Reviews of Geophysics 47: 38 pp.

82

4-SUT266_34(2)_ut64416.indd 82

01/03/17 3:57 pm


doi:10.3723/ut.34.083 Underwater Technology, Vol. 34, No. 2, pp. 83–91, 2017

Technical Paper

www.sut.org

Advances in deepwater structure installation technologies Yi Wang*1, Menglan Duan1, Huaguo Liu2, Runhong Tian1 and Chao Peng1 1 Institute for Ocean Engineering of CUP, Beijing, 102249 2 Kerui Group, Shangdong, 257067 Received 30 May 2016; Accepted 1 November 2016

Abstract New offshore projects are targeting water depths of over 3000 m far from the land, and the preferred option for field development is deepwater structures, which include subsea equipment and pipeline systems. Many publications are focused on the deepwater structure installation technology in order to understand the behaviour of structures during installation and to control the installation process safely. In this paper, the installation solutions backed by engineering tools and numerical simulation methods are presented and discussed for subsea equipment and pipelines, respectively. The corresponding latest advances in the installation technologies are presented, together with their main characteristics and critical challenges. The authors also discuss general trends in future development that may result in further advances. Keywords: Deepwater, subsea equipment, pipeline, installation

1. Introduction In the past decades, offshore exploration and the production of oil and gas have advanced from shallow water into deeper water at an increasing pace. Subsea equipment, together with pipeline systems, have become a substitute for traditional fixed leg platforms in deepwater environments. Up to now, existing subsea equipment and pipelines have been able to operate in water depths beyond 3000 m (Martijn and David, 2014). The term ‘deepwater structure’ covers a very wide range of subsea equipment and pipeline systems, including: manifolds; templates; foundations; pipe line end manifolds (PLEMs); pipe line end terminations (PLETs); subsea umbilical termination assemblies (SUTAs); subsea distribution assemblies * Contact author. Email address: wangyizyn@sina.com

(SDAs); jumpers and flying leads; subsea wellheads and Christmas trees; riser bases; subsea control systems; subsea multiphase flow measurement; subsea compression modules; subsea boosting stations; subsea water injection modules; and subsea separation modules (Bai and Bai, 2012). The deployment of such equipment requires specialised and expensive vessels which need to be equipped with diving equipment for relatively shallow work and robotic equipment for deepwater. ‘Subsea installation’ refers to the installation of deepwater structures in an offshore environment which is a dangerous activity and so heavy lifting is avoided as much as possible. The goal of deepwater structure installation is to maximise economic gain safely using the most reliable and cost-effective solutions currently available (Penati et al., 2015). Different subsea structures require different deepwater installation vessels. Typical deepwater installation vessels include expensive floating drilling rigs, heavy lift vessels, multi-function engineering vessels and pipe-laying barges. For deepwater structures, some of these facilities can be huge reaching a weight of hundreds of tons, with lengths and widths of 50 m, and heights of 30 m. The differences in size, shape and weight of structures generate great challenges to deepwater installation. This paper gives a comprehensive overview of the current state-of-the-art deepwater structure installation technologies and categorises various installation methods based on the latest advances in subsea construction. Three major challenges: lifting/lowering technology; dynamic responses to environmental conditions; and positioning and control techniques, are described here to address the important aspects of deepwater installation technologies.

83

5-SUT268_34(2)_ut64417.indd 83

01/03/17 2:30 pm


Yi Wang et al. Advances in deepwater structure installation technologies

2. Main installation methods Typically, there are six installation methods for subsea equipment (Wang et al., 2012) and three installation methods for subsea pipelines (Bai and Bai, 2014).

2.1. Installation methods for subsea equipment The conventional installation methods for subsea equipment are the drilling riser installation method (DRIM) and the lifting method. With the increased understanding of the mechanical properties of various man-made fibre ropes, allied to the industry’s confidence in fibre rope deployment systems (FRDS), many unconventional deepwater installation methods have been developed and successfully applied. Examples include the sheave installation method (SIM), the pendulous installation method (PIM), the pencil buoy method (PBM), and the wet tow installation method (WTIM). 2.1.1. Drilling riser installation method The DRIM (Fig 1) is an established method for subsea equipment installation (Moreira et al., 1999). Generally, the drilling vessels, which possess dynamic positioning capability and have heave compensation for the main hook, are able to install the subsea equipment (Moreira and Johansen, 2001). The drilling riser systems of drilling vessels are often used to install small facilities such as wellheads, Christmas trees and blow-out preventers (BOPs), but are rarely used to install big facilities such as manifolds because of the size limitation of their moonpools. In the DRIM installation process, the subsea equipment is transported to the operation area by the barge initially and then fixed on the moonpool. The subsea equipment is then lowered to the seabed slowly with the help of a drilling riser. The number of drilling vessels suitable for deepwater installation is limited, because of low availability and expensive day rates. In addition, the connection of drill pipe takes time, and so efficiency is low in deepwater conditions.

Fig 1: DRIM for installation of Christmas tree

2.1.2. Lifting method The lifting method (Fig 2) is the predominant method used for subsea equipment installation, mainly because of the easy availability of multifunction engineering ships; it has been employed in offshore oil and gas fields for decades. In the lifting method installation process, the subsea equipment is transported initially to the installation site by the supply vessels, and then deployed to the installation location slowly using the wire rope which is attached to the crane or winch of the installation vessel (Roveri et al., 1996). Limited by the lift capacity and the wire length of the crane or winch, the lifting method is suitable mainly for conditions where water depth is shallower than 1500 m. The installation vessel must possess the function of dynamic positioning, heave compensation, large lift capacity, large transport deck area and many other features. The weight of the wire rope increases significantly with water depth and introduces much higher requirements of the deployment platform, which may result in a higher cost of installation. 2.1.3. Sheave installation method The SIM (Fig 3), which was first deployed in 2002 to a 1885 m water depth (de Gam Lima et al., 2008), is based on the two-fall configuration of a conventional deployment system so that the installation

Fig 2: Lifting method for installation of manifold

Fig 3: SIM for installation of manifold

84

5-SUT268_34(2)_ut64417.indd 84

01/03/17 2:30 pm


Underwater Technology Vol. 34, No. 2, 2017

ability is increased. In a typical SIM installation process three vessels, including one drilling vessel and two supply vessels, are used. The drilling vessel is mainly used to lift the subsea equipment with its crane and deploy it through the splash zone to a certain depth, where the load of the equipment is transferred to the wire rope. As shown in Fig 3, the wire comes from a supply vessel and goes through the sheave above the subsea equipment up to the drilling vessel. The supply vessel provides the fixed point for the dead end of the wire rope coming from the drilling vessel when lowering the equipment. A second supply vessel maintains an adequate distance from the first supply vessel and, therefore, provides assistance in orienting the subsea equipment thus avoiding any potential twist induced by a two-fall configuration. 2.1.4. Pendulous installation method The PIM (Fig 4) was developed by Petrobras and was first deployed in 2005 to a 1900 m water depth in order to install a large manifold (Wang et al., 2013). In a typical PIM installation process, two small installation vessels are used to launch and deploy the subsea equipment. The first installation vessel uses its crane to lift the structure into the water and through the splash zone, and then transfers the load from the crane to the launch winch wire at a certain water depth. Through a pendulous movement, the load gradually transfers from the launch line to the deployment line of the second installation vessel. Finally, the deployment winch can continue deploying the structure vertically, position it and then land it on the seabed target location. The PIM is a cost-effective method because it uses small conventional deepwater construction or offshore support vessels, without using special rigging and powerful FRDSs. However, it may require an anti-rotation system, such as a counterweight, to control the rotation of the load that may be caused by any hydrodynamic instability during the launch operation.

