K-HN Et Al-1999-Wind Energy PDF

WIND ENERGY
Wind Energ., 2, 25±58 (1999)
Research Towards a Mature Offshore Wind

Article
Energy TechnologyÐGuidelines
from the Opti-OWECS Project
M. KuÈhn,* W. A. A. M. Bierbooms and G. J. W. van Bussel, Institute for Wind Energy,
Delft University of Technology, The Netherlands
T. T. Cockerill and R. Harrison, Renewable Energy Centre, University of Sunderland, UK
M. C. Ferguson, Kvaerner Oil & Gas Ltd, UK
B. GoÈransson, Kvaerner Turbin AB, Sweden
L. A. Harland and J. H. Vugts, Workgroup Oshore Technology,
Delft University of Technology, The Netherlands
R. Wiecherink, Energie Noord West, The Netherlands
Key words: The article reviews the main results of the recent European research project Opti-OWECS
oshore wind (`Structural and Economic Optimisation of Bottom-Mounted Offshore Wind Energy Con-
energy; innovative verters'), which has signi®cantly improved the understanding of the requirements for a
concepts;
economics; large-scale utilization of offshore wind energy. An integrated design approach was
optimization; cost demonstrated for a 300 MW offshore wind farm at a demanding North Sea site. Several
modelling; viable solutions were obtained and one was elaborated to include the design of all major
dynamic components. Simultaneous structural and economic optimization took place during the
behaviour; different design stages. An offshore wind energy converter founded on a soft±soft
operation and
maintenance; monopile was tailored with respect to the distinct characteristics of dynamic wind and
reliability wave loading. The operation and maintenance behaviour of the wind farm was analysed
by Monte Carlo simulations. With an optimized maintenance strategy and suitable
hardware a high availability was achieved. Based upon the experience from the struc-
tural design, cost models for offshore wind farms were developed and linked to a
European database of the offshore wind energy potential. This enabled the ®rst con-
sistent estimate of cost of offshore wind energy for entire European regions. Copyright
*c 1999 John Wiley & Sons, Ltd.
Part I: Background
Introduction
With world attention now focused on the damaging impact of greenhouse gases, wind energy is emerging
as one of the few serious contenders for the large-scale generation of cost-eective, clean energy. The case
for wind energy has been much strengthened in recent years by the signi®cant cost reductions of wind-
generated electricity, together with the substantially improved reliability of modern wind turbines.
There is, however, an important and growing problem which appears to be constraining further
exploitation of wind energy in parts of Europe. Limitations on land use in areas where population density
is high are beginning to hinder the installation of new wind farms. It is conceivable that in some countries
in Northern Europe, public acceptance of onshore wind power projects will reach saturation point in the
*Correspondence to: M. KuÈhn, Institute for Wind Energy, Delft University of Technology, Stevinweg 1, NL-2628 CN Delft, The Netherlands.
Email: ivw@ct.tudelft.nl
Contract/grant sponsor: European Commission.
Contract/grant number: JOR3-CT95-0087.
CCC 1095±4244/99/010025±35 $17.50

Copyright # 1999 John Wiley & Sons, Ltd.
26 M. KuÈhn et al.
Figure 1. Estimated oshore wind energy potential in the European Union 2 (European annual energy consumption
1727 TWh (1989))
future, preventing further land-based development. The exploitation of the huge oshore wind resources,
with considerably less perceived environmental impact than onshore wind farms, will then become crucial
in providing for future energy needs.
Owing at least in part to these factors, attitudes towards oshore wind energy have changed
signi®cantly in the last few years. The European Union's White Paper1 proposes that 24% of electricity
within the European Union should be provided from renewable sources by the year 2010, with possibly
3% provided by wind energy corresponding to an installed capacity of 40 GW. It is unlikely that all this
capacity can be accommodated onshore, but a previous study, supported by the European Joule
programme,2 showed that there are more than sucient oshore wind resources (Figure 1).
Several Northern European countries have ®rm plans for the installation of large oshore wind farms
each with rated capacities in excess of 100 MW, and using technology comparable with the current
generation of megawatt wind turbines. For the more distant future, plans have been developed for
oshore wind farms in the gigawatt range. It is quite conceivable that the installed capacity of oshore
wind power plant may eventually amount to several times that installed on land.
Terminology
In this article a particular terminology is applied which has been developed and used successfully during
the Opti-OWECS project (see Appendix). In order to avoid misunderstandings, there are two conventions
that must be appreciated. Firstly, the acronym ÒWECS' (standing for Oshore Wind Energy Conversion
System) and its synonym òshore wind farms' describe the entire system, i.e. the wind turbines, the
support structures, the grid connection up to the public grid and any infrastructure for operation and
maintenance. Secondly, ÒWEC' (Oshore Wind Energy Converter) is used to refer to a single unit of an
oshore wind farm comprising support structure (i.e. tower and foundation) and the wind turbine
(i.e. aero±mechanical±electrical conversion unit on top of the tower).
Opti-OWECS Project in a Nutshell

The content of this article is based upon results produced during the European research project `Struc-
tural and Economic Optimisation of Bottom-Mounted Oshore Wind Energy Converters' (Opti-
OWECS, JOR3-CT95-0087) lasting from January 1996 to December 1997.
The particular mission of the Opti-OWECS project was to extend the state-of-the-art, to determine
required methods and to demonstrate practical solutions which signi®cantly reduce the electricity cost.
Copyright # 1999 John Wiley & Sons, Ltd. Wind Energ., 2, 25±58 (1999)
Towards a Mature Oshore Wind Energy Technology 27
This will facilitate the commercial exploitation of true oshore sites in a medium time scale of 5±10 years
from now.
The speci®c objectives included:
. a cost estimate and comparison of oshore wind energy converters of dierent sizes and dierent design
concepts;
. an estimate of the cost per kWh of oshore wind energy at sites in dierent regions of the European
Union;
. development of methods for the simultaneous structural and economic optimization of oshore wind
energy converters with due consideration of the site characteristics;
. speci®cation of at least one typical design solution for a bottom-mounted oshore wind energy
conversion system (OWECS).
Achievement of these goals was only possible owing to an international co-operation of industrial
engineers and researchers from the ®elds of wind energy, oshore technology and power management.
This article begins by undertaking a historical survey of developments in oshore wind energy over the
last three decades in order to place the need for the work in context. Results from the project will then be
reviewed in six generic areas:
. consideration of oshore wind farms on a system level;
. development of practical and economic solutions;
. identi®cation of cost drivers for the design solution;
. highlighting the importance of operation and maintenance aspects;
. development of analysis tools for oshore wind farms;
. consistent estimate of energy costs for entire European regions.
For more comprehensive information, reference should be made to the ®nal project reports.3±8
Review of the Development of Offshore Wind Energy to Date

In order to understand both the motivation behind and the diculties associated with the development of
oshore wind energy, a historical review is worthwhile. Emphasis will be placed on two aspects: (i) the
economics of projects and studies and (ii) the design approach adopted.
Economics of Recent Projects and Studies

The utilization of oshore wind power for energy production was proposed for the ®rst time during the
early 1970s9 and considered further by a large number of paper studies during the following two decades.
However, practical demonstrations were hindered both by the expectation of high investment and energy
costs and by the fact that no reliable and suciently large wind energy converters were available on the
market.
Progress with land-based wind energy technology allowed the ®rst, small-scale, oshore wind farms
rated 2±5 MW to be installed in sheltered waters in Northern Europe between 1991 and 1995. While these
prototypes continue to operate well, their economics are often poor and direct extrapolation of such
technologies to a larger scale does not produce a promising prospect. An often stated expectation,
therefore, is that large plant comprised of (multi-)megawatt turbine units will reduce costs and facilitate
exploitation of the enormous potential of oshore wind energy in Europe.
Comparison of a number of studies and projects (Figure 2, Table I) from the 1990s shows a signi®cant
improvement in the costs of electricity generated by oshore wind power with the progress of time. Firstly,
key data of four existing plants all employing turbines of the 500 kW class at sheltered locations are given,
i.e. Vindeby,10 Lely,11 Tunù Knob12 and Bockstigen-Valar (Gotland) ( personal communication with G.
Niklasson, Vindcompaniet, 28 January 1998). Next, some older studies are compared, i.e. the Phase CII13
and the Blekinge study,14 both considering 3 MW machines, and the RES (Renewable Energy System)
Copyright # 1999 John Wiley & Sons, Ltd.
28
Table I. Comparison of energy costs between dierent studies and projects
Study (S)/ No. units Distance
Name of project or study project (P), unit capacity Vhub from shore Water Spec. cost Capacity Energy costsa
and site year (MW) (m s ÿ1) Hhub (m) (km) depth (m) (ECU/kW) factor (ECUct/kWh)
Phase CII, North Sea, UK S'91 711 3 8.3 55 16±21 1900 19% 13
Blekinge, Baltic, SE S'91 98 3 9.0 90 10 15±20 3000 32% 9.1
Vindeby, Baltic, DK P'91 11 0.45 7.5 37.5 1.5 3±5 2150 27% 8.5
RES, North Sea, UK S'93 41 0.4 7.4 33 5 12 4500 33% 16
Lely, IJsselmeer, NL P'94 4 0.5 7.7 41.5 1 5±10 1700 22% 8.3
SK Power, Baltic, DK S'94 180 1 8.2 47 17 8±10 1900 31% 6.7
Tunù Knob, Baltic, DK P'95 10 0.5 7.5 43 6 3±5 2200 34% 6.6
Thyssen, Baltic, DE S'95 140 1.5 7.8 60 4 5±10 1400 27% 6.6
BMFT, Baltic, DE S'95 100 1.2 7.5 60 7 10 1250 31% 5.1
Horns Rev, North Sea, DK S'97 80 1.5 9.2 55 15 5±11 1650 40% 4.9
Scroby Sands, North Sea, UK S'97 25 1.5 8.2 3 1150 31% 4.5
Bockstigen-Valar, Baltic, SE P'97 5 0.55 8 41.5 4 6 1500 33% 4.9
Nearshore, North Sea, NL S'97 100 1 9 60 8 13±17 1900 34% 6.4
Opti-OWECS, North Sea, NL S'98 100 3 8.4 60 11.4 14±19 1250 30% 5.1
9 34% 4.4
a
Energy costs for 20 years loan and 5% real interest rate, no in¯ation, recent exchange rates.
Wind Energ., 2, 25±58 (1999)
M. KuÈhn et al.
Figure 2. Comparison of energy costs between dierent studies and projects
study15 referring to 400 kW units. Finally, attention is directed to seven recent studies based upon wind
turbines in the megawatt or multi-megawatt class, i.e. the SK Power study,16 the Thyssen study,17 the
German BMFT study,18 the proposed project Scroby Sands,19 the prime location `Horns Rev' of the
Danish Plan of Action for Oshore Wind Energy,20 the Dutch Nearshore study21 and the Opti-OWECS
study3 discussed in detail in this article.
Economic comparison of the dierent studies and projects is dicult and possible only at a more or less
qualitative level. Besides the inherent imprecision of paper-based studies, any comparison is complicated
by the presence of signi®cant dierences in price level, economic parameters, exchange rates, environ-
mental and technical conditions. To facilitate rational assessment, a number of speci®c costs and the
energy yield are provided in addition to the raw energy costs. Levelized production costs are all based on a
repayment period of 20 years and a real interest rate of 5% regardless of the economic parameters used in
the original studies.
Despite all the uncertainties, some trends can be seen in the projects and studies:
. A dramatic economic improvement can be observed between, on the one hand, the old studies carried
out in the 1970s and 1980s, of which the Phase CII is representative, and, on the other hand, the small-
scale prototypes and studies from the mid- to late 1990s.
. A `learning curve' can be seen in operation amongst the small-scale prototypes and the latest studies,
thanks to the improving maturity of the technology and increase in the size of both the wind turbines
and the entire wind farms.
. Oshore wind energy, as with wind energy on land, is approaching the cost level of other energy
sources. For instance, typical energy costs based on 5% real interest rate and a repayment period of
20 years of coal- and gas-®red plant range in the order of 3.7±5.5 and 3.1± 4 ECUct/kWh respectively.19
Design Practice in Wind Energy and Oshore Technology

