Installation, operations and maintenance research area 5

Transcription

1 1 Installation, operations and maintenance research area 5 L.W.M.M. Rademakers (ECN)

2 2 RA 5: Installation, operations and maintenance

3 Installation, operations and maintenance research area 5 Author: L.W.M.M. Rademakers With contributions from: E. van de Brug (Ballast Nedam) D. Cerda Salzmann (TUD) E. Echevarria (TUD) A. Gerritsen (STC) P. Kuipers (Lloyd s Register Nederland) T. Obdam (ECN) J. Pasteuning (XEMC Darwind) J. van der Tempel (TUD) T.W. Verbruggen (ECN) J.T. van der Wal (Imares) 1 March, 2010 WE@Sea programme Large-scale wind power generation offshore The work presented in this report is funded by the BSIK-programme of the Dutch Government under project number BSIK Energy research Centre of the Netherlands, ECN P.O. Box ZG Petten The Netherlands RA 5: Installation, operations and maintenance 3

4 4 RA 5: Installation, operations and maintenance

5 Abstract The objective of the research programme Large-scale wind power generation offshore is to provide the knowledge to accelerate offshore wind farm developments in the Netherlands. The research was executed between 2005 and 2010 and covered among others environmental issues, grid connection and integration, financing and markets, operational aspects and the hardware for offshore wind power generation. To execute the research programme, the following research areas were defined: 0. Integration, Scenarios and Monitoring 1. Offshore wind power generation 2. Spatial planning and environmental issues 3. Energy transport and distribution 4. Energy market and finance 5. Installation, operations and maintenance 6. Knowledge dissemination & Training and education This report addresses the results of research area no.5. The objectives of research area 5 were to develop and select cost effective and environmental friendly equipment and procedures for installation, maintenance and operation. This included also the development of the required analysis tools and probabilistic cost models. In connection with the other research areas scenarios had to be developed for fast and cheap installation, operation at low costs and minimum demand for service and maintenance, and easy dismantling. More specific the following objectives should be reached: cost models for analysing and evaluation of scenarios for installation and maintenance including risk management of bad weather conditions; collection and dissemination of operational experience on installation and maintenance (a.o. reports with experiences, logbooks, databases with failure and reliability data, videos) recommendations for improving e.g. turbines, grid connection, ports, vessels, access systems, testing facilities and infrastructure. This report contains the main results of the research carried out within research area 5. After the detailed description of the problems related to the installation, operation, and maintenance of offshore wind farms in Chapter 1 and the international developments of these topics, Chapter 2 contains the objectives of the research and a description of the work programme. Chapter 2 also contains an overview of the projects that have been executed. In Chapter 3 the research results are presented per research topic: 5.1 Installation and access 5.2 Operation and maintenance concepts 5.3 Safety Finally in Chapter 4 the needs for further research are proposed. The annexes of this report contain the summaries of the individual projects that have been carried out in research area 5. RA 5: Installation, operations and maintenance 5

6 6 RA 5: Installation, operations and maintenance

7 Table of Contents 1. Introduction We@Sea and offshore wind power generation Field of interest and links with other research areas Background on installation, operation, and maintenance International developments The problems Installation and Access Operation and Maintenance Concepts Safety How to read the report? Objectives and Work Programme Objectives Definition of research topics Installation and Access Operation and Maintenance Concepts Safety RL5 Partners Highlights and results RL 5.1 Installation and access Installation methods Ampelmann for offshore access RL 5.2 Operation and maintenance concepts O&M Cost Modelling Low cost solutions for load monitoring Design for Redundancy and Fault Tolerant Control RL 5.3 Safety Gaps and future research Installation and access Operation and Maintenance Safety References 43 Project Summaries Optimal integrated combination of foundation concept and installation method PhD@Sea: Ampelmann; a motion-compensating platform for accessing wind turbines Ampelmann, demonstrator PhD@Sea: RAMS for offshore wind farms Development of O&M Cost Estimator for offshore wind farms Load monitoring for wind turbines; Fibre optic sensing and data processing Flight Leader concept for wind farm load counting (phase 1) Dedicated offshore maintenance support tools for XEMC DarWinD wind turbines Nautical Safety Safety and safeguarding offshore wind farms; Integral safety study Health, Safety & Environmental requirements for the Dutch offshore wind industry 89 RA 5: Installation, operations and maintenance 7

8 8 RA 5: Installation, operations and maintenance

9 1. Introduction 1.1 and offshore wind power generation High costs of R&D and relatively long lead times in the development of wind parks rule out a trial-and-error approach and call for a well-founded, conceptual approach covering all critical aspects of wind farm design. The broad range of topics covering design principles to environmental aspects, addressed in this programme requires that they are integrated in order to preserve the overall view and control of the process, prevent sub-optimisations. In the We@Sea programme the following research areas were defined: 0. Integration, Scenarios and Monitoring 1. Offshore wind power generation 2. Spatial planning and environmental issues 3. Energy transport and distribution 4. Energy market and finance 5. Installation, operations and maintenance 6. Knowledge dissemination & Training and education This report addresses the results of research area no Field of interest and links with other research areas The research carried out in research area 5 has mainly a link to research area 1 where design tools were developed. The wind turbine design and its loading pattern has a direct influence on installation, operation and maintenance aspects. 1.3 Background on installation, operation, and maintenance The We@Sea programme was defined in 2002 and 2003 when less than 300 MW offshore wind power was installed. At that time hardly any experiences on installation, operation, and maintenance were available. The wind farms installed at that time in fact consisted of (slightly modified) onshore turbine designs. There was not a standardised method available and proven for installing large amounts of wind turbines in shallow or deeper waters. Only experiences from the offshore oil and gas industry could be used as an example. During the definition of the We@Sea program, it was recognised that offshore wind energy in the Netherlands with the ambition of having 6000 MW installed in 2020 would not become cost effective if onshore turbines would be erected offshore and simultaneously making use of the state-of-the-art experiences from the oil and gas industry for installation, operation, and maintenance. The main reasons are the following. The practices used by the offshore oil and gas industry to install e.g. production plants are based on single pieces or small series whereas wind farms consist of large series of identical plants and thus require different installation processes. The reliability of onshore turbines was such that it would lead to too high maintenance costs and too long downtimes if they would be placed offshore 1. Too many visits will be necessary and each offshore visit will be more expensive than visiting an onshore turbine. 1 Analyses carried out during the definition of the We@Sea program revealed for instance that the costs for operation and maintenance of offshore wind turbines would become 50 to 80% higher per kilowatt-hour than for onshore turbines. RA 5: Installation, operations and maintenance 9

10 The maintainability of the present day onshore wind turbines is such that similar failures offshore will require more repair time and more expensive repair equipment than onshore. The offshore weather conditions are such that not all maintenance actions can be carried out instantaneously. During harsh weather conditions turbines cannot be accessed by personnel and crane ships cannot be used for e.g. installation and repair actions. The harsh weather conditions may lead to long downtimes and thus high revenue losses. Harbours require certain adaptations if they are going to be used for installing large series of wind turbines during good weather seasons, e.g. good access roads, minimum quay sizes for temporary storage of parts and for mooring of installation vessels, and possibilities for (partial) assembly of wind turbines. Offshore wind farms may cause an additional risk to shipping and fishery. Safety procedures being used for installation and maintaining onshore wind turbines strongly differ from the procedures that are common practice in offshore oil and gas which also include safe working with explosive materials. The ambitions of the Dutch government to install, operate, and maintain 6000 MW wind power by 2020 are challenging. With an average size of 3 to 5 MW, at least 1200 turbines should be transported, assembled, installed, and connected to the grid. Each turbine requires typically 5 to 10 visits per year for minor or major repairs. The enormous size of the planned capacity requires a structured approach during all phases in the lifetime to avoid costly errors. These considerations, together with the reasons mentioned above formed the basis for defining Research area 5 (RL5) within the We@Sea program. Within RA5, three research topics can be distinguished. 5.1 Installation and Access 5.2 Operation and maintenance concepts 5.3 Safety At first instance, a fourth and fifth topic Port Development and Dismantling were proposed as research topics but have not been considered further on mainly due to the reduction of the budget. Dismantling has been treated partly in RA 1. Safety was not regarded as a relevant topic during the definition of the We@Sea programme but was added in a later stage. The lack of clear safety regulations for offshore wind farms was expected to become a major bottleneck for the development of new offshore wind farms and We@Sea appeared to be an excellent platform for discussing the issue between all interested parties (developers, operators, turbine manufacturers, certification bodies, and maintenance companies). 1.4 International developments The focus of the We@Sea programme is on generating knowledge to speed up the implementation of offshore wind energy in the Netherlands. Offshore wind energy however is not an industry that is restricted to national boundaries. On the contrary, most companies in offshore wind energy work on a European level or even world wide. This means that knowledge developed in the Netherlands is being used abroad and vice versa. Parallel to the execution of the We@Sea programme several offshore wind farms have been installed in Europe by international consortia and the wind farms are running for some years now. This means that operational experiences become available and that some problems are being solved by industrial parties before national R&D programs have provided the expected answers. A good example for instance is labour safety. It would be preferred that international standards and laws on labour safety are in place before the first offshore wind farm are be taken in operation. However, up to now each new wind farm had to deal with labour safety individually and needed to solve the related problems at a wind farm level. At present, a lot of experience on labour safety during installation and maintenance is available and the international industry is more and more agreeing upon a set of rules and procedures on labour safety. However, the laws and rules still differ per country and harmonisation is still not achieved. Since the offshore wind industry is an international business, it is likely that parts of the research and development topics that were defined within the We@Sea program, was also proposed 10 RA 5: Installation, operations and maintenance

11 elsewhere. At a European level the wind energy Technology Platform was established in 2007 and one of the goals was to define a strategic research agenda that could be used by the European Commission to define stimulation programs for R&D and market deployment. Within the Technology Platform a working group was founded on offshore wind energy and this working group consisted of representatives of utilities, project developers, turbine manufacturers, certification institutes, and R&D institutes. The first release of the strategic research agenda in 2008 included the following topics for offshore wind energy. - Safety, education, and environment - Substructures - new and improved materials and manufacturing methods for cost reduction and deeper waters - Assembly, installation, and decommissioning - transfer of equipment from suppliers across Europe to wind farm sites - installing turbines in a hostile offshore environment - Electrical infrastructure - integration of offshore wind into the grid - improved design tools and life cycle cost approaches - increase reliability and extend the lifetime of components - Turbines - suitable for operating in marine conditions - improving reliability - improve tools for turbine design - Operations and maintenance - development of optimal O&M strategies - development and use of effective access systems - development of condition and risk based maintenance strategies - development of systems to minimise human intervention (better diagnostics and remoteness) As can be seen, the bold printed topics match very well with the topics defined in RA5. The remaining topics are covered in other research areas. It can be concluded that the We@Sea programme touched those topics that are not only relevant for the Dutch situation but also for the European situation During the execution of the We@Sea program, the international industry did come up with new methods for installation and access; new hardware for this has been built and applied. Worthwhile mentioning here is the fact that at the end of the We@Sea program, several project developers are planning far offshore wind farms which require e.g. on site accommodations that allow technicians to stay overnight, or to make use of dedicated supply vessels to transport crews and spare parts. Also attempts have been made to develop new models for assessing O&M aspects and for improving condition monitoring systems. Some results of the We@Sea program, like new access systems and tools for analysing the O&M aspects, have found their way to the international wind industry and are currently being used by international project developers, operators, and turbine designers. 1.5 The problems In this section background information is given on the different topics, mainly on the problems that existed (or still do exist!) at the start of the We@Sea programme and needed to be solved. A typical lay-out of an offshore wind farm is sketched in Figure 1.1. The wind farms consist of a number of turbines, switch gear and transformers (mostly located within the wind farm) and a substation onshore to feed in the electrical power into the grid. The first wind farms are located in shallow waters at short distances from the shore in order to gain experiences with this new branch of industry. Presently, most offshore wind farms are located at distances typically 8 to 30 RA 5: Installation, operations and maintenance 11

12 km from the shore in water depths of 8 to 20 m. Usually monopiles are being used as a substructure and the turbine towers are mounted to the monopiles by means of transition pieces. The size of an offshore wind farm is 50 to 200 MW and consists of turbines with a rated power of typically 1 to 3 MW. Future wind farms are planned further offshore and will consist of larger units, typically 5 MW, and the total installed capacity will be 200 to 500 MW. New and innovative substructures are presently being developed to enable wind turbines to be sited in deeper waters and to lower the installation costs. Figure 1.1: Typical layout of an offshore wind farm, [1] Installation and Access The logistics to put 6,000 MW in operation for the year 2020 are quite challenging. At the start of the We@Sea program, an average turbine size of about 5 MW seemed to be the optimum, so at least 1,200 turbines with 3600 rotor blades should be transported, installed, maintained and dismantled. Between 2008 and 2020, approximately 100 to 150 turbines have to be annually installed. Turbines are being designed in such a way that most of the assembling work has to be done on the site, including the commissioning. Considering the common weather-window of the 6 summer months, this results in 1 installation per day from 2008 on. As can be seen in Figure 1.2, two vessels were needed to install two turbines a day at the Horns Rev project [2]. An exception to this rule is the Beatrice demonstrator project where the wind turbine was installed as one piece, see Figure 1.3. Figure 1.2: Horns Rev project: by means of the two vessels (Ocean Hanne and Ocean Addy) it was possible to erect two turbines per day if weather conditions are optimal 12 RA 5: Installation, operations and maintenance

13 Figure 1.3: REpower 5M turbine assembled onshore and placed as one piece on the substructure [3] An essential difference with common oil & gas offshore operations concerns the large quantity of individual installations per wind farm project. This allows considering new approaches referring to procedures and equipment, combining existing experiences with new technological concepts. And though the time-span for individual installations will have to be reduced, the large quantity of installations nevertheless will ask for a large offshore spread during a long period, which is a new phenomenon in itself. When optimising the installation process one should be careful not to optimise only one part of the entire chain. For each part different solutions are possible but the most optimal one can only be selected if the entire chain is optimised. For instance different projects may require different foundations or sub structures. From the manufacturing point of view perhaps a lattice tower may look optimal, but if such a piece cannot be transported and limit the speed of installation, maybe a monopile is preferred. Next to that, the logistics should be considered like transportation of wind turbine components to the site, or storage and (partial) assembly in the harbour (see Figure 1.4). Figure 1.4: Amalia wind farm: storage of components in the harbour of Ijmuiden [4] RA 5: Installation, operations and maintenance 13

14 At the start of the programme it was recognised that the entire installation chain had to be considered and also that many parts in this chain had to be improved to enable cost efficient installation of large offshore wind farms Operation and Maintenance Concepts All systems and components within the wind farm need to be maintained. Typically, turbines are being visited twice a year and each visit has a duration of 3 to 5 days. In the future it is the aim to improve the turbine reliability and maintainability and reduce the frequency of preventive maintenance to no more than once a year. In addition to the turbine maintenance, also regular inspections and maintenance are carried out for the sub-structures, the scour protection, the cabling, and the transformer station. During the first year(s) of operation the inspection of substructures, scour protection, and cabling is done typically once a year for almost all turbines. As soon as sufficient confidence is obtained that these components do not degrade rapidly operators may decide to choose longer inspection intervals or to inspect only a sub-set of the total population. The maintenance aspects relevant for offshore wind farms are among others: Reliability of the turbines. As opposed to onshore turbines, turbine manufacturers design their offshore turbines in such a way that the individual components are more reliable and are able to withstand the typical offshore conditions. This is being done by reducing the number of components, choosing components of better quality, applying climate control, using automatic lubrication systems for gearboxes and bearings, etc. Often, the turbine control is modified in such a way that not all single failure lead to a standstill. Making better use of the diagnostics and using redundant sensors can assist in this. Maintainability of the turbines. If offshore turbines fail, maintenance technicians need to access the turbines and carry out maintenance. Especially in case of failures of large components, offshore turbines are being modified to make replacements of large components easy, e.g. by making modular designs, or by building in an internal crane to hoist large components, see for example Figure 1.5. Figure 1.5: Examples of internal cranes in the Siemens 3.6 (left) and Repower 5M (right) turbines Weather conditions. The offshore weather conditions, mainly wind speeds and wave heights, do have a large influence on the O&M procedures of offshore wind farms. The maintenance activities and replacement of large components can only be carried out if the wind speed and wave heights are sufficiently low. Preventive maintenance actions are therefore usually planned in the summer period. If failures occur in the winter season, it may happen that technicians cannot access the turbines for repair actions due to bad weather and this may result in long downtimes and thus high revenue losses. Transportation and access vessels. For the nowadays offshore wind farms, small boats like the Windcat, Fob Lady, or SWATH boats are being used to transfer personnel from the harbour to the turbines. In case of bad weather, also helicopters are being used, see Figure 1.6. RIB s (Rigid Inflatable Boats) are only being used for short distances and during very good weather situations. The access means as presented in Figure 1.6 can 14 RA 5: Installation, operations and maintenance

15 also transport small spare parts. For intermediate sized components like a yaw drive, main bearing, or pitch motor it is often necessary to use a larger vessel for transportation, e.g. a supply vessel. A disadvantage of the access systems available at the start of the We@Sea programme is that they cannot be used at high waves. Wind turbines at the North Sea can be accessed say 60 to 70% of the time only! Improved access systems are required for large scale offshore wind. Figure 1.6: Examples of transportation and access equipment for maintenance technicians; clockwise: Windcat workboat, Fob Lady, helicopter, and SWATH boat Crane ships and Jack-up barges. For replacing large components like the rotor blades, the hub, and the nacelle and in some cases also for components like the gearbox and the generator, it is necessary to hire large crane ships, see Figure 1.7. Figure 1.7: Examples of external cranes for replacement of large components; Jack-up barge ODIN (left) and crane ship Sea Energy (right) RA 5: Installation, operations and maintenance 15

16 At the start of the programme hardly any operational experiences of offshore wind farms were available that could be considered as representative for the future offshore wind farms. Therefore, models had to be used to make an estimate the O&M costs and the downtime. Only very few models for analysing the O&M aspects were available and none of them was validated. Input needed for such models was a.o. failure data, on site weather data, and data from access systems, supply boats, and crane barges. Since these data were not available, only best guesses could be made, assuming optimistic, realistic, and pessimistic scenarios. The models focussed on corrective maintenance mainly as follows: Expected annual O&M costs = Expected annual failure frequency * Repair costs To a lesser extent, also preventive maintenance was considered. Condition based maintenance was hardly considered in such models. Condition based maintenance requires a proper determination and prognosis of the health of the system, so-called condition monitoring. Depending on the type of system and component, a wide variety of condition monitoring methods and techniques is available. For instance offline, one can perform periodic visual inspections, analyse oil samples, or carry out vibration measurements periodically. Another option is to permanently install systems that are dedicated to perform certain measurement and analysis tasks and that automatically report analysis results and events (alarms, warnings) to operators. Such systems are referred to as online condition monitoring systems. Another source of information for health determination is operational data from SCADA systems, e.g. the number of starts and stops, development of temperatures over time, or occurrence of alarms. It should be noted that condition based maintenance mainly makes sense if: (1) the design life of the component is shorter than that of the entire turbine - meaning that repair or replacement of the component is foreseen, but it is not clear when; (2) cost savings are expected compared to scheduled and corrective maintenance only; and (3) it is clear that the dominant cause of failure is indeed wear meaning gradual degradation in time towards the end of the design lifetime; this in contrary to abrupt damage due to extreme events, design or manufacturing errors and the like. Data and information resulting from condition monitoring systems (but also from offline inspections and oil samples!) in fact provide information on the remaining lifetime of the components, e.g. a gearbox. In an ideal situation, condition monitoring results from an offshore wind farm should provide information like which group of turbines need maintenance on the short term, say within a few months from now, and which group of turbines require replacement of e.g. gearboxes on the longer term. If the results of condition monitoring measurements are combined with the repair or replacement costs, an estimate is obtained of the (annual) costs for condition based maintenance. At the start of the We@Sea programme it was clear that the models for analysing O&M aspects had to be improved. The same applied to the quality of the input data for these models. The methods used for collecting and analysing data from offshore wind farms seemed inadequate to extract input data for cost modelling and making decisions for optimal wind farm operation. There was a strong need for improving the methods for data collection. Next to that it was also clear that methods for condition monitoring well known in other branches of industry seemed promising for implementing condition based maintenance in wind energy. However, further R&D was needed to select adequate systems for condition monitoring and to develop new means for diagnostics. Since condition monitoring systems provide large amounts of data, it was needed to automate the data processing and to the extent possible combine it with fault tolerant control Safety Safety issues with respect to offshore wind farms can be separated roughly into two groups: 1. Labour safety 2. Safety for fisheries and shipping Within the We@Sea programme it was recognised that the working practices for labour safety of offshore oil and gas and onshore wind energy were both far from optimal for safe working and cost effective construction and exploitation of an offshore wind farm. It was recognised that 16 RA 5: Installation, operations and maintenance