2.1.5. Pencil buoy method The PBM (Fig 5) was developed by Aker Marine Contractors in 2006, and is a cost-effective and safe technique for the installation of subsea equipment (Mork and Lunde, 2007). The PBM comprises three principal steps (Risoey et al., 2007). The first of these involves transferring the weight of the subsea equipment from the crane barge to a pencil buoy, which is a cylindrical structure, in calm inshore waters. A towing operation is then conducted using the pencil buoy to carry the submerged weight of the equipment and rigging to the field. The final step is the transfer of the load from the pencil buoy to a winch on the vessel for lowering to the seabed. The PBM is has several advantages: there is no need for large deck space, no crane is required offshore, and no pendulum motions occur in the air with the risk of slamming loads. This not only minimises the risk but also saves on costs by not having to use the significantly more expensive offshore heavy-lift vessels. The method was originally developed for installing subsea equipment, but it has also been used for recovering structures. 2.1.6. Wet tow installation method The WTIM (Fig 6) was developed by Subsea 7 and performs a submerged tow through the moonpool

Fig 5 : The set-up of PBM

Fig 4: PIM for installation of manifold

Fig 6: WTIM for installation of manifold

85

5-SUT268_34(2)_ut64417.indd 85

01/03/17 2:30 pm


Yi Wang et al. Advances in deepwater structure installation technologies

of an offshore service vessel so as to have a wider operational window for subsea equipment installation (Jacobsen and Leira, 2012). All offshore lifts are eliminated and the maximum crane capacity on-board is utilised in the WTIM, which ensures a safe installation process that, accordingly, is more cost efficient. It is important that the computational models and procedures are validated before the submerged towing operation is carried out because the dynamic behaviour of the equipment and tow arrangement will depend on the hydrodynamic loads.

2.2. Installation methods for subsea pipeline The conventional installation method for subsea pipeline is the S-lay method. Relatively new methods for deepwater pipeline installation are the J-lay method and the Reel-lay method. The S-lay method (Fig 7) takes its name from the shape of the S-configuration of the suspended pipe from the stinger to the seabed (Gernon et al., 1995). Individual lengths of pipe are welded on the pipe laying vessel and deployed horizontally over the stern down a stinger, which supports and guides the pipe as it enters the water. The pipe is held under tension as it leaves the vessel. The length and curvature of the stinger, as well as the available tension, are key elements in maintaining pipe integrity. The technique allows high lay rate, even for large diameter pipes. Hence, it is the technology of choice in many markets today. The J-lay method (Fig 8) is more suitable for deepwater installation, as the pipeline is deployed vertically into the water forming a ‘J’ curve from the surface to the seabed (Cavicchi and Ardavanis, 2003). Several pipes are pre-assembled into ‘strings’ of pipes, which are then upended in a vertical position in a specially built tower on the laying vessel. Here, they are welded together and lowered at a near vertical angle to the seabed (Pertinet and Frazer, 2003). The main advantages of the J-lay

Fig 8: J-lay Method

Fig 9: Reel-lay Method

method are the reduction of the stress imposed on the pipeline and of the distance to the touchdown point, which brings a corresponding reduction in tensioning requirements for pipeline integrity. Moreover, the J-lay method is more accurate in terms of pipe positioning. The Reel-lay method (Fig 9) is generally the most effective way of installing pipelines, particularly for sizes up to around 16 inches (Smith and Clough, 2010). Long pipe segments are welded, tested and coated onshore and then spooled onto a large, usually vertically oriented pipe reel, in one continuous length. Once the Reel-lay vessel is in position, the pipe is unspooled and, in the case of rigid pipe, straightened and then lowered to the seabed as the vessel moves forward. The major advantages of Reel-lay are the high production rate and the controlled welding and inspection conditions onshore. This makes Reel-lay an extremely efficient method for the installation of pipelines in all water depths.

3. The main challenges and advances in deepwater installation

Fig 7: S-lay Method

Dealing with deep or even ultra-deep water installation means that the weight of subsea structure requires special purpose equipment for holding and lowering. The increasing water depth results in more complexity with associated technological challenges.

86

5-SUT268_34(2)_ut64417.indd 86

01/03/17 2:30 pm


Underwater Technology Vol. 34, No. 2, 2017

3.1. Heave compensation for offshore installation Offshore lifting, lowering and holding a load at sea are difficult tasks, especially in rough seas. The hydrodynamic loads imposed upon operating offshore floating vessels are affected by a strong coupling between the vessel and load motions (Hirdaris et al., 2014). Heave compensation (Fig 10) is used to remove the vessel heave motion from that of the load, resulting in the decoupling of the two motions and, therefore, reduced variation in cable tension (Woodacre et al., 2015). Southerland (1970) first presented a paper outlining the examples of heave compensation systems to alleviate the difficulties in payload handling at sea. Since then, three types of heave compensation system have been developed. These have benefited from computational advances, hydraulic advances and control system advances and are discussed in the following subsections. 3.1.1. Passive heave compensation system The PHC system is designed to maintain a constant line tension and requires no input energy to function. A parallel spring–damper, which acts as a mechanical low-pass filter, is placed in series before the load, in order to reduce or isolate any ensuing motion. It is common to use some form of gasbacked accumulator driven hydraulic piston as a spring for PHC. According to Kidera (1983), many early PHCs suffered from cylinder stick problems, where static friction was too large for the load to move easily. Hatleskog and Dunnigan (2006) considered the PHC dynamics for an oil drilling platform. One of the key conclusions they made was that the PHC cannot reduce heave motion coupling to the load by over 80%. This part-failure was one of the driving forces behind the development of Active Heave Compensation in the 1990s. 3.1.2. Active heave compensation The AHC system involves closed-loop control and requires energy input to adjust for the vessel heave amplitude. A mechanically actuated system was

Input ship motion k

c

Passive heave compensator

Reduced output motion

Load

Fig 10: The schematic of a small vessel using a heave compensation system

patented by Blanchet and Reynolds in 1977 and was packaged for retrofit onto cranes (Blanchet and Reynolds, 1977). Barber (1982) presented a patent in which a fixed circuit design was implemented to control heave motion, but the fixed circuit could not be changed. El-Hawary (1982) presented a double-integrator circuit, which was an improved method of correcting heave using sonar data. Jones and Cherbonnier (1990) presented the first patent with a microprocessor controlled AHC system. Korde (1998) performed an in-depth mathematical evaluation of an AHC system whose accelerometer data were used for position and force feedback. Do and Pan (1998) applied a nonlinear model and control scheme to AHC for a drill ship. Kimiaghalam applied transfer function filters to correct for time/phase lag, which was introduced by the hydraulic system or through slow communication of the control system in their AHC system (Kimiaghalam et al., 2001). 3.1.3. Hybrid active–passive heave compensation The HAPHC system features aspects of both PHC and AHC. Hatleskog and Dunnigan (2006) presented a HAPHC system that combines a passive system to hold the bulk of the load, and an active system to assist in further load motion decoupling from vessel heave. Nicoll et al. (2008) simulated attaching a passive heave compensator near the load, with an active system operating at the surface. This system requires the active system to hold the entire load, and if adjustments are required to the passive system, the load must return to the surface.