The design of oshore wind farms should be based, of course, on the accumulated experience of both the
wind energy and oshore technology communities. However, one has to be aware of the dierent
backgrounds and sometimes contradictory principles adopted by these engineering ®elds.
Wind turbine engineering has undergone a rapid development in the last 10 years and is now `coming
of age'.22 During this period the average size of commercial wind energy converters increased by a factor
30 M. KuÈhn et al.
of 10 owing to both better understanding of the technology and demands of the market. The price±
performance ratio of wind energy converters improved signi®cantly. One design philosophy of increasing
interest for the industry includes the introduction of more structural ¯exibility in important components
such as blades, the drive train and the tower. This can reduce dynamic loading and hence also weights and
costs.
Most manufacturers produce between two and four dierent standard machines, which are often
optimized for dierent sets of site conditions and wind regimes. Modi®cations of these base line machines,
to tailor them to speci®c locations, take account of two site parameters only. The average wind speed
in¯uences the fatigue loading and determines the rotor diameter and hub height, while the extreme wind
speed is an important factor in the strength analysis. Dierences in onshore soil conditions are com-
pensated for by making minor variations in the foundation design rather than by using dierent towers.
Wind farm design is mainly driven by the selection of a suitable machine type and size from the range of
standard designs, compatibility with the existing grid infrastructure and noise limitations. Operation and
maintenance aspects are important in order to ensure sucient lifetime and minimize repair costs.
In oshore technology there is great experience with the design of large and unique ®xed structures for
the petroleum industries. These are usually built `®t for purpose' with respect to their particular site and
function. The in¯uence of dynamic response due to wave loading is generally limited by relatively high
structural stiness. Although fatigue is important, it generally takes second place with respect to the
dominant extreme event loading conditions.
Transportation and installation issues are often a main design driver in the oshore industry, since
these costs can be higher than those for the manufacturing of the structure onshore. Reduction and where
possible elimination of underwater inspection and maintenance is essential owing to the dicult access
and the high costs associated with such operations oshore. Other important design aspects include the
safety of personnel working on or travelling to the structures, the environmental impact and provisions to
be made for the removal/dismantling of the structure.
Current Design Approaches for Oshore Wind Farms

The ®rst prototype oshore wind farms and many of the proposed designs within other studies (e.g.
References 16, 18, 20 and 21) employed design approaches which can broadly be categorized as either
`robust and traditional' or `parallel'. In this subsection some aspects from existing plants are given as an
illustration rather than as judgement on the installations themselves.
The robust and traditional approach is organized as two clusters of activity according to the parent
technologies. In other words, the necessary wind energy technology (comprising wind turbine, tower,
farm layout, grid connection design) and oshore technology (related to foundation, marine installation
and submarine cables) are developed using the methods conventionally employed by their respective
industries. The treatment is described as `robust' since conventional design solutions are used (e.g.
standard wind turbine with small modi®cations, sti foundation) and `traditional' because of the direct
application of experience from the normal scope of work of the parties involved.
The second Danish oshore wind farm at Tunù Knob (1995)12 is a good example of the type of design
produced by this approach. To a lesser extent this is also true for the earlier farm at Vindeby (1991).10 In
both cases the attribute `traditional' refers to an onshore design approach being `stretched' to apply to the
oshore situation.
The main objectives of both these projects were to demonstrate the practicality of oshore wind energy,
to investigate any environmental eects and to simply obtain experience rather than produce farms with
good economic performance. As a result, more or less standard onshore wind turbines were employed.
The turbines and towers were installed, in a manner comparable with that used onshore (i.e. separate
lifting of tower, nacelle and rotor), on a specially designed, sti caisson that acts eectively as a small
arti®cial island. Minimization of operation and maintenance costs was aimed for by using well-proven
onshore wind turbine designs marinized with features such as improved corrosion protection, airtight
nacelle and built-in lifting facilities.
In the parallel design approach, overall OWECS design goals are recognized but are only considered at
the individual subsystem level for the wind turbine, support structure, grid connection, etc. Thus the
overall performance is limited to the sum of the separate optimizations of the subsystems. The Dutch pilot
project Lely in the IJsselmeer (1994)11 and the world's ®rst OWEC, the Swedish Nogersund plant (1990),23
can be mentioned as examples of this approach.
At Lely, two promising novel design solutions were applied, ®rstly in the form of a monopile founda-
tion, which supports a standard onshore tower, and secondly in the cable-laying technique, which reduces
the need for a cable-laying ship. Nonetheless, many aspects of the OWECS, e.g. the overall structural
dynamics,24 were not fully considered. This was acceptable because of the sheltered site and the demon-
stration character of the project.
At Nogersund the support structure design, intended as a downscaled prototype for a larger OWEC,
and the installation procedure were adopted directly from onshore procedures to oshore siting. The
entire unit was fully assembled and commissioned prior to towing to the ®nal destination, so that in situ
work was minimized.
Both these design approaches match the objectives of demonstration or near-commercial plant but are
not suitable for the commercial, large oshore wind farms that will be erected by the beginning of the new
millennium. For this purpose a so-called ìntegrated OWECS design approach' is proposed, which will be
discussed in the next section.
Part II: Design Approach and Design Solution

Consideration of Offshore Wind Farms on a System Level
One aspect that provided a kind of `philosophy' for the Opti-OWECS project should be highlighted
before the discussion of any other issues. Speci®cally, it is the opinion of the authors, based on their
experience in this project, that optimum design is facilitated by the consideration of oshore wind farms
on an integral system level.
The rationale underlying this approach is that only the OWECS as one entire system can provide electric
power in a reliable and cost-ecient way over the projected lifetime. Four objectives for an optimum
OWECS design can be stated which are related to the nature of such a system:
. optimum distribution of investment and operation and maintenance (O&M) costs over the entire
OWECS and its lifetime;
. high reliability of OWECS as a whole and of essential subsystems;
. adaptation to economy of scale and partial redundancy of single OWEC units;
. symbioses of experience from wind energy and oshore technology.
Two examples from the Opti-OWECS project illustrate how these objectives can be met during the design
process.
Development and Demonstration of an Integrated Design Approach

During the project a methodology, the so-called ìntegrated OWECS design approach', was compiled in
order to meet the challenge of designing cost-ecient and reliable oshore wind energy conversion
systems.4
The integrated design approach considers the components of an oshore wind farm as parts of an entire
system, i.e. the OWECS. Interactions between subsystems are considered, therefore, in as complete and
practical a form as possible, so that the design solution is governed by overall criteria (i.e. so-called aspect-
systems) such as levelized production costs, adaptation to the actual site conditions, dynamics of the
entire system, structural reliability considerations, installation and commissioning eort, as well as
OWECS availability. Although such an approach will most likely produce an innovative design solution,
the evolutionary nature of technical progress and the commercial risks inherent to innovative solutions
32 M. KuÈhn et al.
Figure 3. Multi-level control system within the integrated OWECS design approach
must be kept in mind. In particular, it would be a serious mistake to interpret it as a justi®cation for the
application of unproven wind turbine design to the harsh oshore conditions.
Figure 3 illustrates the multi-level control system* of the integrated OWECS design approach within the
framework of a more general methodology, which was developed independently by other researchers in
the ®eld of large, complex civil engineering projects.25
On the one hand, the organization of the design of such systems requires a decomposition of the system
with respect to its requirements. The interrelations between the bullets in the upper part of Figure 3
represent the friction between problem and solution. Clustering of requirements, related to the system as a
whole rather than certain components, results in aspect-systems. They facilitate eective goal control by
the project manager.
On the other hand, the organization of the structural design is improved by a decomposition of the
system with respect to subsystems (design clusters) and successively into elements. In theory the sub-
systems should be governed by maximum internal and minimum external interrelations. However,
practical considerations such as the involved disciplines, organizations, materials, etc. are often of equal
or greater importance. Cluster leaders, located at subsystem level, carry out the control of the actual
design work that takes place on the engineering level.
For a less abstract representation of the above approach the reader is referred to a latter part of the
article. In that section a detailed description is given of the ®rst application of the methodology during the
development of a typical design solution for an oshore wind farm.
Overall Dynamics of Oshore Wind Energy Converters

The integrated approach can be demonstrated in the analysis of the overall dynamic behaviour of oshore
wind energy converters (OWEC). Overall dynamics should be considered from the outset of the design
*The used technical terms àspect-systems' and `multi-level control system' are quoted from Reference 25.
Figure 4. Reduction of calculated fatigue loading due to integrated modelling of wind and wave response (right) with
respect to superposition of separate models (left and middle)
work. Obviously the complexity and the extension of the applied models and investigations will be limited
at the beginning and can be more sophisticated near the end, as indicated by the following examples.
During the feasibility study a check is recommended on the compatibility of support structure and wind
turbine concepts with respect to dynamic characteristics, aerodynamic damping and global fatigue loads.
Also the compatibility of the support structure concepts and the site conditions, including water depth,
hydrodynamic fatigue and extreme load characteristics and ice loading, should be checked. The next
conceptual phase is suited for sensitivity analyses with respect to the soil conditions, the assessment of the
ratio between wind and wave response, parameter studies on combined fatigue and simultaneous
optimization of wind turbine aspects (e.g. rotor speed, blade and machinery layout and rotor diameter)
and support structure concepts (e.g. stiness, hub height). Because of the substantial computational
eorts involved in time domain analyses, rapid methods in the frequency domain are of great advantage
here. During the detail design phase, sucient data should be available for a detailed dynamic analysis
with an integrated OWEC model in the time domain which fully takes into account interactions between
subsystems and the non-linearities in the wind turbine part. Further aspects to be considered include the
®ne tuning of the design and investigations of the variability of site parameters within the wind farm.
As an illustration of the possible bene®ts, Figure 4 compares the eect of treating the OWEC with two
separate models for the wind turbine and the support structure with one integrated model comprising
both subsystems. The diagram shows the result of a fatigue analysis of a monopile support structure
suering simultaneous aerodynamic and hydrodynamic loading. So far, little is known about how to
combine the results of separate analyses in a rigorous manner. A simplistic, clearly conservative but safe
method for combining the results of separate analyses would be the in-phase superposition of damage
equivalent loads due to wind and waves. This approach is far from adequate and does not account for the
interaction of the wind turbine and support structure, neglecting the aerodynamic damping developed by
the rotor dependent on both operational states and wind conditions.
In contrast, the simultaneous wind and wave response of the integrated model is signi®cantly lower
because of two particular features. Firstly, proper consideration of the aerodynamic damping brings a
lower wave response. Secondly, the partial cancellation of the wind and wave response is accounted for by
the fact that the actual phase relation between both load components has been considered.
A further treatment of this complex subject is beyond the scope of this article and the reader is referred
to Reference 24 for full details.
Development of Practical and Economic Solutions