17 there is a big need for setting up new rules for labour safety and working towards harmonisation of these rules in the different countries. Unfortunately, within the We@Sea programme no new knowledge has been generated on labour safety. This topic will not be discussed here any further and only the safety for fisheries and shipping will be considered. Research results presented during the conference Offshore Wind Farms and the Environment in Billund, Denmark (Sep. 2004) showed that no significant damage to nature and environment had been recorded from studies on Danish offshore wind turbine parks. An important aspect of offshore wind farms that has not been evaluated was the safety aspect of these farms for other users of the space (and the air space above), as well as the associated topic of securing offshore wind farms. Shipping accidents reported the past have shown they were often the result of bad seamanship, rather than of the technical state of the vessel. Several organisations have pleaded for traffic guidance systems for ships that should operate similar to air traffic guidance systems. Such a system is operational for instance in the English Channel and has resulted in a remarkable decrease in shipping accidents. 1.6 How to read the report? This report contains the main results of the research carried out within research area 5. Chapter 2 contains the objectives of the research and a description of the work programme to meet the intended objectives. It also contains an overview of the projects that have been executed. In Chapter 3 the research results are presented per research topic: 5.1 Installation and access 5.2 Operation and maintenance concepts 5.3 Safety Finally in Chapter 4 the needs for further research are proposed. The annexes of this report contain the summaries of the individual projects that have been carried out in research area 5. RA 5: Installation, operations and maintenance 17

18 18 RA 5: Installation, operations and maintenance

19 2. Objectives and Work Programme 2.1 Objectives The objectives of Research area 5 defined initially were to develop and select cost effective and environmental friendly equipment and procedures for installation, maintenance and operation, and dismantling. This included also the development of the required analysis tools and probabilistic cost models. In connection with the other research areas scenarios should be developed for fast and cheap installation, operation at low costs and minimum demand for service and maintenance, and easy dismantling. More specific the following objectives should be reached: cost models for analysing and evaluation of scenarios for installation and maintenance including risk management of bad weather conditions; collection and dissemination of operational experience on installation and maintenance (a.o. reports with experiences, logbooks, databases with failure and reliability data, videos) recommendations for improving e.g. turbines, grid connection, ports, vessels, access systems, testing facilities and infrastructure. A crucial aspect would be the development of ports. In the end, the ports should enable testing facilities to reduce the commissioning time and an adequate infrastructure. Within RA5, four research topics were defined to meet the above mentioned objectives. 1. Installation 2. Procedures and tools for cost-efficient maintenance and repair 3. End of life: dismantling, transportation, recycling 4. Port development 2.2 Definition of research topics After the We@Sea programme was submitted for the first time in 2004, the Ministry of Economic Affairs decided that a strong reduction in the budget was necessary. All research areas had to lower their ambitions and within RA5 the participants and the management team of We@Sea agreed on leaving out topic number 3 End of life: dismantling, transportation, recycling. It was expected that this topic did not have the highest priority at that time. Leaving out other topics would break up the coherence between the different topics. Later on, the subject dismantling became part of RA1. After the We@Sea programme was approved by the Dutch government, the We@Sea partners were invited to submit project proposals that would solve the problems defined in one of the three research topics. This was done for all three topics, however not all projects have been executed as planned. Within the topic Port development only one proposal was submitted but it never has started. On the contrary, during the regular progress meetings of RA5 it was agreed that the topic Safety was not included in the original programme but was considered as relevant. The lack of clear safety regulations for offshore wind farms was expected to become a major bottleneck for the development of new offshore wind farms and We@Sea appeared to be an excellent platform for discussing the issue between all interested parties (developers, operators, turbine manufacturers, certification bodies, and maintenance companies).it was therefore decided to leave out Port development and replace it by the topic Safety. For the latter topic, three project proposals were submitted and executed. During the execution of the We@Sea programme including the RA5 projects it was decided to restructure the three research topics which led to the following classification. 5.1 Installation and access 5.2 Operation and maintenance concepts 5.3 Safety RA 5: Installation, operations and maintenance 19

20 2.3 Installation and Access For the topic Installation and Access, the objective was to develop cost effective equipment and procedures for installing and accessing offshore wind turbines on a large scale. Installation should be faster and cheaper than the present approach and risks due to possible bad weather conditions should be quantifiable. Access systems should be developed to transfer personnel and spare parts even under harsh conditions. The cost effective specialised installation equipment and procedures would be derived from different scenarios studies, considering costs, risks, and time-span. Among others this would lead to functional specifications for installation vessels and recommendations for port development, testing facilities, infrastructure, and assembly facilities. Within RA5.1 the following projects and PhD programme have been executed. Project number Title Project leader Contact person Optimal foundation and installation methods for ffshore wind turbines Ballast Nedam E. v.d. Brug PhD@Sea: Offshore access TU-Delft D. Cerda Salzmann Ampelmann, demonstrator TU-Delft J. van der Tempel 2.4 Operation and Maintenance Concepts The objective for the topic O&M was to develop cost effective scenarios for operation, maintenance and repair. With the new scenarios it was expected that wind farms could have an availability above 95 % at much lower costs. To analyse the O&M aspects and to quantify availability and downtime, new tools had to be developed that were based on techniques already used in other branches of industry (terms like risk based maintenance, asset management, decision support tools, integrated logistic support, are often used for this). Within this topic 5.2, various scenario studies for different wind farms would be executed and based on the results advise would be given to (1) turbine designers to improve the design of their turbines, to (2) operators to develop optimal O&M scenarios, and to (3) developers of vessels to design adequate access systems. Next, the idea was to develop databases with reliability figures, access systems, and weather conditions since these data are difficult to obtain and form the starting point for the scenario studies. To optimise the O&M scenarios it was foreseen that data coming from measurement programs, SCADA 2 systems, and condition monitoring systems, and failure and repair data had to be used to enable condition based maintenance. If needed, also the methods for diagnostics and condition monitoring had to be improved or new methods had to be developed. Finally, research was planned to minimise the number of visits by improving the control system and applying fault tolerant control. Within RA5.2 the following projects and PhD programme have been executed. Project number Title Project leader Contact person PhD@Sea: RAMS for offshore wind farms TU-Delft E. Echevarria Development of O&M Cost Estimator for offshore wind farms ECN L. Rademakers Load monitoring and optimisation of O&M ECN T. Verbruggen Flight Leader (phase 1) ECN T. Obdam Offshore maintenance support tool XEMC-Darwind J. Pasteuning 2 SCADA system = Supervisory Control And Data Acquisition system 20 RA 5: Installation, operations and maintenance

21 2.5 Safety The topic Safety was not foreseen at first instance since it is not really a research topic. However during the execution of the We@Sea project it was recognised more and more that both labour safety and nautical safety had to be considered more in depth. We@Sea appeared to be an excellent platform for discussing the issue between all interested parties. W.r.t. labour safety the aim was to harmonise procedures applied on the one hand to onshore wind energy and on the other hand to offshore oil and gas. Furthermore it was intended to give recommendations to harmonise the labour safety rules in the different European countries. W.r.t. nautical safety it was the aim to identify additional risk for shipping and fishery for the wind farms planned in the North Sea and to come up with mitigating measures if needed. Within RA5.3 the following projects have been executed. Project number Title Project leader Contact person Nautical Safety STC A. Gerritsen Safety near Offshore Wind Farms Imares J.T. van der Wal HSE-topics Lloyd's Register Nederland P. Kuipers 2.6 RL5 Partners At the beginning of the We@Sea program, the following parties showed interest in participating actively in RA5. R&D institutes and consultants: ECN, TU-Delft, KEMA, WMC, Greenpeace Industry: Dynamar, GeniusVos, Ballast Nedam, Fabricom, GustoMSC, Fugro Engineers, Smart Tower, Vestas (and NEG Micon), Multiwind, Rheden Steel, Shell, Siemens At the end of the We@Sea progam it is concluded that the majority of the work has been carried out by ECN and TU-Delft and to a lesser extent by KEMA, Ballast Nedam, GeniusVos, GustoMSC, Fugro, and Shell. STC, Imares and Lloyd s Register joined RA5 later one when the research topic Safety was added. RA 5: Installation, operations and maintenance 21

22 22 RA 5: Installation, operations and maintenance

23 3. Highlights and results In this chapter the main results of the research are presented. The results have been grouped per research topic and will not be discussed per project. Detailed project results can be read in the project reports (see also Section 5 with references). 3.1 RL 5.1 Installation and access Within this research topic, three projects have been executed, one on the development of new foundation methods to reduce the installation costs, and two on the development of the Ampelmann, an innovative system to be placed on a vessel which allows personnel to access turbines even under harsh conditions Installation methods The approach to develop cost effective foundation methods was to carry out parameter studies and cost analyses for different sizes of wind turbines, various water depths and North Sea soil conditions, and different distances to the harbour ( ) [5]. Four foundation concepts have been analysed MW wind turbine with a steel monopile and a transition piece placed at average soil conditions. The option with a drilled concrete monopile was also considered MW wind turbine with a concrete tripod on piles placed at worst case soil conditions (see Figure 3.1) 3. 8,5 MW wind turbine with a gravity based foundation placed at average soil conditions. 4. 8,5 MW wind turbine with a steel jacket structure placed on worst case soil conditions. For the parameter studies a water depth of 21 and 32 metres was assumed. Average soil conditions were defined as a seabed of send which is very dense at 20 metres depth and worst case soil conditions are defined as a seabed of sand and clay. The distances to the shore were varied between 24 metres and 135 metres. The cost analyses showed that the foundation costs increase with increasing complexity of the site (viz. larger turbines, larger wind farms, farther offshore, deeper waters and worse soil conditions). At sites with a small complexity, the steel monopile is the preferred option. Installation can be done with smaller jack-ups (including sufficient crane capacity) and monopiles and transition pieces can be transported by barges. Sites with a medium complexity could use the new concept of the drilled concrete monopile. Installation can be done by using a heavy lift vessel and a vertical drill, the concrete monopiles can be transported floating. At sites with a high complexity there are two possible foundation concepts, either the gravity based foundation, transported by heavy lift vessels, or the steel jackets, transported by barges and installed by jack-ups or transported and installed by self propelled jack-ups with sufficient pay load to carry a number of jackets. Some remarks to the parameter studies should be considered. The calculations are based on new investments, companies already owning equipment or production facilities will favour their concept, mostly being part of an existing business model and last but not least commercial considerations will as well have their impact and lead to shifts in the results presented here. RA 5: Installation, operations and maintenance 23

24 Figure 3.1: Concrete tripod on piles Figure 3.2: Development of foundation costs as a function of the complexity of the site 24 RA 5: Installation, operations and maintenance

25 3.1.2 Ampelmann for offshore access During the start-up phase of the program, a new idea was born at the TU-Delft for accessing wind turbines safely even under harsh weather conditions and it was named the Ampelmann (as easy as crossing the street). It was decided that the fundamental approach for the development had to become part of the We@Sea programme by the definition of a PhD programme ( ) [6]. The scope of the programme ranged from the early conceptual design studies up to and including the first offshore tests at the North Sea. The development of the Ampelmann also included the construction of a demonstrator [7]. The majority of this development was sponsored by the industry and partly by We@Sea ( ). The principle of the Amplemann is a motionless deck on a vessel. The deck is mounted on top of a Stewart platform, a mechanism (often used) for flight simulators) that can provide motions in all six degrees of freedom using six hydraulic cylinders. The Stewart platform is fixed on the ship deck. To keep the transfer deck motionless, a sensor continuously measures the motions of the ship deck. The cylinders of the Stewart platform are controlled in such a way that all ship motions are counteracted, thereby creating a stable and motionless transfer deck. The principle of the Amplemann with a gangway to the wind turbine mast is given in Figure 3.3. Figure 3.3: Artist impression of the Ampelmann system The main driver within the design and development of the Ampelmann system was safety. The Ampelmann safety philosophy led to a fully redundant system design enabling full motion compensation, and thus safe access, in sea states up to 3 meters. During the first phase of the PhD program, scale and dry tests were performed, see Figure 3.4. Finally the design was made into a full-scale prototype with all redundancies thoroughly tested. Finally, the Ampelmann system was taken offshore to prove its function: to provide safe access to an offshore wind turbine. The Ampelmann system is the first system ever to provide a full motion-compensating platform to enable safe offshore access. The system has been thoroughly tested in offshore conditions and proved to provide safe access in sea states up to 3 meters, see Figure 3.5. After its successful test results, the Ampelmann system has become commercially available. RA 5: Installation, operations and maintenance 25

26 Figure 3.4: Dry testing set-up (left) and wet testing (right) Figure 3.5: Offshore demonstration of the Ampelmann The Ampelmann can be installed on any vessel of the SMIT Bronco class and larger. The deck area required is a minimum of 8 x 8m and installation can be completed in minimum of 8 hours, when deck space is abundant or 16 hours when fitting is tighter. During the Demonstrator project, the Ampelmann has been tested for motion compensation up to H s = 1.5m. During transit a sea state was encountered of H s = 2.0m and in simulation mode the Ampelmann platform could compensate these motions with residual motion of less than 20cm on the SMIT Bronco. A measurement campaign on board the supplier Far Splendour, a 70m Dynamic Positioning (DP) vessel for supplying offshore platforms, the motion behaviour in H s = 3.0m was recorded. 3.2 RL 5.2 Operation and maintenance concepts Within RA5.2 models have been developed to estimate the costs of offshore wind farms during the planning phase but also during the operational phase. Such models are being used by project developers and developers of offshore wind turbines to assess the costs and downtimes for various O&M strategies and to assess the impact of design choices. The model developed to 26 RA 5: Installation, operations and maintenance

27 assess the O&M aspects in the planning phase, the ECN O&M Tool [8], [12], [13], offers the possibility to take into account preventive and corrective maintenance. The method to be used in the operational phase (the OMCE = Operation and Maintenance Cost Estimator) makes use of data generated by the wind farm during the first years of operation and allows to consider not only preventive and corrective maintenance but also condition based maintenance [8],[9], [10], and [11]. Both the ECN O&M Tool and the OMCE approach have been used by the wind turbine manufacturer XEMC Darwind to optimise the design of their direct drive 5 MegaWatt offshore turbine [14]. In order to take into account condition based maintenance, it is important to apply adequate diagnostics. Within RA5.2 methods have been developed to monitor the mechanical loads on the main components of single wind turbines [16] and to translate these loads to all wind turbines in an offshore wind farm [17], [18], [19], [20]. Finally, a strategy has been proposed to increase the turbine availability by developing a design methodology that equips wind turbines with reconfigurable capabilities that allow the turbines to continue operating, despite faulty components O&M Cost Modelling At a general level, maintenance can be subdivided into preventive (or calendar based) maintenance, corrective maintenance, and condition based maintenance. 1. Calendar based maintenance: the effort and cost are usually determined by one or two visits per year. After 3 or 4 years the calendar based maintenance costs can be somewhat higher due to e.g. oil changes in gearboxes. The input for calendar based maintenance is determined as the number of repair days with associated costs for labour, equipment, spares, etc. and usually derived from the service manuals with prescribed intervals and procedures. 2. Unplanned corrective maintenance: Costs due to random failures which are more difficult to predict. At the beginning of the wind farm operation the corrective maintenance costs can be somewhat higher than expected due to teething troubles. The input for estimating the effort for unplanned corrective maintenance is based on the failure rates of components and the associated repair costs (mainly equipment, personnel, spares, consumables, revenue losses) 3. Condition based maintenance: It might be that major overhauls have to be carried out, for instance due to unexpected wear out of components designed for the lifetime (e.g. replacement of gearboxes or pitch drives). This type of maintenance is not foreseen initially, but when it has to be carried out during lifetime it generally will be planned, hence it is categorized as condition based maintenance. The input for estimating the effort for condition based maintenance is derived from the number of components that are expected to fail in the next coming year(s) and the associated repair costs. The assessment if components are about to fail should come from a.o. inspection results or condition monitoring measurements. Before the We@Sea programme started, a first version of the ECN O&M Tool was available with a focus only on the corrective maintenance part. The long term average annual O&M costs for offshore wind farms were estimated as: Annual O&M costs = Annual failure frequency of components * Repair costs. The repair costs consist among others of the costs for personnel, equipment, spare parts, and revenue losses. The revenue losses are based on the downtime and take into account the difficulties to access the wind turbines during periods of bad weather. The repair process assumed to estimate the repair costs is given in Figure 3.6. RA 5: Installation, operations and maintenance 27

28 time Event Activities Failure of wind turbine Time To Repair (TTR) T_logistics T_repair T_travel T_wait T_travel T_mission Repair equipment, crew and spare parts ready for take off Weather condition suitable for take off Repair crew arrives at WT and repair is started Repair completed Repair crew arrives at base Alarm is send to central operation office and next responsible engineer is informed about damage; equipment, crew and spare parts are arranged. Due to bad weather conditions the repair team can not depart Repair team travels to damaged WT's Repair is carried out Repair team travels back to base Figure 3.6: Repair process The ECN O&M Tool has been implemented in two MS-Excel sheets: WaitingTime.xls to determine the annual (or seasonal) average waiting time (T_wait) as a function of the mission time (T_mission), and the allowable significant wave height (H s_max ) and wind speed (V w_max ); and CostCal.xls to determine the annual (or seasonal) average downtime and costs. WaitingTime.xls Offshore equipment can be used or repair actions can be carried out if the wind and wave conditions are below certain values. Based on wind and wave data for a selected location the program WaitingTime.xls determines when the weather conditions are suitable for carrying out certain repair actions and calculates the average time one has to wait before a suitable weather window will occur after a failure. The program uses time series with three hourly wind and wave data as input. The program results in second or third order polynomials for the mean value and the standard deviation of the waiting time as a function of the duration of the maintenance activity. In Figure 3.7 an example is given of such a polynomial. Processing of Data Date H s [m] V w [m/s] 600 YM6 IJmuiden munitie stortplaats: Hs=1,5; Vw=12; year Max values: 1, :00 0,35 2, :00 0,4 2, :00 0,41 1, :00 0,37 2, :00 1,39 3, :00 1,29 1, :00 1,31 3, :00 1, :00 1,38 7, :00 1, :00 1,44 6, :00 1,69 6, :00 1,41 8,7 M ean value waiting tim e [hrs] Mean 2nd order 3rd order Mission time [hrs] Figure 3.7: Example of determining relationship between average waiting time and mission time The example represents the annual average waiting time as a function of the mission time at the location "IJmuiden Munitiestortplaats" and is based on 11 years of measured data. The mission can be carried out up to a significant wave height of Hs = 1.5 m and a wind speed of Vw = 12 m/s. Similar polynomials can also be generated per season and for different weather limits Hs and Vw. 28 RA 5: Installation, operations and maintenance

