Planned intervention as a maintenance and repair strategy for offshore wind turbines

Transcription

1 Journal of Marine Engineering & Technology ISSN: (Print) (Online) Journal homepage: Planned intervention as a maintenance and repair strategy for offshore wind turbines A Karyotakis & R Bucknall To cite this article: A Karyotakis & R Bucknall (2010) Planned intervention as a maintenance and repair strategy for offshore wind turbines, Journal of Marine Engineering & Technology, 9:1, To link to this article: Published online: 01 Dec Submit your article to this journal Article views: 241 View related articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 27 November 2017, At: 09:02

2 Planned intervention as a maintenance and repair strategy for offshore wind turbines A Karyotakis and Dr R Bucknall, Department of Mechanical Engineering, University College London, London, UK The UK is planning to develop a significant offshore wind turbine power generation capacity to contribute towards the reduction of greenhouse gas emissions from the electricity generating sector.the maintenance and repair of high numbers of offshore wind turbines represents a major challenge to ensure a reliable yet cost effective source of electricity. This paper examines planned intervention as a possible operation and maintenance (O&M) strategy for large offshore wind farms. Computer-aided simulation employing the Monte Carlo method is used to examine the planned intervention maintenance strategy in terms of wind farm availability, cumulative energy output and production cost of energy. AUTHORS BIOGRAPHIES Alexander Karyotakis gained an MSc degree in Mechanical Engineering from University College London (UCL), and is currently pursuing a PhD degree in mechanical engineering at UCL having joined the Offshore Power Research Group in His research interests are focused on the reliability of wind turbines and CO 2 emissions of offshore wind farms. He is an Associate member of the IMarEST. Richard WG Bucknall received a PhD degree in electrical engineering at the Royal Naval Engineering College, Plymouth, UK. Currently, he is a Senior Lecturer and Course Director of the MSc Marine Engineering and the MSc Power Systems Programs at University College London (UCL). He joined UCL as a Lecturer in 1995 and was promoted to Senior Lecturer in Dr Bucknall is a member of the Department s Marine Systems Research Group and leads the Offshore Power Research Group. He is a member of the IMarEST and Society of Naval Architects and Marine Engineers. INTRODUCTION Offshore wind farm development has a short history of less than 30 years with the first offshore wind farm being built off the coast of Denmark in Today, there are 25 operational offshore wind farms with another four in their last stages of installation (see Table 1) with the majority of these consisting of only a modest number of wind turbines. For the future, greater numbers of larger offshore wind farms having more powerful wind turbines are being planned, including 10GW for the UK as part of its energy development strategy. 1,2 It has been reported that offshore wind farms face significant operation and maintenance (O&M) challenges 3, 4 mainly because, unlike land-based installations, the wind turbines are located in remote locations where access is heavily dependent upon the weather and sea-state but also because the reliability of offshore wind turbines has shown itself to be worse than for onshore wind turbines primarily because of the exposure of wind turbines to the hostile maritime environment. For the large scale offshore wind farms planned for the future the O&M strategy can be expected to become an even greater engineering challenge. 5,6 The current O&M practice adopted for existing offshore wind farms is reactive response. Essentially, when a component fails and the wind turbine becomes non-operational a maintenance expedition is launched at the first opportunity to carry out the repair, this being an additional visit to the wind turbine for routine preventive maintenance purposes. Failed components are generally repaired in situ or by exchange so the reactive response maintenance strategy ensures the wind turbine is returned back to full operational status as quickly as No. A Journal of Marine Engineering and Technology 27

3 Location Country Capacity (MW) No of turbines Type & rating Year of install n Vindeby Denmark 5 11 Bonus 450kW 1991 Lely Holland 2 4 NedWind 500kW 1994 Tuno Knob Denmark 5 10 Vestas 500kW 1995 Irene Vorrin Holland NordTank 600kW 1996 Bockstigen Valor Sweden 3 5 WindWoeld 500kW 1998 Blyth UK 4 2 Vestas 2MW 2000 Middelgruden Denmark Bonus 2MW 2000 Utgrunden Sweden 10 7 GE-Wind 1.4MW 2000 Yttre Strengung Sweden 10 5 NEG Micon 2MW 2001 Horns Rev Denmark Vestas 2MW 2002 Arklow Bank Ireland 25 7 GEWE 3.6MW 2003 North Hoyle UK Vestas 2MW 2003 Nysted Denmark Bonus 2.2MW 2003 Samso Denmark Bonus 2.3MW 2003 Scroby Sands UK Vestas 2MW 2004 Frederikshaven Denmark Various 2004 Kentish Flats UK Vestas 3MW 2005 Ronland Denmark Bonus & Vestas 2006 Barrow UK Vestas 3MW 2006 Q7-WP Holland Vestas 2MW 2006 Breitling Germany Nordex 2.5MW 2006 Ems-Enden Germany Enercon 4.5MW 2006 Beatrice UK 10 2 Re-Power 5MW 2007 Egmond ann Zee Holland Vestas 3MW 2007 Burbo Bank UK Siemens 3.6MW 2007 Robbin Rigg UK planned Lynn Dowsing UK planned Alpha Ventus Germany planned Lillgrund Bank Sweden planned Table 1: Existing offshore wind farms and those planned or under construction in Europe. (Listed by year of installation). 