The integrated design approach, as described in the previous section, and OWEC design tools, developed
and re®ned during the Opti-OWECS project, were applied to the design of a typical oshore wind farm.
34 M. KuÈhn et al.
The practical solutions presented in this section are based upon technology which is currently available or
which will become available in the near future.
Rationale
An OWECS design solution was developed against the background of three objectives. Firstly, the
improved understanding of the principles underlying the design of OWECS gained during the course of
the project was to be demonstrated by a practical solution. Application of promising innovations for
large-scale utilization, e.g. novel installation methods, consideration of operation and maintenance
aspects, integrated design approach, etc., was for this research project more important than achievement
of the absolute economic optimum. Secondly, during the design process, areas of remaining poor
understanding were to be identi®ed and appropriate solutions were to be developed. Finally, the economic
feasibility of large OWECS was to be investigated.
The integrated OWECS design approach concerns the total design trajectory. Consequently, not only
the ®nal result is described, but also the way in which it was achieved.
Project Identi®cation
The initial phase of the project identi®cation comprised three aspects:
. establishment of the project groupÐan international co-operation of industrial engineers and
researchers from the wind energy ®eld, oshore technology and power management;
. determination of the project conditions (i.e. objectives and work programme);
. formulation of the particular design conditions, e.g. wind turbine size of 3 MW or larger, neglect of
onshore grid connection aspects, etc.
Feasibility Study
During the ®rst step of the feasibility study a broad inventory of all relevant aspects and concepts was
made and preselections for the conceptual designs were identi®ed. Furthermore, a terminology appro-
priate to OWECS was established in order to promote smooth communications (see Appendix).
The identi®cation of seven distinct reference sites in Northern European waters was carried out in
parallel with further investigation of the subsystem concepts and the other essential features of overall
dynamics and operation and maintenance. The water depth of the considered sites ranged from 8 to 25 m.
The majority of the sites can be characterized as real oshore rather than inshore or nearshore sites, as is
the case for existing oshore wind farms.
Based upon a qualitative OWECS evaluation, the following subsystem concepts were selected for
further development:
. The 3 MW, 80 m diameter NaÈsudden development line was chosen as reference turbine since it is the
only one in the multi-megawatt league with a reasonable operational track record. A recent design
study26 demonstrated that a commercial cost level can be achieved even for turbines of this size. In
addition, the turbine rating of about 3 MW extends the state-of-the-art of other oshore projects
currently under development. Another justi®cation for this choice was the availability of the entire set
of design data and the industrial interests within the project consortium.
. two wind turbine concepts (gearedÐ®xed speed; direct driveÐvariable speed);
. rotor variants with diameters between 80 and 100 m and dierent rotor speeds;
. distinctly dierent combinations of support structure con®guration and installation procedure
(Figure 5), dynamic characteristics (i.e. sti±sti, soft±sti, soft±soft) and site (i.e. North Sea, Baltic
Sea);
. state-of-the-art options for grid connection and wind farm layout.
Figure 5. Examples of support structure concepts (GBS gravity based system)
Conceptual Design
The conceptual design phase was carried out mainly in parallel with work on subsystem development and
with the extension of the OWECS tools for cost modelling, simulation of the operation and maintenance
behaviour, structural reliability considerations and overall dynamics.
Improved knowledge of particular OWEC aspects gained during this phase led to the consideration of a
monopile support structure, in addition to the gravity-based monotower and the gravity-based lattice
tower (lower row in Figure 5). The extreme design wave together with a relatively low water depth resulted
in a very signi®cant heave force on the gravity-based monotower, which required an expensive amount of
ballast. Furthermore, both the soft±sti monotower and the sti±sti lattice tower suered considerable
aerodynamic fatigue loads owing to their fundamental eigenfrequency being close to the blade-passing
excitation frequency. In particular, the stress concentration at the tubular joints and the sti character-
istics of the lattice design, chosen well beyond the range of dynamic wave excitation at the exposed North
Sea site, showed con¯icts with the dynamic wind turbine loading.
Several results were directly related to the integrated approach:
. integrated development of support structure concepts and installation procedure;

. simultaneous optimization of wind turbine (rotor speed, blade layout) and support structure (i.e.
structure stiness) with the main goal of reduction of aerodynamic fatigue loads;
. consideration of overall dynamics of OWEC in the support structure design;
. development of optimal O&M strategies based on Monte Carlo simulations;
. development of structural reliability analysis for an OWEC support structure.
Next, the cost model developed elsewhere during the project was used to evaluate dierent oshore
wind farms assembled from the subsystem concepts developed during the conceptual design, for the seven
preselected sites from the feasibility study. For the same OWECS concept the energy costs between the
seven compared sites, which were all considered initially as promising, diered by a factor of up to 1.6.
Furthermore, a reduction of the energy cost of about 20% was found by application of a support
structure of the monopile variety instead of a gravity-based monotower type.
The economic performance together with some other criteria led to the selection of the ®nal OWECS
concept and the related site. Table II describes the chosen con®guration.
36 M. KuÈhn et al.
Table II. Main data of the design solution

Main design data
Farm capacity 300 MW, i.e. 100 times 3MW
Wind turbine Kvaerner Turbin WTS 80M (3 MW, 80 m)
Support structure Soft±soft monopile
Oshore grid connection AC submarine cables 24/150 kV
Array eciency 93% (uniform spacing 10D)
Transmission eciency 96%
Availability 96.5%
Net annual energy yield 787 GWh/year
Site data
Location Dutch North Sea, near IJmuiden
Assumed annual wind speed (60 m) 8.4 m s ÿ1 (A 9.5 m s ÿ1, k 2.2)
Distance from shore 11.4±18.6 km (15 km from centre)
Water depth 14±19 m (LAT)
Economic data
Wind turbine cost 170 MECU
Support structure and OWEC installation cost 118 MECU (incl. 25 MECU installation cost)
Oshore grid connection cost 77 MECU (incl. 47 MECU power collection cost)
Project management cost 2% of total capital cost
Total capital costs 372 MECU (1240 ECU/kW)
Operation and maintenance cost 9 MECU/year
Decommissioning cost 10% of initial capital
Economic lifetime 20 years
Real interest rate 5%
Levelized production cost (LPC) 5.1 ECU ct/kWh (4.4 ECUct/kWh for 9 m s ÿ1 at 60 m)
Detail Design
During the detail design phase the selected concept was developed further and interactions between the
subsystems were fully considered.
This integration facilitated several innovations, which represent signi®cant steps towards a mature
oshore wind energy technology:
. An appreciably high oshore farm availability of 96.5% was achieved through a rational approach,
comprising a gradual improvement of the turbine's reliability and maintainability with respect to the
current onshore state-of-the-art and an innovative operation and maintenance solution. The latter
included both an optimized operation and maintenance strategy and a cost-ecient solution to the
`cranage problem' in the form of a permanently and quickly available self-propelled modi®ed jack-up.
The section on operation and maintenance contains more details of these solutions.
. Close co-operation between structural design and dynamic simulations of the OWEC facilitated a soft±
soft monopile design even for 20 m water depth (LAT) in a demanding southern North Sea environ-
ment with ®rm soil conditions. Balancing the combined aerodynamic and hydrodynamic fatigue loads,
reduction of the water-piercing cross-section and consideration of weight and installation issues were
the main reasons for this choice. This is remarkable because the fundamental eigenfrequency of 0.29 Hz
lies well below the rotor frequency of 0.37 Hz but just in the region of signi®cant dynamic wave
excitation extending between 0.04 and 0.5±1 Hz.
. Consideration of tower and foundation as one system resulted in a more appropriate and economic
design than the separate treatment of tower and foundation.
. Signi®cant cost reductions were possible by innovative installation methods for gravity-based systems
which were ¯oated and towed as entire OWEC unit to their ®nal destination. Conceptual solutions have
been developed for both ¯oating and lifting installation of monopile structures. The variety of oshore
equipment available, including jack-ups, crane vessels, barges, etc., opens up several opportunities
Figure 6. Innovative features of the design solution for a 300 MW oshore wind farm
which have to be judged case by case. Also, open questions remain, e.g. on the sea fastening for the
vertical transportation of the tower and wind turbine unit.
. The aerodynamic eciency of the wind farm, the cable costs of the grid connection and the space
requirement of the OWECS were well balanced.
Placement of the OWEC transformer in the nacelle was found to be optimal after consideration of wind
turbine, support structure, grid connection and maintenance aspects. Most important for this choice was
that expensive modi®cations of either the tower diameter or the access platform are avoided and that the
dynamic characteristics remain nearly unchanged.
It is worth noting that neither during the detail design phase nor after the ®nal evaluation of the design
solution were major revisions of the design required. The main reason for this was that the conceptual
design had already been carefully examined with respect to technical feasibility and economic
performance.
The innovative features of the design solution are illustrated in detail in Figure 6.
Economic Performance of Design Solution

The detail design phase was concluded with an economic analysis and parameter study on important cost
drivers that con®rmed the viability of the solution. Costing was based on the actual design speci®cation
and was independent from the OWECS cost models discussed elsewhere in this article. An overview of
main data is given in Table II.
Based upon economic parameters usually applied by public sector utilities, i.e. 20 years loan and 5%
real interest rate,20 the levelized production costs were determined at 5.1 ECUct/kWh. This energy cost is
related to a rather conservative estimate of the average wind speed of 8.4 m s ÿ1 at 60 m height. With a
38 M. KuÈhn et al.
higher wind speed of 9 m s ÿ1, as adopted by other recent studies (e.g. References 2 and 21) for very similar
sites, the energy costs are signi®cantly lower, being reduced to approximately 4.4 ECUct/kWh.
For the chosen site, onshore grid connection costs, which have not been considered so far, would
account for only relatively small additional costs, at the level of a few per cent of the energy cost, so long as
no major grid reinforcements have to be undertaken.
It can be concluded that oshore wind energy is fully viable, and energy costs of the order of 5 ECUct/
kWh (20 years loan, 5% real interest rate) are achievable at sites with good wind conditions and if high
wind farm availability is ensured. The main cost drivers of oshore wind energy will be dealt with in the
next section.
Identi®cation of Cost Drivers for the Design Solution