3.2. Self-weight of steel wire The properties and performance of conventional steel wire are well known, and there is a clear set of standards and guides for its use in offshore installation applications. However, steel wires are reaching their technological limit owing to both limited availability of raw materials and limited performance because of the self-weight increasing with the water depth. This makes conventional wire rope systems inefficient and impractical on most deepwater installation operations (Torben and Ingeberg, 2007). At 3000 m water depth, the weight of a 5 inch steel wire is about 170 t, which means the equipment capacity will be less. Increasing the size and weight of structures to be installed implies the need for high diameter wires characterised by high strength. An approach could be to perform multi-falls systems aimed at varying and optimising the rope diameter, but it would be a time-consuming procedure. The multiwire and segmented length options have achieved practical application, which needs the qualification

87

5-SUT268_34(2)_ut64417.indd 87

01/03/17 2:30 pm


Yi Wang et al. Advances in deepwater structure installation technologies

Fig 11: Double capstan principle FDRS

of strength capacity and to be rationally based on safety factors (Crawford et al., 2011). Several FRDSs (Fig 11), including synthetic fibre ropes and high modulus polyethylene material (HMPE) ropes, have been designed to tackle typical fibre rope criticalities like overheating, abrasion and cutting caused by bending issues. Some of these systems are based on the single capstan winch principle and the double capstan principle, others on the traction winch principle (Frazer et al., 2005; Torben et al., 2011). FRDSs have the advantages of being essentially neutrally buoyant in sea water, and having a high strength-to-weight ratio, good elongation properties and dynamic toughness. These features eliminate the impact of the self-weight of the wire ropes on the lifting and lowering capacity of a deepwater installation system. The main fibre options for deepwater applications are aramid, polyester and HMPE, while other possible candidate materials are highstrength zylon, vectran and nylon. Fibre ropes are not the best technical solution for shallow-water installation operations, but seem to be the sole solution for frontier scenarios characterised by heavy payloads and ultra-deep waters.

3.3. Hydrodynamic response evaluation technology During a free-falling operation and a submerged towing operation, the dynamic behaviour of the installation structure will depend upon the hydrodynamic loads, which act on the components of the structure. The magnitude of these forces will affect the deflection angle of the deploy wire, and the forces will also set the operational limits for permissible sea states. Hence, it is important that the computational models and procedures are validated before such an operation is carried out. Typically, modern commercial software is used for numerical analysis to evaluate the hydrodynamic

response of a structure during the installation process. When simulating the PIM or WTIM by using modern commercial software, correct estimates of the hydrodynamic coefficients are necessary, especially the added mass and drag coefficients. However, the coefficients for arbitrary shapes of subsea structures are incomplete in commercial software – only coefficients for simple geometrical shapes are readily available. To address this, Fernandes and Mineiro (2007) introduced two methods, the frequency limit method and the constant acceleration method, for assessing the inertial properties of bodies with complex geometries. Another method for hydrodynamic response evaluation is to simplify the 3 degrees of freedom (DOF) structure to 1 DOF. Jacobsen and Leira (2012) presented a simple 1-DOF system for submerged towing of a subsea template by WTIM, which is integrated in the time domain using the Newmark-beta method. In this model, the tension in the towing wire can be estimated by superposition of quasi-static drag forces and dynamic forces. Mirzaeisefat and Fernandes (2013), Fernandes and Armandei (2014), and Fernandes and Mirzaeisefat (2014, 2015), simplified the complex manifold to a hinged flat plate so that the flow-induced rotation in the free-falling process during PIM could be simulated. This is important for installation safety by controlling the oscillatory behaviour of the structure.

3.4. Positioning and control techniques The specialised installation vessels operating in harsh marine environments have to maintain a fixed position in relation to a fixed point on the ocean bed, despite having to contend with the forces of sea currents, waves and wind acting on them. This is because a small change in the position of the vessel may lead to accidents in the sea. Typically, the dynamic positioning (DP) system is used to allow the vessel to maintain position so that the installation process can take place safely and efficiently. DP is a computer-controlled system that maintains the vessel’s position automatically by using its own propellers and thrusters. The Classification Societies have issued rules for dynamic positioned ships, described as DP1, DP2 and DP3 (Desai, 2015). DP1 has no redundancy, which means loss of position may occur in the event of a single fault. DP2 has redundancy so that loss of position should not occur from a single fault of an active component, but may occur after failure of a static component. DP3 also has redundancy so that loss of position should not occur from any single failure, including a completely burnt subdivision or flooded watertight compartment.

88

5-SUT268_34(2)_ut64417.indd 88

01/03/17 2:30 pm


Underwater Technology Vol. 34, No. 2, 2017

The earliest DP systems were designed using conventional proportional–integral–derivative (PID) controllers in a cascade to suppress the wave-induced motion components. A model-based control concept, which exploits stochastic optimal control theory and Kalman filtering techniques, has been employed to address the DP problem by Balchen et al. (1980). Morishita and Souza (2014) proposed a procedure for attenuating the control law based on the observer backstepping methodology. Donaire and Perez (2012) presented a port-Hamiltonian framework to design a nonlinear set-point-regulation controller with integral action. Benetazzo et al. (2015) presented a solution to guarantee a fault-tolerant robust control for the dynamic positioning of an over-actuated offshore supply vessel.

3.5. Protection against large external pressure The protection against large external pressures in ultra-deep water has resulted in new applications involving thick steel walls, higher strength grades and even other concepts for subsea pipelines. A solution to thick steel walls was proposed by Palmer in the early 1990s and is to install a water filled pipeline to remove the load from external pressure during installation. DNV proposed an X-Stream pipeline concept based on this idea (Venås, 2012). It involved a combination of already established fieldproven technological control of the pressure differential during installation and operation through an inverted high integrity pressure protection system with double block and bleed valves. Another novel concept to reduce minimum steel wall thickness was the development of ‘sandwich’ pipes, which consists of two steel pipes with concrete or another suitable material in-between (Castello and Estefen, 2008).

4. General tendencies in future development The increase of field development in deep water and the installation of deepwater structures in colder and harsher environments will bring great challenges for installation technologies.

4.1. Powerful deepwater installation vessels Typical deepwater installation vessels include drill vessels, heavy lift vessels, subsea construction vessels and pipe laying vessels. Installation of new floating production platform topsides such as semi-submerged platforms (SEMIs), tension leg platforms (TLPs) and float production storage offloading (FPSOs) will be installed increasingly either by modular-lifting at the quayside where they ar fabricated onshore or floatover installation in inshore, sheltered waters. Therefore, anticipated new builds are being designed in the lifting range of 1500 to 5000 t with their main

intended usage either for deepwater construction or pipe laying. The new deepwater installation vessels should be equipped with a series of multi-functional deepwater tools so that they can be powerful enough for deepwater field installation of up to 3000 m water depth.

