Final Report Summary - RELIAWIND (Reliability focused research on optimizing Wind Energy systems design, operation and maintenance: Tools, proof of concepts, guidelines & methodologies for a new gener
- benchmarking past failure rate and downtime data from operational WTs;
- measuring current failure rate and downtime data from WTs, their sub-systems' assemblies, subassemblies and components in operational wind farms (WFs), comparing those results with the public domain benchmark;
- identification of six critical sub-systems in a modern WT;
- identification of five failure modes in those six critical sub-systems;
- understanding failures and failure mechanisms in a modern WT.
- definition of a logical architecture for the detection, location, diagnosis and prognosis of failures in critical systems of a modern WT;
- demonstrating the project findings in a WF maintenance model;
- carried out a training needs analysis and prepared training materials presented at 13 sessions to more than 240 wind industry professionals;
- prepared a programme of standardisation and specification for the collection and analysis of reliability information in the wind industry to IEC committees;
- disseminated consortium knowledge through a website, posters, conference and journal papers at the EWEC 2009, 2010 and 2011, visual media and at more than 18 other international events;
- completed 78 internal deliverables.
Project context and objectives
On 8 and 9 March 2007 the European Union (EU) Council of Ministers agreed: 'Renewable energy will supply at least 20 % of the EU's energy demand by 2020'. Provided sufficient emphasis is placed on technological, research and development (R&D) and market development, wind power can make the most substantial contribution to this target. The geography of Europe and current technological evolution towards 2020 mean that offshore wind energy will play a key role in achieving this target.
Current offshore wind operation and maintenance (O&M) costs are too high, requiring high feed-in tariffs to encourage private investors to make the business case to enter the market. This project aims to change this paradigm by encouraging offshore wind to be deployed with similar operational performance and O&M costs as onshore through better availability and lower cost of energy.
Based on successful experience in the aerospace sector, the RELIAWIND Consortium has changed that paradigm by jointly and scientifically studying how reliability affects the design, operation and maintenance of WTs, leading to a new generation of offshore, and onshore WTs for the market beyond 2015.
Ten top partners took part in this ambitious project, each of them leaders in technical and operational disciplines in the wind power generation value chain:
- This included the wind industry itself, Gamesa (project coordinator), Alstom Wind Power, LM Wind Power, Hansen Transmissions, ABB and SKF.
- Technology experts were GL Garrad Hassan and PTC-Relex Reliability Software and Services.
- Academia was represented by Durham University, United Kingdom and SZTAKI, a Research Institute of the Hungarian Academy of Sciences in Budapest.
The project aimed to achieve better availability and lower cost of energy for WTs, through the deployment of new systems with reduced maintenance requirements and increased availability. To this end, the project proposes an architecture directed at a modular design more immune to environmental conditions, permitting the replacement of components simply and quickly; to improve component monitoring systems and achieve more accurate component diagnosis; and to develop preventive maintenance algorithms for failure anticipation. These new technologies will be integrated in future generations of WT components, WTs and WFs.
Main scientific and technological results and foreground
The main results of the project can be summarised as follows:
- Benchmarking from the public domain the normalised failure rate and downtime of operational WTs and their major assemblies in the field (Work package (WP) 1: Field reliability analysis).
- Measurement of WT normalised mean times between failure (MTBF) and mean times to repair (MTTR) for WTs, their sub-systems, assemblies, subassemblies and components, in operational WFs and comparing those results with the public domain benchmark (WP1: Field reliability analysis).
- Identification of six critical sub-systems in a modern WT (WP1: Field reliability analysis).
- Understanding failures of sub-systems, assemblies, sub-assemblies and components in a modern WT and their mechanisms (WP2: Design for reliability).
- Identification of five failure modes for each of the six critical sub-systems (WP2: Design for reliability).
- Definition a logical architecture for the detection, location, diagnosis and prognosis of failures critical sub-systems, assemblies, sub-assemblies and components in a modern WT and specification of a modern WT health monitoring system (WP3: Algorithms).
- Demonstration of the principles of the project findings in a WF maintenance model (WP4: Applications).
- Carried out a training needs analysis and on that basis prepared training materials which were presented at 13 sessions to more than 240 wind industry professionals who were internal and external partners of the consortium (WP5: Training).
- Prepared a programme of standardisation and detailed specification for the collection and analysis of reliability information in the wind industry and forwarded that proposal to national and international IEC committees (WP6: Dissemination).
- Dissemination of new knowledge generated by the consortium in posters and conference papers at the European Wind Energy Conferences of 2010 and 2011, in journal articles, on the website, through visual media and at more than 18 other international events, including cooperation with other international initiatives such as Sandia working group (WP6: Dissemination).