1,5,11,12 practically possible. 7 This approach to maintenance for offshore wind farms has been shown through experience to be effective in maintaining high levels of availability but it is expensive because many failures are unpredicted and it is costly to initiate and carry out repairs especially at short notice. 7, 8 In contrast, the reactive response maintenance strategy is reported as being a suitable strategy for onshore wind turbines, where remoteness has much less impact on accessibility and largely unaffected by weather and the availability of specialist transport, eg, ships and helicopters. 7 Investigations reported in this paper show that the reactive response maintenance strategy is neither practical nor is it desirable for large offshore wind farms when the cost of electricity generation must be competitive primarily because access is dependent upon environmental conditions, ie, weather and sea state. 9,10 O&M costs are known to account for up to 30% of the price of energy produced by offshore wind farms, 4,13,9,14 so reducing the maintenance costs would inevitably mean greater competitiveness of electricity generated by offshore wind turbines in the electricity market. This paper has therefore set out to examine the O&M challenges for future offshore wind farms and in particular investigate an alternative O&M strategy known as the planned intervention maintenance strategy. Computer-aided simulations of planned intervention maintenance strategy scenarios have been developed using the Monte Carlo method and their performance validated and verified using appropriate practical data. Simulations have also been run examining different planned intervention strategies on the likely impact on wind turbine availability, cumulative energy output, and the production cost of energy. REVIEW OF CURRENT O&M PRACTICES FOR OFFSHORE WIND FARMS It is reported that for offshore wind power the O&M costs account for typically 23% of the project s total expenditure as summarised in Fig 1, 6,15,14 or alternatively this may be expressed as accounting for between 25 30% of the cost of the energy produced. 4,15,13,9 For an equivalent onshore project, the O&M costs as a percentage cost of the energy is estimated to be between 5 10%. 9 The main reason for such a difference may be attributed to the impact of operating a wind turbine in the marine environment where the wind turbines are stressed in a harsher maritime environment and their accessibility for repair and maintenance restricted by weather and sea-state conditions, and also by their distance to shore which affects accessibility, meaning expedition costs to visit offshore wind turbines tend to be more expensive than visits to onshore wind turbines. 16 The following notable studies are relevant to the argument that there is a need for a re-evaluation of O&M strategies for offshore wind farms: 28 Journal of Marine Engineering and Technology No. A

4 Hau 7 suggests that in order to limit the complex and expensive work at sea then a preventive maintenance strategy should be a key objective when considering the offshore wind farms. He concludes that O&M strategies must be investigated and quantified to indentify suitable strategies for future offshore wind farms. Fig 1: Offshore wind farm typical cost breakdown 6,15,14 Musial and Butterfield 10 suggest that much of the success of onshore wind turbines can be attributed to the ability of designers to reduce capital costs of wind turbines and maintaining acceptable levels of reliability recognising that onshore wind farms benefit from ease of accessibility so the reactive response O&M strategy can be adopted to ensure revenue is maximised. When considering offshore wind farms however, the authors explain that O&M at sea is significantly more challenging, more time consuming and more costly. They conclude by suggesting the reactive response maintenance strategy has not yet shown itself to provide the best economic solution for offshore wind farms. Kooijman et al 13 suggest that the reactive response maintenance strategy as currently applied to existing offshore wind farms is a major factor when explaining why they are economically uncompetitive compared with onshore wind turbines. They suggest an increase of turbine availability is needed to maximise revenue and conclude that the optimisation of wind farm O&M strategy is needed to improve competitiveness so large scale offshore wind energy projects may compete successfully in a liberalised European Energy Market. Van Bussel 16 shows that the cost of maintenance of offshore wind farms is far higher than equivalent onshore wind farm projects, and suggests that availability of offshore wind turbines is compromised in the marine environment. He concludes that a re-design of the maintenance strategy is required to achieve economically viable operation of offshore wind farms. Jacquemin et al 17 concludes that the design of costeffective O&M strategies will represent a significant challenge for the offshore wind energy industry. Their paper suggests that the O&M of offshore wind farms is inherently multi-variant and