29 CostCal.xls The program CostCal.xls is being used to determine the long term annual (or seasonal) costs for O&M and the associated downtime. The program initially focused on unplanned corrective maintenance but during the program, a module has been included to incorporate also preventive maintenance actions. The program uses among others the following input: 1. weather windows and waiting time polynomials as generated with WaitingTime.xls; 2. wind turbine and wind farm information such as number of wind turbines, capacity factor of the wind farm, investment costs of turbines, costs of technicians, length of working day, etc.; 3. failure behaviour of the turbines and the repair actions which are foreseen; 4. characteristic values of access systems (weather limits, costs, mobilisation time, etc.); 5. preventive maintenance actions with costs and long term fixed costs like annual contracts. During the execution of the We@Sea program, the ECN O&M Tool has been validated by Germanischer Lloyd and 18 licenses of the software have been sold worldwide to utilities, project developers, and wind turbine manufacturers and more than 25 offshore wind farms have been analysed with this tool. Based on the feedback of the users, several updates have been made and sent to the users. The users of the tool first set up a baseline scenario. Based on the output they identify the weak spots in the proposed O&M strategy and come up with alternatives. Parameters like the capabilities of vessels, cranes, or reliability of components can be changed, helicopters can be used instead of small access boats, or hotel ships can be considered to avoid travelling to the turbines on a daily basis. This process is sketched in Figure 3.8. Figure 3.8: Schematic representation of the ECN O&M Tool for determining the O&M costs and downtime of an offshore wind farm. RA 5: Installation, operations and maintenance 29

30 Most effort however w.r.t. to the development and implementation of cost estimation tools has been put on the development of the Operation and Maintenance Cost Estimator (OMCE). The entire OMCE concept was partly based on a question of an operator: Can you tell us how many gearboxes we need to replace in the next two years and how much will it cost? It was clear that this question was very relevant, but very difficult to answer and therefore ECN initiated the idea of developing the OMCE as a tool that could be used by operators of large offshore wind farms to determine the costs of O&M for the next coming years. The OMCE is intended to be used during the operational phase of a wind farm, e.g. to make reservations for future O&M budgets or to estimate the O&M needs at the end of the warranty period. The structure of the OMCE is presented in Figure Failure rate INFO -Repair strategy Unplanned Corrective Maintenance Raw data Event list Structured data Calendar Based Maintenance OMCE Calculator Annual O&M Costs -Time to failure INFO (Repair strategy) Condition Based Maintenance Figure 3.9: OMCE concept showing the data flow from raw data to estimated O&M costs The OMCE developed within the We@Sea programme roughly consists of two parts: (1) the OMCE Building Blocks to process the operational data, and (2) the OMCE Calculator to estimate the future costs. 1) The OMCE requires feedback of operational data of a specific wind farm under consideration, such as O&M data, SCADA data, data from measurement campaigns, and data from condition monitoring programs. Data about failures, repair actions, the vessel usage, spare parts, and weather conditions are analysed to estimate the effort for unplanned corrective maintenance. Data from condition monitoring and load measurements are analysed to estimate the effort for condition based maintenance. For this purpose four so called OMCE Building Blocks (BB) have been specified, each covering a specific data set. - BB Operation and Maintenance: to process operation and maintenance data and to determine failure rates of components and successive repair actions. - BB Logistics: to determine which equipment, crews, and spare parts are needed for certain maintenance and repair actions. - BB Loads & Lifetime: to determine the loads on mechanical components as a basis to prioritise inspections for the most heavily loaded parts (see also Section 3.2.2). 30 RA 5: Installation, operations and maintenance

31 - BB Health Monitoring: to determine the time to failure and repair, based on among others results of condition monitoring programs, inspections. The main objective of these Building Blocks in fact is twofold. First of all the Building Blocks should generated (updated) input parameters for the OMCE Calculator and second, the Building Blocks should provide information in a more general way that can be used by operators to assess the adequacy of the applied O&M strategy. 2) The OMCE Calculator will be used to assess the O&M cost for the coming period of say the next 1, 2 or 5 years, based on the results of the BB s. The OMCE calculator considers three types of maintenance: (1) unplanned corrective maintenance (2) condition based maintenance, and (3) calendar based maintenance. Initially it was foreseen that the data generated by the wind farm could be processed immediately by the Building Blocks; however it was concluded that the format in which the data are available require an intermediate step and need to be structured first. In Figure 3.9 this is indicated as the conversion from raw data (stored in many different files, reports, and databases) to the event list in which the different events (e.g. a component failure) are linked to the individual maintenance actions (e.g. resets, repairs, replacements) and the usage of personnel, equipment and spares. Within the We@Sea program, the following project results have been achieved. the structure of the OMCE Calculator as shown in Figure 3.9 has been developed and described; the functional specifications of the OMCE Calculator have been described; the technical requirements of the OMCE Calculator have been described; the model description of the OMCE Calculator has been reported the technical specifications of the OMCE Building Blocks have been determined and described; the specifications for the Event List to structure raw data before they can be processed by the BB s have been described. The actual programming of the OMCE Calculator was outside the scope of the project but is foreseen within the D OWES (Dutch Offshore Wind Energy Services, project. The first demo version of the OMCE Calculator will become available mid Some results of the OMCE are being used already by wind farm operators. For instance: the knowledge to collect, analyse, and report wind farm data in accordance with the specifications of the OMCE Building Blocks has been applied for several wind farms. Early 2009, when the OMCE developments within We@Sea were nearly completed, XEMC Darwind, developer of a direct drive 5 MegaWatt offshore wind turbine (see Figure 3.10) started to use the OMCE approach to set up a framework for data collection to ensure that the collected data can be used for cost modelling purposes. Also, the ECN O&M Tool has been used by XEMC Darwind to optimise the turbine design w.r.t. to offshore O&M. XEMC Darwind aims at only one visit per turbine per year for both planned and unplanned maintenance. XEMC Darwind and ECN carried out the following work packages within the We@Sea program. 1. Structured identification of technical risks in an offshore wind turbine design and proposals for modifications addressing these risks. 2. Optimisation of the turbine design by minimising the expected kwh-costs using the ECN O&M tool. 3. Adaptation of a dedicated offshore maintenance support tool (OMCE) for repair and service planning during the operational phase of a wind farm. RA 5: Installation, operations and maintenance 31

32 Figure 3.10: Artist impression of the of the 5MW XEMC Darwind DD The risk assessment programme to identify, prioritise, and address technical risks proved to be very fruitful. The large number of issues found underlines that generally even a strong existing wind turbine concept needs re-evaluation for offshore purposes. Several design improvements have been realised -at relatively low costs- which will result in increased availability. The brainstorm sessions of the system-, design-, process- and supplier FMECA (Failure Mode, Effects, and Criticality Analysis) caused a considerable increase of knowledge on offshore wind turbine design. With respect to structuring the risk assessment programme the timing of the FMECA and the correct ranking of the failures played an important role. The FMECA should be performed early in the design path, after finishing the conceptual phase, but far before the detailed design is finalised. When the severity is estimated during the ranking of the failures, it is important that offshore conditions are taken into account. Regarding the addressing of the risks identified, it should be noted that any known failure mode, which may possibly lead to a major intervention (i.e. offshore Jack-up crane needed), should be never left unaddressed because of high repair costs and loss of yield involved. Modifications become especially necessary when the system analysed contains unproven technology, since the occurrence rate of the failure is unknown. The solution here is either a re-design to a degree that the failure mode is eliminated, or the development of a back-up solution to be ready and available when the failure occurs. Some risks of the turbine design are related to long lead items, for example the main bearing, and in those cases, a change of the design is not the most preferred option for reducing a technical risk. The risk can be mitigated when the detection of the failure is improved. Condition monitoring programs in which measurements are performed on vibration, temperatures, stresses, etc. help to understand and identify trends in an early phase of the failure. Subsequently, appropriate measures can be taken to delay wind turbine shutdown, for example by switching to reduced power generation mode. 2. Design optimisation can also be carried out by quantifying the effect of the design options on lifetime costs. In this way subjectivity is avoided and decisions can be made on a more solid basis. Very often assumptions will have to be made on cost levels when building the 32 RA 5: Installation, operations and maintenance

33 baseline model. Therefore optimisations of design features will automatically be based on comparative analysis and not on absolute costs. Since the purpose is to choose between design options, a comparative analysis meets the objective. Concerning the method followed for cost optimisation, the ECN O&M Tool proved to be a powerful instrument for estimating the lifetime O&M costs of an offshore wind farm. Especially the process of generating input to the O&M cost model proved to be as valuable as the final result because an offshore way of thinking has to be adopted. After the extensive baseline model is filled in, a new calculation can be done almost instantly by changing only relevant to re-design parameters. Sensitivity analysis on crucial results is strongly advisable in order to find what the relation between certain factors and resulting costs is. It was found that certain design modifications can reduce O&M costs and revenue losses significantly, in the order of 1 to 5%. A considerable amount of improvements related to the possibility of postponing repairs until the next already scheduled maintenance period, were also proposed based on the results. 3. With respect to maintenance optimisation during operation of a wind farm using OMCE, it was found that a structured collection of operational data is crucial for successful implementation. The data collection should be based upon a well defined Functional Breakdown Structure (FBS). The breakdown method as proposed by VGB PowerTech [15] Reference Designation System for Power Plants (RDS-PP) Application Explanations for Wind power Plants was chosen because it is in line with international developments. The Functional Breakdown Structure should be used to arrange the maintenance data (valuable service technician s input) with sufficient possibilities for automatic processing. Consequently the major part of information to be stored has to be pre-defined. The processed maintenance data can be used to build an event list which serves as an input to the OMCE Calculator but also for engineering optimisation purposes Low cost solutions for load monitoring In order to apply condition based maintenance as a means to optimise the O&M strategy of an offshore wind farm, it is necessary to apply adequate diagnostics. Within the We@Sea program, the emphasis of these diagnostics has been on low cost solutions to measure (1) the loads on wind turbine components, mainly the rotor blades and the tower, with optical fibres instead of copper strain gauges (Fibre Optic Blade Monitoring) and (2) a method to convert the loads measured at turbines strategically chosen within an offshore wind farm to the loads of all other turbines within the wind farm (Flight Leader concept). Fibre Optic Blade Monitoring has been selected as a promising method for monitoring the fatigue and extreme loads in rotor blades (and the tower). As opposed to the frequently used copper strain gauges, the optical sensor should have the following advantages: more reliable, less drift and more stability of the signal, and longer lifetimes. A system for fibre optic blade monitoring consists of sensors in the blade (and tower), a read out unit (interrogator) in the hub, a measurement system in the nacelle or tower bottom, software for data processing and generation of key figures for operators, and a connection to the internet for data transfer, alarm messages and remote control, see Figure The optical sensors are based on Bragg grating techniques. Optical sensors based on Bragg grating techniques are being used for a long time already but with limited success in wind turbines. The measurement techniques also show some disadvantages and the We@Sea programme was intended to overcome most of these issues. The most important ones are given below. - A practical one is that people working in this area are not familiar with optical components, which imply that handling should not require additional skills. - The sensors should be compensated for temperature effects. - The reliability, accuracy, temperature sensitivity, and robustness of most interrogators are insufficient for operating in wind turbines for a long period of time. - The total measurement chain is sensitive to temperature effects, polarisation, hysteresis, and attenuation. - The measurement system generates large amounts of (sometimes erroneous!) data that should be reduced to key figures that can easily interpreted by wind farm operators. - The total system is too expensive. RA 5: Installation, operations and maintenance 33

34 Figure 3.11: General layout of a blade monitoring system based on optical sensors for strain measurements Within the We@Sea program, the following has been achieved. 1) An interrogator has been selected that met most of the specifications for wind turbine applications listed below: Strain resolution : 1 Strain accuracy / stability : better than 5 Maximum strain level : Frequency : >16 Hz Sensors per blade : 4 stain sensors / 4 temperature sensors An inventory has been made of available interrogators followed by a first selection made in a desk study. This resulted in three candidates, which matched with the user requirements for a large extend. These interrogators have been tested at ECN which resulted in the choice of a device which meet the requirements for a large extend and which is also considered as suitable for further experiments and development tests for sensors. The accuracy and resolution sufficient for the application, but also the selected showed strong sensitivity for polarization. The price level is still above target. 2) For the sensors, also a desk study has been performed. This showed that no sensor was compliant with the user requirements. The strain is often not measured directly, while installation and replacement were difficult. However, this desk study also resulted in a new idea with respect to a new approach. A sensor assembly was designed and built. This sensor is: easy to install (typically: instrumentation of all three blade roots can be done in one day); easy to replace (plug-and-play); no on-site calibration and special knowledge required. Endurance tests for the blade-sensor bonding have been completed successfully and the entire sensor assembly is now being tested, see Figure A request for patent is presently pending. 34 RA 5: Installation, operations and maintenance

35 Figure 3.12: Endurance testing of the sensor assembly at WMC 3) Specifications for the data analysis software are ready and the programming is ongoing within the D OWES program. The overall data structure for the blade monitoring software is depicted in Figure 3.13; two main processes can be distinguished. - An on-line module which continuously collects and processes the relevant data from the measurement system and subsequently stores the results in a database. - Reporting module, which provides online access to the database and which generates periodic reports, e.g. frequency plots, load spectra, warnings, extreme loads, scatter plots with statistical values, and capture matrices. The software is designed in such a way that it can be used not only for processing data coming from optical measurement systems but in fact for all sorts of load measurement systems as long as the data acquisition system meets some minimum requirements w.r.t. the file format. It is expected that the sensor assembly with the interrogator and measurement system will be build in an operating turbine end of The software for data analyses will be tested offline also end of 2010 and online early Endurance tests of the complete system will be finished end of Both sensor assembly and data analysis software will become commercially available early RA 5: Installation, operations and maintenance 35

36 Figure 3.13: Overall structure data processing and information retrieval Once the loads on a single wind turbine can be measured accurately over a long period of time (e.g. by means of optical fibres) the Flight Leader concept enables to transfer these loads to the loads acting on all wind turbines within an offshore wind farm. Instead of equipping all turbines with mechanical load measurements only a few reference (Flight Leader) turbines are extensively instrumented. Using data from these turbines relations between standard SCADA parameters and load indicators, which are representative for the ageing or degradation of a certain wind turbine component, are established. Once such relationships are determined these can be combined with SCADA data from the other turbines in the wind farm. This enables the determination of the accumulated loading on all turbines in the farm. The principle is presented in Figure Figure 3.14 Illustration of the flight leader concept; the load measurements performed on the flight leader turbines (indicated by the red circles) are used to establish relations between load indicators and standard SCADA parameters; these relations are combined with the SCADA data from all other turbines in the wind farm in order to estimate the accumulated loading of all turbines in the farm. 36 RA 5: Installation, operations and maintenance

37 A demo version of a software model (see Figure 3.15), making use of neural networks, has been developed in MATLAB. The software includes all aspects of the Flight Leader concept and is intended to be used by operators of offshore wind farms and can be applied to process the SCADA data and mechanical load measurements from an (offshore) wind farm. The main output of the model is a comparison of the accumulated mechanical loading of all turbines in the offshore wind farm. This information can subsequently be used to optimise O&M strategies, for example by prioritising the inspection or replacement of certain components on the heavier loaded turbines. Empirical DB module SCADA data Mechanical load measurements Categorisation module Pre-processing Characteristic load module Accumulated load & output module Generation of output Data input Simulated DB Figure 3.15 General structure for the flight leader computer model. The developed software model has been implemented at the ECN Wind turbine Test location Wieringermeer (EWTW) with 5 Nordex N80 turbines. Several analyses have been performed where the main goal of the research was to assess if the Flight Leader principle can be accurately applied in practice. In addition to this the research had the goal to determine what method is best used for characterising the relation between SCADA parameters and load indicator, which is essentially the core of the Flight Leader principle. Furthermore it has been tried to identify the contributors to insecurities in the predictions of the flight leader software. In an example is shown where turbine 2 was used as the fully instrumented reference turbine. The blade root flapwise bending moments have been derived in accordance with the Flight Leader concept and it can be concluded that turbine 4 has observed 8 % higher fatigue loads. Figure 3.16 Load accumulation of all five turbines relative to load accumulation of turbine 2 for the blade root flapwise bending moment RA 5: Installation, operations and maintenance 37

38 Finally a second analysis has been performed with the goal of evaluating whether the proposed Flight Leader principle can also be applied to make accurate estimations of the load accumulation for all turbines in a large offshore wind farm, where wave-induced loading and wake effects play an important role. For this purpose use has been made of data from the Offshore Wind farm Egmond aan Zee (OWEZ). At present the Flight Leader is ready to use and can be implemented online at an offshore wind farm. It should be noted here that the Flight Leader research in fact comprises the research behind the OMCE Building Block as described in Section Design for Redundancy and Fault Tolerant Control Within the PhD programme RAMS (=Reliability, Availability, Maintainability, and Servicability) for Offshore Wind Turbines a strategy has been to increase the availability of offshore wind turbines by developing a design methodology that equips the wind turbines with reconfigurable capabilities that allow them to maintain operation despite faulty components, even at reduced capacity if necessary, until maintenance can arrive. That is aiming to achieve fault tolerance at not significant additional cost by designing for reconfiguration. The proposed methodology is a new design philosophy and it is based on two concepts: functional redundancies and reconfiguration. When a component is placed in the design it has a function to accomplish. However, a component has the potential of behaving in more than one way. These potential behaviours can be used to perform functions that are supposed to be performed by a faulty component. A functional redundancy exists when these potential behaviours match the required unperformed function. Re-configuration is intended to be a builtin capability to be used as a repair strategy, based on these potential functional redundancies provided by the components. The proposed design philosophy consists of four steps: 1. the first step is to build an FBS (Functional Behaviour State) model; 2. the second step is to identify the critical functions based on a reliability analysis; 3. the third step is to find potential functional redundancies targeting those functions; and 4. the fourth and last step is to design allowing re-configurability of the wind turbine to continue operation despite critical faults. The proposed design philosophy has been applied to several functions of different sub systems of a wind turbine, e.g. power regulation, braking, and yawing. Within the wind farm, there are different levels at which this functional redundancy concept can be applied. The possible solutions can range from using information from adjacent wind turbines to setting up different operational modes. However, up to now, the design philosophy did not lead to an example of a design modification of a wind turbine or a sub system. It is expected that on the longer term the re-configuration concept to achieve self-maintained wind turbines is an interesting and promising approach to reduce downtime, failure events, maintenance visits, and to maintain energy output possibly at reduced rate until the next scheduled maintenance. Possibly, some results are not ideal for permanent practice or long periods of operation but in some cases they represent solutions to particular failure situations. It is expected that the classical design philosophy changes by finding and incorporating functional redundancies. The expected outcome is a more flexible and robust system, allowing alternative configurations which are potential solutions to overpass faults. 3.3 RL 5.3 Safety The research on safety has focussed on the impact of the construction and operation of wind farms on shipping and fishery. Offshore wind farms influence safety at se, e.g. by occupying space that was previously freely available for shipping and fisheries. Furthermore wind turbines are obstacles with which ships can collide. The dangers here are mutual; both the wind turbines can suffer damages from such an incident, as well as the ships. In case a ship is carrying a hazardous cargo, e.g. chemicals or oil, there is a threat to the environment as well. 38 RA 5: Installation, operations and maintenance