With 78 internal deliverables completed, the RELIAWIND project's main goal was to usher in a new generation of more efficient and reliable WTs, providing practical results to be used WT design, operations and maintenance. The work of the consortium was divided into six WPs as follows.
WP 1: Identification of critical failures and components
The overall objective was to identify the critical sub-systems, assemblies, sub-assemblies and components of WTs and their failure modes based on processing WT historical operational data and carrying out a failure modes effects and criticality analysis (FMECA). The first work consisted of establishing the state of art and the methodology to be employed. Using the established methodology the field data was analysed and interpreted.
Public domain data
An assessment of the current state of the art in measuring operational WF reliability  showed that a number of quantitative studies of WT reliability have been carried out in the last 10 years. The Dutch research programme DOWEC, which has been among the pioneers of the quantification of WT reliability figures, has presented some interesting studies -. Further reliability analyses have used data from existing commercial and public databases. Relevant results have been achieved by research carried out by various authors -.
The objectives of these studies have been to extract information from the field to understand WT reliability from a statistical point of view and provide a benchmark for further analysis. These previous works have presented relevant results; however consideration must be made to the method of recording and reporting these data, which are discussed by Spinato et al. . The objective of the RELIAWIND field study was to build a database of downtime events from a number of WFs containing a large number of WTs .
Field study data
The primary aim was to present the results of the project field study. The methods of the field study were established and the approach was taking account of all operational data that is recorded at modern WFs, including:
- 10-minute average SCADA data;
- fault / alarm logs;
- work orders / service reports; and
- O&M contractor reports.
These sources are discrete and are not designed to easily allow reliability information to be extracted; a substantial effort was invested in connecting these sources. All the downtime events at each WT in the study were identified and then tagged according to a common taxonomy .
Data for this study were provided by WT manufacturers who were members of the RELIAWIND consortium and by others operators who were members of the RELIAWIND users' working group.
Relex field study data
The FMECA allowed an analysis of the failure rates for the generic RELIAWIND WTs, carried out according to the taxonomy used for the FMECA as the results of the above- mentioned field data. This modelling described the procedure, method and data sources needed to build a WT reliability database and then to predict the WT reliability and FMECA of different designs using Relex Studio software.
Data included in the study
450 wind-farm months' of WT data were added to the field study database, comprising around 350 WTs operating for varying lengths of time. This is in the form of 35 000 downtime events, each one tagged within the standard taxonomy.
WP2, Understanding failures and their mechanisms, design for reliability
This part of work aimed at identifying best design practices and providing guidelines that can enhance overall WT reliability and availability, namely describing the procedure, method and data sources needed to build a WT reliability database and then predict WT reliability by means of a failure modes effects and criticality analysis (FMECA) of different WT designs using Relex Studio software.
The WT reliability prediction and reliability block diagram (RBD) models presented in the analysis are based on two generic WT configurations, R80 and R100. For the R100, a component FMECA was also performed.
Reliability analysis procedure
The procedure for WT design for reliability analysis was applied to the WT and can be used as a classical reliability design analysis during the design and redesign phases. The reliability design analysis can be performed on the overall systematic level as well as the sub-system levels.
The aim of overall systematic level reliability analysis was to integrate the whole system reliability model using common reliability analysis procedures which require a WT system functional block diagram specification and sub-system reliability model specifications.
The sub-system level reliability analysis builds a reliability model for each sub-system in order to analyse the sub-system reliability to:
- investigate the interaction of the sub-system models on the whole system;
- develop a design guideline for sub-system design teams;
- define optimum sensing devices and locations for characterising sub-system failures.
WP3, Logical architecture definition for advanced WT health monitoring
Using the results of the field data in WP1 and design analysis in WP2, WP3 has targeted the critical sub-assemblies identified in WP2 to develop algorithms for the detection, location, diagnosis and prognosis of faults in those sub-systems designed to raise WT reliability and availability.
WP3 identified the vital need to simplify and aggregate the valuable data coming from WTs and unlock a data overload stasis that is gripping the industry. That can initially be done by developing algorithms for the detection, location, diagnosis and prognosis of faults in those sub-systems shown to be unreliable to operate on the data collected from WT SCADA and CMS systems to raise reliability and availability.
A number of algorithms were developed to detect, locate, diagnose and prognose the fault behaviour in the six sub-systems described above. Software systems were developed to apply these algorithms to data made available by partners to demonstrate their fault detection efficacy. A number of monitoring processes were thereby developed.