multi-criteria and they promote the idea of developing new tools to help quantify the various O&M strategies to mitigate risks. The paper suggests that the risk factors are mainly related to adverse weather conditions, poor reliability, reduced accessibility and the cost of specialised equipment required for certain O&M operations offshore. DEFINITION OF PROPOSED O&M STRATEGIES The summaries of notable publications given in the previous paragraph suggest that the O&M strategy is a significant factor to be considered when evaluating the economic viability of future offshore wind farms. Further, it is observed that the O&M processes of offshore wind farms are compromised by weather and sea-state conditions 4 and their distance to shore poses a further additional challenge and expense. It is a paradox that the greatest winds are available in open water at sea yet it is this environment which posses the greatest challenge to achieve high reliability and for the O&M of offshore wind turbines. A general conclusion reached in much of the published work is that the reactive response maintenance strategy is unlikely to be economically attractive for future large offshore wind farms because significant maintenance resources will be needed. An alternative O&M strategy is the planned intervention strategy which may be suitable when considering the O&M of significant numbers of wind turbines in large offshore wind farms. A planned intervention maintenance strategy adopted for offshore wind turbines could potentially offer greater effective use of maintenance resources and therefore be more economically attractive; however, a compromise is needed with increased downtime (as turbines await their turn for repair) reducing revenue against reduced O&M costs. The reliability of the offshore wind turbines is an important factor to consider when balancing costs and revenues. A planned intervention maintenance strategy for offshore wind farms would involve scheduling visits to each wind turbine at specified times across the year with the scheduled visits being largely determined by the reliability of the wind turbines and the weather conditions. An offshore wind farm using a planned intervention maintenance strategy would mean each wind turbine receiving as many visits as necessary to maintain its availability above a specified level which itself is determined by considering the economics. The required availability level should be determined by balancing the maintenance resource costs, eg, manpower, transportation means (ships and helicopters), etc, and the cost of downtime, ie, the loss of revenue from the sale of electricity. Planned intervention is a maintenance strategy whereby repair and maintenance periods are planned to occur at particular times so that there are fixed intervals between each repair and maintenance periods. This means that should wind turbines become non-operational between the planned intervention times will remain non-operational until the next planned intervention occurs. Wind turbines functioning correctly will be preventively maintained (serviced) at the next planned intervention visit regardless. The planned intervention maintenance strategy means that repairs and No. A Journal of Marine Engineering and Technology 29

5 maintenance will effectively occur at the same time at a time planned in advance so allowing optimal use of maintenance equipment and resources. Whilst for large scale offshore wind farms the planned intervention strategy offers reduced operational costs through better management of the O&M resources on the other hand, the periods between repair and maintenance need very careful consideration otherwise extended unavailability of a wind turbine will reduce revenues to unacceptable levels. Comparing the planned intervention strategy with the reactive response maintenance strategy for offshore wind farms then there are both advantages and disadvantages: Elimination of unplanned repair events. Unplanned maintenance events account for a significant proportion of maintenance expeditions for existing offshore wind farms using the reactive response maintenance strategy (typically between 50 70%). 6,15 The planned intervention maintenance strategy deals only with planned events and by engaging advanced planning is able to effectively use maintenance resources. Unplanned events when using a planned intervention maintenance strategy are simply ignored. Weather dependency. The planned intervention maintenance strategy allows scheduling of maintenance visits at periods of benign weather conditions and therefore avoids cancellations or postponing expeditions due to unexpected weather. For example, planned maintenance visits to offshore wind farms could be scheduled to occur only between May and October for the North Sea region where weather and sea state data indicate more suitable conditions. The effect of weather on the planned intervention maintenance strategy could be minimised, which in turn could decrease maintenance costs. Reduced maintenance spare parts. Reactive response maintenance strategy usually results in the need to carry large inventories of the spare parts as it is difficult to predict when the parts will be needed. 