Generic Analysis of Costs Associated with OWECS
Collation of data available in the literature shows that the major cost associated with all existing oshore
wind energy conversion system projects and studies is the expense of the initial investment required to
establish a wind farm. This cost itself has three approximately equal components: the turbine machinery,
the support structure and the electrical equipment/grid connection. Parameters related to the site
in¯uence the precise contribution each of these makes to the overall cost. The distance to shore has the
greatest eect on the investment, with the turbine becoming the increasingly dominant expense the closer
the farm is to the shore. Both the water depth and the design wave height at any location have a less
signi®cant, but still important, in¯uence on the costs. Considering the energy cost, the annual wind speed
shows the most signi®cant eect compared with all other site parameters.
Aside from the initial investment, the next largest expense confronting the owner of an OWECS is the
cost of its operation and maintenance (O&M). Indeed, earlier studies27 have shown that operation and
maintenance could account for as much as one-third of the cost of the electricity produced. Furthermore,
there appears to be a strong linear relationship between the ®nancial burden of O&M and the resulting
electricity price, such that even relatively small changes in O&M costs could have a substantial in¯uence
on the economic viability of an OWECS project.
Cost Breakdown and Parameter Study for the Opti-OWECS Design Solution
The cost breakdown of the energy cost for the Opti-OWECS design solution is shown in Figure 7. As with
earlier studies, operation and maintenance costs were found to make a signi®cant contribution. The
Figure 7. Contributions to energy costs of Opti-OWECS design solution (annual mean wind speed 8.4 m s ÿ1, 20 years
loan, 5% real interest rate)
origins of this result were traced to the discovery that optimum energy costs required high OWECS
availability, which could in turn only be achieved with a continual high O&M eort and permanent
disposal of heavy maintenance equipment.
The energy cost exhibited a perhaps surprisingly high sensitivity to changes in the energy yield
(Figure 8). This has a number of consequences. Firstly, excellent wind conditions are essential for the
economic viability of OWECS. Secondly, great care must be exercised in estimating long-term wind
conditions if reliable energy cost predictions are to be made. Thirdly, only if high availability is
guaranteed can the full potential of exposed sites be exploited. Finally, it would be economically worth-
while to invest a relatively large amount of capital in producing a reliable OWECS with a high energy
output. For instance, an increase in energy production of only 10% would bring economic improvements
even if it required an increase in turbine costs of up to 30%.
The relation between the average wind speed and the levelized production cost (Figure 9) demonstrates
once again the paramount importance of operation and maintenance. The curve relating the two ¯attens,
or even turns upward, at high wind speeds because of the lower availability associated with more
demanding environmental conditions.
Figure 8. Sensitivity of normalized energy cost on (isolated) variation of subsystem cost, O&M costs and energy yield
respectively
Figure 9. Normalized energy costs and farm availability against mean annual wind speed
40 M. KuÈhn et al.
Figure 10. OWECS availability and energy cost as a function of the reliability of the design for several site conditions
Highlighting the Importance of Operation and Maintenance Aspects

Identi®cation of Operation and Maintenance as Main Cost Driver
Two striking dierences with onshore wind farms are the largely reduced accessibility by bad weather
conditions (wave height, wind speed and visibility) and the costs of transport or lifting operations.
Therefore, the operation and maintenance (O&M) aspects of oshore wind farms are analysed in a
comprehensive way.
Figure 10 gives one example of the application of the O&M design tool which was developed during the
project. Firstly, the OWECS availability is shown in the area plot as a function of both the reliability of
the design and the accessibility of the wind turbines. A site without restriction with respect to accessibility
(i.e. onshore) reaches 97% availability for a state-of-the-art design increasing up to 99% for an extremely
reliable design under an adequate maintenance strategy. As weather conditions get worse, availability may
fall down to 53%±89% depending upon the failure rate and maintenance approach of the turbines.
Secondly, also shown in the ®gure is the calculated energy costs (20 years loan, 5% real interest rate) as
a function of failure rate and accessibility (line plot). The assumption is made that the cost at 100%
accessibility increases from 0.06 to 0.065 ECU/kWh with increasing reliability. Harsher weather con-
ditions will simultaneously lead to decreased accessibility and increased gross energy yield. With state-of-
the-art reliability, cost values below 0.05 ECU/kWh are not reachable, since increased gross energy yield is
more or less compensated by decreased availability. However, improved reliability will ®nd its payback in
terms of signi®cantly reduced energy cost.
Concept Lines for Future Wind Turbines and Associated Maintenance Strategies
Introduction. In the present section a number of concept lines for future oshore application are
developed. Such concept lines represent dierent compromises between the desirable features of turbines
intended for oshore use. Some designs attempt to minimize the amount of maintenance required, some
focus on simplifying and therefore reducing the costs of maintenance, while others neglect maintenance
reduction altogether, aiming to maximize the energy yield.
The concepts are intended to represent something of what the `next generation' of wind turbines might
look like. Each, therefore, will be described in relation to the base case, which is itself a ®ctitious machine.
The following concept lines have been identi®ed:
. Base Case wind turbine;
. Disposable wind turbine;
. Reliable (reduced failure rate) wind turbine;

. Advanced Control (lightweight) wind turbine.
After some elaboration of the concept lines an attempt will be made to make some comparisons and
projections. This will include the various maintenance strategies applicable to the concepts. Finally, the
Opti-OWECS design solution with respect to O&M will be presented.
Base Case Wind Turbine. Using commercially available onshore wind turbines as a starting point, it is
possible to derive a generic state-of-the-art design with respect to failure rate and maintenance demand.
This is possible using information from databases in which the operational track record of a large number
of wind turbines was gathered. From the various sources available a database from the state of Schleswig-
Holstein in Germany28 was selected. The large number of failure modes were, for the present purpose,
grouped into six failure classes, each having a dierent character with respect to failure rate, repair time,
spare part requirement and demand for external cranage.
For three popular designs in the 500/600 kW class, of which at least 25 identical wind turbines had been
in operation for more than 1 year following the commissioning period, the failure rates and modes were
taken and grouped into these six failure classes. The average failure rate of all 110 wind turbines taken
together turned out to be 2.3 failures/year. In order to establish the Base Case reference point for OWEC
designs with respect to O&M, the value for the most reliable of the three designs in each of the six classes
was taken. This resulted in a total failure rate of 1.79 failures/year and a regular service visit every
6 months. In fact, this is already a kind of forward-looking approach. Although the Base Case failure
rates may be achievable through modi®cations to standard 500/600 kW designs, they are certainly not yet
proven design practice for the larger wind turbines foreseen in oshore applications.
Disposable Wind Turbine. The primary goal underlying a disposable no-maintenance concept is the
production of the lowest capital cost machine consistent with a predictable life span. No provision will be
made for any maintenance at all, allowing the adoption of a very tightly integrated component arrange-
ment. This concept is formulated with a view for use with a no-maintenance strategy, where failed
machines are abandoned and no attempt is made at repair. Disposable machines would have to be
designed to ensure that the expected lives of all their components are very similar. Combining components
with dierent lives would unnecessarily increase the capital cost.
Reliable Wind Turbine. This concept is based on the assumption that engineering improvements will
reduce the failure rate of the machine. A signi®cant improvement in reliability, however, necessitates a
substantial increase in the machine capital cost as well. This reduced failure rate, reliable concept will
represent an incremental development of existing technology.
Advanced Control Wind Turbine. An alternative approach to oshore design is to neglect the reduction
of maintenance demand and attempt only to optimize the balance between the machine capital cost and
performance. The advanced design aims to achieve good survivability oshore by alleviating high loading
through bending and compliance rather than resisting with pure strength. This could be achieved through
innovative control features such as rapid and independent blade feathering to reduce the eects of gusts.
Variable speed control systems can also have a role in load alleviation and improvement of the aero-
dynamic eciency. Furthermore, it is conceived that the capital costs of an advanced turbine will be
comparable with a more conventional design, where its potential energy yield will be higher through its
advanced control and compliant behaviour.
Comparison of the Design Concepts. A qualitative comparison of the four concept lines is depicted in
Figure 11.
42 M. KuÈhn et al.
Figure 11. Qualitative comparison of the perceived advantages and disadvantages of the proposed wind turbine concepts
(B, Base Case; D, Disposable turbine; R, Reliable turbine; A, Advanced Control turbine)
The disposable concept is likely to be very risky for the developer. The economics is dominated by
achieving very low or even non-existent expenditure on O&M. Thus any de®ciencies that cause unexpect-
edly high failure rates will wreck the economic calculations in a devastating fashion.
The reduced failure rate, reliable design is at the opposite end of the capital cost scale from its
disposable counterpart. The reduction of failure rates will probably be a costly exercise. It will evidently
never exhibit the absolute O&M economy that the disposable machine aims for. In return for the extra
capital expenditure its O&M cost will be signi®cantly lower than the base case. Being essentially derived
from current technology, the risk associated with a reduced failure rate machine is likely to be fairly low.
Building machines adapted from tried and trusted designs is safer than adopting radical innovations.
It is dicult to speculate usefully about the economic features of an advanced control lightweight
machine. Without an in-depth analysis the only sensible conclusion at present is that its O&M demand
will be more than for the other machines. The potential energy capture (assuming the same availability as
the base case turbine) is the highest. Also fairly clear is the risk involved with an advanced projectÐit
would be considerable because of the sheer quantity of development work necessary.
Maintenance Strategies. Within discussions about oshore application it is sometimes proposed to

provide no maintenance at all to an OWECS. Then it is argued that it might be advantageous to apply
major overhaul say only every 5 years, replacing failed turbines as whole units. With the failure rate of the
base case it may be concluded that such a way of `maintaining' a wind farm is not viable. Even when the
failure rate is reduced drastically, to say 10% of the Base Case total failure rate, the overall availability of
the OWECS in the ®fth year of operation is only 45%. Reducing the overhaul interval to once every
3 years means an availability in the third year of 64%, which is probably still too low to make operation
economic.
A reduction of the (total) failure rate to a value signi®cantly lower than 0.2 failures/year would be
necessary to obtain a reasonable availability (say always above 80%) without undertaking any corrective
maintenance over a period of 3±5 years. It is clear, in this light, that OWECS adopting no-maintenance
strategies are not feasible options for the near future. Evidently the same conclusion can be drawn for the
development of the disposable wind turbine design line.
For the other design lines, including the Base Case, a more or less conventional maintenance strategy
has to be applied in order to keep the availability of the OWECS at a suciently high level.
Apart from the repair of failures (`corrective maintenance'), machines need also regular care (`preventive
maintenance'). Base Case preventive maintenance is assumed to take place every 6 months. This should be
interpreted as a nominal time interval (comparable with the regular milage service of a car). For oshore
application it is assumed that such service work takes place when the weather permits a visit to the turbine.
Of course this in turn assumes that sucient repair crews are available to take advantage of the ®rst
possibility of access. Otherwise it may still aect the availability of the OWECS in a considerable way.
With the use of a Monte Carlo simulation tool for the O&M operations within a wind farm (described
in a following section) it is possible to evaluate the consequences of modi®cations in the strategies in
relation to the dierent design lines. This is demonstrated in the following example.
Assume that the Reliable design has a 20% decreased failure rate with respect to the Base Case, whereas
the failure rate of the Advanced Control design is increased by 20%. Furthermore, assume that the
Advanced Control design has a potential energy yield which is 10% higher than the Base Case and the
Reliable design. The simulated weather involves a moderate 18% storm percentage. The (single) travel
time to the OWECS for the repair crew is 1 h. For repairs requiring the replacement of heavy components
there is one (external) crane available for hoisting actions within the OWECS. Furthermore, a number of
cost assumptions are made with respect to maintenance operations.
The columns of Figure 12 show the energy output as a function of the deployment of crew. The
potential energy yield of the farm (100 wind turbines rated at 1.2 MW, assuming 100% availability) equals
325 GWh/year (and 357 GWh/year for the Advanced Control design). With one crew, consisting of two
people working for 12 h per day over 7 days per week (thus eectively consisting of four persons), the
availability of the wind farms turns out to be 74%, 86% and 63% respectively for the three consecutive
designs. Note that a longer work period per day is not very realistic when the problems of assessing the
wind turbine and working in the dark are taken into account.
With the simultaneous deployment of two crews, onshore availability ®gures (96%±98%) are recovered
again, except for the Advanced Control design (88% availability). Owing to the assumed 10% gain in
potential energy output, the actual output of the farm equipped with Advanced Control wind turbines is
however equivalent to the OWECS equipped with one of the other two designs.
It can be seen from Figure 12 that in all three cases it seems favourable to have two maintenance crews
working in the wind farm when only O&M costs are regarded. For a wind farm equipped with the
Advanced Control design it is however still advantageous to deploy the third crew as well. The costs of
O&M do increase from 0.015 to 0.0154 ECU/kWh, but as a consequence the yearly energy yield is also
increased from 318 to 342 GWh/year. This means that the extra expenditure of 0.5 MECU for the extra
Figure 12. Energy output and O&M costs related to crew deployment and wind turbine design
44 M. KuÈhn et al.
crew results in an electricity gain of 24 GWh/year. This leads to a pro®t when the price for the extra kWh
exceeds 0.021 ECU/kWh, which is evidently the case.
Operation and Maintenance Solution for the OWECS