4.2. Installation of subsea factory The subsea factory is seen as vital to realising future opportunities for exploiting resources from greater depths and in colder and harsher environments. This will benefit from the technological breakthrough in subsea processing, including gas-to-liquid or liquidto-liquid separation, gas compression or boosting single or multi-phase and re-injection, together with remote control technology that allows for export to any offshore facility. As field development technology is moving towards the subsea factory, a standalone subsea production system on the seabed implies increasing weights to be installed, as well as more control to meet stringent installation targets. 4.3. Harsh and remote environment As oil and gas resources near coastal areas have been exploited fully, remote locations such as arctic or subarctic environments are being put on the development agenda. In harsh and remote arctic environments, the main challenges to face are the short operating season and weather window, as well as the severe and unpredictable weather and marine conditions. The shortness of the operating season means offshore installation and support operations must be executed with the maximum effectiveness and coordination. Working in severe weather and marine conditions means the installation vessels must be able to resist the severe dynamic loads induced by these conditions. Therefore, the availability of an adequate installation spread, in terms of seakeeping, ice-class and productivity, is just one of the challenging factors to be dealt with.

5. Conclusion This paper gives a comprehensive overview of deepwater structure installation technologies, and categorises various installation methods into two kinds based on the characteristic of installation structure. It describes six installation methods for subsea equipment and three installation methods for pipelines in order to address the important aspects of subsea installation technologies. The main challenges and advances of the deepwater structure installation are also discussed in order to better understand subsea oil and gas industry at the present stage, and evolving treans are considered to provide reference for future development.

89

5-SUT268_34(2)_ut64417.indd 89

01/03/17 2:30 pm


Yi Wang et al. Advances in deepwater structure installation technologies

Acknowledgments This paper is funded by the Key Laboratory of Shandong Province Offshore Petroleum Equipment, Shandong ShengLi Petroleum Equipment Industrial Technology Research Institute (grant no. KRKFJJ-03), National Key Research and Development Plan (Grant no 2016YFC0303700) and Science Foundation of China University of Petroleum, Beijing (2462015YQ0412).

References Bai Y and Bai Q. (2012). Subsea engineering handbook. USA, Gulf Professional Publishing, 956 pp. Bai Y and Bai Q. (2014). Subsea pipeline design, analysis, and installation. USA, Gulf Professional Publishing, 824 pp. Balchen G, Jenssen NA, Mathisen E and Saelid. (1980). A dynamic positioning system based on Kalman filtering and optimal control. Modeling, Identification and Control 1: 135–163. Barber NR (1982). Control means for motion compensation devices. US Patent No. 4349179. Benetazzo F, Ippoliti G, Longhi S and Raspa P. (2015). Advanced control for fault-tolerant dynamic positioning of an offshore supply vessel, Ocean Engineering 106: 472–484. Blanchet JP and Reynolds TJ. (1977). Crane hook heave compensator and method of transferring loads. US Patent, No. 4003472. Castello X and Estefen SF. (2008). Sandwich pipes for ultra deepwater applications. In: Proceedings of the Offshore Technology Conference, 5–8 May, Houston, Texas. Cavicchi M and Ardavanis K. (2003). J-lay installations lessons learned. In: Proceedings of the Offshore Technology Conference, 5–8 May, Houston, Texas. Crawford M, Crawford P and Shand D. (2011). Application and evolution: 17.5 Te, 85 Te and 250 Te to 3000 meters using a drum winch approach, HMPE fiber rope and efficient offshore vessels. In: Proceedings of the Offshore Technology Conference, 2–5 May, Houston, Texas. de Gam Lima JMT, Kuppens ML, Da Silveira PF and Stock PFK. (2008). Development of subsea facilities in the Roncador field (P-52). In: Proceedings of the Offshore Technology Conference, 5–8 May, Houston, Texas. Desai N. (2015). Dynamic positioning: method for disaster prevention and risk management. Procedia Earth and Planetary Science 11: 216–223. Do KD and Pan J. (2008). Nonlinear control of an active heave compensation system. Ocean Engineering 35: 558–571. Donaire A and Perez T. (2012). Dynamic positioning of marine craft using a port-Hamiltonian frame work. Automatica 48: 851–856. El-Hawary F. (1982). Compensation for source heave by use of a Kalman filter. Ocean Engineering 7: 89–96. Fernandes AC and Armandei M. (2014). Phenomenological model for torsional galloping of an elastic flat plate due to hydrodynamic loads. Journal of Hydrodynamics 26: 57–65. Fernandes AC and Mineiro FPS. (2007). Assessment of hydrodynamic properties of bodies with complex shapes. Applied Ocean Research 29: 155–166. Fernandes AC and Mirzaeisefat S. (2015). Flow induced fluttering of a hinged vertical flat plate. Ocean Engineering 95: 134–142.

Frazer I, Perinet D and Vennemann O. (2005). Technology required for the installation of production facilities in 10,000 ft of Water. In: Proceedings of the Offshore Technology Conference, 2–5 May, Houston, Texas. Gernon GO, Kenney TD, Harrison G and Prescott CN. (1995). Installation of deepwater pipelines utilizing S-Lay methods. In: Proceedings of the Offshore Technology Conference, 1–4 May, Houston, Texas. Hatleskog JT and Dunnigan MW (2006). Heave compensation simulation for non-contact operations in deep water. In: Proceedings of the IEEE OCEANS 2006 Conference, 18–21 September, Boston, Massachusetts. Hirdaris SE, Bai W, Dessi D, Ergin A, Gu X, Hermundstad OA, Huijsmans R, Iijima K, Nielsen UD, Parunov J, Fonseca N, Papanikolaou A, Argyriadis K and Incecik A. (2014). Loads for use in the design of ships and offshore structures. Ocean Engineering 78: 131–174. Jacobsen T and Leira BJ. (2012). Numerical and experimental studies of submerged towing of a subsea template. Ocean Engineering 42: 147–154. Jones AB and Cherbonnier TD. (1990). Active reference system. US Patent, No. 4962817. Kidera E. (1983). At-sea handling and motion compensation. In: Proceedings of the IEEE OCEANS 1982 Conference, 29 August–1 September, San Francisco 766–770. Kimiaghalam B, Homaifar A and Sayarrodsari B. (2001). An application of model predictive control for a shipboard crane. Proceedings of the 2001 American Control Conference, 25–27 June, Arlington, USA, 929–93. Korde UA. (1998). Active heave compensation on drill-ships in irregular waves. Ocean Engineering 25: 541–561. Martijn D and David R. (2014). Deepwater development strategy. In: Proceedings of the Offshore Technology Conference, 5–8 May, Houston, Texas. Mirzaeisefat S and Fernandes AC. (2013). Stability analysis of the fluttering and autorotation of flow-induced rotation of a hinged flat plate. Journal of Hydrodynamics, Ser. B 25: 755–762. Moreira JRF and Johansen TE. (2001). Installation of subsea trees in Roncador field at 1800 m water depth using the drill pipe riser. In: Proceedings of the Offshore Technology Conference, 30 April-3 May, Houston, Texas. Moreira JRF, Rovina PS, Couto P and Neumann B. (1999). Development and installation of the drill pipe riser, an innovative deepwater production and completion/ workover riser system. In: Proceedings of the Offshore Technology Conference, 3–6 May, Houston, Texas. Morishita H and Souza C. (2014). Modified observer backstepping controller for a dynamic positioning system, Control Engineering Practice 33: 105–114. Mork H and Lunde J. (2007). A cost-effective and safe method for transportation and installation of subsea structures – the pencil buoy method. In: Proceedings of the SPE Offshore Europe Conference, 4–7 September, Aberdeen, Scotland. Nicoll RS, Buckham BJ, and Driscoll FR. (2008). Optimization of a direct drive active heave compensator. In: Proceedings of the ISOPE 18th International Offshore and Polar Engineering Conference, 6–11 July, Vancouver, Canada, 241–248. Penati L, Ducceschi M, Favi A and Rossin D. (2015). Installation challenges for Ultra-deep water. In: Proceedings of the 12th Offshore Mediterranean Conference and Exhibition, 25–27 March, Ravenna, Italy. Perinet D and Frazer I. (2003). J-Lay and steep S-Lay: complementary tools for ultradeep water. In: Proceedings of