18 For large scale wind farms this maintenance strategy could result in excessive stock holding costs that will increase the maintenance costs of offshore wind farms. On the other hand, the planned intervention maintenance strategy would use reliability levels and planned inspections that can help to predict when spare parts will be needed. There are fewer surprises and more spare parts can be purchased using just-in-time practice, which reduces the maintenance costs of the offshore wind farm. Grouping failures together. Considering the planned intervention maintenance strategy then repairs and/or replacement of components and their planned inspection could be grouped together. One maintenance expedition could therefore be used to service multiple offshore wind turbines. On the other hand, there are time constraints (light conditions) and working constraints for safety reasons that may limit the number of wind turbines serviced per expedition. 9,10 Nonetheless, this advanced scheduling of the repair, replacement and inspection may result in reduction of maintenance resource costs. Revenue may be severely compromised. If a wind turbine fails then it would be repaired as quickly as possible with the reactive response maintenance strategy but this is not the case with the planned intervention maintenance strategy since repairs will only occur at the next planned intervention. With the latter strategy greater revenue may be lost should downtime be for extended periods. Clearly, the planned intervention maintenance strategy represents a significant departure in the repair and maintenance of offshore wind turbines to the current reactive response maintenance strategy. Each strategy is likely to give different revenues and profitability with the reliability of the offshore wind turbines being a crucial aspect in the analysis. METHODOLOGY Earlier investigations by Van Bussel, 16 and also by Giorsetto and Utsurogi, 19 suggest that the analysis of O&M strategies for offshore wind farms is not simple, mainly because of the interactions of the various stochastic processes such as weather, the reliability of wind turbines in the offshore environment, unknown time to repair in the marine environment, etc. 16 In the case of offshore wind farms, the number of variable parameters increases, as compared to onshore wind farms, due to additional factors introduced as a consequence of operating in the marine environment. 16,19 A deterministic approach to modelling the planned intervention maintenance strategy would therefore not produce a sufficiently accurate description of the processes and may offer erroneous results, because the model would have to be constrained to a fixed set of input values. 16 In the investigations that follow the Monte Carlo method is used to simulate the O&M processes of offshore wind farms for the planned intervention maintenance strategy, in order to take into account the stochastic behaviour of the parameters mentioned above. Monte Carlo simulations can be used for system reliability and availability modelling, using suitable computer simulation programs, since the Monte Carlo method allows simulations with minimal constraints as regarding the nature of input assumptions on parameters, eg, variable failure and repair times of systems can be simulated. 20,24 For offshore wind farms, the use of the Monte Carlo simulation allows an understanding of the variable of input parameters, such as the time to failure and time for repairs to statistically predict output performance for a given maintenance strategy, eg, planned intervention maintenance strategy. The main aim of using Monte Carlo simulations is to understand statistical variability of the failure rates and time to repair of different offshore wind turbine components using reliability data. Every component of the wind turbine will have a failure rate, ie, the inverse of time to failure, and a repair time at the beginning of its operation, ie, at beginning of year one. Every time a component fails and is repaired a new failure rate is assigned to that component and hence to the offshore wind turbine. This process is performed for every failure at every maintenance period of the project, and is simulated over a large number of iterations needed for the Monte Carlo analysis, in order to obtain the statistical output performance 30 Journal of Marine Engineering and Technology No. A

6 variability of the availability of the wind farm, the cumulative energy output and the production cost of energy. To predict the performance of the planned intervention maintenance strategy applied to offshore wind farms, new algorithms were needed and these were developed and implemented using the Mathworks Matlab software. ECONOMIC MODEL To calculate the cost of energy produced by offshore wind farms the use of a standard discounting calculation, the Levelised Production Cost (LPC) technique, has been adopted. The justification being that this technique is the method proposed by the IEA (International Energy Association), 21 to compare the competitiveness of wind farm projects. Other studies reporting on wind farm cost calculations have also used this technique. 22,15,23,25 The LPC is the cost price of production per unit of energy, expressed in actualised nominal money. 