The above example clearly shows the importance of a well-adapted maintenance strategy with respect to
machine design in order to obtain an optimal OWECS design solution. Monte Carlo simulations are used
extensively to optimize the design solution presented in the Opti-OWECS project. As a starting point the
mechanical design of the WTS 80 (3 MW) machine was chosen. Since this is not a commercial series built
machine ( just two of them have been realized until now), it was regarded not very sensible to use the
maintenance track record of this machine. Instead the above-described integrated failure modes and rates
of the Base Case design (representing a premium engineered commercial 500±600 kW onshore wind
turbine), added up to a total of 1.79 failures/year, were taken as a starting point. With the details of the
design taken from the machine itself, an assessment was made with respect to the possible reduction of the
failure rates within an acceptable cost band.
It was concluded that further reduction of the number of failures requiring external cranage could not
be obtained. The associated Base Case failure rate of 0.32 lifting operations per year was therefore
maintained. Failure rates regarding the design of the hydraulic systems could be reduced by 30% by
careful re-examination. The major reduction in failure rates could however be obtained by application of
( further) redundancy in the electrical and electronic systems and a signi®cant increment in the number of
sensors used for remote control and reset. The total (design) failure rate for the WTS 80M (the OWEC
version) achieved is then 1.02 failures/year. This means an average corrective maintenance interval of
some 18 months. The regular preventive maintenance period is determined at a ¯exible once per year for
minor service (ready within 1 day) and a major service overhaul every 5 years (nominally 3.5 days).
For the wind farm consisting of 100 wind turbines of the type WTS 80M the cranage demand adds up
to about 27 operations each year. From Monte Carlo simulations it becomes clear that for this number of
operations the only possible approach to achieve a good OWECS availability and economics is the
purchase of cranage facility. The present design solution opts for a modi®ed self-propelled jack-up
platform. Jack-ups are commonly used for the exploration of oil®elds and provide a stable platform for
maintenance work. In the design solution a crane is added to the legs of the jack-up and in moderate sea
depth (typically until 30 m) the remaining length of the leg above the sea level together with the crane's jib
is enough to achieve the required lifting height.
Part III: Analysis Tools and Cost Estimate for European Regions
Development of Analysis Tools for Offshore Wind Farms
Tool for Cost Analysis and Preliminary Design
In parallel with the design work a computer-based cost model for the economic assessment of certain
OWECS concepts was developed. The model allows investigation of the eect of changes in important
parameters on the cost of energy from an OWECS, including the wind speed, the support structure height,
the size of the turbine and the distance from the shore.
The model was validated through re-evaluation of some well-known OWECS proposals and some real
farms. Its predictions were compared with published data for the validation cases and exhibited an
acceptable correspondence.
A rigorous quanti®cation of the accuracy range of the cost model's estimates was beyond the scope of
the project and would require a probabilistic analysis. An indication of the range can be obtained by
considering two types of uncertainty: that due to the basic input data and that introduced by the cost
model itself.
Uncertainty due to basic data aicts any calculations based on measured information. For onshore
sites the typical uncertainty range for energy yield predictions is between 10% and 20% even when time
series data have been collected over an extended period. At many of the oshore locations investigated
with the model, no long-term measurements of the actual site are available. The uncertainties in the
energy yields estimated by the cost model are likely to be substantially greater than those typical onshore.
Similar considerations apply to the variability and the limited knowledge of other environmental para-
meters that in¯uence the cost estimates, including water depths, wave heights, foundation conditions, etc.
It is also important to keep in mind that the model undertakes simpli®ed calculations and there may be
locations with peculiar localized conditions that breach the model's assumptions. In these cases the model
will not produce meaningful results.
The uncertainty introduced by the cost model is dicult to estimate and relates to the cost values built
into the model. Whilst speci®c material costs for the wind turbine and the support structure are quite well
predictable, this is not the case with many of the components, including submarine cables and the power
transmission system. Such materials cannot be ordered o the shelf and their cost will depend on the order
size and negotiation with the manufacturer. Finally, prices for marine operations are largely determined
by the availability of dedicated equipment and are therefore normally given only on a day-by-day basis.
In essence, the uncertainty in the absolute costs predicted by the OWECS cost model, assuming precise
input data, is thought to be of the order of magnitude of engineering cost estimates for feasibility studies
(typically some 30%±40%). Much of this error, though, is likely to be systematic in nature, in¯uencing all
calculations equally. By way of example, the model might tend to overestimate the cost of the support
structure at a particular location by a certain amount, say a%. However, most of the support structures
designed by the model at most locations will be subject to a similar overestimate of around a%. Thus the
value of the model as a comparative tool is better than that suggested by the uncertainty in the absolute
costs. As the purpose of the code is to provide directives for the selection of dierent sites and evaluation
of preliminary design options, these high levels of systematic uncertainty are not regarded as an obstacle
to its successful use.
One particular feature of the cost model is that, as far as possible, many of the calculations are
undertaken on what could be termed a `®rst-principles' basis. In other words, the model reaches its cost
estimates by actually designing, albeit in a highly simpli®ed manner, an OWECS meeting parameters
speci®ed by the user, and then costing the result. A further feature of the model is the ¯exibility in
con®guration that it incorporates. Options are available to allow both detailed examinations of well-
de®ned OWECS concepts and more general study of broad trends.
Within the model an OWECS is treated as a series of interdependent subsystems. The major subsystems
are designed and costed as follows.
Turbine. As the model deals mostly with the oshore-speci®c aspects of OWECS, it can basically only
consider pre-designed turbines, i.e. the detailed optimization of the turbine components has to be done
externally. If none of the commercial or pre-production designs, included within the model as `reference
machines', meets the user requirements, the model can estimate sensible values for the major turbine
parameters. For the time being, speci®ed rated power is used as design driver in such a case, whilst future
releases of the code might adopt speci®c torque as governing criterion.
Support Structure. The support structure design and costing routines represent the most complex part of
the model and consumed a substantial proportion of the development eort. For any OWECS the
support structure is designed from scratch using a simpli®ed algorithm formulated during the project. Full
account is taken of wind fatigue loading, ultimate strength requirements, including ice loads if necessary,
and the need to avoid resonance. A major restriction, purely in the context of the cost modelling, is that
hydrodynamic fatigue is ignored, because the interaction between dynamic wave response and design
properties as well as the combination of wind and wave fatigue was beyond the scope of implementation.
As a consequence, soft±soft designs, most prone to dynamic ampli®cation of wave loading, are not
considered. By default, the routines produce a soft±sti design, falling back on a sti±sti design if that
proves necessary. Facilitates are included to allow a limited automatic optimization of certain design
46 M. KuÈhn et al.
parameters. The model has greatest capability with monotower designs, using either pile or gravity
foundations and a variety of installation techniques. Lattice-type support structures can currently be
modelled in a very limited way, although work is continuing to extend this aspect of the model.
Operation and Maintenance. Several operation and maintenance `scenarios' are incorporated in the
model, based on cost and availability data produced by the Monte Carlo code for the simulation of
operation and maintenance behaviour developed during the course of the project.
Grid Connection. A semi-independent module computes all electrical and economic details of the power
cabling, both within the wind farm and for the connection to the shore. The consideration of the onshore
grid connection costs is currently limited to cable costs only, which are assumed to be linearly dependent
on the distance between shore and public grid. For the results presented in this article, the model was set
to compute the cost of delivering power to the beach, and no account was taken of any costs involved in
connecting from the beach to the power grid.
The Opti-OWECS cost model has been developed to run within the Microsoft Excel 5.0 (or better)
spreadsheet system to ensure wide accessibility to the wind energy community. While a spreadsheet oers
a convenient means of analysis, the `cell-by-cell' nature of the calculation technique makes code
development dicult, and highly prone to errors resulting from mis-referenced cells. In a complex sheet
such errors are almost impossible to ®nd. Fortunately, Excel 5 incorporates a highly capable program-
ming language known as `Visual Basic for Applications' (VBA) that allows code to be developed in a
fashion comparable with that used for any conventional high-level language. As much as possible of the
model was written using VBA in a structured manner, making any future upgrades a much easier prospect
than with the authors' previous models. A major advantage of the approach is that the usual, intuitive
spreadsheet interface can be used for data input and output, with the casual user being unaware of the
`machinery' that underlies the calculations.
GIS Cost Model

A `GIS' or Geographical Information System is essentially a computer-based mapping and data
manipulation tool. Spatially varying data, such as the water depths within a region, can be stored along
with information ®xing the data geographically, e.g. the latitude and longitude of a particular feature.
Such spatial data may be stored in a number of formats. Of greatest interest here is the raster format,
wherein any area under consideration is divided into a regular array of cells, and each cell is allocated a
value. It is implicitly assumed that this single value is representative of the entire geographical area
encompassed by the cell.
In comparison with more conventional mapping techniques, the great advantage of a GIS is its ability
to perform mathematical manipulation of the stored spatial data. This facility has been employed here to
couple the output of the cost model, as described above, with a raster format spatial database of oshore
conditions compiled in a previous Joule project.2 This composite system can be used to plot contour maps
showing how the cost of energy produced by a range of OWECS concepts would vary as a function of
geographical location. While the results presented in this article are con®ned to Northern European
waters, there is no reason in principle to prevent the system being used further a®eld, so long as appro-
priate environmental data are available.
The approach employed in the development of the GIS model is illustrated in Figure 13. Although the
diagram, as well as the methodology, seems straightforward, it must be stressed that a large number of
design and site parameters have complex, interrelated in¯uences on the cost of OWECS-produced energy.
Considerable simpli®cations had to be made in the calculations even to implement the methodology
shown. As an example, the calculations performed with this model neglect the costs involved in con-
struction of an onshore grid connection. Furthermore, only one type of support structure (the monopile)
was considered, even though the rest of the project investigated a range of bottom-mounted support
structure concepts. It is unlikely that the monopile structure is a good choice of support in all the regions
Figure 13. Approach followed in the comparison of energy costs at European sites
investigated. Additionally, it was assumed that the power transmission cable is directed to the nearest
shore even if it turns out to belong to a small and possibly uninhabited island. Cost estimates in regions
such as Northern Scotland or near the Wadden Sea, where there are many such small islands, require
careful interpretation therefore.
It is clear that the GIS model is subject to a number of limitations in its present form. However, the
reader should keep in mind that this part of the work was to a large extent motivated by a desire to
develop a `®rst-draft' methodology for GIS-based economic assessment of OWECS, and indeed to
demonstrate that such calculation procedures were possible at all. These objectives have been entirely
satis®ed. Moreover, it has proved possible to apply the system to regions of Northern Europe and produce
economic evaluations with at least a qualitative validity. The authors intend to continue development of
the model to address its limitations.
Code for Simulation of Operation and Maintenance Behaviour