90

5-SUT268_34(2)_ut64417.indd 90

01/03/17 2:30 pm


Underwater Technology Vol. 34, No. 2, 2017

the Offshore Technology Conference, 30 April–3 May, Houston, Texas. Risoey T, Mork H, Johnsgard H and Gramnaes J. (2007). The pencil buoy method – a subsurface transportation and installation method. In: Proceedings of the Offshore Technology Conference, 30 April–3 May, Houston, Texas. Roveri EF, de Oliveira MC and Moretti MJ. (1996). Installation of a production manifold in 2000 ft water depth offshore Brazil. In: Proceedings of the Offshore Technology Conference, 6–9 May, Houston, Texas. Smith SN and Clough AJ. (2010). Deepwater pipeline and riser installation by the reel-lay method. In: Proceedings of the Offshore Technology Conference, 3–6 May, Houston, Texas. Southerland A. (1970). Mechanical systems for ocean engineering. Naval Engineers Journal 82: 63–74. Torben SR and Ingeberg P. (2011). Field pilot of subsea equipment installation in deep water using fibre rope in two-fall arrangement. In: Proceedings of the Offshore Technology Conference, 2–5 May, Houston, Texas.

Torben SR, Ingeberg P, Bunes Ø, Bull S, Paterson J and Davidson D. (2007). Deployment system for ultradeepwater installations. In: Proceedings of the Offshore Technology Conference, 30 April–3 May, Houston, Texas. Venås, Asle (2012). Reduction of deep-water pipeline costs. Ship & Offshore Magazine No. 3, 2012. Wang A, Yang Y, Zhu S, Li H, Xu J, and He M. (2012). Latest progress in deepwater installation technologies. In: Proceedings of the 22nd ISOPE International Offshore and Polar Engineering Conference, 17–22 June, Rhodes, Greece. Wang A, Zhu SH, Zhu XH, Xu J, He M and Zhang C. (2013). Pendulous installation method and its installation analysis for a deepwater manifold in South China Sea. In: Proceedings of the 23rd ISOPE International Offshore and Polar Engineering Conference, 30 June–5 July, Anchorage, Alaska, USA. Woodacre JK, Bauer RJ and Irani RA. (2015). A review of vertical motion heave compensation systems. Ocean Engineering 104: 140–154.

91

5-SUT268_34(2)_ut64417.indd 91

01/03/17 2:30 pm


th

14 AN

A NU

L

MCE Deepwater Development 2017 “ D elivering D eepwater Value, Ef f icienc y & Flexibilit y �

NH Grand Hotel Krasnapolsk y 3 -5 April, 2017 Ams terdam

w w w. MCE DD.com

Organized by

Sponsors:

In Par tnership with

Suppor ted by

Of ficial Media Suppor ter:


doi:10.3723/ut.34.093 Underwater Technology, Vol. 34, No. 2, pp. 93–98, 2017

Technical Paper

www.sut.org

Detecting the Ma‘agan Mikhael B shipwreck Deborah Cvikel*1, Ole Grøn2 and Lars Ole Boldreel2 1 Leon Recanati Institute for Maritime Studies, University of Haifa, Haifa 3498838, Israel 2 Department of Geosciences and Natural Resource Management, University of Copenhagen, Østervoldgade 10, DK 1250 Copenhagen K, Denmark Received 13 April 2016; Accepted 11 October 2016

Abstract A shipwreck, designated as Ma‘agan Mikhael B, was discovered in 2005 by divers about 70 m from the shoreline and at a depth of 1 m, embedded in sandy seafloor sediments. Soon after, the shipwreck was lost in the sand. In May 2015, it was located in a survey with a chirp sub-bottom profiler, and a water-jetting survey in August 2015 confirmed its position. This paper discusses the detection of the Ma‘agan Mikhael B shipwreck using a chirp sub-bottom profiler. Keywords: Chirp sub-bottom profiler, Ma‘agan Mikhael B, shipwreck

1. Introduction Kibbutz Ma‘agan Mikhael is located on the Mediterranean coast of Israel, about 35 km south of Haifa (Fig 1). The coast is straight and exposed, and there is a chain of four small islands 100 m–150 m from the shoreline. The seafloor is shallow and sandy, with a dynamic sediment regime. The site is well known for the discovery of the 400 BC Ma‘agan Mikhael shipwreck in 1985 (Linder, 2003; Kahanov, 2011). The remains of another wooden ship were discovered by two kibbutz members, N. Helfman and Y. Batzir, diving in 2005. They reported the observation of six framing timbers protruding from the sand, ceramic sherds and stones, possibly ballast. However, soon after its initial discovery, the shipwreck was covered with sand and it disappeared, and only an approximate location of the wreck was reported. While side-scanner and multi-beam systems are well-suited for detecting and mapping shipwrecks that lie partly exposed above the seafloor, obtaining * Contact author. Email address: dcvikel@research.haifa.ac.il

detailed information about shipwrecks buried and covered by seafloor sediments requires a different solution. To be able to distinguish shipwrecks located below the sediment surface, as well as the buried parts of shipwrecks with elements of their construction visible above the seafloor, the use of instruments that can penetrate the seafloor sediments, either physical (probes, cores, or water-jetting) or remote sensing systems, such as sub-bottom profilers, are necessary. Remote-sensing has been applied in maritime archaeology since the 1960s. A pioneering and wellknown example of the maritime archaeological application of sub-bottom profilers was Edgerton’s use in 1968 of a ‘mud pinger’ (5 kHz and 12 kHz) to locate the 1545 shipwreck of Mary Rose below the seafloor (McKee, 1973). Throckmorton et al.’s (1973) use of a 5 kHz EG&G ‘penetrating pinger’ in combination with a magnetometer and a side-scan sonar for locating anomalies resulting from the naval battle at Lepanto in Greece, should also be noted as an important early advance. The same is true of Meissner and Stümpel’s (1979) recording of the sediment-embedded Viking ship ‘wreck 1’ in Haithabu harbour prior to its excavation in 1979. Another case dating from 1989 is interesting because it demonstrated that the shape of a medieval-period shipwreck, embedded in seafloor sediments at Sundekilen, Norway, could be successfully recorded with a conventional echosounder (Simrad EA 300P) due to its absorbance of the acoustic signal. The acoustic result confirmed the previous record of the shipwreck’s outline, using probes (Nævestad, 1991). Quinn et al. (1997, 2002) used a chirp system to record shipwrecks embedded in the seafloor sediments – for instance, the