25 In other words, the LPC expresses the production cost in terms of current purchasing power, whereas future costs and revenues are discounted. The equations to calculate the LPC are: 25,21 or TotalCost of Project LPC= Total Energy Produced LPC = where: Ctot is the total expenditures for year t, Etot is the total energy production for year t, ELoss is the energy lost due to downtime, repair and maintenance activities, TEC is the economic lifetime of the project and r is the interest rate. Expanding equation (2) we get: 22,25 LPC = TEC t= 0 TEC t= 0 I tic C tot + ( ) 1+ r ( E E tot Loss ) T EC t= 0 t C + C + C O& M S R + C 1+ r t DEC ( 1+ r) T EC ( E E tot Loss ) t= 0 ( ) where: I tic is the total investment cost of the offshore project (CAPEX), usually incurred in year zero. C O&M is operations and maintenance costs, which are incurred each year. C S unpredictable costs, eg unexpected environmental damage. C R is the component costs associated with C O&M C DEC is the decommissioning costs, which are the cost associated with dismantling the offshore wind farm and are incurred in the year after the last year of operation, ie, year 21. t (1) (2) (3) The unpredictable costs Cs have been assumed to be zero since there is no easy way to calculate them. Cr costs are absorbed into the capital investment cost Itic, ie, initial spares costs and also into maintenance costs CO&M for ongoing spares costs. Decommissioning costs, which must be estimated because no offshore wind farm has yet been decommissioned, can be determined as a percentage of capital investment costs, the percentage value being 2.5% of the capital investment cost. 15 For simplicity the decommissioning costs can be added to the capital investment cost, so re-writing equation (3) including the above explanations: LPC = I T C EC O M + t t ( & = 0 1+ r) T EC ( E E tot Loss ) total t= 0 Equation 4 is therefore used to represent the basic economic model that may be used to investigate the cost implications of the planned intervention maintenance strategy. RELIABILITY MODEL FOR OFFSHORE WIND FARMS The availability of an offshore wind turbine is defined as the percentage time it is capable of producing and selling electricity to market from an engineering integrity perspective, ie, it is actually capable of operating should their be appropriate wind speeds. It is therefore a function of wind turbine reliability and its repair and maintenance strategy. With a proper service organisation and by ensuring that regular maintenance actions are performed on time, the operators of current onshore wind turbines produce energy with wind farm availability of up to 98%. 6,4 In situations where access limitations impinge maintenance expeditions, as happens in offshore projects, there may be unacceptable downtime and reduced wind turbine reliability directly affecting the availability levels of the wind farm. It is therefore essential to assess the maintenance demand of an offshore wind farm in conjunction with other design parameters such as the reliability of the wind turbines, in order to achieve the required availability of the wind farm for minimal expenditure. For significant numbers of offshore wind turbines then such high availability levels as experienced on onshore wind farms, can only be achieved by using the reactive response maintenance strategy provided sufficient resources of maintenance teams, vessels, helicopters, etc, are available. For a planned intervention maintenance strategy, the availability level of the wind farm is compromised for the benefit of optimised expeditions and reduced costs for repairs and maintenance of the wind turbines. The differences between onshore and offshore wind farms as explained earlier show that the offshore wind turbines are subject to higher stresses, both environmental and power utilisation, due to their location in the marine environment. Higher stressed operation generally results in higher failure rates in offshore wind turbines compared to onshore wind turbines. (4) No. A Journal of Marine Engineering and Technology 31

7 Published data on the reliability levels of offshore wind farms does not exist, therefore in order to construct the reliability model it was necessary to use published data on the failure rates of onshore wind turbines to calculate the failure rate of offshore wind turbines. This has been achieved by noting that onshore and offshore wind turbines are of similar type and by taking into consideration the higher environmental and power utilisation stresses that they endure. Table 2 shows the different environmental conditions and the environmental stress factors applied. By setting the general environmental conditions for onshore wind turbines as general purpose and ground based, ie, K1 onshore = 1, then the environmental stress factor for offshore wind turbines should be in the range between marine sheltered and marine exposed, ie 1.5<K1 offshore<2. The marine sheltered definition relates to components that are located in the marine environment but not exposed, whilst marine exposed is used for components that are fully exposed to the marine environment. The analogy for an offshore wind turbine is that the marine sheltered components of the wind turbine are those components within the nacelle, eg, the gearbox, generator, drive train etc, whilst the marine exposed components are represented by the blades, blade tips, antennas, wind vanes, tower and other components on the