A Monte Carlo program simulating the O&M behaviour of an oshore wind farm was developed. With
the help of this program it was possible to investigate various possibilities for deployment of maintenance
hardware, crews and O&M strategies. As to equipment, a distinction can be made between crew transport
(e.g. vessel and helicopter) and lifting equipment (e.g. modi®ed jack-up, crane vessel and built-in lift
system).
The program simulates the O&M behaviour of an OWECS over a lifetime period by following the state
of each component of the wind farm one time step at a time. At the start of each simulation run the failure
rates of the used wind turbines and the O&M strategy are de®ned. Further, the number of crews and
shifts, the kind and quantity of equipment and the site parameters are speci®ed. Dierent maintenance
strategies can then be evaluated by changing the input parameters, e.g. the time interval between
preventive maintenance visits. A random number generator acting on assumed probability distributions
simulates stochastic events, such as the occurrence of failures (of the wind turbine components) and the
state of the weather. At the end of the simulation run the total O&M costs, the achieved availability and
the produced energy of the wind farm are presented as output.
Application of Met±Ocean Database from the Oshore Technology

An essential step in the design of an optimized OWECS is to acquire a large data set of environment
measurements at the intended location as state-of-the-art in the oshore oil and gas industry. The
48 M. KuÈhn et al.
database should contain information on all relevant meteorological and oceanographical (met±ocean)
parameters over a long period, say 25 years. These data may be obtained from large measurement
programmes or from hindcast techniques. The latter techniques rely on a numerical modelling of the wind
and atmospheric pressure ®elds for an area; these ®elds are well available over long periods of time and
enable derivation of met±ocean parameters for grid points covering the whole area of interest. The present
project used the NESS database,29 kindly made available by Shell International Exploitation and
Production BV, Den Hague.
The information of the database is used for the following design considerations:
. the resistance to be provided to withstand the accumulated responses due to all circumstances experi-
enced over the entire lifetime of the structure, i.e. fatigue damage;
. the ultimate strength to be provided by the OWEC to withstand the maximum response expected to
occur during its service life, i.e. extreme response;
. to derive statistics for the met±ocean conditions in which installation and maintenance operations on
the OWEC can be performed, the so-called `weather windows';
. to assist in the prediction of the energy yield.
Structural Reliability Analysis of a Support Structure

The symbiosis of wind energy and oshore technology required for economic exploitation of the oshore
wind energy resource can be furthered through the application of structural reliability methods. For
an example OWEC support structure located at a demanding North Sea site it turned out that applying
such a method reduced the extreme design loading by about 40% relative to the conventional design
approach.
In conventional design practice for oshore structures the environmental conditions are determined on
the basis of independent estimates of extreme wave, extreme current and extreme wind conditions, each
having a return period of e.g. 50 years. These conditions are next assumed to occur at the same time and to
act in the same direction. This results in environmental design conditions and a corresponding global load
condition with a very long return period and an unnecessarily conservative design. Recent advances in
oshore engineering have led to a reliability-based design method that takes correlation between the
environmental conditions into account, which allows matching of structural design with prede®ned risk
criteria. This approach consists of four steps (Figure 14):
1. De®nition of the environment in terms of storm events, in which wave, current and wind conditions
are correlated instead of in independent environmental conditions. An essential requirement is the
availability of a large database containing information on the simultaneous occurrence of wind, waves
and current at the intended site during a long period (e.g. 25 years).
2. Determination of the long-term distribution of the extreme response of the structure during an
arbitrary storm. A storm consists of a succession of sea states, each with its associated wave, current
and wind conditions. The most straightforward manner to determine the (non-linear) response
behaviour in a particular sea state is, in principle, to perform a time domain simulation using a FE
model. This has to be repeated for many dierent realizations of the same sea state, for all sea states in
a storm and for all storm events in the database. The huge computational eort involved can be
drastically reduced by application of the technique of constrained random simulations,30 but remains
substantial when a lot of sea states need to be considered. By appropriate combination of the
individual response distribution from each simulation the long-term distribution of the extreme
response during an arbitrary storm can be determined.
3. Determination of the long-term distribution of the extreme response during the lifetime of the
structure. The results of step 2 are now combined with the probability that a storm will actually occur,
using the storm arrival rate derived from the database.
Figure 14. Determination of the extreme response distribution with a desired return period based upon the structural
reliability method
4. Determination of the probability of failure of the structure in a given lifetime by combining the result
from step 3 with information on the ultimate strength of the structure. When the lifetime is longer than
the duration of the database, this inevitably requires extrapolation, which should be done with care.
The results of this step make it possible to perform an economic risk evaluation or, alternatively, to
determine an environmental load level for structural design which meets a prede®ned reliability level.
Extension of Design Tools for Oshore Wind Energy Converters

OWEC dynamics played a pronounced role during the Opti-OWECS project. Analytical tools previously
developed for OWEC in the scope of JOUR 0072,2 and at Delft University of Technology,24,31 as modal
analysis and time domain simulations were extended using state-of-the-art methods in both wind energy
and oshore technology. Furthermore, considerations of the overall dynamics through the application of
an OWEC design tool formed an integral part of the design process, rather than merely a means of
checking the ®nal design solution. In fact, this approach reaped considerable bene®ts by producing a more
cost-ecient design solution than would otherwise have been the case, e.g. a soft±soft lightweight
monopile instead of a stier and heavier support structure, and also by allowing more reliable design
calculations, e.g. by consideration of inherent uncertainties of the environmental conditions. Moreover,
the design of the large two-bladed wind turbine pro®ted from the tailored dynamics of the entire OWEC
through adjustments in support structure stiness, rotor speed and blade design.
50 M. KuÈhn et al.
Consistent Estimate of Energy Costs for Entire European Regions

Overview
While the enormous potential of oshore wind energy in Europe is now well known,2 in general only a
poor understanding has been developed of the cost drivers important in exploiting this vast resource. The
majority of data available in the literature were cost estimates for speci®c OWECS concepts at particular
locations, which were dicult to compare rationally as a result of the diering assumptions each
embodies.
The GIS model described earlier has been used to estimate, on a consistent basis, how the production
cost of electricity from bottom-mounted OWECS might vary at all technically feasible sites within the
northern regions of the EU. Most environmental data for the calculations were taken from the JOUR
0072 database,2 kindly made available by Garrad Hassan and Partners and Germanischer Lloyd. A
limited number of other data sources were also employed.
It should be emphasized that this part of the work has a more general basis than most of the other data
presented in this article. In particular, the absolute levels of cost presented are heavily dependent on a
large number of assumptions implicit to the work. No great reliance should be placed on those absolute
values therefore. Instead the results should be regarded as a comparison of sites based on the reference
OWECS concepts that have been developed during the project as a whole. Thus, by way of an example, a
site with a predicted energy cost of 6±7 ECUct/kWh should only be regarded as (a) less economically
attractive than a site associated with an energy cost of 5±6 ECUct/kWh, (b) more economically attractive
than a site at 7±8 ECUct/kWh and (c) much more attractive than a site with an estimate of 9±10 ECUct/
kWh. While the cost estimates presented in this section do have some quantitative validity, it would be a
serious mistake to regard them as absolute values.
Regions Considered and Reference OWECS

Calculations were undertaken for certain northern regions of the European Union, including the waters of
Great Britain, Belgium, the Netherlands, Germany and Denmark. These regions were selected because
they are generally perceived as being hospitable to bottom-mounted oshore wind farms, and because the
principal expertise of the project partners lies within them. In order to satisfy the objective of comparing
the potential of dierent types of sites and technology, two distinct classes of oshore wind farm were
considered.
The ®rst, comprising 100 OWEC each of 1.5 MW rated capacity, was intended to represent the ®rst
medium-scale, commercial OWECS that might be erected at the beginning of the next century. Com-
parable OWECS are already proposed for speci®c locations in Danish and Dutch waters. Sites with good
wind conditions but only a moderate depth of water and reasonable proximity to the shore are considered
most suitable for this concept.
The second class of wind farm employed very large, multi-megawatt OWEC with a rated capacity of
4 MW each. Again the farm consisted of 100 units. Such oshore wind farms are considered suitable for
more remote and exposed locations in comparison with their smaller brethren, and construction is
conceived within a much longer time frame, perhaps 10 years from now. Such a development period will
be required to achieve the same high level of reliability assumed in the study for both reference OWECS.
Results
Cost maps for two dierent oshore wind farms are shown by way of illustration for the region of Belgian
and Dutch waters in Plate 1. Energy costs are based on an economic life of 20 years with a 5% real interest
rate. Blank areas of the maps signify locations at which the reference OWECS cannot be built either
because the water depth is too large for the monopile support structure (i.e. greater than 30 m) or because
the location is already used for other purposes (e.g. shipping lanes, oil platforms, wildlife reserves).
The results con®rm that the average wind speed and distance from shore are the most important
parameters in¯uencing energy cost. Most Danish North Sea locations, for example, although in a more
Energy cost map for 100 x 1.5 MW OWECS
Legend
11) Land
12) 2-3 ECUc/kWh
13) 3-4 ECUc/kWh
14) 4-5 ECUc/kWh
15) 5-6 ECUc/kWh
16) 6-7 ECUc/kWh
17) 7-8 ECUc/kWh
18) 8-9 ECUc/kWh
19) 9-10 ECUc/kWh
10) 10-11 ECUc/kWh
11) 11-12 ECUc/kWh
12) 12-13 ECUc/kWh
Energy cost map for 100 x 4 MW OWECS