93

6-SUT262_34(2)_ut64414.indd 93

24/02/17 12:43 pm


Cvikel et al. Detecting the Ma‘agan Mikhael B shipwreck

Fig 1: Location map of Ma‘agan Mikhael and the Ma‘agan Mikhael B shipwreck site (N. Yoselevich)

wrecks of Mary Rose and its related scour marks, as well as the French frigate, La Surveillante. More recently, additional advances and improvements have been made in recording shipwrecks and other cultural features embedded in seafloor sediments using chirp-based systems (Grøn and Boldreel, 2014; Grøn et al., 2015; Grøn and Hermand, 2016). In the case of early sub-bottom profilers, with only one transducer used for both emitting and receiving the signal, water depth was a critical factor. The vibration from signal emission had to be halted before the transducer could receive signals. This took some time, which meant that the first part of the reflection from the seafloor could not be received in very shallow water. Normally, about 2 m of water was required between the transducer and the seafloor. Recently, modern systems with separate emitting and receiving transducers have produced reasonably good images with only a few decimetres of water below the fish (Grøn et al., 1998; 2007; Grøn and Boldreel, 2014). It is generally assumed that subbottom systems have trouble penetrating sandy seafloor sediments. However, experience shows that a Teledyne Chirp III, sweeping the frequency interval 2 kHz–23 kHz, penetrates sandy sediments well (Boldreel et al., 2010; Grøn et al., 2015). The present study aims to present the equipment and methodology used to detect and verify the location of an ancient wooden shipwreck in shallow water (later designated the Ma‘agan Mikhael B shipwreck). Experience shows that an interdisciplinary approach is essential in order to obtain useful results (Grøn et al., 2015). Therefore, the procedure involved seismic sub-bottom profiling to map anomalies representing potential archaeological features,

and these were later verified or rejected by diver investigation involving water-jetting. The paper describes the Chirp sub-bottom profiling survey and water-jetting survey conducted to locate the shipwreck, and concludes with a discussion of the results.

2. Chirp sub-bottom survey A compressed high intensity radar pulse (CHIRP) acoustic seismic sub-bottom profiler survey was carried out off the shore of Ma‘agan Mikhael in May 2015 in order to locate the shipwreck observed in 2005. The instrument used was a Teledyne Chirp III, which had previously demonstrated good ability to distinguish archaeological objects with restricted horizontal dimensions (e.g. wooden poles) and wooden shipwrecks in shallow water, even with a sandy seafloor (Grøn and Boldreel, 2014; Grøn et al., 2015). Owing to a westerly wind causing wave action, the recordings were quite disturbed, showing the rather even seafloor as a wavy line on the seismic recordings and distorting embedded horizontal archaeological features accordingly. Because of the hard seafloor and the shallow water, the first multiple also reduced the visibility of archaeological features in the recorded data. Since ideal survey conditions with a calm sea are rare in most places, those conducting surveys generally attempt to: • widen the window for weather conditions under which useful data can be recorded (surface noise reduction, heave compensation, among others); • improve the ability to interpret such sub-bottom data for archaeological features in general, through training.

94

6-SUT262_34(2)_ut64414.indd 94

24/02/17 12:43 pm


Underwater Technology Vol. 34, No. 2, 2017

This involves recording known features and verification or rejection of anomalies interpreted as archaeological features. Data were gathered during 4 hrs of recording in very shallow water, sometimes less than 1 m deep, with the outboard motor almost touching the sand at a speed of about 1 knot. A navigation precision of ±10 cm was obtained with C-Nav calibration of the differential global positioning system (DGPS). As standard procedure, the DGPS antenna was placed exactly above the centre of the fish, which was mounted on one side of the boat resting on a fender. This eliminated the problem of an offset between the DGPS antenna and the transducer – the point at which the data are recorded. It also solved the problem of bubbles from the propulsion disturbing the signal recording and avoided disturbance caused by the noise from the hull. Experience shows that precise and accurate positioning is essential, as it saves a significant amount of expensive diving time when verifying potential archaeological anomalies that have been recorded. It also facilitates precise 3D reconstruction of the features recorded in the seafloor sediments prior to underwater excavation. Data stored by the Teledyne Chirp III system are processed entirely automatically, while interpretation processing is normally restricted to signal amplification (time variable gain). During the recording, signal power and gain are the main factors that vary on the basis of the profiles displayed on the screen. The different signal configurations have very little effect on the data recorded. Data interpretation is

based on direct visual recognition of morphological elements that may represent artificial features, and which deviate from what appear to be geological features. There is wide variation in the density of sediment and other surrounding sediment characteristics, as well as in the embedded wood, depending on its species and degree of preservation. Therefore, it is impossible to produce useful models which permit surveys to take in to account how human artefacts can be distinguished from their surrounding sediments (Grøn and Boldreel, 2014). The type of interpretation employed is based on an ongoing systematisation and testing of the anomalies observed. A realistic result depends on obtaining a high degree of positive verifications of the distinguished anomalies, while it appears unrealistic to reach a percent record of hits. The first interpretation of the recorded data from Ma‘agan Mikhael resulted in the registration of five features as possible sections of shipwrecks (darker lines in Fig 2). Four were located in a concentration ~25 m long N-S (possible shipwreck C), and a fifth (shipwreck B) was located 50 m NW of shipwreck C (Figs 2 and 3). The location of the 30 lines recorded is shown in Fig 3.

3. Results of water-jetting survey Water-jetting is a method used to locate archaeological artefacts beneath the seabed. Water is forced under pressure through a metal pipe (~2.5 cm in diameter), which is pushed into the seabed as deep

Fig 2: Plan of locations of possible sections of shipwrecks. The track numbers appear next to each line. The dark lines are the result of the first interpretation of the seismic profiles, and the lighter lines are the result of the second interpretation after shipwreck B was located. Also shown are possible shipwreck sites C and D. UTM coordinates are in zone 36N

95

6-SUT262_34(2)_ut64414.indd 95

24/02/17 12:43 pm


Cvikel et al. Detecting the Ma‘agan Mikhael B shipwreck

Fig 3: The sub-bottom survey off Ma‘agan Mikhael: track (white line), shipwreck B and possible shipwreck C (based on Google Earth)

as possible. Objects located under the sand, such as fragments of wood or other delicate materials, are driven upwards by the water pressure and emerge. They are then collected underwater by the divers, and are brought to the surface to be later studied and analysed. Therefore, to verify the possibility of a shipwreck, after the seismic sub-bottom survey a water-jetting survey was conducted in August 2015 focusing on shipwreck sites B and C. The team consisted of students and volunteers, assisted by staff of the maritime workshop of the Leon Recanati Institute for Maritime Studies, University of Haifa. There were no significant findings in site C. However, in shipwreck site B the survey exposed wooden fragments, apparently of two different species, an 18-cm-long piece of rope, and a pinecone. A fragment of wood and a sample of the rope were dated by 14C AMS analysis to the range of 650–885 AD, which includes the late Byzantine and early Islamic periods in the region. The site was designated as the Ma‘agan Mikhael B shipwreck, and may potentially provide additional evidence for the construction techniques of ships of the period, which are now a topic of major interest.