nacelle and the outer plane of the nacelle housing. Fig 2: Failure rates of onshore wind turbines per component for Germany, Denmark and Sweden 27,28,26,29 Fig 2 shows the failure rates against components in onshore wind turbines collected in Germany, Denmark and Sweden. These failure rates have been used in a reliability model to calculate an accurate range of failure rates for offshore wind turbines. The mathematical expression of the model is of the form: 30 λ λ K * K * p A (5) XA O 1 2 where: λ XA λ o K 1, K 2, K n p(a) = ( ) ( ) is the predicted failure rate for a system, for failure mode A. Considering a wind turbine λ XA is the calculated offshore wind turbine failure rate. is the base failure rate of the system, in ideal conditions of minimal stress levels. are the stress factors (environmental, power utilisation, etc) for different conditions that the system is used in. 30 is the proportion of failure rate for every failure mode of the system. By adding the proportions of failure rate for all the failure modes of the system then p(a) becomes unity. General environmental conditions Environmental stress factor K1 Ideal, static conditions 0.1 Vibration free, controlled conditions 0.5 General-purpose, ground-based 1 Marine, sheltered 1.5 Marine, exposed 2 Road 3 Rail 4 Air 10 Missile 100 Table 2: Environmental stress factors for mechanical systems. 30 Percentage of component K2 nominal rating Table 3: Rating stress factors for mechanical systems 30 Additionally for reliability studies, differences between onshore and offshore wind turbine applications must be accounted for in terms of a power rating stress factor (K2), which when applied to failure rate accounts for the difference in the capacity factors between offshore and onshore wind turbines. Typical onshore values are between 18 28%, 1,6,31 with the equivalent (existing) offshore wind farm capacity factor being around 34%. 1,6 In future, offshore wind farms capacity factors of around 45% are expected. 1,6 The percentage of a component s nominal rating is presented in Table 3, a range of the power rating stress factor (K2) for offshore application can be calculated based on the capacity factor of a specific site. ANALYSIS AND RESULTS A typical future offshore wind farms as proposed in Round 2, eg, the London Array offshore wind farm, is expected to have the characteristics given below: 5,11 Minimum number of wind turbines 175. Unit power rating between 3.6 MW. Capacity factor of wind farm of up to 45%. Project lifetime 20 years. Discount rate applied 5%. These characteristics were used to form the case study simulated below. 32 Journal of Marine Engineering and Technology No. A

8 Verification of computer models The purpose of verification was to carry out a sensitivity analysis to give added confidence to the developed models. The verification of the computer programs developed to simulate the planned intervention maintenance strategy model was achieved by establishing a baseline offshore wind farm to carry out a sensitivity analysis on the input parameters of the model. Each input parameter of the planned intervention maintenance strategy model is varied through a range of values in order to investigate how the model reacts. An example of this process is presented in this paper for the capacity factor and the discount rate. Fig 3 shows the effect that the capacity factor of the wind farm has on the cumulative energy output and LPC. Four different values were used for the simulations: 1,6,31 18%, which represents the typical onshore wind farm minimum capacity factor, 28%, which represents the typical onshore wind farm maximum capacity factor 34.2%, which is the average capacity factor for the existing offshore wind farms 45%, which is the expected capacity factor for far offshore wind farms and 100%, this value does not reflect reality because wind is unlikely to blow constantly to give the same power output for 100% of the time. This value having been chosen to show how the model reacts to extreme boundary values. By increasing the capacity factor of the wind farm, it can be observed from Fig 3 that the cumulative energy output is increasing and consequently the LPC of energy is falling. This observation shows the advantages that offshore wind farms have over equivalent onshore projects and how the future development in the offshore wind sector will affect the unit cost of energy produced. Fig 3:The effect of the capacity factor on the cumulative energy output and LPC The discount rate of a project plays a key role for its financial viability, and its significance can be observed in Fig 4. In this figure the discount rate of the offshore wind farm project is varied with three different values 2, 6 and 12 to examine how the economic model reacts. It becomes clear by observing Fig 4 that as the discount rate increases the LPC of energy is rising and the financial viability of the project is therefore becoming more and more difficult. The results presented in Figs 3 and 4 were anticipated according to the background theory and the work presented in the previous paragraphs, which validates the developed algorithms. Fig 4:The effect of the interest rate on the costs of energy production Planned intervention scenarios comparison Two different scenarios of the planned intervention