Legend
11) Land
12) 2-3 ECUc/kWh
13) 3-4 ECUc/kWh
14) 4-5 ECUc/kWh
15) 5-6 ECUc/kWh
16) 6-7 ECUc/kWh
17) 7-8 ECUc/kWh
18) 8-9 ECUc/kWh
19) 9-10 ECUc/kWh
10) 10-11 ECUc/kWh
11) 11-12 ECUc/kWh
Plate 1. Energy cost map for OWECS employing 100 x 1.5 MW units (top) and 100 x 4 MW units (bottom) in Belgian
and Dutch waters (energy cost without onshore grid connection, for 20 years loan and 5% real interest rate)
Copyright © 1999 John Wiley & Sons, Ltd Wind Energy, 2 (1999)
demanding environment, show better performance than Baltic sites. Higher net annual energy yield more
than compensates the increased investment and O&M costs for the exposed sites.
Likewise the western shores of the British Isles are economically more attractive than the `less windy'
eastern shores. Despite the improved wind conditions found further oshore, energy costs increase
signi®cantly with distance from the shore owing to reduced turbine availability and the increased cost of
constructing long power transmission cables. However, more complex relations do arise, for example,
between the Cumbrian coast of Great Britain and the Isle of Man, where some oshore sites are
potentially more interesting than nearshore locations.
One advantage of using GIS for this type of work is that it enables statistical comparisons of data to be
carried out with comparative ease. Figure 15 compares the cost distribution of the total capacity of all
technically feasible wind farms as a function of cost level for both types of reference plant in the Belgian±
Dutch region. An additional distinction based on the distance of the sites considered from the shore has
been introduced to try to account in a simple way for the fact that non-technical reasons, such as
objections from the general public, may prevent full exploitation of inshore locations. For the purposes of
Figure 15. Distribution of proportion of installed capacity available at particular cost levels employing dierent size of
OWEC unit and wind farm rating in Belgian and Dutch waters: (a) inshore sites; (b) oshore sites (values normalized
with total capacity employing 1.5 MW units further than 5 km from shore; energy cost without onshore grid connection,
for 20 years loan and 5% real interest rate)
52 M. KuÈhn et al.
this analysis, medium-scale OWECS are arbitrarily considered to be ìnshore' when less than 5 km from
the shore, with the boundary being extended to 10 km for large-scale farms.
For both wind farm types the distribution of the capacity starts at approximately the same cost level,
but the centre of gravity lies at lower costs for the larger-scale OWECS than for smaller-scale wind farms.
The ®gure clearly shows that a greater proportion of capacity is available for lower cost levels if larger-
scale wind farms are used. Employing multi-megawatt machines does, however, not have much in¯uence
on the absolute minimum cost level for which electricity can be produced by an OWECS. The bene®t
oered by the larger OWECS is visible at more mediocre locations, which for instance tend to be more
remote or have greater water depths. Furthermore, the speci®c power per square kilometre is higher when
larger turbine units are used.
Exploitation of sites with the best economic performance can start therefore on a near- to medium-term
timescale using more mature `megawatt technology'. There would seem to be little economic advantage in
waiting for the development of large-scale multi-megawatt turbines before constructing wind farms at the
best locations. For OWECS with rated capacities in the range of several hundred MW, especially those
located at sites not having the very best environmental conditions, development of multi-megawatt
technology may oer advantages.
Part IV: Conclusions and Recommendations

Main Conclusions
Consideration of Oshore Wind Farms on a System Level
. The economic exploitation of the oshore wind energy potential requires a symbiosis of wind energy
and oshore technology. Although this seems obvious, progress towards this symbiosis is by no means
evident and requires the minds of both communities to be open. Design methodologies recently
developed for complex civil engineering systems are also adequate for oshore wind farms. Within the
integrated OWECS design approach the goal of the design process can be controlled eectively on the
system level by overall criteria such as levelized energy costs, adaptation to site conditions, installation
and commissioning eorts, availability, overall dynamics, etc.
. The dynamics of an entire oshore wind energy converter (OWEC) are complex but their consideration
from the outset of the design solution oers a signi®cant potential for cost reduction, especially if
fatigue (whether or aerodynamic, hydrodynamic of combined origin) is governing.
Development of Practical and Economic Solutions

. In several ®elds, e.g. support structure design, OWEC installation, operation and maintenance,
OWEC dynamics, design approach, structural reliability considerations, the study demonstrated new
propositions which can contribute signi®cantly to a mature oshore wind energy technology.
. Oshore wind energy is fully viable and energy costs of the order of 5 ECUct/kWh (annual mean wind
speed 8.4 m s ÿ1 at 60 m, 20 years loan, 5% rate of interest, without costs for onshore grid connection)
are achievable at sites with good wind conditions and good availability. This estimate is based upon
currently available technology constructed within an innovative, integrated design approach con-
sidering the entire system. Such cost levels depend to a large extent on the OWECS availability
achieved. At hostile oshore sites, cost values close to those achieved onshore will be possible only if the
failure rate of the wind turbines is reduced by a factor of (at least) two.
. Signi®cant cost reduction is possible by using innovative installation methods considering the entire
OWEC unit. Conceptual solutions have been developed for ¯oating installation of gravity-based
designs, as well as for both ¯oating and lifted installation of monopile structures.
. The dynamics of economic support structures have to be tailored with respect to the sometimes
opposing requirements from the wind turbine and the oshore environment.
Identi®cation of Cost Drivers for the Design Solution

. The major driver for the energy costs of large-scale OWECS found is the average wind speed of the site.
While this may seem an unsurprising, indeed obvious conclusion, it is dicult to ®nd it in the published
literature, which more often refers to water depth and distance from shore as design drivers. Other
important environmental parameters are the distance from the shore, the distance from electricity grid
infrastructure and the water depth. Of the `controllable' drivers, the most signi®cant are the operation
and maintenance strategy adopted for the farm, along with the rated capacity of both individual
OWEC units and the whole wind farm. The design wave height and soil conditions found at the
OWECS can have a smaller, although still signi®cant in¯uence on the farm economics.
. Reduced wind shear and the costs involved in the support structure, OWEC installation and main-
tenance lead to a lower optimum hub height at sea in comparison with the land situation.
. Gravity-based support structures might be economically unattractive owing to the large amount of
ballast required to counteract the quite considerable heave forces generated by extreme waves in shallow
waters.
. Although it is not a sensitive parameter, the most economic turbine spacing oshore is considerably
larger than on land.
Highlighting the Importance of Operation and Maintenance Aspects

. The availability of wind farms at real oshore sites employing commercial wind turbines without
signi®cantly improved reliability and without optimized operation and maintenance solutions may be
unacceptably low, e.g. 70% or even less.
. Reliability of oshore wind energy converters and operation and maintenance solutions should be
optimized with respect to the (levelized) energy production costs rather than to either capital or
operation and maintenance costs.
. Operation and maintenance costs mainly related to the wind turbine can account for up to 30% and
more of the energy costs.
. The operation and maintenance problem can be tackled by a rational approach using realistic
assumptions for failure rates, Monte Carlo simulations of the O&M behaviour of the system, etc.
Development of Analysis Tools for Oshore Wind Farms

. Estimates of investment as well as energy costs of oshore wind farms are much more complex than for
their land-based counterparts; therefore, any cost prediction requires careful interpretation and
dierent studies are dicult to compare. Nonetheless, an OWECS cost model valid for a speci®c
con®guration is an extremely useful tool during the conceptual design phase and for the comparison of
dierent sites.
. If extreme loads are governing, signi®cant savings in support structure costs are possible by application
of state-of-the-art methods for structural reliability analysis, such as have recently become available in
the oshore engineering community for the oil and gas industry.
. Time domain simulations of the overall dynamics of OWEC are an important design tool. Proper use
can result in less risky and more economic solutions.
. Support structure design for demanding environments, certainly if fatigue becomes important, requires
the use of high-quality databases on the correlation of wind and wave conditions as is state-of-the-art in
oshore technology.
Consistent Estimate of Energy Costs for Entire European Regions

. Energy costs of entire regions in Northern European waters have been estimated in a consistent
manner.
54 M. KuÈhn et al.
. Results of the comparison of energy costs in dierent regions con®rm the huge oshore wind energy
resources within the EU also reported by previous work. While the absolute values of the ®gures in the
energy cost maps are at best tentative in nature, it is concluded that a signi®cant proportion of these
resources will be exploitable on a commercial or near-commercial basis within the near future
(assuming that technical innovation within the wind energy industry continues its current rapid
progress).
. Comparisons are made of OWECS concepts based on turbines with rated capacities of both 1.5 and
4 MW at many real locations around Northern Europe. Using a larger-scale wind farm does not
in¯uence the absolute minimum cost level for which electricity can be produced at prime sites in a
signi®cant way. Instead, the bene®t oered by large-scale OWECS is that greater proportions of sites
are available for lower cost levels. Therefore, exploitation of the sites with superior performance can
start on a near- to medium-term timescale based on mature `megawatt technology'. For OWECS with
rated capacities in the range of 400 MW and larger, especially those located at sites not having the very
best environmental conditions, development of multi-megawatt technology may oer advantages.
. Windy North Sea sites oer better economics than most sites in the more sheltered Baltic Sea.
Recommendations for Future Work

Design of Oshore Wind Farms and Development of the Technology
. The integrated OWECS design approach should be further developed, e.g. by application to real plant.
. The OWECS guidelines of Germanischer Lloyd were, as far as is known by the authors, used for the
very ®rst time in a design situation. Application proved to be successful; nonetheless, veri®cation,
further development and clari®cation of particular issues should be addressed in the near future.
. The integration of very large oshore wind farms in the international electricity system will result in
speci®c design requirements. Thus large OWECS should be designed with due consideration for the
level of the overall energy system rather than only for the level of the particular oshore wind farm. The
consequences of this demand on e.g. wind turbine design, need careful investigation.
. Installation of large multi-megawatt OWEC remains an area of potential cost reduction and further
investigations are recommended on both ¯oating and lifted installation.
. Apart from further cost reduction in the ®eld of operation and maintenance, techniques for more cost-
ecient cable laying and cable connection to the OWEC units should be developed. Further cost
reduction and economic bene®ts are expected from development of industrial infrastructure for
construction of OWEC units, HVDC transmission, ®nancial instruments for exploitation.
. The importance of O&M aspects should be considered as a main design driver, especially for the wind
turbine. Even further advanced design approaches, e.g. design for RAMS (Reliability, Availability,
Maintainability and Serviceability), should be developed.
Further Development of Analytical Methods