39 The study had the character of an integrated safety study. It answers some of the questions towards the risk involved with offshore wind farms for shipping and how to mitigate this risk. By furthering our knowledge of the above topics the study should also be helpful towards the issue of insurance of offshore wind farms and where accountability and liability are likely to be placed in case of an accident. The aim of the study was to decrease the risk of incidents with or as a result of the offshore wind farms. The proposal was to perform an integrated study of external threats for offshore wind farms, where the following elements were distinguished: probability * impact of damage as a result of shipping, including ships adrift and cargo items lost at sea; securing of offshore wind farms with transponders (to be developed), including storage and retrieval of transponder data. This topic also includes the use of radar to record vessel movement; how to address the issue of interference of OWP and air traffic safety. An inventory of possible situations (What if scenarios) has been build. An important tool that was used for this, were stakeholder interviews. The risk aspects of the situations were developed further to allow a preliminary assessment of (possible) cumulative safety effects of offshore wind turbine parks. Next, the developed risk aspects are used in preparation for a scenario analysis using the Process Hazard Analysis method (or PHA). In this method the situation with its risk aspects forms the start of a chain: Situation (or What if ) Hazard Consequences Safeguards Recommendations. A number of models suited for performing a more in-depth study of safety issue s have been identified. Possible safeguards and available mitigation measures are elaborated. The study was led by IMARES and partners involved in the study were TNO-MEP, TNO-FEL, Stichting De Noordzee, and Ecofys. The study has resulted in a single deliverable [22]. Following the proposed methods an inventory of possible situations (What if ) has been build, based primarily on a series of stakeholder interviews. Amongst the stakeholders that were interviewed were authorities, non-governmental organizations, operators and mariners. The risk aspects of the situations were developed further to allow a preliminary assessment of (possible) cumulative safety effects of offshore wind turbine parks. Especially during the construction phase cumulative effects are expected. These effects are associated with the high number of ship movements of traffic to and from the site that will need to cross shipping lanes. For the rest effects are expected to behave linearly. The Process Hazard Analysis method showed that the events for each (sub)scenario could be largely split up into three categories: shipping, air transport and technical failure of equipment within the wind farm (including the cable corridor). The most frequent events identified by the analyses are an increase in shipping traffic during construction of the wind farm, an increase in traffic crossing shipping lanes and a reduction in manoeuvring space available for shipping in the vicinity of the offshore wind turbine park. As far as accidents an increased risk for collisions, both between ships and wind turbines as well as between ships is noted. The consequences of possible accidents related to offshore wind farms were listed and showed that the release of hazardous substances and damage to the wind turbines are the most common events according to the scenario analysis. Loss of cargo, loss of navigational capabilities by a vessel and the loss of one or two vessels are also possible consequences. The scenario analysis also listed a number of mitigation measures (safeguards). The measures mostly frequently identified included measures for properly signalling the presence of wind farms to shipping, adjustment to the procedure for approaching a port (mainly in cases where a wind farm is located close to crossing shipping lanes) and the introduction of procedures to regulate the traffic associated with the wind farm. This final measure would have to include procedures both for the construction phase as well as for maintenance related traffic during normal operation. Finally a frequently suggested measure to prevent technical failures to the equipment and infrastructure of a wind farm is to have good and preventive maintenance procedures in place. As far as air traffic incidents are concerned, no mitigation measures are suggested. A number of models suited for performing a more in-depth study of safety issues were identified. Possible safeguards or available mitigation measures are elaborated. Here much attention is given to the possibilities of integrating systems as AIS (Automatic Identification System) and RA 5: Installation, operations and maintenance 39

40 radar systems. Conclusions were drawn and recommendations were formulated. An important point here is the realisation that the main risks are different during the construction phase from once the wind turbine park is operating normally. Most safeguards are geared towards ensuring that the wind turbine park will be detected by ships, including fishing vessels, by introducing sufficient signalling systems. Especially under foul weather conditions these combined signals need to complement each other. 40 RA 5: Installation, operations and maintenance

41 4. Gaps and future research Clearly, research area 5 of the We@Sea programme has significantly contributed to the generation and dissemination of knowledge to make the installation, operation, and maintenance of offshore wind farms more cost effective. However, not all issues have been solved. On the one hand due to the effect that not all objectives of the original We@Sea programme have been met, on the other hand not all issues were addressed initially and during the execution of the programme new issues showed up. As mentioned already in Section 2.2, the first reason why the original We@Sea objectives have not been met is because less subsidies were obtained than asked for. The topic End of life: dismantling, transportation, and recycling has not been considered further on but is still a topic for further R&D. Even outside the We@Sea programme this topic is not really considered seriously. The most probable cause for the lack of interest is that all parties are focussing on low cost installation and operation. The dismantling of wind farms is too far in the future. The second reason why the original objectives have not all been met is that, although the We@Sea consortium consisted of industrial parties mainly and only few R&D institutes, most project proposals were submitted by R&D institutes and also the majority of the work has been executed by the R&D institutes. In general it can be concluded that if the industrial parties would have had more involvement, the issues labour safety, installation and access, and port development could have been treated in more detail. The emphasis of the work related to these topics requires a practical approach, rather than the scientific approached of the R&D institutes. For the topic Operation and maintenance concepts the emphasis was indeed on the scientific approach and it can be concluded that the We@Sea programme has provided relevant results to minimise O&M costs and downtimes and that these results have found their way to the practical implementation for wind farm development and operation. 4.1 Installation and access Within RA 5.1, an attempt has been made to develop an approach for installing wind turbines in large series in a cost effective manner. However, during the execution of this task it turned out that the problem was more complex than originally foreseen. The result of the project was a proposal for an optimal support structure for certain combinations of: water depth, soil conditions, and distances to the shore. However, RA 5.1 did not lead to optimised installation vessels, new foundation designs, or new installation methods (e.g. assembling onshore vs. assembling offshore). These issues are still to be investigated and crucial for further reducing the costs to install offshore wind farms. Furthermore, issues like floating offshore turbines for deep waters has not been treated at all but could become of relevance for far offshore applications. 4.2 Operation and Maintenance The emphasis has been put on the development of a new means for turbine access and egress (Ampelmann), development of tools for estimating O&M costs of offshore wind farms and to a lesser extent on the development of low cost methods for load monitoring and fault tolerant control. Within the We@Sea programme, no attention has been paid to actually optimising the O&M procedures, e.g. by using operational experiences and using the models. Especially for far offshore wind farms where the use of Floatels, mother ships, or a permanent basis is considered to reduce the travel time for personnel, a Smart Maintenance approach is required. In the next coming years time and effort should be spent on actually implementing the We@Sea results. Data should be collected, analysed and fed back, and simulations should be carried out to determine the most optimal strategies for high availability at low costs. Low cost load monitoring methods should be applied in addition to the other methods of health monitoring that are already common practice to change from preventive and corrective maintenance to RA 5: Installation, operations and maintenance 41

42 condition based maintenance. Furthermore, new and low cost means for access and egress under harsh conditions are still required. 4.3 Safety The topic Nautical Safety has been addressed extensively within the programme. Most of the risks for shipping and fishery were identified and quantified and mitigating measures were proposed. It is recommended to indeed apply these measures and to verify if these measures indeed lead to a better safety. However the HSE (Health, Safety, and Environment) received only very limited attention within the We@Sea programme. HSE during installation, operation, and maintenance is a very important issue. Adaptation of HSE procedures from oil and gas may lead to too strict (and thus expensive) working procedures; onshore wind energy procedures are possibly not strict enough. Moreover, the HSE rules may differ from country to country and harmonisation is required. Perhaps HSE is not really an R&D topic but a harmonised approach will have an impact on the technical solutions for installation of turbines, access and egress, and maintenance. 42 RA 5: Installation, operations and maintenance

43 5. References [1] [2] [3] [4] [5] E. van de Brug, et al: Optimal Integrated Combination of Foundation Concept and Installation Method Executive Summary- ; Ballast Nedam, October 2009 (to be published) [6] D.J. Cerda Salzmann and J. van der Tempel: Ampelmann The New Offshore Access System ; Delft University of Technology DUWIND, Paper presented at the Offshore Wind Energy Conference 2009, Stockholm [7] J. van der Tempel, D.J. Cerda Salzmann, J.M.L. Koch, F.W.B. Gerner, and A. Göbel: Ampelmann Demonstrator: Completion of a motion compensation platform for offshore access and review of different applications ; Final Paper, Delft University of Technology DUWIND Offshore Engineering, July 2008 [8] L.W.M.M. Rademakers, H. Braam, T.S. Obdam, R.P. v.d. Pieterman; Operation and Maintenance Cost Estimator (OMCE)Final Report ; ECN-E , October 2009 [9] L.W.M.M. Rademakers, H. Braam, T.S. Obdam, R.P. v.d. Pieterman; Operation and Maintenance Cost Estimator (OMCE) to Estimate the Future Costs of Offshore Wind Farms ; ECN-M , Paper presented at the Offshore Wind Energy Conference 2009, Stockholm [10] L.W.M.M. Rademakers, H. Braam, T.S. Obdam; Estimating Costs of Operation and Maintenance of Offshore Wind Farms ; ECN-M , Paper presented at the EWEC 2008, Brussels [11] T.S. Obdam, H. Braam, L.W.M.M Rademakers, P.J. Eecen; Estimating Costs of Operation and Maintenance of Offshore Wind Farms ; ECN-M ; Paper presented at the European Offshore Wind Energy Conference 2007, Berlin, Germany [12] Rademakers, L.W.M.M.; Braam, H.; Obdam, T.S.; Frohböse, P.; Kruse, N.: Tools for Estimating Operation and Maintenance Costs of Offshore Wind Farms: Stat-of-the-art ; ECN-M ; Paper presented at the EWEC 2008, Brussels, Belgium [13] L.W.M.M. Rademakers, Braam, H, M.B. Zaaijer, and G.J.W. van Bussel: "Assessment and Optimisation of Operation and Maintenance of Offshore Wind Turbines", Proceedings EWEC 2003, Madrid June 2003 [14] O. Brug, E. Koutoulakos, R. V.d. Pieterman, and T. Verbruggen: Dedicated Offshore maintenance Support Tools for XEMC Darwind Wind Turbines ; Document #: DWD_Wpe70001_We@Sea.Report, XEMC Darwind, [15] VGB Power Tech: Guideline Reference Designation System for Power Plants, RDS- PP; Application Explanations for Wind Power Plants ; VGB 116 D2, First edition 2007 [16] T.W. Verbruggen: Load Monitoring for Wind Turbines; Fibre Optic Sensing and Data Processing ; ECN-E , October 2009 [17] Obdam, T.S., Rademakers, L.W.M.M., Braam, H.l.: Flight Leader Concept for Wind Farm Loading Counting: Final Report ; ECN-E , October 2009 [18] Obdam, T.S. et al.: Flight Leader Concept for Wind Farm Loading Counting: Offshore evaluation ; ECN-M , Paper presented at the EOW 2009, Stockholm, Sweden [19] Obdam, T.S.; Rademakers, L.W.M.M.; Braam, H.: Flight Leader Concept for Wind Farm Loading Counting and Performance Assessment ; ECN-M ; Paper presented at the EWEC 2009, Marseille, France RA 5: Installation, operations and maintenance 43

44 [20] T.S. Obdam, L.W.M.M. Rademakers, H. Braam; Flight Leader Concept for Wind Farm Load Counting: First Offshore Implementation, ECN-M Paper presented at the OWEMES conference, Brindisi, May 21-23th, [21] E. Echavarria: RAMS for Offshore Wind Farms ; PhD Thesis, Delft University of Technology DUWIND; to be published [22] J.A. van Dalfsen et al: Veiligheid en beveiliging van offshore Windturbineparken; Integrale veiligheidstudie ; IMARES report C072/09, September 2009 (in Dutch, with an English summary) 44 RA 5: Installation, operations and maintenance

45 Project Summaries Optimal integrated combination of foundation concept and installation method Organisation Ballast Nedam, The Netherlands Report title Optimal integrated combination of foundation concept and installation method Author E. van de Brug Date October 2009 Being a Dutch R&D project this study primarily focuses at specific conditions that occur in the section of the North Sea continental shelf that belongs to the Netherlands. However, some conclusions might offer added value for other offshore sites and countries as well. The first phase of the project focuses on alternative foundation concepts. In the second phase these foundation concepts were combined with installation scenarios, based upon different types of equipment and a range of distances between the onshore basis and the offshore site. Basic data During the first phase of the R&D project, site conditions were investigated as part of a desktop study. As one main outcome both an average site (Case A) and worst site (Case B) were selected for the Dutch part of the North Sea. Site conditions Case A Case B Water depth 21,0 m 32,0 m Wave (Hm0 / period) 7.0 m / 10.0 s 9.5 m / 10.5 s Soil Profile Average Worst case Case A. Average Soil Profile Depth below Soil type (kn/m 3 ) (degree) Cu (kpa) seabed and up Sand, silt Sand, dense Sand, very dense Case B. Worst Soil Profile Depth below Soil type (kn/m 3 ) (degree) Cu (kpa) seabed and up Sand Clay Sand, dense Apart from actual site conditions the single most important input variable is the wind turbine itself. A state of the art 5.0 MW turbine and a possible future 8.5 MW turbine size were both defined. 5.0 MW Turbine Rotor diameter: 126m Hub Height: 90m Top Head Mass (nacelle + blades): 350 tonnes Tower mass: 500 tonnes RA 5: Installation, operations and maintenance 45

46 8.5 MW Turbine Rotor diameter: 160 m Hub Height: 110m Top Head Mass (nacelle + blades): 800 tonnes Tower mass: 1000 tonnes Wind loads are both based upon and derived from existing turbine models. As a next step distances to four main supply harbours were selected based upon the existing Dutch offshore wind developments. Cluster Harbour Min. Distance (km) Max. Distance (km) Average Distance (km) Noord1 Eemshaven Noord 2 Eemshaven Midden IJmuiden Zuid Rotterdam Phase One: Foundation concepts The combination of the two pre-selected site conditions and the two different turbine sizes resulted into four concept variations (I IV): A. A-WEC1: 5.0 MW wind turbine facing average site conditions; B. A-WEC2: 8.5 MW wind turbine facing average site conditions; C. B-WEC1: 5.0 MW wind turbine facing worst-case site conditions; D. B-WEC2: 8.5 MW wind turbine facing worst-case site conditions. Hereafter four different concepts were preselected and engineered: A. A-WEC1: 5.0 MW wind turbine facing average site conditions Steel monopile (ø 7000 x 63 /90; 680 tonnes) with Transition piece (ø 7300 x 65; 360 tons); C. B-WEC1: 5.0 MW wind turbine facing worst-case site conditions Concrete tripod on piles (12000 x x 12000; 7,875 tonnes); B. A-WEC2: 8.5 MW wind turbine facing average site conditions Concrete gravity based foundations (33000 x 33000; 7,750 tonnes) D. B-WEC2: 8.5 MW wind turbine with worst-case site conditions Steel jacket structure (36000 x 36000; 650 tonnes) Based upon market developments, a steady growth in the offering of 5.0 MW turbines and foundations and new foundation concepts (i.e. Drilled Concrete Monopile), the foundation list has been expanded. This enlargement both concerns new concepts like the Drilled Concrete Monopile as well as changes in dimensions of the above-mentioned foundation type variations. The list with alternative solutions is shown below and includes estimates on individual foundation manufacturing prices. It should be noted that the cost estimate calculations were conducted for (comparative) evaluating purpose only. 46 RA 5: Installation, operations and maintenance

47 Concept Weight (tonnes) MW Manufacturingcosts Man. costs / MW A1. Monopile m 0.58 m MW A2. Concrete monopile m 0.36 m 5.0 MW B. Concrete gbs m 0.40 m 8.5 MW C1. Concrete tripod m 0.62 m 5.0 MW C2. Concrete monopile m 0.66 m 5.0 MW C3. Concrete gbs m 0.45 m 5.0 MW C4. Steel jacket m 0.57 m 5,0 mw D. Steel jacket 8.5 mw m 0.38 m Phase Two: Installation concepts During the second project phase installation methods were investigated, including cumulative equipment costs calculated on the basis of investment costs, operating cost and workability expressed actual working hours per year. This workability is further based upon a maximum operational limit related to Significant Wave Height (Hs) as shown in the graph below. Higher operating limits (sea state) in general lead to a simultaneous increase in investment costs (more robust installation vessel), see the table below. For several foundations and distances to harbours scenarios were calculated as shown in the underneath table. The graph shows the direct costs of the offshore spread. Excluded are project management costs, onshore site costs, onshore equipment, etc. The graph s main function is to compare the costs of different offshore spreads. The installation of a jacket foundation is calculated in a similar manner. Due to the component weight of the other alternative concepts the number of possible scenarios proved rather limited. For instance not every offshore spread is capable to hoist loads up to 8000 tonnes. Regarding installation concept choices a main outcome is that both the steel monopile and jacket foundations installation process turn out to be the most cost efficient. However, key precondition is that it has to be performed by a jack-up with sufficient crane capacity and a modest payload. Sea transport is most efficient when conducted by barges. Simultaneously, there are no indications that extra investments in higher payloads and higher operational workability will turn out successful measures. Adding up fabrication costs and installation costs provides an the overall system costs overview. RA 5: Installation, operations and maintenance 47

48 Summary of conclusions Foundation concept Monopile / transition piece steel Jacket steel Gravity Based Foundation concrete Tripod on piles concrete Drilled concrete monopile A common monopile & transition piece foundation is the proven foundation type for turbines up to 5 MW installed in shallow water conditions. Installation can be performed best by relatively small jack-up platforms with modest payload and equipped with a large crane (1000 tonnes), plus combined with barge transport. Critical but manageable point is the offshore loading of the monopile. A high number of annual installations up to 200 foundations per year is possible. The all-in-one vessels (self propelled jack-ups / vessels) prove 50% - 60% more expensive regarding their net cycle time costing. The higher workability of a self propelled jackup improves this performance but the cumulative costs are still 20-40% higher. A high payload is recommended due to a reduced transport time, especially for wind farms located at longer distances from a supply harbour. Installation characteristics similar to the monopile installation spread. Installation costs are 30% to 60% higher due too a higher cycle time, whereas the maximum number of annual installations is with about 130 approximately 35% lower. Especially at greater water depths and by employing larger turbines the jacket proves to be a cost efficient overall solution requiring an acceptable amount of construction steel. Main disadvantage of this foundation type is that a single dedicated vessel must perform both the offshore transportation and installation activities. This because there are no methods available yet for successful offshore loading, or foundation floating and submerging. As a result the number of annual installations will be 50-70% lower compared to a monopile foundation based installation method. In addition only a limited number of suitable installation vessels is available. Main advantage is that the fabrication costs are low compared to steel structures, especially if the investment costs are spread over a large number of foundations. Combines disadvantages linked to steel jackets and a concrete gravity based foundation type and is therefore not a recommended option. A new foundation technology concept/method developed in parallel with the WE@SEA project for a Swedish R&D Project. The aim of this new concept is to combine the monopile s structural behavior with a cost efficient use of concrete as a construction material. The drilled concrete monopile can be installed by employing a heavy lift vessel fitted with a vertical drill system. Within this calculation exercise the bulk of foundation development investment costs have been dedicated to a small number of projects. This results in installation costs being a factor 3 to 4 higher compared to the steel monopile installation costs. Nevertheless the overall picture is positive, particularly for mid-size wind projects. For projects facing high-wave design characteristics, new technology innovations are required in order to reduce wave impact, and thus the concrete monopile diameter and weight. Due to the high cycle time (incl. drilling) the maximum number of annual installations is limited to approximately 100 foundations. 48 RA 5: Installation, operations and maintenance

49 The way ahead The main aim of the R&D project was to find a best-suited concept for offshore foundations in the near and distant future, but this ambitious objective could not be achieved. Main reasons are that even in the rather small North Sea section studied, key parameters tend differ to such an extent that it proves impossible to combine these variables into one overall concept. Within the Dutch part of the North Sea, met ocean conditions, soil conditions and the distance to main supply harbours all differ. In addition, each individual offshore wind farm encompasses its own distinct characteristics in terms of size, type and number of turbines. Combining these factors results into to unique set of variables, with either minor, average or great complexity as illustrated below. Nevertheless, some general statements regarding optimised concept combinations can be made. At sites with a minor complexity, a steel monopile proves the preferred option. Installation activities can be performed by smaller size jack-ups provided the fitting of sufficient crane capacity, and an offshore logistical system comprising barge transport of the monopiles and transition pieces. For sites with average complexity the drilled concrete monopile solution can prove a viable alternative for equivalent steel monopiles. The installation method as a precondition has to consist of a heavy lift vessel with vertical drill system, while the concrete monopiles are transported by tugs in horizontal floating position. At sites with a high complexity there are two possible foundation concepts, either the gravity based foundation, transported by heavy lift vessels, or the steel jackets, transported by barges and installed by jack-ups or transported and installed by self propelled jack-ups with sufficient pay load to carry a number of jackets. Exceptions to these rules do always occur. The above calculations are based on new investments, and companies already owning equipment or production facilities favouring their own concept. These are often part of an existing business model, whereas last but not least commercial considerations no-doubt have their impact too and can cause shifts into the above illustration. RA 5: Installation, operations and maintenance 49