The interpretation of seismic sub-bottom recordings made for archaeological purposes requires specialist knowledge about the relevant archaeological features, as well as about how such features appear in sub-bottom recordings. As previously demonstrated (Grøn et al., 1998; 2007; Quinn, 2011), precise position control can be obtained through conscious use of the technological possibilities available. The considerably lower sailing speed for archaeological recordings – best below 1 knot – than for geological recordings (often 3–5 knots) results in further details being observed, which are normally not observable in geological recordings. However, a necessary condition is that the boat attitude is sufficiently stable, otherwise low speed may degrade the received signal. Precise navigation is an important factor for saving expensive diving time in the verification phase, since it is often impossible to determine the precise character of observed potential archaeological anomalies before they have been checked by divers, water-jetting survey or other methods. The water-jetting survey at the Ma‘agan Mikhael sites was carried out after the seismic sub-bottom survey, and focused on sites B and C. Since site B had produced only one possible shipwreck signature in the first interpretation, the seismic data were reinterpreted to extract more information. This resulted in the further registration of three faint possible shipwreck sections, apparently from the same shipwreck (B), at 1 m depth and more than 0.5 m into the fine quartz sediment (Figs 3–6). Furthermore, the data revealed an E-W-orientated series of shipwreck-like features (D) (Figs 2 and 7), which are now interpreted as representing either a former phase of a now submerged bed of a watercourse, or contamination left after the excavations of the 400 BC Ma‘agan Mikhael shipwreck in 1988–1989. Hundreds of sandbags and metal remains were left in and around this excavation (a)

4. Discussion The findings from the Ma‘agan Mikhael sites demonstrate that good results can be obtained in sandy sediments. Compared with diving surveys, the subbottom profilers apparently have a much greater potential for revealing shipwrecks embedded in the seafloor and not visible above it. The present study shows that archaeological analysis, with its focus on detection of small-scale anomalies, differs significantly from geological/geophysical interpretation.

(b)

Fig 4: (a) Section of possible shipwreck C in Line 3 (Fig 2); and (b) the section marked with white line

96

6-SUT262_34(2)_ut64414.indd 96

24/02/17 12:43 pm


Underwater Technology Vol. 34, No. 2, 2017

(a)

(b)

site (Breitstein, 2003). The sections of possible shipwrecks (Figs 4–6) show how difficult the wreck sites are to distinguish in these recordings. For detecting wooden shipwrecks embedded in the seafloor, the Teledyne Chirp III appears to produce reliable results as increasingly more experience is accumulated. The better seismic sub-bottom profilers provide such good resolution that the interpretation of their recordings allows some degree of characterising the buried elements (Quinn, 2011).

5. Conclusions Fig 5: (a) Section of possible shipwreck C in Line 4 (Fig 2); and (b) the section marked with white line

(a)

(b)

Fig 6: (a) Section of verified shipwreck B in Line 15 (Fig 2); and (b) the section marked with white line

(a)

The seismic data produced during the chirp sub-bottom profiler survey presented in this paper indicated three possible shipwreck sites off the shore of Ma‘agan Mikhael. A water-jetting survey of two of them was conducted in order to verify the seismic data, and located the possible remains of a shipwreck at site B. Half of the anomalies distinguished as potential shipwrecks were thus verified as such, which seems to be a good result considering the recording conditions. The chirp data were therefore of great value in identifying and outlining the remains of the shipwreck hidden in the sandy seafloor sediments. With increasing training in interpreting the profiles recorded under different conditions, this method is becoming more efficient for detecting maritime archaeological targets such as shipwrecks and pole structures embedded in the seafloor, even in very shallow water.

Acknowledgments The chirp sub-bottom profiler survey was supported by the Israel Science Foundation (grant no. 1899/12), and conducted with the aid of A Yurman and M Bachar from the maritime workshop of the Leon Recanati Institute for Maritime Studies, and P T Jørgensen of the Department of Geosciences and Natural Resource Management, Copenhagen, to whom the authors are grateful. Thanks are due to A Bar, N Helfman, M Cohen, S Cohen, D Moskovich, I Ogloblin and T Simhaev for their help during the water-jetting survey (IAA permit G-81/2015).

(b)

References

Fig 7: (a) Possible shipwreck section in line 15 through feature D, which is here alternatively interpreted as a section through a former watercourse displaying slight meandering – the northern and southern outer limits of its bed, as well as its latest course, looking like a concave shipwreck feature (Fig 2); and (b) the section marked with white line

Boldreel LO, Kuijpers A, Madsen EB, Hass C, Lindhorst S, Rasmussen R, Nielsen MG, Bartholdy J and Pedersen JBT. (2010). Postglacial sedimentary regime around northern Sylt, South-eastern North Sea, based on shallow seismic profiles. Bulletin of the Geological Society of Denmark 58: 15–27. Breitstein S. (2003). Logistic and Operational Aspects of the Excavation. In: Linder E and Kahanov Y. (eds). The Ma‘agan Mikhael Ship. The Recovery of a 2400-Year-Old Merchantman. Final Report, Volume I. Jerusalem: Israel Exploration Society and University of Haifa, 8–23.

97

6-SUT262_34(2)_ut64414.indd 97

24/02/17 12:43 pm


Cvikel et al. Detecting the Ma‘agan Mikhael B shipwreck

Grøn O and Boldreel LO. (2014). Chirping for Large-Scale Maritime Archaeological Survey: A Strategy Developed from a Practical Experience-Based Approach. Journal of Archaeology 2014. Grøn O and Hermand J-P. (2016). Settlement archaeology under water: practical, strategic and research perspectives. In: Proceedings of the IEEE/OES Acoustics Underwater Geosciences Symposium (RIO Acoustics), 29–31 July 2015, Rio de Janeiro, Brazil. Grøn O, Hoffmann G, Brunn H and Schietzel K. (1998). The Use of Acoustic High Resolution Sub-bottom Profilers for Geo-Archaeological Survey. Results from Jungshoved, Kerteminde Firth and Haithabu/Hedeby. In: Wesse A. (ed). Studien zur Archäologie des Ostseeraumes. Von der Eisenzeit zum Mittelalter. Festschrift Für Michael Müller-Wille. Neumünster: Wachholtz Verlag, 151–160. Grøn O, Nørgård Jørgensen A and Hoffmann G. (2007). Marine Archaeological Survey by High-Resolution SubBottom Profilers. Norsk Sjøfartsmuseums Årbok 2007: 115–144. Grøn O, Boldreel LO, Cvikel D, Kahanov Y, Galili E, Hermand J-P, Nævestad D and Reitan M. (2015). Detection and mapping of shipwrecks embedded in sea-floor sediments. Journal of Archaeological Science: Reports 4: 242–251. Kahanov Y. (2011). Ship reconstruction, documentation, and in situ recording. In: Catambis A, Ford B and Hamilton D. (eds). Oxford Handbook of Maritime Archaeology. Oxford: Oxford University Press, 161–181.

Linder E. (2003). The Discovery and Preliminary Exploration of the site. In: Linder E and Kahanov Y. (eds). The Ma‘agan Mikhael Ship. The Recovery of a 2400-Year-Old Merchantman. Final Report, Volume I. Jerusalem: Israel Exploration Society and University of Haifa, 3–7. McKee A. (1973). The search for King Henry VIII’s ‘Mary Rose’. In: Blackman DJ. (ed). Marine Archaeology. Colston Papers vol. Xxiii. London: Butterworths, 185–202. Meissner R and Stümpel H. (1979). Miniseismik in der Archäeometrie. UMSCHAU in Wissenschaft und Technik 13.1: 412–415. Nævestad D. (1991). Kulturminner Under Vann. Vurdering av de nye tiltak i forvaltningen. Oslo: Norsk Sjøfartsmuseum. Quinn R. (2011) Acoustic Remote Sensing in Maritime Archaeology. In: Catambis A, Ford B and Hamilton D. (eds). Oxford Handbook of Maritime Archaeology. Oxford: Oxford University Press, 68–89. Quinn R, Bull JM, Dix JK and Adams JR. (1997). The Mary Rose site—geophysical evidence for palaeo-scour marks. International Journal of Nautical Archaeology 26.1: 3–16. Quinn R, Breen C, Forsyth W, Barton K, Rooney S and O’Hara D. (2002). Integrated geophysical surveys of the French frigate La Surveillante (1797), Bantry Bay, Co. Cork, Ireland. Journal of Archaeological Science 29.4: 413–422. Throckmorton P, Edgerton HE and Yalouris E. (1973). The Battle of Lepanto search and survey mission (Greece), 1971–1972. International Journal of Nautical Archaeology 2.1: 121–130.