maintenance strategy have been developed PM1 and PM2 that determine the total availability of the offshore wind farm, the cumulative energy output and the LPC of energy: PM1 is a planned intervention maintenance strategy with one summer maintenance period per year. PM2 is a planned intervention maintenance strategy with two maintenance periods per year, one in May and one in October. A comparison is given between PM1 and PM2 in Fig 5, where the variables are plotted against the mean time to failure (mttf in years) of the wind turbines. From the reliability model it is shown that the current offshore wind turbines experience an mttf of between 0.25 and 1. The range of mttf between 1 and 2.5 represents the expected reliability levels of future offshore wind turbines. Considering wind farm availability versus the wind turbine mttf in Fig 5 then the two curves representing the PM1 and PM2 scenarios show how the wind farm availability increases with increasing wind turbine mttf. The PM2 scenario yields higher availability levels compared to the PM1 scenario, this being explained by the fact that PM2 scenario has twice as many scheduled maintenance visits to the offshore wind farm each year. Now consider the energy output versus wind turbine mttf in Fig 5. The PM2 curve intersects the PM1 curve at a wind turbine mttf of 0.33 years. For wind turbine mttf levels lower than 0.33 years the PM1 scenario yields higher energy out- No. A Journal of Marine Engineering and Technology 33

9 put compared to the PM2 scenario, whilst for wind turbine mttf levels higher than 0.33 years the PM2 scenario yields higher energy output. These curves can be explained by considering the energy losses during the scheduled maintenance periods /kwh and for PM2 between /kwh, which are found to be lower as compared against to the onshore wind farm values. This indicates that for higher wind turbine reliability levels the planned intervention maintenance strategy tends to produce results in terms of LPC of energy that are directly comparable with onshore wind farms, and in some cases even lower. Fig 5: Comparison between the two planned intervention maintenance strategy scenarios, PM1 and PM2 Now consider the LPC of energy versus the wind turbine mttf in Fig 5, which is presented in two different graphs for two wind turbine mttf ranges, ie 0.25 mttf 0.5 and 0.5 mttf 2.5. Considering the LPC of energy for 0.25 mttf 0.5, the PM1 curve intersects the PM2 curve at a wind turbine mttf of 0.35 years. For wind turbine mttf levels lower than 0.35 years the PM1 scenario yields lower LPC of energy compared to the PM2 scenario, whilst for wind turbine mttf levels higher than 0.35 years the PM2 scenario yields lower LPC of energy, this being a result of the energy output change discussed in the previous paragraph. Considering the LPC of energy for 0.5 mttf 2.5, then the PM2 curve shows significantly lower results as compared to the PM1 curve, this being explained by the higher energy output as the wind turbine mttf increases. This observation indicates that as the wind turbine reliability increases then the PM2 scenario would be preferred over the PM1 scenario. Comparison against published data In order to assess the results shown in Fig 5 they are compared against data obtained from onshore wind farm projects, as listed in Table 4. Considering the wind turbine mttf range of 0.25 mttf 0.5, the LPC of energy for PM1 ranges between /kwh and for PM2 between /kwh, which are found to be higher as compared to the onshore wind farm values. On the other hand, considering the wind turbine mttf range of 0.5 MTTF 1, the LPC of energy for PM1 ranges between CONCLUSION According to the available literature on O&M of offshore wind turbines, this paper considers the technical challenges that future offshore wind projects will face. It is pointed out that the O&M strategy adopted from the operators of onshore and existing offshore wind farms may not satisfy the economic needs of future offshore wind farms. A number of studies have concluded that a re-design of the repair and maintenance practices is essential, in order for future offshore wind farms to be technically and financially viable. The Monte Carlo methodology adopted developing algorithms for computer based programs to simulate the planned intervention maintenance strategy as a possible solution to the technical challenges of O&M for the future offshore wind farms. From the simulation of the proposed solution a number of significant observations have been concluded: The adoption of planned intervention offers a viable solution to the economical challenges that O&M strategies pose for future offshore wind farm developments and would minimise the weather dependency and optimise the maintenance practices of the operators of future offshore wind farms. The number of planned intervention intervals in a year for a future offshore wind farm should be a product of a careful examination of each specific site in terms of capacity factors, weather-related accessibility, reliability of wind turbines and LPC of energy, otherwise the wind farm could experience extensive unavailability and revenue loss. The planned intervention maintenance strategy could achieve LPC of energy directly comparable and in some cases more economical than onshore wind farm projects when taking into consideration the reliability levels of the offshore wind turbines. REFERENCES 1. Douglas Westwood Limited (DWL) The world offshore renewable energy report Renewables, UK. Onshore wind farms Year Capacity factors % % Table 4: Levelised production cost of energy (Values in Pounds ( ) per kwh of energy produced) Journal of Marine Engineering and Technology No. A