50 Ampelmann; a motion-compensating platform for accessing wind turbines Organisations Delft University of Technology, The Netherlands Report title Ampelmann; a motion-compensating platform for accessing wind turbines Report number WE@SEA Author David Cerda Salzmann Date September 2009 Due to the increasing number of offshore wind farms that have been built in the past years, near-shore locations for new wind farms are becoming increasingly scarce. As a consequence, wind farms are gradually being placed farther from shore, where wind speeds are higher and these new (potential) locations typically allow for larger size wind farms comprising a substantial number of turbines. However, such sites are commonly located in deeper water and subject to more demanding wave conditions compared to already operational offshore wind farms. With regard to operations and maintenance this can pose a formidable challenge. A key aspect is accessibility, defined as the percentage of total available time a turbine can be accessed. Downtime losses Whenever an offshore wind turbine requires a corrective maintenance action, the installation will remain unavailable for electricity production until the failure has been remedied. Reduced accessibility, most probably due to rough wind and wave conditions, can result into elongated downtime periods thereby reducing the turbine s availability. Decreased availability translates into a power production loss and therefore also loss of revenue. Over 90% of all maintenance actions require the transfer of personnel and/or parts that can either be carried or require to be hoisted by a turbine s permanent onboard crane. In the offshore wind industry personnel transfer with the aid of helicopters is not common due too a combination of safety related reasons and high costs. Furthermore each turbine as an essential precondition has to be fitted with a hoisting platform. Based upon these considerations offshore wind turbines are generally accessed by dedicated vessels. Safe transfer is commonly accomplished by intentionally creating frictional contact between a vessel s bow and a turbine foundation s accommodating boat landing arrangement, aimed at virtually eliminating vessel translations at the point of contact. The main disadvantage of this latter access method is a limitation to moderate wave conditions. Based upon industry comments, a fair estimate of these limiting wave conditions appears to be a significant wave height H s = 1.5m. Future wind farms at locations with more demanding marine conditions will be faced with a significantly reduced accessibility when employing current vessel-based access methods. That in turn is due to a maximum significant wave height as a main transfer-limiting factor. Wave data In order to examine the accessibility of typical offshore wind farm sites as a function of their limiting sea state, two Dutch offshore locations with their wave data readily available were selected. These include the IJmuiden Munitiestortplaats (YM6) and the K13a platform (K13). The former is situated approximately 37 km from shore, and the latter much further at about 100km distance. Scatter diagrams indicating the annual sea states distribution of both locations were applied to determine year-round accessibility for fictive wind farms. The YM6 location is representative for sea conditions found at already operational wind farm sites: Offshore Windpark Egmond aan Zee (OWEZ) and Prinses Amalia Windpark (previously named Windpark Q7). Both sites are situated in each other vicinity and thus exposed to similar wave conditions. Current access methods are typically limited to a significant 1.5m wave height and result into a 68% accessibility figure, see Table 1. At the K13 location, this accessibility figure reduces to 60% for the same access limit. Table 1 further indicates that when the access-limiting significant 50 RA 5: Installation, operations and maintenance

51 wave height can be raised to 2.0m or even 2.5 meters, a huge accessibility gain can be achieved at both operational wind farm sites. By contrast, a H s increase from 2.5m to 3.0 meters relatively shows a much smaller effect and one can question whether this would justify the likely additional costs involved. Table 1 Year-round accessibility for different limiting sea states at two offshore sites Year-round accessibility [%] for different limiting sea states Location Distance shore = H s,li = H s,li = H s,li = H s,li = H s,li YM6 37 km K km With the anticipated offshore wind farms growth, especially at locations further from shore with an inherently rougher wave climate, there appears a clear industry need for developing a safe vessel-based access system. Such a system requires a built-in capability to facilitate offshore wind turbine maintenance with a high accessibility rate, preferably up to a significant wave height H s = 2.5m. The Ampelmann In order to create a safe transfer system, it would be ideal to have on a vessel a transfer platform for which the vessel motions can be compensated in all six degrees of freedom. This in order to make the platform motionless in comparison to the fixed object that needs to be accessed: the offshore wind turbine. A gangway between transfer platform and turbine will then enable service personnel to walk safely from the vessel to the offshore structure and vice versa. Systems that can create motions in all six degrees of freedom exist in the form of flight simulators. The moving part of these simulators is an assembly comprises of a cockpit and video screens. This assembly is set in motion by a configuration of six hydraulic cylinders known as a hexapod or a Stewart platform, Figure 17. Due to the application of six cylinders, these platforms can move in a controlled manner in all six degrees of freedom. This principle seems ideally suited for effectively eliminating all motions when mounted on a vessel, and after replacing the cockpit and video screens by a transfer deck. A prerequisite for motion compensating is to continuously make available accurate real-time vessel motions measurements, and a control system that converts the motion sensor data stream into Stewart platform control signals. Thus by combining Stewart platform technologies and motion sensors data active motion compensation can be achieved in all six degrees of freedom. System requirements This basic concept was invented during a Wind Energy Conference in Berlin in The inventors christened their concept Ampelmann after das Ampelmännchen (Figure 18), the typical little male figure with the hat that until today features all former East Berlin traffic lights. RA 5: Installation, operations and maintenance 51

52 Figure 17 Flight simulator with Stewart platform Figure 18 Das Ampelmännchen Figure 19 Artist s impression of the Ampelmann system Prior to Ampelmann system development, the following system requirements were determined: Highest possible safety standards; Ship-based system, applicable on a wide range of vessels; No adaptation to wind turbines necessary; Provide accessibility H s 2.5m. It has to be examined first whether the different technologies combined in the Ampelmann system, i.e. the Stewart platform and motion sensor, would allow for a sufficiently fast and accurate motion control. This in turn is necessary to create a motionless upper deck on a moving vessel. As a proof of concept principle, a series of scale model tests were performed successfully. The set up comprised a small sized Stewart platform (cylinder stroke of 20 cm) in combination with an Octans motion sensor (consisting of three accelerometers and three fibreoptic gyros) and custom-developed support software. This proof of concept was conducted in two separate stages. First by placing the system on top of a larger Stewart platform (Figure 20) aimed at testing and performance enhancement by controls fine-tuning. Thereafter the system was mounted upon a 4-metre long vessel placed in a wave basin with a built-in capacity to excite the vessel with regular and irregular waves (Figure 21). Dry and wet tests performed with the scale Ampelmann model gave improved insight in the function of all combined technologies. In random wave fields, the Ampelmann scale model succeeded to keep the upper Stewart platform nearly motionless with residual motions limited to less than 10mm. A Mooring Lines B Roll Dampers C Octans D Wave height Figure 20 Dry test setup Figure 21 Wet test set-up in wave basin 52 RA 5: Installation, operations and maintenance

53 Safety considerations The next step was to develop a prototype aimed at proving the Ampelmann concept for the purpose of transferring personnel under real-life offshore conditions. This objective represented three new main challenges: Turn the integral Ampelmann system inherently safe for personnel transfer; Create a system that counteracts motions of a sea-going vessel in H s = 2.5m; Prove its full operational use under offshore conditions: Easy Access. Although Stewart platforms with cylinder strokes exceeding 1m are common in flight simulator use, their application under highly demanding offshore conditions is new. To tackle the first main challenge the following safety philosophy was decided upon: Operation must continue after a single component failure; This ride-through-failure mode must continue to function for at least 30 seconds. A safety based design procedure was then developed in order to create a system that is inherently safe whilst meeting the other two challenges: full motion compensation in predefined sea states and easy access. This procedure is shown in Figure 22 and defined into four subobjectives: Verify the strength of all structural components; Create Stewart platform that can compensate vessel motions in H s = 2.5m; Ensure full redundancy of the Ampelmann system (motion control); Prevent possible failures due to human errors. Safety Based Structural Strength No failure of structural components Ship Motion Compensation Motion Control Operational Procedure Full motion compensation No failure of critical components No failure due to human errors Lloyd s Register: Design Appraisal Fabrication Survey Load Test Optimized Stewart platform design for full motion compensation in Hs = 2.5m All main components redundant Trained operators Easy access Ampelmann Safety Management System (ASMS) monitors all system functions, takes mitigation measures and warns operator Figure 22 Safety based design procedure Full redundancy To ensure full Ampelmann system redundancy, a Failure Modes and Effects Analysis (FMEA) was performed aimed at identifying potential failures on all system components and examine the impact of each failure. For all impacts that can cause either direct or indirect Stewart platform malfunctioning or other hazardous situations, measures were taken to reduce failure occurrence and/or their impact. These iterations were performed for all components until a system design emerged where component or computational failures could no longer inflict any unsafe effects. In other words, after any failure the Ampelmann system is capable to continue its full functioning for at least 30 seconds. One FMEA outcome was that all critical system components had to be made redundant, and all redundancies were individually tested to prove their ride-through-failure capacity. RA 5: Installation, operations and maintenance 53

54 In order to link all potential component failures to operational procedures, several HAZID (Hazard Identification) meetings were held with main stakeholders during the Ampelmann prototype development process. The outcome of these meetings led to the drafting of an Ampelmann Safety Management System (ASMS). In this extensive spreadsheet-based model, all possible failure modes were linked to a specific warning level. These warnings are only visible to the operator, who can assess whether a person being transferred can finish the activity before the system is returned to its settled position. Only in case a multiple failure occurs this is relayed to all crewmembers, simultaneously making alarm lights to start flashing and sirens releasing sound. When that happens a person amidst a transfer operation has 5 seconds left before the system will retract itself from the structure. He/she has the choice to either complete the transfer or step back and hold on tight. During these 5 seconds also the operator has an option for manually aborting the operation. To prevent failures due to human errors all Ampelmann operators will receive thorough training. Personnel being transferred will receive a safety instruction, but basically only need to watch a traffic light mounted on the transfer deck. In case the operator switches on green light, this serves as a starting sign for safe passage over the gangway to access an offshore wind turbine. Stewart platform design The Stewart platform architecture should enable vessel motion compensation in H s =2.5m sea states. In addition, axial forces in the hydraulic cylinders caused by the transfer deck and gangway should be minimised in order to optimise power requirements and costs. Finally, mechanical platform singularities should be avoided. Mechanical singularity within a platform can be defined as the configuration or pose of a mechanism that causes unpredictable behaviour. Such an unwanted situation can and must be prevented by examining all possible platform poses. A dedicated design process was developed to determine a dedicated Stewart platform architecture best adapted for the Ampelmann system. This part was accomplished by first determining many possible architecture options, all limited by different boundary conditions: cylinder stroke length and size limits. Furthermore, ultimate load cases for the Ampelmann application were to be established. Then a calculation procedure was performed for all proposed platform architectures aimed at determining the motion envelope, calculate cylinder forces and as a check for singularities. From vessel motion simulations, it was found that within the motion envelope the vertical excursions were a key system criterion. After analysing different design options the best-suited platform architecture could be selected. The systematic process for determining this optimised platform architecture is shown at flowchart Error! Reference source not found.. This methodology yielded the final Stewart platform architecture as well as a clear procedure for determining future architectures for Ampelmann systems in case design requirements are being altered. Size Limits Stroke length Load cases Calculation Procedure Other Architecture Parameters Design Considerations Preferred Architecture Figure 23 Flowchart to determine Stewart platform architecture 54 RA 5: Installation, operations and maintenance

55 Subsequently, the required motion compensation capacity of the already selected Stewart platform architecture was examined for three different vessel types. For this purpose, vessel motions for all six degrees of freedom were simulated on these vessels in different sea states. The results are presented in Figure 24. The illustration clearly shows that the initial project objective to enable full motion compensation in a sea state of H s = 2.5m is being met when the Ampelmann is mounted at a 50-metre long vessel. Type vessel: Anchor handling tug Dimensions: 24m x 10m x 2.75m Displacement: 120 tons Type vessel: Multi purpose vessel Dimensions: 50m x 12m x 3.80m Displacement: 900 tons Type vessel: Offshore support vessel Dimensions: 70m x 16m x 5.60m Displacement: 4000 tons Figure 24 Motion compensating capacity of the Ampelmann system on different vessels Testing and certification After Ampelmann prototype assembly, a series of tests were performed aimed at verifying proper system functioning: Motion Tests; Redundancy Tests; Motion Compensation Tests; Operation and Emergency Simulation; Operational tests. During these motion tests, first the Stewart platform s entire motion envelope was verified (Figure 25). Subsequently the control system was further fine-tuned until sufficiently high motion control accuracy was achieved. The next step was to verify full system redundancy, which process included simulating failures for each individual component. In addition checking whether a redundant component takes over its functionality in case of a failure, and confirming built-in warning capabilities by the Ampelmann Safety Management System. Figure 25 Motion test Figure 26 Motion compensation test After completing these motion and redundancy tests, the Ampelmann system was loaded on a barge to perform motion compensation tests outside the Port of Rotterdam (Figure 26). In a sea state of H s = 1.5m the residual motions measured on the transfer deck were less than 40mm heave and less than 0.5 degrees roll and pitch, confirming the Ampelmann system functioning. After returning onshore, the gangway was mounted onto the transfer deck and operational procedures as well as emergency cases were simulated as shown in Figure 27. As a final test, the Ampelmann was installed on the SMIT Bronco to test offshore access in the OWEZ wind farm off the Dutch coast. A personnel transfer operation is demonstrated in Figure 28. RA 5: Installation, operations and maintenance 55

56 Figure 27 Operation Simulation Figure 28 First operational test In addition to the extensive test sequence, the Ampelmann system structural strength was to be certified. Lloyd s Register performed fabrication surveys on all structural components during the production phase. Furthermore a full design appraisal was conducted. A final load test involved applying a 450kg mass attached to the tip of a fully extended gangway, a process witnessed by Lloyd s Register.. Commercial product The Ampelmann development process has resulted into a proven fully redundant transfer system. After the successful transfer demonstration trial at OWEZ, the Ampelmann has now become commercially available and was already applied in several different offshore projects. These projects include an offshore platform decommissioning (Figure 29), where motion compensation was achieved in sea states up to H S = 2.8m, and specific support during offshore wind turbine transition piece installation (Figure 30). A second Ampelmann system has been operational since the summer of Spread over all projects combined, these Ampelmann s by the end of 2009 had provided over personnel transfers. Figure 29 Ampelmann at platform decommissioning Figure 30 Ampelmann at transition piece installation 56 RA 5: Installation, operations and maintenance

57 Ampelmann, demonstrator July 1, 2008 Authors: Dr.Ir. Jan van der Tempel Ir. David Cerda Salzmann Ir. Frederik Gerner Ir. Jillis Koch Ir. Arjan Göbel Section Offshore Engineering, Delft University of Technology The Delft University of Technology earlier this year successfully completed the Ampelmann Demonstrator research project. The main aims were further development of this novel motion compensation platform that as a unique capability enables safe and easy offshore wind turbine access. The project s main objective was to build and optimise the new technology up to a demonstrator phase, a process encompassing full operational testing of the offshore access system under harsh maritime conditions. Additional project aims were the development of dedicated safety features and operational procedures and to validate the Ampelmann s motion compensation capabilities under realistic offshore conditions. Project sponsors include Dutch specialist organisations Delft University of Technology, We@Sea, oil and gas company Shell, SMST Designers and Constructors, and maritime services specialist Smit. The current safe-access limit for vessel-based personnel transfer to offshore wind turbines is a significant wave height (Hs) of 1.5m. Denoted Hs = 1.5m, this value represents 73% of typical Dutch North Sea wave conditions. The phrase significant wave height itself is defined as the mean height (from crest to trough) of the highest third of the waves in a sample. Equally relevant, significant wave height as a term is widely regarded the approximate equivalent to visually observed wave height. The Ampelmann has been designed for a standard 50-metre long vessel and enables full compensation to Hs = 2.0m, representing 85% of typical Dutch North Sea wave conditions. An Ampelmann system option is to transfer service personnel under Hs = 2.5m (93%) wave conditions, an operational offshore condition whereby some degree of platform motion is accepted. As part of the Demonstrator project, a substantially smaller vessel (25m length) has been employed. The offshore demonstrations finally aimed at proving full motion compensation capabilities up to Hs = 1.0m, and to simultaneously test a reduced compensation mode for higher sea states. Advanced features The Ampelmann platform design features a 6-metre base diameter, while the hydraulic cylinder stroke measures a maximum 2 metres. The gangway system has been designed in such a manner that after connecting the device to a turbine, the gangway is capable to rotate and translate freely. This substantial degree of movement freedom enables the system to fully compensate for small motions of the Ampelmann transfer deck during normal operation, and larger motions during operation in the ride-through-failure mode. Maximum gangway length is 16m and includes a telescopic section that measures 6m. During personnel transfer operations the vessel will always maintain a mean 7-metre distance from an offshore wind turbine foundation. Regarding safety, the Ampelmann system has been designed with sufficient built-in backup and redundancy capabilities to enable continued operation in the event of a single component or system failure. This ride-trough-failure mode must be retained for minimal 10s. However, depending on specific failed component(s) and/or system(s) and the nature of ongoing offshore activities during a failure incident a ride-trough-failure mode might even have to remain active for more than 30 seconds. Perhaps most important, during a ride-through-failure mode service staff that are amidst an offshore transfer operation should have sufficient time to either finish or RA 5: Installation, operations and maintenance 57

58 abort their installation access or descend procedure. That involves either a safe return to the Ampelmann seat or to escape in opposite direction to the (still) connected offshore wind turbine structure. When a turbine access procedure is completed or aborted, the Ampelmann operator can retract the gangway and bring the system into stationary position. This procedure does contain built-in programmed time limits, after which expiry the system automatically takes over the Ampelmann s return into safe position. All component failure modes have been evaluated as part of a failure mode effect analysis or FMEA. This evaluation in addition aimed at incorporating sufficient redundancy capability, the adequate spacing of regular system checks, and the design of a state-of-the-art Ampelmann maintenance strategy. In addition, all operational execution modes have been examined with the aid of a HEMP analysis method. This methodology primary aims at identifying all critical process steps and any necessary safety and other precautions to be taken by the crew on board of a vessel equipped with an Ampelmann system. Design, building and testing The Ampelmann Demonstrator design process commenced during September Final design was completed by December 2006, followed by the ordering of main components early January Ampelmann system assembly was completed in only four days soon after component delivery during May The Demonstrator unit was tested offshore on June 27 th and continued at July 11 th. Compensation capabilities tested during a Hs = 1.5m sea state simulation returned excellent results with only 40mm heave and 5 degrees roll and pitch residual motion values recorded. As a next step optimising and concept finalizing activities were performed on the Ampelmann Demonstrator during the summer of 2007, followed by a series of onshore tests until December 7 th. Final demonstrations were conducted for a selected audience comprising representatives from Shell, SMIT, wind turbine supplier Vestas, certification agency Lloyd s Register, and Dutch government Staatstoezicht op de Mijnen. The next day, the Ampelmann Demonstrator was fitted to a SMIT owned vessel named Bronco that ours later left for the port of IJmuiden. On Friday, December 14th, a first safe and successful North Sea transfer was completed to and from a 3 MW offshore wind turbine. This WTG 03 Vestas turbine forms part of the 108 MW 36- turbine OWEZ offshore wind farm, jointly owned by Shell and Dutch energy utility Nuon. The Bronco returned to Rotterdam on December 21 st, where Ampelmann demobilization was completed the same day. Conclusions The Ampelmann Demonstrator project met all goals and objectives set earlier. These include the design, building and offshore testing of an Ampelmann Demonstrator unit capable to meet compensation test under Hs = 1.5m conditions and a successful offshore transfer in Hs < 0.5m. Operational durability testing and access in higher wave conditions will be tackled in future projects. Summarised the project can be considered a success as it has fully proved that motion compensation can indeed be performed under realistic offshore conditions. In other words: With the aid of an Ampelmann offshore wind turbine access can indeed become as easy as crossing a street. Finally, the research team headed by Delft University of Technology project leader Jan van der Tempel during project execution received support from a vast number of experts from various academic and professional disciplines. These experts originate from inside and outside the university, while additional support was provided continuously by the above mentioned sponsor partner companies. 58 RA 5: Installation, operations and maintenance