. With respect to the high importance of operation and maintenance aspects the O&M simulation tool
should be further developed, e.g. improved weather simulation, wind turbine reliability model.
. The cost modelling and cost analysis of OWECS should be extended, e.g. by additional support
structure and wind turbine models, re®nement of the GIS-based cost model.
. Dynamic considerations particular to OWEC should be taken into account already at an early stage of
the design process. Although very successful, the analysis approach of the overall dynamics of OWEC
should be further re®ned and veri®ed by measurements on suitable oshore wind energy converters of
dierent type and at distinct sites.
. The structural reliability methods should be applied not only to the support structure under extreme
conditions but also to the entire OWEC under dierent operating conditions.
. The risk involved in large oshore wind farms should be reduced by improving the understanding of the
oshore wind conditions and associated other met±ocean parameters in order to facilitate more reliable
predictions of the energy yield, optimum design and optimum strategy for both installation and
operation and maintenance.
Acknowledgement
The European Commission has supported the project in the scope of the framework of the Non Nuclear
Energy Programme JOULE III (Research and Technical Development) under grant JOR3-CT95-0087.
References
1. European Commission, White Paper for a Community Strategy and Action Plan, Energy for the Future:
Renewable Sources of Energy, Campaign for Take-o, European Commission, Brussels, 1997.
2. H. G. Matthies, C. Nath, T. E. Schellin, A. D. Garrad, M. A. Wastling, D. C. Quarton, J. Wei, M. Scherweit and
T. Siebers, Study of Oshore Wind Energy in the ECÐJOULE I (JOUR 0072), Verlag NatuÈrliche Energien,
Brekendorf, 1995.
3. M. KuÈhn, W. A. A. M. Bierbooms, T. T. Cockerill, M. C. Ferguson, B. GoÈransson, L. A. Harland, R. Harrison,
G. J. W. van Bussel, J. H. Vugts and R. Wiecherink, Òpti-OWECS ®nal report. Vol. 0: Executive Summary',
Institute for Wind Energy IW98.139R, Delft University of Technology, 1998.
4. M. KuÈhn, L. A. Harland, W. A. A. M. Bierbooms, T. T. Cockerill, M. C. Ferguson, B. GoÈransson, G. J. W.
van Bussel and J. H. Vugts, Òpti-OWECS ®nal report. Vol. 1: Integrated design methodology for oshore wind
energy conversion systems', Institute for Wind Energy IW98.140R, Delft University of Technology, 1998.
5. M. KuÈhn (ed.), T. T. Cockerill, L. A. Harland, C. SchoÈntag, G. J. W. van Bussel and J. H. Vugts, Òpti-OWECS
®nal report. Vol. 2: Methods assisting the design of oshore wind energy conversion systems', Institute for Wind
Energy IW98.141R, Delft University of Technology, 1998.
6. T. T. Cockerill, R. Harrison, M. KuÈhn and G. J. W. van Bussel, Òpti-OWECS ®nal report. Vol. 3: Estimate of
costs of oshore wind energy at European sites', Institute for Wind Energy IW98.142R, Delft University of
Technology, 1998.
7. M. C. Ferguson (ed.), M. KuÈhn, W. A. A. M. Bierbooms, T. T. Cockerill, B. GoÈransson, L. A. Harland, G. J. W.
van Bussel, J. H. Vugts and R. Hes, Òpti-OWECS ®nal report. Vol. 4: A typical design solution for an oshore
wind energy conversion system', Institute for Wind Energy IW98.143R, Delft University of Technology, 1998.
8. T. T. Cockerill, Òpti-OWECS ®nal report. Vol. 5: User guide OWECS cost model', Institute for Wind Energy
IW98.144R, Delft University of Technology, 1998.
9. W. E. Heronemus, `Power from the oshore winds', 8th Ann. Conf. of the Marine Technology Society, 1972,
pp. 435±466.
10. F. Jensen and F. Olsen, Òshore wind farm at Vindeby, Lolland. Final report', THERMIE WE357/87DK, 1997.
11. H. J. Kouwenhoven, `Lely windturbine project. Final report', THERMIE WE23/89NL, Energie Noord West,
1996.
12. P. S. Madsen, `Tunù Knob oshore wind farm', European Union Wind Energy Conf., GoÈteborg, May, 1996,
pp. 4±7.
13. P. B. Simpson, `Costs of oshore energy from bottom-mounted two-bladed HAWTs', Prospects for Oshore
Wind Energy: the State of the Art and Future Opportunities, Harwell, 1993.
14. À study of oshore based wind power; the Blekinge project summary report', 1991.
15. RES (Renewable Energy Systems), `Feasibility study for a prototype oshore wind turbine', ETSU W/35/0021/
REP/A, 1993.
16. F. A. Olsen, N. Raben and F. V. Jansen, `Feasibility study on a large-scale oshore wind farm located in the
Baltic Sea', European Wind Energy Conf., Thessaloniki, October 1994, pp. 1018±1023.
17. E. Hau, WindkraftanlangenÐGrundlagen, Technik, Einsatz, Wirtschaftlichkeit, 2nd edn, Springer, Berlin, 1996
(in German).
18. V. Diedrichs (ed.), Oshore-Windenergiesysteme, FoÈrdervorhaben des BMFT Nr. 0329645 A und B, Vulkan
Engineering, Bremen, 1995 (in German).
19. D. J. Milborrow, Cost, Prices and Values, Vol. 2 of Wind Energy, the Facts, Commission of the European Union,
1999.
56 M. KuÈhn et al.
20. SEAS (ed.), Action Plan for Oshore Wind Farms in Danish Waters, Elkraft v/SEAS Wind Power Department,
1997.
21. Novem, Haalbaarheidsstudie Demonstratieproject Nearshore Windpark, Novem, 1997 (in Dutch).
22. D. C. Quarton, `The evolution of wind turbine design analysisÐa twenty year progress review', Wind Energ., 1,
5±24 (1998).
23. K. J. MaÊnsson, `Pilot plant for oshore windpower', European Wind Energy Conf., Amsterdam, 1991,
pp. 446±448.
24. M. KuÈhn, `Dynamics and design optimisation of oshore wind energy conversion systems', Doctoral Thesis,
Delft University of Technology, 1999.
25. H. A. J. de Ridder, `Design and Construct of complex civil engineering systems: a new approach to organisation
and contracts', Doctoral Thesis, Delft University of Technology, 1994.
26. Kvaerner Turbin AB, Development Study IIIÐFinal Report, Kvaerner Turbin AB, Kristinehamn, 1997.
27. T. T. Cockerill (ed.), `Structural and economic optimisation of bottom-mounted oshore wind energy
converters', First Technical Report of Joule Project JOR3-CT95-0087, 1996.
28. W. EggergluÈb (ed.), Wind Energie VIIIÐPraxisergebnisse 1995, Landwirtschaftkammer Schleswig-Holstein,
Rendsburg, 1996 (in German).
29. D. J. Peters, C. J. Shaw, C. K. Grant, J. C. Heideman and D. Szabo, `Modelling the North Sea through the
North European storm study', Paper OTC 7130, 1993.
30. L. A. Harland, Èxtreme response analysis of oshore structures', Doctoral Thesis, Delft University of
Technology, 1999.
31. M. KuÈhn, `Simulation of oshore wind turbines under stochastic loading', European Wind Energy Conf., 1994,
pp. 845±851.
32. M. KuÈhn, `Terminology, reference systems and conventions to be used within the Opti-OWECS project',
Institute for Wind Energy IW96.097R, Delft University of Technology, 1996.
Appendix: OWECS Terminology
Oshore wind energy is a fairly young and multidisciplinary ®eld. Unfortunately, no uniform terminology
exists and misunderstandings can occur quite easily. Therefore, within the Opti-OWECS project one has
agreed upon a particular terminology, conventions and reference systems in order to make the internal
and external communication more eective. Here a short version of Reference 32 is presented.
Preface
In principle, the common practice concerning notation and convention within the considered disciplines,
i.e. wind energy technology, oshore technology and engineering economics, should be used. However,
harmonization is required in the description of the entire system and its components, the interfaces
between subsystems and the structural design. The two former aspects are treated in this appendix.
Offshore Wind Energy Conversion System (OWECS)

oshore wind energy conversion system (OWECS)
Entire system, comprising (usually) several wind energy converter units, for conversion of wind energy
into electric power, including the wind turbines, the support structures, the grid connection to the
power delivery point and operation and maintenance aspects. Note that the environment, i.e. air, water
and soil as well as the utility grid, is not considered as a part of the OWECS.
oshore wind farm

Synonym for OWECS.
Figure 16. Subsystems of the (bottom-mounted) OWECS
Subsystems of the OWECS

Subsystems of an OWECS (Figure 16) comprise either physical parts of the system, e.g. wind turbine,
support structure, or important aspects such as operation and maintenance aspects.
oshore wind energy converter (OWEC)
Single unit of the OWECS comprising wind turbine and support structure.*
wind turbine (WT)
Component of an oshore wind energy converter that transforms wind energy into electric power on
generator voltage or AC recti®er voltage, comprising rotor, nacelle with entire interior, control and
safety system and electrical turbine system.
support structure (bottom-mounted)
Structure that supports the wind turbine and transfers the loading into the soil. Hence the support
structure comprises both the tower and the foundation.
grid connection and wind farm layout
This comprises two main parts that are considered for convenience as one subsystem. Firstly, the
electrical system that takes the power provided at the turbine connection points and collects it at the
wind farm collection point(s) and successively transmits it to the onshore connection point with the
public grid. Secondly, the physical arrangement of the OWEC units.
operation and maintenance aspects
Auxiliary facilities, equipment and strategy required for operation, maintenance, control and
administration of an OWECS.
Boundaries of OWECS Subsystems

wind turbine and support structure
The ®xed end of the yaw mechanism of the nacelle is de®ned as the boundary between the (horizontal
axis) wind turbine and the support structure. All geometric and dynamic conditions are expressed with
respect to the reference frame of the support structure.
wind turbine and grid connection
The turbine switch gear or circuit breaker at the tower base is de®ned as the boundary between the
electrical system of the turbine and the grid connection. The voltage at the connection point
corresponds to the generator or the inverter (if any). Although a transformer might be installed at the
wind energy converter unit, it is regarded as part of the grid connection.
*No plural of the abbreviation ÒWEC' should be used; instead one may use ÒWEC units', the full spelling or the singular form of ÒWEC', if
possible.
58 M. KuÈhn et al.
grid connection and utility grid

The power of the OWECS is provided as three-phase AC at the voltage level of the utility grid to which
the wind farm is connected. In the absence of other explicit conventions the connection point is situated
at the ®rst dry location onshore regardless of the actual grid infrastructure on land.
The main components of the OWECS subsystems are de®ned by Figure 17.
Figure 17. Components of the OWECS subsystems

K-HN Et Al-1999-Wind Energy PDF

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

K-HN Et Al-1999-Wind Energy PDF

Загружено:

Авторское право:

Доступные форматы

WIND ENERGY

Wind Energ., 2, 25±58 (1999)

Research Towards a Mature Offshore Wind

CCC 1095±4244/99/010025±35 $17.50

Opti-OWECS Project in a Nutshell

Review of the Development of Offshore Wind Energy to Date

Economics of Recent Projects and Studies

Figure 2. Comparison of energy costs between dierent studies and projects

Design Practice in Wind Energy and Oshore Technology

Current Design Approaches for Oshore Wind Farms

Part II: Design Approach and Design Solution

Development and Demonstration of an Integrated Design Approach

Overall Dynamics of Oshore Wind Energy Converters

Development of Practical and Economic Solutions

Figure 5. Examples of support structure concepts (GBS  gravity based system)

. integrated development of support structure concepts and installation procedure;

Table II. Main data of the design solution

Economic Performance of Design Solution

Identi®cation of Cost Drivers for the Design Solution

Highlighting the Importance of Operation and Maintenance Aspects

. Reliable (reduced failure rate) wind turbine;

Maintenance Strategies. Within discussions about oshore application it is sometimes proposed to

Operation and Maintenance Solution for the OWECS

GIS Cost Model

Code for Simulation of Operation and Maintenance Behaviour

Application of Met±Ocean Database from the Oshore Technology

Structural Reliability Analysis of a Support Structure

Extension of Design Tools for Oshore Wind Energy Converters

Consistent Estimate of Energy Costs for Entire European Regions

Regions Considered and Reference OWECS

Energy cost map for 100 x 4 MW OWECS

Part IV: Conclusions and Recommendations

Development of Practical and Economic Solutions

Identi®cation of Cost Drivers for the Design Solution

Highlighting the Importance of Operation and Maintenance Aspects

Development of Analysis Tools for Oshore Wind Farms

Consistent Estimate of Energy Costs for Entire European Regions

Recommendations for Future Work

Further Development of Analytical Methods

Appendix: OWECS Terminology

Offshore Wind Energy Conversion System (OWECS)

oshore wind farm

Figure 16. Subsystems of the (bottom-mounted) OWECS

Subsystems of the OWECS

Boundaries of OWECS Subsystems

grid connection and utility grid

Figure 17. Components of the OWECS subsystems

Вам также может понравиться

Figure 2. Comparison of energy costs between dierent studies and projects

Design Practice in Wind Energy and Oshore Technology

Current Design Approaches for Oshore Wind Farms

Overall Dynamics of Oshore Wind Energy Converters

Figure 5. Examples of support structure concepts (GBS gravity based system)

Maintenance Strategies. Within discussions about oshore application it is sometimes proposed to

Application of Met±Ocean Database from the Oshore Technology

Extension of Design Tools for Oshore Wind Energy Converters

Development of Analysis Tools for Oshore Wind Farms

oshore wind farm