10 2. Department of Trade and Industry (DTI) Future Offshore, a strategic framework for the offshore wind industry. A report published by the DTI in November Van Bussel G and Bierbooms W The DOWEC offshore reference wind farm: analysis for transportation for operation and maintenance. Wind Engineering, Vol 27, No 5, 2003, pp Van Bussel G and Henderson A State of the art and technology trends for offshore wind energy: Operation and Maintenance issues. Proceedings of offshore wind energy special topic conference, Brussels, Belgium, Dec Part of the CA-OWEE (Concerted action on offshore wind energy in Europe) project. 5. British Wind Energy Association (BWEA) Prospects for offshore wind energy. A report written for the European Union. Altener Contract XVII/4.1030/Z/ London. 6. CA-OWEE (Concerted action on offshore wind energy in Europe) Offshore wind energy Ready to power a sustainable Europe. Final report. Dec Commission under contract number NNE Hau E Wind turbines, fundamentals, technologies, application, economics. Springer, Berlin EWEA (European Wind Energy Association) Wind Energy The Facts. Earthscan, London Van Bussel G, et al Assessment and optimisation of operation and maintenance of offshore wind turbines. European Wind Energy Conference, Madrid, Spain Report ECN WIND: ECN-RX Musial W and Butterfield S Energy from offshore wind. National Renewable Energy Laboratory, US Department of Energy. Offshore Technology Conference, Houston Texas, May Douglas Westwood Limited (DWL) The world offshore renewable energy report, Renewables, UK. 12. Breton S and Moe G Status, plans and technologies for offshore wind turbines in Europe and North America. Renewable Energy Journal, Vol 34, pp Kooijman M, et al Large scale offshore wind energy in the North Sea a technology and policy perspective. Energy research centre of the Netherlands. Wind Engineering Journal. Vol 28, pp Elkinton C, et al Offshore Wind Farm Layout Optimization (OWFLO) Project: Preliminary Results. 44 th American Institute of Aeronautics and Astronautics (AIAA) Aerospace Sciences Meeting and Exhibit, Reno, Nevada, January 9 12, Martin Kuhn, et al A typical design solution for an offshore wind energy converting system. Opti- OWECS (Offshore wind energy conversion systems). Final Report. Vol 4. The Netherlands. 16. Van Bussel G Operation and maintenance aspects of large offshore wind farms. Proceedings of the 1997 European Wind Energy Conference, Dublin, Ireland, October pp Jacquemin J, et al Multi-dimensional optimisation of O&M provision for offshore wind projects. European Offshore wind conference. Operation and maintenance session. Berlin Mobley K, et al Maintenance Engineering Handbook. 7 th Edition. McGraw-Hill publications. 19. Giorsetto P and Utsurogi K Development of a new procedure for reliability modelling of wind generators. IEEE Transactions on Power Apparatus and Systems. Vol PAS-102, No 1, pp O Connor P Practical Reliability Engineering. 3 rd Edition revised. John Wiley and Sons. England. 21. Tande JO Estimation of cost of energy from wind energy conversion systems. IEA (International Energy Agency). Recommended Practices for Wind Turbine Testing. Edition Schontag C Optimisation of operation and maintenance of offshore wind farms. Report IW R. Institute for Wind Energy. T.U. Delft, The Netherlands. November Van Bussel G and Zaaijer M Reliability, availability and maintenance aspects of large-scale offshore wind farms, a concepts study. Proceedings of MAREC (Marine Renewable Energy Conference) 2001, Newcastle, UK, March Martinez W and Martinez A Computational Statistics handbook with MATLAB. Chapman & Hall/CRC, Boca Raton, Florida, Jaaijer M Overall cost modeling of the DOWEC lifecycle in a wind farm. Part of the DOWEC (Dutch Offshore Wind Energy Converter) project. Delft University of Technology, March 2003, pp Tavner P, et al Reliability analysis for wind turbines. Wind Energy, Vol 10, issue 1, pp ISET (Institut für Solare Energieversorgungstechnik). Annual Institute Report. Development and Results for years 2002, 2003, 2004, 2005 and Ribrant J and Bertling L Survey of failures in wind power systems with focus on Swedish wind power plants during IEEE transactions on energy conversion. Vol 22, No 1, March Tavner P, Spinato F, van Bussel G, Koutoulakos E Reliability of different wind turbine concepts with relevance to offshore application. European Wind Energy Conference (EWEC). Brussels. April Davidson J The reliability of mechanical systems. IMechE (Institute of Mechanical Engineering). Guides for the Process Industries. Wiley. pp Schreiber M, et al Offshore wind energy in the North Sea a study for the Greenpeace. October DTI (Department of Trade and Industry)/Ernst and Young Reform of the renewables obligation: Impact of banding the renewables obligation costs of electricity production. Supporting reports that accompany the Energy White Paper. April No. A Journal of Marine Engineering and Technology 35