59 RAMS for offshore wind farms Organisation Delft University of Technology, The Netherlands Report title RAMS for offshore wind farms Authors Erika Echavarría Date September 2009 RAMS for offshore wind farms As part of a policy plan to meet national and European renewable wind power targets, the Netherlands aims at installing 6000 MW offshore wind capacity by As part of this target and by assuming the application of 2 6 MW class wind turbines, 1000 to 3000 wind turbines have to be installed and maintained at the North Sea. This research project is dedicated towards operation and maintenance (O&M) concepts, with a special focus at offshore wind farm Reliability, Availability, Maintainability, and Serviceability (RAMS) issues. Offshore operational wind farm conditions differ from those experienced onshore especially due to their exposure to adverse weather conditions, remote locations, and often difficult to nearimpossible accessibility during certain periods of the year. The resulting challenges call for dedicated O&M strategies and more reliable wind systems capable to operate continuously. A common approach to improve systems reliability is by incorporating component redundancies. Even though this may turn out be a good solution in some cases, it can also result in undesirable repercussions and other unwanted consequences. These potentially include mass increase, a demand for additional component space, and higher overall systems cost. An alternative approach applied in this research project is to analyse functional capabilities of existing components aimed at allocating new originally unintended functions and/or functionalities. A new design philosophy to increase availability The report introduces a new design philosophy to increase availability of wind turbines based on functional redundancies and reconfiguration. It describes the functional redundancy concept and presents the new design philosophy implemented in a software tool named KIEF (Knowledge Intensive Engineering Framework), which as such primarily supports functional representation. It contains an overview of the reconfiguration concept as a maintenance strategy, even though further research will still be needed. As indicated earlier, demanding offshore conditions present new and huge challenges with respect to O&M activities. Main differences compared to average onshore wind sites are linked directly to issues like accessibility, foundations, transportation and installation equipment. That in turn reinforces the need for new maintenance strategies, specific equipment and vessels, inventory and storage approach, etcetera. Offshore wind turbines availability for generating electricity is negatively affected by adverse weather conditions, remote site locations, and poor accessibility during certain periods of the year. Insufficient installation reliability is another unwelcome phenomenon. The inaccessibility issue, especially during winter, makes it virtually impossible to achieve offshore an availability level within the 98% range. The latter is a state-of-the-art value achieved for onshore wind turbines by applying conventional maintenance methods. For example when a relative simple failure occurs in an onshore wind farm turbine, it can often be remedied within an acceptably short period. However, when a similar failure occurs in a wind turbine operating at the North Sea during winter, a remedying period can easily extent to days or even weeks. Such a period is typically characterised by of strong wind speeds, corresponding to a high wind power yield potential. Therefore, an offshore failure at the wrong time can easily result in big revenue loss. RA 5: Installation, operations and maintenance 59

60 O&M costs Generally harsh offshore wind conditions do not only complicate wind turbine access, but also significantly raise O&M costs. While onshore O&M cost figures are usually in the 10-15% range of total kwh cost, they typically represent a much higher 25-30% figure for offshore wind turbine applications. In addition, corrective maintenance costs are roughly a factor two higher compared to those for preventive maintenance. This is due to the fact that for some offshore wind farms, expensive turbine access options like helicopters may sometimes have to be called in for corrective O&M purpose. Furthermore, a narrow workable weather window usually hampers offshore wind turbine maintenance activities during winter. The available time span is sometimes less than a single service personnel shift lasts. That in turn may necessitate additional service personnel trips to complete a given maintenance procedure. These examples clearly emphasise some main causes behind unacceptably low availability levels associated with several of today s operational offshore wind turbines, and simultaneously highlight the need for RAMS based research. Additionally, a general lack of reliability records for offshore wind farms combined with a lack of estimates on reduced availability thereby negatively affect project financing, and will among others lead to higher interest rates and insurance costs. As the installation of 6000 MW offshore wind power capacity in the North Sea represents a major overall challenge, RAMS as a dedicated support tool can provide specific methodology for enabling a more cost-effective O&M. In this research project reliability is defined in analogy to IEEE formulating: The ability of a system or component to perform its required functions under stated conditions for a specified period of time. Compared to onshore employment, offshore wind turbine reliability has a much higher impact on systems availability due to accessibility related issues impacting operation during parts of the year. Statistical data Rapid wind market growth encouraged wind turbine manufacturers to up scale their products with increasingly shorter intervals. Wind turbines introduced 20 years ago were much smaller and rather uncomplicated compared to modern 2-3MW turbines equipped with advanced technologies. Therefore, 20-year lifetime reliability statistical data can probably only be obtained from those wind turbine models that are not commercially available anymore. This in turn turns it difficult for the wind industry to obtain accurate data on actual reliability performance issues. In general, literature is consistent in their conclusions on a general lack of reliability-based data, reinforced by the diversity of methods applied for data collection. Other remarks on technology development are also available in literature. One comment on the effect of introducing new models in the market, suggests that mean failure rate increases have been found in these databases. During the past rapid wind industry growth period, manufacturers were forced to design and produce commercial wind turbines at competitive costs. They thereby looked towards other industries where technology is proven and components are even selected off the shelf. However, in some cases, results show lower reliability figures than initially anticipated, mainly because wind turbine operational conditions are not yet fully understood. Sometimes these conditions differ substantially from those known in other industries. For instance, gearboxes show high failure rates despite their broad application in many other industrial fields. Offshore wind turbines further operate under different conditions compared to onshore installations and therefore face other application requirements. Reliability and maintenance issues under these circumstances for instance cannot be treated like those known for onshore sites where accessibility is normally not regarded a main constraint. 60 RA 5: Installation, operations and maintenance

61 Design topologies New designs and solutions aimed at remedying specific reliability problems therefore need to be developed. And concepts that did not seem advantageous in the past may prove optimal solutions for (future) offshore application, and vice verse. However, reliability data on less common design concepts are limited. Wind turbine design has evolved in time by introducing changes at all technological levels, and wind turbine manufacturers in the past have explored a wide variety of alternative solutions. They did for instance develop different design topologies, such as vertical or horizontal axis of rotation and upwind or downwind rotor placement. These wind technology developers also designed different control strategies and found alternative solutions for smaller components such as brake systems incorporating so-called blade tip brakes. Today, manufacturers generally adopted a limited number of main technologies they consider as offering optimised results for common onshore conditions. For offshore wind farms by contrast, wind turbine reliability levels remain uncertain. There are for instance no generally available maintenance records that cover a significant period, while the wind industry simultaneously commenced at a relatively late stage to systematically document and process available data. Onshore wind turbine reliability data figures are usually restricted to complete installations and main components like rotor, gearbox, and generator. However, data on subcomponents can hardly be traced back in these statistics. Generally, such data are helpful to study wind turbine technology patterns, determine individual components reliability, and improve failure modes understanding. These database sources themselves are rather limited in origin but offer the advantage of containing data on wind turbine makes & models that have been operating for over 10 years. Today s best-known wind turbine databases include WMEP and LWK from Germany, and WindStats Newsletter published in Denmark. Additional databases originate from Finland (VTT), and Sweden (Elforsk). Comparing data In the past several attempts were made to compare data provided by these different databases. However, the initiatives all faced similar challenges when comparing different topologies put at various sites and with information gathered in different manners. Some for instance evaluate reliability without simultaneously comparing different technologies used, like indicating a choice for induction or synchronous generators. In other cases, authors present complete wind turbine failure rates for each operational year, whereas some other studies provide a wind turbine lifetime analysis curve. Individual issues considered include for instance downtime as a function of failure type and rate, icing effects, and wind speed related effects on reliability. This research project includes a reliability analysis for different wind turbine subsystems at a time scale, by making use of a WMEP database. A main research focus is at maintenancerelated data provided on an annual basis, more specifically on failure rates per operational year. The emphasis thereby is on technology trends for different components. Wind turbine reliability aimed at enhancing component reliability can be achieved by different methods like by applying higher quality materials, superior design (methods), more extensive testing, and/or the use of redundant components. Additional optimising options include improved systems robustness, and the sufficient testing of components and/or proven technologies originating from other industries in wind turbines. Systems reliability can be further optimised by methods such as failure mode, effect and criticality analysis (FMECA), fault tree analysis (FTA), event tree analysis (ETA), and reliability centred maintenance (RCM). Of key importance in this respect remains to fully understand failure causes and their consequences prior to proceed with fixing these issues aimed at increasing systems availability. Availability One definition of availability is that a wind turbine must be always ready for operation between its cut-in and cut-out wind speeds. Availability, and therefore energy generation potential can be increased for example by applying optimized maintenance strategies. The latter can for instance RA 5: Installation, operations and maintenance 61

62 be achieved through introducing built-in fault supervisory controller diagnosis capabilities, and/or by operating dynamic controllers for specific subsystems. Additional methods to improve wind turbine availability do include superior installation access, by maintaining a strategic parts inventory, and/or through incorporating a condition monitoring system. Further possible optimising measures include novel offshore access systems such as the Ampelmann, and a range of design measures focused especially at costs reduction, enhancing performance, and/or improved transport logistics. However, no measure or solution can assure 100% availability, and attempts to approach this 100% goal typically incur a substantial costs penalty. Most solutions share in common that they aim at correcting and/or preventing failures, which in turn necessitates maintenance and/or diagnosis visits. Accessing installations is then compulsory, and for offshore applications this requirement substantially increases maintenance costs due to high offshore transportation and equipment costs. Access methods research is one of the available options to target availability optimising issues. Another proven approach is fault tolerance, which enables a system to continue operation after a failure incident did occur. Maintenance is defined as a manner/method employed to repair a broken or failing/failed system or component. Serviceability in turn refers to a planned schedule for systems maintaining & servicing. It typically includes a number of annual visits, and the specific service nature (electrical, mechanical, combined, et cetera). Serviceability itself can be viewed best as carrying out maintenance procedures. Wind farm maintainability and serviceability involves a range of different but often interlinked actions and aspects, including wind turbine concept, procedures and O&M methods, vessel types, et cetera. All strategies to improve reliability, maintainability, and serviceability ultimately focus at obtaining higher availability and consequently optimised energy production. Reconfiguration The reconfiguration methodology is based upon a design process aimed at equipping wind turbines with required capabilities to effectively deal with faults and failures in a new and innovative manner. Main purpose is to maintain operation and simultaneously achieve some degree of physical independence from O&M crews. This methodology, regarded a first step towards building/operating self-maintaining wind turbines and as a new design philosophy is based upon two concepts: functional redundancies and reconfiguration. Each component thereby has a specific function to accomplish but with the potential of fulfilling multiple functions. For instance, a pitch system is used to regulate power output, but is also employed for aerodynamic braking purpose. This multiple behaviour potential can be employed to perform functions that are originally performed by a faulty component. Functional redundancy occurs when potential behaviour matches a required unperformed function. Reconfiguration is intended to become a built-in capability applied as a repair strategy, and is as such based upon potential functional redundancies provided by individual components. Possible solutions range from reusing information from adjacent wind turbines, such as wind speed and direction, up to setting up different operational modes like re-wiring, reconnecting, changing parameters or control strategy. The methodology is not intended for materials or component optimising. Nevertheless, it may incorporate additional components as part of an overall design effort to achieve reconfigured capabilities. Summarised, the methodology aims at finding systematically functional redundancies within the system that, with the aid of reconfigurable design functions will generate provisional repair strategies by allowing alternative operational modes. 62 RA 5: Installation, operations and maintenance

63 Self Maintenance Machine concept Originating from the 1990 s the Self Maintenance Machine (SMM) concept aims at designing and building a machine that applies embedded intelligence together with built-in sensors and actuators. Such an SMM could perhaps not operate at full capacity but at designated level (i.e. 50%), thereby adding robustness, and fault tolerance capabilities. The SMM methodology was successfully applied for a photocopying machine. However, there is little information available on further SMM developments. This research project adopts the SMM concept and applies it to wind turbines aimed at finding innovative alternative O&M solutions, especially under harsh offshore conditions. One of the main envisaged contributions is bringing a functional redundancy concept closer to the mechanical offshore wind turbine domain, where a self-maintenance capability reduces the need for (all-year) installation accessibility. Additional added benefits linked to this design methodology include a contribution towards a reduced installation stoppage rate, and a reduction in the number of failure events. Other potential benefits include reduced installation downtime, fewer failure incidents, and maintaining power output capability until the next planned maintenance visit. Finally these functional redundancies can provide fault tolerant solutions and bring flexibility into the system s operation at no significant additional weight and/or cost penalties. A new design philosophy presented as part of this research project aims at increasing wind turbine availability based upon applying already built-in functional redundancies and reconfiguration capabilities. The outcome is an enabling design methodology that can effectively cope with faults by providing alternative operational modes in terms of wind farm sub-system reconfiguration. As a new application it has the potential to gain in future importance. That because offshore wind turbine energy output goes up as part of a continuous size up scaling trend from 5 6MW today to perhaps 8 10 MW in the next 4 8 years. RA 5: Installation, operations and maintenance 63

64 Development of O&M Cost Estimator for offshore wind farms Organisations Energy Research Centre ECN, The Netherlands Report title Operation and Maintenance Cost Estimator (OMCE) Project numbers ECN Authors L.W.M.M. Rademakers H. Braam T.S. Obdam R.P. v.d. Pieterman Date February 2009 Operation and Maintenance Cost Estimator (OMCE) Within the project Operation and Maintenance Cost Estimator (OMCE) models have been developed, with a built-in capability to estimate offshore wind farm O&M costs during the planning as well as operational phase. Such models are already applied by project developers and offshore wind turbine suppliers alike, and aim at assessing costs and downtime linked to various O&M strategies. A second related application is to gain improved insight into the impact of specific technology design choices. Named ECN O&M Tool one model was developed to assess wind turbine O&M aspects in the planning phase, and it offers advanced possibilities to address preventive and corrective maintenance issues. The method to be used in the operational phase (the OMCE = Operation and Maintenance Cost Estimator) makes use of data generated by a given wind farm during the initial years of operation. One of the options is to expand preventive and corrective maintenance costs and optimising issues with condition based maintenance. Maintenance focus 1. Calendar based maintenance: effort and cost are usually determined by one or two visits per year. After 3 or 4 years the calendar based maintenance costs can be somewhat higher due to among others gearbox oil changes. Input for calendar-based maintenance is defined as the number of repair days with associated costs for labour, equipment, spares, et cetera, and is usually derived from service manuals with prescribed intervals and procedures; 2. Unplanned corrective maintenance: Costs due to random failures that are more difficult to predict. At the wind farm operation start corrective maintenance costs can turn out somewhat higher than expected due a variety of issues often referred to as teething problems. Inputs for estimating the effort required for unplanned corrective maintenance is based upon component failure rates and the associated repair costs (mainly equipment, personnel, spares, consumables, and revenue loss); 3. Condition based maintenance: It might turn out that major overhauls are required, for instance due to unexpected and accelerated wear in turbine components originally designed for a 20-year lifetime (e.g. gearbox or pitch drive exchange). The nature of this specific maintenance type was not foreseen initially, but when the need arises it generally can be planned ahead. Hence it is categorized as condition-based maintenance. Input for estimating the required effort for condition-based maintenance is derived from the number of components that is expected to fail within the coming year(s) and their associated remedying costs. The actual assessment to determine whether components are about to fail should among others originate from inspection results and/or condition monitoring measurements. 64 RA 5: Installation, operations and maintenance

65 Even before the program commenced, a first version of the ECN O&M Tool (Figure 1) had become available to the wind market. Its product focus was only on unplanned corrective maintenance. Long-term mean annual offshore wind farm O&M costs were estimated as: Annual O&M costs = Annual component failure frequency x Repair costs Combined costs Repair costs among others comprise the combined costs for personnel, equipment, spare parts, and revenue loss. These latter losses are based upon actual downtime and as such need to take into account all challenges linked to wind turbine access especially during late autumn, winter and early spring. The ECN O&M Tool consists of two spreadsheets with routines programmed in the Visual Basic software tool, and is very easy to familiarize with. As becomes clear from Figure 1, inputs consist of wind farm specific failure data, plus data on vessel characteristics (i.e. costs, and maximum allowable weather access conditions), and site-specific weather data. Figure 1: Schematic representation of the ECN O&M Tool for determining the O&M costs and downtime of an offshore wind farm. During the We@Sea program execution period ECN s ECN O&M Tool has been expanded and optimised by incorporating specific user-based feedback information. As a result the tool is now capable to accept calendar based maintenance, and as a second benefit has been turned more user friendly. In 2007 certification agency Germanischer Lloyd issued a validation statement for the ECN O&M Tool. Today more than 10 worldwide licenses have been sold to among others utilities, project developers, consultants, and wind turbine manufacturers. In addition over 25 offshore wind farms were analysed by employing this tool, whereas a majority of these wind farms are presently under development and some already under construction. Tool licensees can share their experiences during so called user days, a recurring event that was for the first time held in RA 5: Installation, operations and maintenance 65

66 OMCE development However, a majority of efforts with regard to cost estimation tools development and implementation have been put into an Operation and Maintenance Cost Estimator (OMCE). This OMCE conception was initiated partly based upon a specific operator question: Can you tell us how many gearboxes we will need to exchange within the next two years and what this cost us? Without doubting the relevance of this question, it is simultaneously extremely difficult to answer. ECN therefore decided to develop OMCE as a dedicated tool capable to determine future O&M costs some years ahead, with a main business focus at large offshore wind farm owners/operators. OMCE application is dedicated at a wind farm operational phase, for instance in support of prepare future O&M budget reservations and/or for estimating end of warranty period O&M needs. The OMCE process structure is presented in Figure 2. Figure 2: OMCE concept showing the raw data to estimated O&M costs data flow The OMCE roughly comprises two parts: (1) OMCE Building Blocks that process all operational data, and (2) OMCE Calculator for estimating future costs. In contrast to the ECN O&M Tool, the OMCE is not only a calculation tool. It in fact covers a full process from data collection during the wind farm operation to costs estimating. To a certain extend one specific module of the OMCE, the OMCE Calculator, can be compared with the ECN O&M Tool. However, this former module does not only deal with corrective and preventive maintenance but is also dedicated to condition based maintenance. As a simulation tool, it by comparison possesses more advanced built-in capabilities like assessing the influence of specific limitations in vessels and spare parts availability, during planning, and by prioritising certain maintenance operations. Operational data feedback OMCE Building Blocks require operational data feedback of a specific wind farm under evaluation. This data package is either O&M-related, SCADA-related, or derived from measurement campaigns and condition monitoring tools. Data on failures, repair actions, vessel usage, components, and weather conditions are thereby analysed as specific inputs for estimating unplanned corrective maintenance efforts. Data originating from condition monitoring and load measurements are by contrast analysed to estimate necessary condition based maintenance efforts. For this purpose four so-called OMCE Building Blocks (BB) have been specified, each covering a specific data set: 66 RA 5: Installation, operations and maintenance