98

6-SUT262_34(2)_ut64414.indd 98

24/02/17 12:43 pm


Extreme Life of the Sea By SR Palumbi and AR Palumbi Published by Princeton University Press

Softcover, 2015 ISBN 9780691169811 240 pages

Extreme Life of the Sea is a book that capitalises on a winning father and son team from the Palumbi family. It is refreshing in many ways in that it is chock full of meticulous sound scientific fact from scientist father Stephen, and is penned by the writer son Anthony in such a way that exposes the reader to what could be described as an onslaught of enjoyable yet complex science without even realising it. The style in which it is written makes it incredibly accessible and frankly a joy to read. All too often, popular science books are polarised between dry scientific writing and over sensationalised popular writing styles, but not here: the balance is every bit as impressive as the examples of the extreme marine life that it documents and brings to life. The book categorises the extremes in a simple manner, such as the earliest, the hottest, the coldest, the oldest and the deepest examples of what the oceans have to offer. Each one introduces the reader to some well-known and some rather obscure organisms intertwined with the often strange behaviours

and impressive, somewhat jawdropping evolutionary adaptations that make them so fascinating. The book starts with a great introduction to life on Earth and introduces things like the Cambrian explosion and the treasure trove of fossils found in the Burgess Shale. This is followed by the archaic era of the trilobites and the nautilus, through to the coelacanths and the organisms we are more familiar with today. The journey into modern oceanic life then starts with bacteria and moves swiftly into some of the deepest lifeforms, introducing those surreal worlds of hydrothermal vents, bioluminescence and deep-sea gigantism. The authors do not let the deep-sea chapter overshadow the next – the shallowest. While explaining habitats such as the coastal, inter-tidal and mangroves, they still manage to portray these as equally bizarre, and indeed extreme, as lesser known habitats, but their familiarity often makes this hard to remember sometimes. Appreciation of the organisms that seem so well known is again challenged in the ‘oldest’ chapter, delivering astonishing facts about whales, turtles, black corals and, of course, the immortal jellyfish. The ‘fastest sprints and longest journeys’ chapter is self-explanatory and really drives home the impression of astonishing feats that occur every day across our oceans. The book does not simply list these, but through some incredibly illustrative writing, really makes the reader think. Again, such writing will make even the most battlehardened marine biologist find a new sense of appreciation of marine life.

www.sut.org

Book Review

doi:10.3723/ut.34.099 Underwater Technology, Vol. 34, No. 2, pp. 99–100, 2017

The ‘hottest’ and the ‘coldest’ chapters bring to life arguably the most impressive adaptation marine life have been forced to make – from surviving life as hydrothermal vent fauna, to the antifreeze in the blood of Antarctic ice fish. Perhaps the most interesting chapter was on future extremes. Wading through what must be forests of new findings and interpretations, the authors tease out a compelling story about what is possibly about to happen on our planet. Having by now read the rest of the book, there is some comfort in knowing that marine life has managed to hit other impending disasters head-on, and whatever the future holds, life will adapt. How it might adapt, and what is left may not be recognisable, but adapt it will – from small-scale physiology to macro-ecological patterns. The chapter serves as a springboard to where marine science should be heading right now and in the foreseeable future – one might say inspiration in the face of adversity. It is also important to add that this work is extensively footnoted and well referenced, leaving the reader in no doubt that the content is contemporary and correct, and a road map of where to find out more should they wish. Furthermore, it is frequently illustrated with impressive colour plates in the centre. Alongside the hard facts and intricate concepts, the authors use many a metaphor and simile to convey often complex concepts in a highly humorous manner. It is refreshing to see writing that is trying, successfully, to convey challenging biological concepts with fun and humour. The result is a book that should be thoroughly recommended to

99

7-SUT263_34(2)_ut64427.indd 99

24/02/17 12:43 pm


SR Palumbi and AR Palumbi. Extreme Life of the Sea

the modern undergraduate, as well as postgraduates and early career researchers, to provide a background and sounding board for scientific inspiration. To be perfectly honest, it should also be read by more senior professionals

in marine biology as a reminder of just how bizarre, surreal, astonishing and often downright weird marine life can be, and why anyone choses this as a career. In a nutshell, it is a great and enjoyable book that

will inspire readers and leave them with an incredible sense of appreciation of the natural world. (Reviewed by Dr Alan Jamieson, Newcastle University)

100

7-SUT263_34(2)_ut64427.indd 100

24/02/17 12:43 pm


UT2 and UT3 The magazines of the Society for Underwater Technology

UT2

Underwater Vehicles Oceanography

UT2

UT2

February March 2015

August September 2015

June July 2015 2105

Subsea Power Distribution Underwater Vehicles Sonar

1

T H E M A G A Z I N E O F T H E S O C I E T Y F O R U N D E R W A T EUT2 R T February E C H N O LMarch O G Y 2015

T H E M A G A Z I N E O F T H E S O C I E T Y F O R U N D E R W A T E R T E CUT2 H N April O L OMay G Y 2015

Subsea Engineering Underwater Vehicles

1

1

T H E M A G A Z I N E O F T H E S O C I E T Y F O R U N D E R W A TUT2 E R August T E C H September N O L O G Y 2015

UT2 covers a focused range of underwater subjects including offshore, marine renewables, subsea engineering, ocean resources, diving and manned submersibles, underwater science and robotics. The magazine is represented at all the many exhibitions around the world at which the Society both co-organises and attends. Furthermore, the magazine is distributed at the many subsea training courses that are organised by the Society, ensuring it reaches tomorrow’s engineers and technologists.

December 2012

UT3 The magazine of the Society for Underwater Technology

UT3

UT3 February March 2014

January 2013

Excavation and Trenching Underwater Intervention

1

UT2 December 2012 mber 2012

Communications Underwater Vehicles

1

T H E M A G A Z I N E O F T H E S O C I E T Y F O R U N D E R WAT E R T E C H N O LO G Y

UT3 January 2013

1

T H E M A G A Z I N E O F T H E S O C I E T Y F O R U N D E R W A T E RUT2 T EFebruary C H N O LMarch O G Y 2014

UT3 is the online magazine of the Society for Underwater Technology, and covers the subsea industry. It consists of the content of the print magazine UT2, greatly expanded with other information.

UT2 and UT3 are available online at http://issuu.com/ut-2_publication

09C-SUT-34(2)-IBC.indd 1

24/02/17 12:20 pm


Showcase your energy solutions at the UK’s largest renewable energy and sustainability event 10-11 May 2017, SECC Glasgow • Meet face to face with the largest group of buyers in renewables • Develop business with 7,000+ visitors • Join 400+ leading suppliers & build your sales pipeline • Make valuable contacts at sector-specific free networking events

Contact the Sales Team today to secure your preferred stand location. +44 (0)20 8439 5560 all-energy@reedexpo.co.uk

Market sectors represented:

Learned Society Patron:

Host city:

In association with:

Organised by:


Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.