67 - BB O&M: Processes O&M-related data for determining component failure rates and successive repair actions required; - BB Logistics: Determines what equipment, crews, and spare parts/components are required for certain maintenance and repair actions; - BB Loads & Lifetime: Determines mechanical component loads as a primary input for prioritising inspections focused at the heaviest stressed parts; - BB Health Monitoring: Determines time to failure necessitating repair, a process among others based upon condition monitoring results and/or inspections. The main objective of these Building Blocks is in fact twofold. First, all Building Blocks should generate (updated) input parameters for the OMCE Calculator. Second, these Building Blocks should provide information in a more general manner, aimed at reapplication by operators for assessing the overall adequacy of a given O&M strategy. OMCE Calculator is applied for assessing O&M cost for say the next 1, 2 or 5 years, and is based upon BB results. The OMCE calculator considers three dedicated maintenance types: (1) corrective unplanned, (2) condition based, and (3) calendar based. It was envisaged initially that all data generated by a given wind farm could be processed instantly by these Building Blocks. However, it had to be concluded later that the available data format requires an intermediate step, and needs to be structured first. This process is visualised as a conversion from raw data (stored in many different files, reports, and databases) to an event list in Figure 2. In this latter list all different events (i.e. component failure) are linked to individual maintenance actions (i.e. resets, repairs, exchange), personnel employment, and equipment and spares usage. Project achievements Within the We@Sea program scope, the following project results have been achieved: The OMCE Calculator structure as shown in Figure 2 has been developed and described; Functional OMCE Calculator specifications were described; Technical OMCE Calculator requirements have been described; An OMCE Calculator model description has been reported; OMCE Building Blocks technical specifications have been determined and described; Event List specifications for raw data structuring before processing by the BB s were described. The actual OMCE Calculator programming was not part of the project scope, but is envisaged within the D OWES (Dutch Offshore Wind Energy Services, project. A first OMCE Calculator demo version will become available by mid 2010, and a commercial version early Wind farm operators do already apply some OMCE results. For instance: the knowledge to accumulate, analyse, and report wind farm data in accordance with OMCE Building Blocks the specifications has been applied for several wind farms. RA 5: Installation, operations and maintenance 67

68 Load monitoring for wind turbines; Fibre optic sensing and data processing Organisation Energy Research Centre ECN, The Netherlands Report title Load monitoring for wind turbines; Fibre optic sensing and data processing Report numbers ECN Author T.W. Verbruggen Date 2009 Wind turbine components condition monitoring is already considered a key wind industry topic and one with growing importance. For component like bearings, gearboxes and other rotating equipment, such technologies and methods are ready available but these generally originate from different (industrial) applications. The development of dedicated condition monitoring tools for rotor blades is hampered by the fact that no reliable measuring systems to accurately determine rotor blade actual loads are available yet. Simultaneously turbine owner/operator interest in this specific area increases fast and for more than one reason. A first is linked to a wish of gaining fresh insight into actual rotor blade loads. The second reason is an aim to reduce these dynamic blade loads by applying new and more advanced control techniques. For loads measurement, optical techniques are often referred to as the most promising solution available. This because of their long fibers lifetime expectancy and a proven insensitivity to electrical interference, combined with stable performance and together turning regular recalibration superfluous. Optical measurement systems have been in use already for several years, but results so far not always proved convincing and therefore require further development. Unfortunately, the systems applied by ECN initially showed lots of failures, often causing measurement interruptions. In addition fitting sensors in the components appeared time consuming and the jobs had to be conducted by specially trained employees. The sensors connection itself also proved prone to failure, whereas necessary repair jobs turned out laborious due too the special nature and bonding techniques required. Huge analyzing effort Another general disadvantage linked to condition monitoring systems is that a large volume of data is generated, which in turn requires a huge analyzing effort plus specific know-how for proper results interpretation. This together adds to both cumulative work load and overall costs. For ECN s fibre optic sensing and data processing project, the following objectives aimed at improving load measurements applicability and better monitoring of rotor blades have been laid down: Selecting an interrogator capable to meet load monitoring requirements; The selection or (in-house) sensor development suitable for this application, an integrated process that includes installation and handling; Development of a processing tool, which operates automatically and generates only the required data. During a desk study aimed at interrogators selection an inventory search followed by a first product pre-selection were conducted. This resulted in three product candidates, each to a large degree matching user requirements. These interrogators were all tested by ECN. That test outcome resulted in the choice of a device that not only meets most pre-set requirements, but is also considered suitable for future experiments and sensors development tests. Both accuracy and resolution are regarded sufficient for the specific application, while as an added benefit the selected interrogator showed a strong sensitivity for polarization. However, the product price level is still above target. 68 RA 5: Installation, operations and maintenance

69 User requirements Another desk study has been performed for the sensors selection. This applied research revealed that no available sensor offered full compliance with user requirements. Materials strain is for instance often not measured directly, while installation and/or replacement proved both difficult. However, the desk study also resulted into a fresh project approach idea. During the next step, a sensor assembly was designed, built and tested. This sensor does meet the following requirements: Easy-installation of the device within a rotor blade structure, whereby actual assembly averages minutes only. This favourable result implies that a turbine can be fully instrumented with the load measurement system within a day, thereby greatly minimizing turbine down time; Design for easy maintenance whereby sensor failure can easily be detected, and replacement taking only about five minutes. Sensor calibration can be preformed during production, which adds a Plug & Play built-in capability. Temperature compensation is included with respect to sensor properties; Wind turbine technicians trained for conducting regular operation and maintenance (O&M) activities can perform installation and service, with no special additional skills required; Compliant with respect to resolution and accuracy performance and non-sensitive to drift. Several parties have already shown interest for this patent pending sensor. Automatic data processing Load measurement systems do generate large quantities of data, which in practice adds difficulty for process operators. One distinct project output is an automatic data processing concept capable to provide key figures to the operator regarding installation upkeep cost effectiveness enhancement decisions. The concept itself is described in a product specification, while the software is in the development stage. The software offers the following distinct features: A measurement system that converts strain measurements into bending moments. A welldefined interface between measuring system and data processing unit that enables software application in combination with other measurement systems; The software automatically generates load spectra and equivalent loads in accordance with relevant IEC standards, thereby taking into account operational conditions; Measurements are validated automatically; Measurements can be processed directly, which eliminates the need to store large measuring data quantities; The software enables loads comparison between individual turbines in order to prioritize O&M and installation inspection schemes; Although an interface between individual wind turbine and the measurement system is required with respect to PLC-signals, the data processing itself is well suited for all wind turbine types; Data processing includes automatic warning and detailed event analysis. Further development The software development did not yet result in a field ready application version, but many of the key requirements are in place already. At the moment the development process is especially hampered by complexities linked to large load variations, combined with unreliability of measuring data. And although the measurement system should generate reliable data, faults can show strong effects on final analysis results. This in turn implies that a lot of effort still needs to be dedicated towards this specific part of the data processing task. RA 5: Installation, operations and maintenance 69

70 Flight Leader concept for wind farm load counting (phase 1) Organisations Energy Research Centre ECN, The Netherlands Report title Evaluation of the Flight Leader Research Area RL5 Report numbers ECN-X Authors T.S. Obdam L.W.M.M. Rademakers H. Braam Date October 2009 Evaluation of the Flight Leader Previous research has indicated that wind turbine power output, and especially rotor blade load fluctuations, strongly depend on whether or not an installation operates in the wake of other wind farm turbines. These observations imply that turbine loading in large (offshore) wind farms is location/position specific; turbines located in the centre of a given wind farm operate more often in the wake of other installations compared to turbines located at a wind farm edge. It is therefore expected that during wind farm operational lifetime some specific components will degrade faster on these turbines experiencing higher loads, compared to their equivalents subjected to more moderate loading. This specific kind of information might offer sufficient reason to adjust maintenance and inspection schemes according to turbine loading, instead of assuming similar degradation behaviour for all individual wind farm installations. Degradation behaviour When major overhaul of a specific component is planned, turbines experiencing higher components loads can be awarded first exchange priority, which can potentially result into substantial O&M cost savings. A most obvious method to gain insight into individual turbine loads in a given (offshore) wind farm, is by instrumenting all critical components. However, in practice after wind farm completion actual component load levels are only measured in very few occasions, while such measurements are relevant for model verification or for detecting (unexpected) peak loads. The main reason for not measuring these loading effects are this it represents a comprehensive undertaking both in terms of cost as well as time, especially if all turbines need to be measured. Figure 0 Illustration principle flight leader concept. Load measurements performed on the flight leader turbines (indicated by red circles) are used to establish relations between characteristic loading and standard SCADA parameters. These relations are then combined with SCADA data from all other wind farm turbines for estimating their accumulated loads. 70 RA 5: Installation, operations and maintenance

71 Within the project context a so-called Flight Leader concept has been developed aimed at making available accumulated loading estimates on critical components of all individual offshore wind farm turbines and at acceptable costs. Basic idea behind the Flight Leader concept is that only a few offshore wind farm turbines need to be instrumented for mechanical loads measuring. These turbines are labelled Flight Leaders. By using these Flight Leader turbine measurements, relations can be made visible between characteristic loading and standard SCADA parameters (e.g. wind speed, yaw direction, pitch angle, etc.) measured at all turbines. Once such relationships have been determined for wind farm reference turbines (Flight Leaders), and these in turn can be combined with SCADA data from all remaining units. The methodology thus enables accumulated loads determination of all wind farm turbines (Figure 1). Objectives and approach Main project goal was to develop a cost effective method for determining mechanical loads imposed on individual wind farm turbines. Secondary project goals include: Development of a software model demo version capable to process both SCADA and mechanical loads data from an offshore wind farm and predict total load accumulation of all wind farm turbines; Flight Leader concept feasibility and accuracy evaluation by applying measuring data from an existing onshore wind farm; Investigate added benefits linked to including aero-elastic simulation results in the empirical Flight Leader model. A first project action was drafting functional Flight Leader software specifications functionality describing what should be included in the software model. Based upon this information input actual software development could commence by writing functional and technical specifications. These describe different software processes in detail and also serve as a detailed programming guideline. Parallel to Flight Leader software specifications drafting, in addition the measuring capability of ECN s Wind turbine Test site Wieringermeer (EWTW) infrastructure has been greatly expanded. As a first step, a measurement plan that involved describing a detailed description of all planned instrumentation activities was drafted. Based upon this information package, two Nordex N80 turbines were instrumented and all relevant infrastructure-related results compiled in a detailed report. This document includes implementation details of measured data from ECN s LTVM database on five N80 turbines, including formulae applied for calculating pseudo-signals. In an additional project effort a detailed wind turbine model was programmed in ECN s aeroelastic code PHATAS, followed by model evaluation and tuning through actually applying EWTW wind farm measurement data. Furthermore, by applying ECN s wake analysis program FluxFarm and wind field generation software SWIFT, various 3-D wind fields (both free-stream and wake conditions) covering the entire wind turbine operational range were generated. These wind fields in fact served as input for the aero-elastic simulation program. Next, by using simulation runs, output relations between SCADA parameters and load indicator were derived and these in turn have been compared with corresponding relations resulting from measurements. Another project task was finding out whether these simulations are indeed well suited for accurate wind turbine load accumulation prediction. Based upon detailed technical specifications a Flight Leader software demo version was programmed in MATLAB and after its completion a start was made with concept evaluation by applying EWTW wind farm data. After introducing some changes based upon experiences and insight gained with testing and evaluation, the updated Flight Leader software was reemployed for evaluating the Dutch OWEZ offshore wind farm. Measurements One main project added benefit is that the measurements infrastructure at ECN s Wind Turbine test site Wieringermeer (EWTW) has been greatly expanded. At present the following data are being retrieved continuously from all five N80 turbines: RA 5: Installation, operations and maintenance 71

72 Maintenance sheets; SCADA data (134 signals, 10-minute statistics), all turbines. These data are obtained from Nordex on a daily basis; Measured SCADA data (25 Hz); o Turbine operational mode; o Wind speed; o Wind direction; o Power output; o Generator speed; o Yaw direction; o Pitch angle. Mechanical load measurements at two N80 turbines (No. 6 & No. 8): o Blade root bending moments; o Tower bottom bending moments; o Tower top torsion; o Main shaft torque and bending moments; o High-speed shaft torque. A second main project result is the completion of a detailed Nordex N80 turbine aero-elastic module programmed in ECN s Program for Horizontal Axis wind Turbine Analysis and Simulation (PHATAS). The model has been tuned and validated by employing measurements derived from ECN s EWTW wind farm. Third, a demo software model version was programmed in MATLAB. The software incorporates all Flight Leader concept aspects and is focused at offshore wind farm operators for processing offshore wind farm SCADA data and mechanical load measurements. A main model output is an accumulated offshore wind farm turbine mechanical loads comparison. These valuable data can subsequently be reapplied to optimise O&M strategies, for example by prioritising inspection visits or specific component exchange on the worst loaded turbines. The general flight leader computer model structure is shown at flowchart Figure 2. Figure 2. General flight leader computer model structure Flight leader model input data are accumulated from a given offshore wind farm. Two different data types can thereby be distinguished: 1. SCADA data from all turbines; 2. Mechanical load measurements only from flight leader turbines. Both statistical data sets should be accumulated at 10-minute intervals. Mechanical loading impact Wind turbines unfortunately under certain circumstances do not operate in normal power production mode. Furthermore, when being part of a (offshore) wind farm, wind turbines not always do experience free-stream wind conditions. Both latter operational conditions are expected to impact mechanical loading. A first flight leader model step therefore was to categorise each dataset timestamp in one of the five pre-defined turbine states j and three pre- 72 RA 5: Installation, operations and maintenance

73 defined wake conditions k as possible combinations. Possible combinations are indicated in Table 1. A data categorisation step output example is shown in Figure 3. Table 1. Possible turbine states & transitional modes and wake condition combinations ID Turbine state or transitional mode j Normal power production 2.1 Parked/Idling 3.1 Start-up 4.1 Normal shutdown 5.1 Emergency shutdown Wake condit ion k Freestrea m Partial wake Full wake Not Applic able Figure 3. Data categorisation step results example for the five EWTW wind farm turbines. Load case ID s refer to Table 1. Empirical database After all available data have been categorised flight leader turbine measurements can be reapplied to establish relations between (standard) SCADA parameters and load indicators. These in turn are indicative for a certain component s (fatigue-related) damage, or otherwise aging and/or degradation processes. As mentioned above, these relations are expected to differ for all identified turbine states & transitional modes and wake conditions. Therefore, relations between SCADA parameters and load indicators need to be determined for each possible combination displayed in Table 1. The software model further makes it possible to characterise relations by using more traditional methods such as interpolation or multivariate regression but also through employing artificial neural network techniques. An example of the software s empirical database module is shown in Figure 4. RA 5: Installation, operations and maintenance 73

74 Figure 4 Flight Leader s empirical database module for establishing relations between SCADA parameters and load indicators. In the period directly following offshore wind farm commissioning only very limited measured data are available. Taking that into account it might be beneficial to incorporate aero-elastic simulation results into the flight leader model. This is particularly interesting for those situations with low probability occurrence, such as emergency shutdowns or extreme wind speeds (i.e. during storm). Estimating load indicators Next step is estimating load indicators at all individual offshore wind farm turbines. This can be achieved by combining SCADA data accumulated at all turbines, with all relations between SCADA parameters and load indicators stored in the empirical database. As part of this latter process optionally also results from aero-elastic simulations can be incorporated. Now assume a situation whereby for a given turbine during a certain time period no SCADA data can be made available. Under such conditions load indicators can neither be estimated with an empirical nor a simulation database. In order to ensure a fair comparison with regarding total accumulated loading, the software tool also contains a procedure for handling missing data. Finally, a last remaining Flight Leader model activity is a capability to generate and display desired comparative output data on accumulated offshore wind farm mechanical loading of all turbines. This output as a model precondition needs to be made readily available for several main load indicators (i.e. blade root bending and tower bottom bending, or main shaft torque). Besides the above indicated main outputs the software model can in addition calculate and display various accumulated loading breakdowns. For instance the contribution of each turbine s state or transitional mode, or wake condition to total accumulated loading. Furthermore, load accumulation per time period can be studied. These output results can substantially contribute towards gaining better insight in offshore wind farm the performance including as to what specific operating conditions show the greatest impact on individual wind turbine loading. An example of such dedicated output data is depicted in Figure RA 5: Installation, operations and maintenance

75 Figure 5. Flight Leader software output generation example displaying contributions of individual load cases to total load accumulation. The Flight Leader software demo version has already been applied for various analyses trials using data from both the onshore EWTW and offshore OWEZ wind farms. The main project goal was to determine whether the Flight Leader concept is a feasible support tool that can be applied successfully for accurate wind farm turbine load accumulation estimates. Secondly, what method proves best suited for characterising relations between SCADA parameters and load indicator, essentially the core behind the Flight Leader principle. Furthermore attempts were made to identify main contributing causes to explain current uncertainties in Flight Leader software predictions. Prediction accuracy A general main conclusion is that Flight Leader software prediction accuracy has proven relatively high. The two biggest prediction errors occur during parked/idling and emergency shutdown load condition. But for these situations the resulting error does not impose significant impact upon comparing total loads accumulation, since hardly any load is accumulated with the turbine in parked or idling condition. However, in the latter case prediction errors may have considerable effect since for some load cases emergency shutdowns can significantly contribute to total load accumulation. Since these errors are mainly caused by limited specific data availability, it might for these specific load cases for increased prediction accuracy be beneficial to incorporate aero-elastic simulation results in Flight Leader software. OWEZ offshore wind farm trials data confirm the encouraging results derived from the onshore analysis, which above all indicates that the Flight Leader concept is indeed applicable for large offshore wind farms. Several methods for characterising the relations between SCADA parameters and load indicator have been investigated too. A main conclusion is that neural networks are the best-suited method for characterising these relations. Especially the prediction accuracy for new data from RA 5: Installation, operations and maintenance 75

76 other wind farm turbines) is significantly higher compared to other methods (e.g. second order polynomial or partial least squares). When applying electric power output as a load indicator it appeared that prediction errors occurred, which in turn could be attributed to the fact that these N80 turbines are all equipped with un-calibrated or individually different nacelle anemometer types. A similar conclusion could be drawn when performing the same analysis but by applying OWEZ wind farm data. This latter trial clearly illustrates that calibrated nacelle anemometers (and other sensors) are an absolute prerequisite for reliable flight leader principle application. Further research & developments Since 2008 Flight Leader project results have already been presented at several workshops and conferences. Flight Leader evaluation using both data of an onshore and offshore wind farm have been performed in Offline manner. Hereby a dataset was retrieved from an existing wind farm, with the measurement infrastructure not optimised for Flight Leader application. In future the Flight Leader concept should be evaluated based upon Online implementation. During wind farm construction each Flight Leader turbine location should be carefully selected. Furthermore, all turbines should be equipped with calibrated sensors in order to ensure accurate Flight Leader predictions. After commissioning a given wind farm, at regular intervals (every week or on monthly basis) measurement data should be retrieved and fed into the Flight Leader software. This is aimed at updating load accumulation prediction for all wind farm turbines and is preferably combined with additional data/results from inspections or condition monitoring systems. The combination of functions substantially enhances components condition assessment capability and therefore opportunities for adjusting O&M schemes accordingly. This offline analysis has shown that the Flight Leader concept offers a cost-effective method for assessing wind farm turbine loads accumulation. The relations between standard (SCADA) signals and load indicators thereby form a key to Flight Leader concept application. The more accurate these relations, the more reliable calculations on accumulated loading become. But before these relations can be established it needs to be decided what standard signals should be applied for load indicators estimation. Until now this has been performed with a trial-anderror approach. This is far from ideal especially since wind farm operators might not possess detailed knowledge on wind turbine behavioural aspects. Therefore, an automated procedure should be integrated in the software, a support tool capable to employ some statistical methodology for selecting an appropriate SCADA parameter set for estimating given load indicator values. If it turns out that no automatic procedure can be developed, at least a library containing a preferred set of SCADA input signals for a number of load indicators should be developed. Aero-elastic simulations It will be worthwhile to apply aero-elastic simulations for assessing the importance of different SCADA parameters in relation to estimating given load indicator values. Added benefit of simulations is the fact that a large number of input signals is readily available, but were not measured yet on EWTW turbines. With these simulations it becomes possible to identify whether certain (currently not measured) parameters can be put into use for estimating load indicator values at greater accuracy. If this is indeed found to be the case a next recommended step would be to investigate if and in what manner it will turn out possible to measure these parameters on a modern multi-megawatt class wind turbine. During research efforts conducted so far relatively simple fatigue-based load indicators have been applied. It is thereby commonly accepted that for both wind turbine blades and tower materials fatigue damage is the driving degradation mechanism. Drivetrain component degradation is by comparison less understood and therefore a relevant future research topic. During Flight Leader software development several assumptions were made. Finally it has to be verified whether these assumptions are correct or not and also their influence on Flight Leader software output data needs to be assessed in greater detail. 76 RA 5: Installation, operations and maintenance

77 Dedicated offshore maintenance support tools for XEMC DarWinD wind turbines Organisation XEMC DarWinD, The Netherlands In association with energy research Centre ECN, The Netherlands Report title Dedicated offshore maintenance support tools for XEMC DarWinD wind turbines Report numbers Authors Oscar Brug (XEMC Darwind) Efstathios Koutoulakos (XEMC Darwind) René van de Pieterman (ECN) Theo Verbruggen (ECN) Date August 2009 The research project was conducted in close collaboration between ECN and emerging offshore wind turbine manufacturer XEMC Darwind under the embrella. It summarizes main results of a 5 MW offshore wind turbine design effort. Technical risk assessments and an O&M modelling tool developed by ECN were both employed for the project. In addition, a support tool aimed at the operational phase of an offshore wind farm with a main focus at advanced maintenance planning and including recent accumulated field records (OMCE) is presented. XEMC DarWinD technology thereby serves as an exemplary case focused at developing best practice tools for the offshore wind industry. Civil engineering Offshore wind farms generally involve higher capital costs compared to their onshore equivalents, largely due to expensive support structures and for instance installing the power export cables buried into the seabed. Moreover, offshore wind turbine access can be difficult especially during winter, whereby accessibility primarily depends on weather conditions and access system employed. The number of unplanned visits due to turbine failure and planned maintenance visits should therefore ideally be reduced to an average of about one annual visit per installation. XEMC DarWinD aims at commercially introducung its new 5 MW wind turbine concept designed for optimised system reliability, maintainability and availability and indeed one annual service visit. This DD115 is an up scaled and optimised version of the former 2 MW Zephyros turbine, developed in the Netherlands early this century. Conceptually both sister products share several main design features: A Direct-Drive Permanent Magnet generator; Functional integration of generator and rotor bearing, a single rotor bearing concept, and the application of circular stiffeners; Partial passive generator cooling; Cooling, controls, transformer, and power converter placed in lower tower section; 3kV generator (XEMC DarWinD only) for efficient power transmission from nacelle to the lower tower; No additional shell needed for nacelle and hub, castings provide housing. XEMC DarWinD s design approach `Less-is-More` focuses at fewer components, but each with increased robustness aimed at reducing the number of failure incidents and to minimise overall O&M requirement. Even though the turbine concept featuring a permanent magnet type generator is potentially well suited regarding offshore operational reliability, maintainability and availability, marine application optimisation and sub-systems redesign remain to be essential preconditions. That latter process is mainly driven by failure mode analysis, whereas in order to reach an envisaged single annual O&M visit target, dedicated offshore maintenance support tools are required too. Within this context ECN develops a dedicated software instrument for optimising operational RA 5: Installation, operations and maintenance 77

78 offshore wind farm maintenance aspects. As part of this software development a structured methodology has been formulated that focuses at allocating the most favourable maintenance strategy for further optimising specific wind turbine designs. Ultimate target is to achieve the highest possible offshore availability at optimised lifecycle costs. Objectives As part of the collaboration between ECN and XEMC DarWinD the latter provides specific technical input, whereby the DD115 serves as a case study example. Project objectives include: 1. Conducting a technical risk inventory with the aid of a Failure Mode and Effect Analysis or FMEA that includes proposals for remedying modifications by addressing these risks. A design process requires a substantial number of choices to be made, whereas process main outcomes need to be reviewed with respect to reliability and maintenance aspects; 2. Design optimisation aimed at minimising expected kwh-costs by employing the ECN O&M Tool. Some design alternatives necessitate additional analysis methods to be applied to for further evaluating (potential) optimising solutions. A cost model for maintenance cost prediction serves as support tool during the optimisation process. 3. Development of an offshore maintenance support tool OMCE dedicated towards repair and service planning during a wind farm operational phase. Optimised and costeffective offshore wind farm maintenance requires a planning tool that preferably utilises historic wind turbine data. Equally important is to make available a complete maintenance data set covering full turbine lifetime, which in turn serves as a main planning input tool. However, accumulating that kind of dedicated database for existing onshore and offshore equipment is a well-known bottleneck. For an emerging turbine manufacturer amidst a prototype development phase this poses an even bigger challenge. Operational experience accumulated with offshore wind farms is still limited. From the early 1990 s a first generation kw wind turbines appeared, followed from 2000 by larger size installations and several major offshore wind farms. Unfortunately, in a number of reported cases significant technical failures occurred soon after wind farm commissioning, sometimes resulting in main components replacement especially affecting transformers, generators, and gearboxes. Weather window According to wind industry sources, elongated offshore installation downtime is directly linked to poor accessibility at times when remedying actions are most needed. Waiting for a sufficient operational weather window often proves a critical parameter negatively affecting installation downtime. Timely barge/vessel availability is another constraint, since heavy lifting equipment like jack-up s required for component replacement jobs, might already be employed elsewhere. In almost all cases, targeted turbine availability was estimated higher than actual figures achieved. This mismatch can easily result into unexpected yield and thus revenue loss, underlining the significance of estimating O&M parameters as accurately as possible. Given actual offshore-specific O&M characteristics, accurate offshore wind farm O&M cost estimations cannot be obtained with direct calculation methods. To overcome this shortcoming several dedicated models have been developed. These usually calculate main parameters especially availability, O&M costs and downtime associated to component and/or systems failures. Currently available software tools include: ECN O&M Tool applied for calculating offshore wind farm O&M costs during planning phase; Modeling Windfarm Capex & Opex (MWCOST) incorporating Sloop Technology. This development by BMT Fluid Mechanics Limited includes an extension of SLOOP software package applied for determining offshore oil & gas installation O&M parameters; 78 RA 5: Installation, operations and maintenance

79 A model developed by partners Frans de Jong and Mecal targets at estimating onshore wind farm O&M costs; CONTOFAX software tool (1996) developed at the Delft University of Technology (NL) by Christian Schöntag and Gerard van Bussel. The tool is applied for estimating offshore wind farm O&M parameters by means of Monte-Carlo simulations. However, the above tools are being applied exclusively for estimating wind farm availability and O&M costs during their planning phase. The OMCE tool instead focuses at the operational wind farm phase. Risk Priority Number XEMC DarWinD has applied a risk assessment program during the DD115 design phase. Main target was to develop a highly reliable and easy to maintain offshore wind turbine capable to operate 20 years with at least 95% availability. This prompted a necessity to identify and reduce technical risks, by applying a Failure Mode and Effect Analysis (FMEA) tool that already proved its worth in multiple industries including automotive and aerospace. FMEA can be applied during the design phase, as part of a process or systems development, and within supplier partnerships. XEMC DarWinD initiates FMEA to identify technical risks at several project stages. Subdivided into systems, design, and/or process FMEA s, possible failures are identified and ranked by severity, occurrence and detect ability. Each failure is given a `score` based upon these characteristics, known as Risk Priority Number (RPN). Failures with the highest RPN qualify for a follow up, which can either involve a redesign, a planned test to determine failure probability, or an improved method to detect a failure at an early stage. The assessment process commenced with a system FMEA involving the entire wind turbine. It aims at obtaining a clear picture on high-risk areas and failure modes within the overall system and exposes interfaces between individual subsystems and highlights turbine interactions with the offshore environment. A design FMEA usually commences with a brainstorm session involving engineers and field experts. With this group of professionals a particular design aspect is discussed, whereby technical risks are identified. During a follow-up ranking session engineers and field experts involved determine each failure RPN by estimating severity, occurrence and detect ability according to a scorecard. Such a scorecard contains scales running from 1 (no effect) to 10 (catastrophic) and provides guidelines for determining failure impact. A scorecard enables for instance determining the impact of specific failures to humans and/or the environment, to wind turbine functioning, damage caused (costs), loss of earnings (kwh), or repair actions needed (time/costs). XEMC DarWinD has initiated a supplier partnership program aimed at creating awareness on consequences linked to operating a supplier s product in the harsh marine environment. Key is establishing long-term relationships with main suppliers. That in turn as an envisaged prime benefit should ultimately result in optimised components and subsystems offered by each individual supplier, instead of having to relay on off-the-shelf products. The core of this supplier partnership program is a supplier FMEA, with an individual supplier product design being analysis topic. Identifying potential failures External field experts and suppliers together with XEMC DarWinD engineers and other specialists discuss design functionality and interfaces at systems level thereby identifying potential failures. To be effective, a session on a specific topic involving a brainstorm needs to be thoroughly prepared, and should only be executed after completing the internal design FMEA. After a supplier session, feedback to all participants is crucial for building necessary awareness on a specific offshore wind turbine component design. RA 5: Installation, operations and maintenance 79

80 Process Failure Mode and Effect Analysis (PFMEA) as a method is capable to analyse each step or phase in a physical process (production, transportation, assembly, or installation). It specifically aims at identifying potential failure causes that might negatively impact product quality. A PFMEA should be performed just after completing the first assembly process design stage. It also needs to be conducted in such a manner that any improvements to a given product, process, or quality system can be implemented preferably before a prototype is actually being assembled. PFMEA s have been performed on three main components/sub-systems: nacelle and rotor assembly, and the bottom tower section accommodating a majority of the power electronic systems. In total about 1100 possible failure modes were identified, but most of these are unlikely to occur due too DarWinD design choices and/or production tool(s). Furthermore, detail quality checks during assembly can prevent several potential failures to actually occur. In total about 60 possible failures with unacceptable risk were assessed. Of these, approximately 50 failures relate to the drive train concept and might negatively affect yield potential. However, it is remarkable indeed that about 40% of these possible failures are directly linked to component damages inflicted during assembly. PFMEA results will be brought into the assembly process as an optimising measure. Once final design details are known a PFMEA will be performed again, this time with a focus at the latest assembly process steps. Example: FMEA on main bearing and lubrication system A DD115 multi-row roller type main bearing features a continuous oil lubrication system and has been subjected to a design as well as supplier FMEA. During a main bearing lubrication unit design FMEA, about 70 potential failure modes were identified. Many of these issues are related to insufficient oil flow that can in time cause severe bearing damage. A combination of high severity, medium occurrence, and low detect ability resulted in a high RPN. Several causes were identified, including grid outage, clogged hoses, hose leakage, oil pump failure, and alterations in oil properties due to for instance changing environmental conditions. These findings resulted into a design modification suggestion, which involves adding a flow sensor to the system enabling a continuous oil flow monitoring resulting in improved detect ability. In case oil volume flow is reduced for whatever reason, the installation will be switched off automatically before serious damage does occur. However, during the supplier FMEA session objections were made to the suggested remedying solution arguing that a flow sensor is generally known to trigger unnecessary alarms. A recommended alternative solution was to integrate a simple oil pressure sensor at a carefully selected spot within the flow loop, which as a controls component enjoys a favourable 1.0 x 10-6 failures rate per hour. The latter solution, provided it is carefully positioned within the flow loop, will address all failure modes identified without affecting overall turbine reliability. Offshore design optimisation with ECN O&M Tool Any design choice (i.e. drive and control system) and offshore turbine upkeep strategies affect electricity-generating costs. All design and turbine upkeep variables should be optimised in order to maximise lifetime benefits of the wind farm owner/operator. A Levelised Production Cost (LPC) method originating from the International Energy Agency (IEA) has been adopted as main economic assessment parameter for quantifying many of these latter aspects. LPC is defined as present value of all costs divided by the discounted electricity production value, and represents average lifetime cost for producing one unit (1 kwh) of electric power. An LPC includes all wind farm lifecycle aspects from manufacturing and installation to operation and dismantling. 80 RA 5: Installation, operations and maintenance

81 When considering alternative design optimising options, both initial investment and lifetime O&M expenses need to be taken into account. However, calculating especially offshore wind farm maintenance costs involves several stochastic parameters, which include site accessibility limitations due to adverse weather conditions and vessel availability. For that reason, XEMC DarWinD applies the ECN O&M Tool that has proven to be capable of generating O&M cost estimates for offshore wind farms (and not individual wind turbine level) before a designs has been finalised. After a baseline model has been defined, a fresh calculation can performed almost instantly. When considering different maintenance strategies it is crucial that only a limited number of parameters are adjusted simultaneously, enabling a clear root cause identification process revealing both potential cost increase or decrease (trends). A sensitivity analysis can thereby prove useful as an instrument providing additional project insight. However, since the ECN O&M Tool is a model, it should be noted that all results must be carefully examined in order to make certain that they correspond to realistic values. Despite this call for caution it was found that dedicated design optimisation measures by quantifying costs effects, offer the potential to reduce both O&M costs and revenue losses by 1-5%. Most of these measures are either focused at easing maintenance and/or options to postpone repairs until a next scheduled maintenance visit, and as such do not require extended re-engineering effort. Installation upkeep One of the other objectives of this We@Sea funded project is to provide a kick-off for developing the Operation & Maintenance Cost Estimator or OMCE software tool. Once fully tested and optimised this tool is aimed to become a commercial product available to any offshore wind turbine manufacturer. The instrument should further facilitate a choice of upkeep strategies for the medium to long term. Finally, the risk assessment program aimed at identifying, prioritising and addressing technical risks proved fruitful. The large number of issues found underlines that generally even a strong existing wind turbine concept needs re-evaluation for offshore purposes. Several design improvements could be realised at relatively low costs, resulting in increased availability. The system-, design-, process- and supplier FMEA brainstorm sessions thereby resulted into a substantially expanded offshore wind turbine design knowledge base. Since XEMC Darwind actively involves suppliers in FMEA sessions, this fresh knowledge was shared with many wind industry partners raising awareness on reliability and maintainability issues under demanding offshore conditions. FMEA timing is thereby crucial and the assessments should be performed at an early design path stage after finishing the conceptual phase, but well before completing detail design. RA 5: Installation, operations and maintenance 81

82 Nautical Safety Organisation STC BV, The Netherlands Report title Nautical Safety Report number Authors A. Gerretsen, STC C. Westra, ECN Date May 2010 Offshore wind power development in Europe is expected to continue growing fast, from 35,000MW by 2020 to 150,000MW in A main Dutch government goal outlined in the programme Schoon en Zuinig - Clean and Efficient - is to generate by % of the national energy use from renewable sources. This objective translates into 6000MW cumulative wind power capacity in the Dutch section of the North Sea, and this should form a basis for further future growth. Future Dutch offshore wind activities during the next decade will concentrate especially at locations 20 80km distance from shore. Geographically these wind farm sites are allocated between IJmuiden and Den Helder and to the North of the Wadden eilanden (Frisian Islands). A key challenge is not only to meet these ambitions in time, but also to generate maximum added value with these and foreign projects. That in turn automatically results in additional shipping traffic and increased additional offshore activities. The Dutch section of the North Sea is also one of the busiest in terms of shipping movements and one of the most intensively used maritime regions in the world. Besides further intensify shipping offshore wind power expansion will at the same time put increasing pressure upon available free North Sea space. That in turn will affect North Sea nautical safety in an area that during the past decades due to sustained efforts has become one of the safest shipping regions of the world. Four sections As part of the We@Sea programme a dedicated research project has been conducted aimed at gaining insight into the nautical safety as implemented in coastal countries situated around the southern part of the North Sea. The project comprises four sections: A literature and Internet search with a focus at the current status regarding available know-how on nautical safety. In parallel, access to publicly available government documents has been facilitated within legal boundaries of the Wet Openbaarheid van Bestuur (WOB); An inventory of nautical safety legislation as implemented in the other European nations; Verification of the insight gained with relevant stakeholders; Formulation of a next stage research projects on nautical safety. In order to determine whether the construction of large offshore wind farms has a negative impact on nautical safety, additional information reinforcement is required during policy decisions. Especially crucial zoning decisions linked to offshore wind farms nautical safety in the Dutch Exclusive Economic Zone (EEZ) have to be supported and reinforced by validated factual data. Summary and main conclusions By taking an Internet and literature search as starting point dedicated applied research has been conducted into nautical safety aspects in and around offshore wind farms. In addition, a number of stakeholders have been interviewed and that in turn formed the basis for continued research. 82 RA 5: Installation, operations and maintenance

83 Main conclusions In all countries situated at the southern part of the North Sea a 500-metre safety zone is maintained around offshore structures viewed as obstacles; Legal shipping passage distances vary per country. But independent of individual approaches a two nautical mile passing distance measured from shipping lanes within a recognised Vessel Traffic Services (VTS) area appears common; In the United Kingdom shipping passage distances are being determined on the basis of a risk assessment conform the As Low As Reasonable Possible or ALARP principle. Other nations use different approaches; The United Kingdom boasts the most extensive and best documented approach, whereas the least documented situation was found in Belgium; The Netherlands also has a risk assessment procedure in place, but only at high level and focused at estimating collision risks between a vessel and wind farm or between vessels. An additional focus is at changes in shipping traffic intensity due to wind farm construction. Recommendations It is essential to standardise passage distances between shipping and wind farms in order to ensure a smooth and gradual transfer between EEZ s of different adjoining countries; In order to achieve such a standardization a continued research effort that builds upon current report findings is necessary; Optimising passage distances should be feasible once a number of practical and systematic field trials have been completed; Concentrating offshore wind farms and their surroundings in a single VTS area will contribute to enhanced overall nautical safety. Nautical safety As a term nautical safety is not easy to define exactly, but its functioning is among others affected by: A. General parameters - Room for manoeuvring; - Vessel and navigation equipment condition; - Overall quality and experience of the crew; - Physical conditions, especially area width and depth; - Hydro metrological conditions; - Shipping density; - Availability of VTS capability; - Additional real-time available information of the area. B. Parameters linked to the presence of wind farms - Limited room for manoeuvring; o Reaction time; o Time to vessel full stop; o Vessel full stop time in relation to space and distance; o Turning space with and without engaged engine power; - Limited visibility, by radar and visual means; - Effects upon communication and positioning systems. Provided responsible authorities are well prepared, it is feasible to anticipate at changes to a given status quo, which in turn can show a positive impact upon nautical safety. Unsafe conditions do especially occur under conditions whereby unexpected events differ from event expectations. Adequate prediction of these differences is essential as a method to enhance overall nautical safety. This improved status can be achieved by: - Conducting measurements meteorological, hydrological, regarding water depth, and positioning; - Predictions based upon experience as well as modelling; RA 5: Installation, operations and maintenance 83

84 - Training - crew, shipping pilots, Benefit Resources Management (BRM), Environmental Resources Management (ERM), calamities; - Conducting simulations at macro and micro level. Past nautical research during determining the positioning of harbour access channels and deep channels has among others shown that there are substantial overall benefits associated to investments in focused applied research. This includes simulations, personnel training, and building upon experience. C. Possible additional safety enhancing mitigating measures - Markings conform Aids to Navigation rulings by the International Association of Lighthouse Authorities (IALA); - Daily positioning markings conform IALA rulings; - Adequate lightning provision; - Safety corner markings according to Interocean American Shipping (IAS) standards; - Radar positioning at corner points; - Integration within VTS area; - Providing tug boat assistance; - Radar information exchange to vessels in the area; - VHF radar communication; - Compulsory AIS markings for all vessels; - Promoting AIS marking application by for instance offering reductions to insurance premiums and subsidies; - Generating a dummy AIS signal from VTS systems. Figure 1. Radar image of two offshore wind farms off the Dutch coast near IJmuiden on a ship s radar 84 RA 5: Installation, operations and maintenance

85 Figure 2. Safety margin during starboard sailing Figure 3. Available space for emergency stop manoeuvring that requires approximately 7 vessel lengths RA 5: Installation, operations and maintenance 85

86 Figure 4. Shipping passage at a Danish offshore wind farm near Copenhagen 86 RA 5: Installation, operations and maintenance