First, the logical architecture of an advanced WT health monitoring system was defined. This architecture meets the related wind industry and other standards; moreover improvements for the wind industry related standard were suggested. The most critical failures for the most critical components were defined and possible controller mitigating actions have were described.
Next, a practical approach to visualise the SCADA signals and alarms on time domain was devised. A specific algorithm to evaluate the mean residual life (MRL) for the WT components was developed. As a result of establishing new algorithms for failure detection, location and prognosis, improved FMECA applications were developed.
A Maintenance task description template (MTDT) was developed that includes all the variables that have influence on the availability / maintainability. The algorithm for planning maintenance activities based on the previous tasks was developed in order to achieve the maximum operational availability with the pre-defined resources and logistical constraints. Aspects like weather conditions, spare parts availability, or capacity of maintenance teams, were taken into account when preparing the schedule. SCADA based data reports were defined to support maintenance related decision making at an operator and WF owner level. Establishment and direct ordering of SCADA reports to different WT working situations, e.g. failure cases.
Consequently, a set of maintenance decision-making tools were defined based on the failure detection, location, and prognosis. These tools will enable maintenance planners to feedback the maintenance schedule and optimise maintenance resources on the short- term and medium-term horizons.
WP 4, Demonstration of the principles of project findings
The algorithms developed in WP3 were then applied to information collected from WT SCADA and maintainer information and applied to the maintenance strategy developed in WP3 to allow incipient faults in key WT sub-assemblies to be identified early and preventative maintenance to be scheduled appropriately so that downtime is reduced and availability raised. A set of software for this purpose has been developed by SZTAKI.
The wind industry is currently collecting an enormous amount of data from WFs in the field but to date very little has been done to process and use this data to improve operational performance. The stimulus to this appears to be the migration of WTs offshore where access costs are high and downtimes extended. It is clear that if a method can be developed to process the large amount of data available from WTs in WFs against an ordered maintenance strategy then substantial improvements in reliability and availability should be achievable.
Under this task the tools and algorithms developed in WP3, based on the principles of WP2, were integrated to establish a practicable condition monitoring system. This system was to be suitable for application to a WT unit, manageable by the operator and suitable for integration into existing SCADA products, which delivers genuine potential for planning condition based maintenance.
The system was developed as a software platform for the multi-agents developed in WP3. It was designed using software components to include the new functionalities and technologies and to be capable of being subjected to overall system tests based on real turbine experience collected in WP1. The multi-agent platform was developed for the coexistence and exchange of information among the different agent types: failures detection, location, diagnosis and maintenance generation and maintenance sequencing. Next the platform coordinating agents were developed. The consistent software (SW) platform demonstrator is able to simulate and optimise the operation and maintenance concept, by maximizing WT availability and optimising CoE for WFs.
The ultimate aim was to devise practical applications to validate the developments against well-known cases that allow evaluating the capacities of the methodology, to identify the necessary improvements and to compare them against the results obtained nowadays by other methods. This feedback allowed fitting / calibrating the performance of the multi-agent platform.
In this task four types of applications of the platform were defined and realised:
- wind turbines;
- multiple wind turbines in diverse locations;
- WFs; and
- several WFs connected to a net of transport.
The work included study, design and development of the graphical user interface (GUI) that provides information of the state and the results of the system of detection of anomalies and diagnostics for equipment. The system man machine interface (MMI) as well as the user graphical interfaces for the supervision, management and control of the multi-agent system were developed.
An important interface with a decisions making support system was established. This interface includes the following principal characteristics: graphical support to the different alternatives for decision making, simulator of results depending on every solution that a decision maker considers, comparator of results of the simulations, intelligent adviser in line, continuous scanner of the system data bases to show the previous situations most similar to the current one and the solutions that were adopted, and capacity of interaction with the MMI of the system.
WP 5, Training internal and external partners
To train internal and external partners within the consortium in the methods being developed by the consortium.
It has been clear in the work of this project of the importance of making design, manufacturing, operations and maintenance staff aware of the importance of raising WT reliability and the methods by which it may be done. The consortium developed a set of training materials which were aimed at these staff and deployed in the offices of consortium partners.
Consortium has energetically trained its own partners, including more than 240 engineers directly involved in WT design, but also business development managers and sales staff who see the commercial benefits of high reliability products.
WP 6, Dissemination of new knowledge
To disseminate the results of the consortium as widely as possible.
The issues developed in this consortium are well-known in other industries, such as aerospace and automotive. Therefore a key outcome from the consortium was to ensure that reliability and availability was as well-publicised in the wind Industry as those topics are in the aerospace and automotive industries.
The consortium disseminated its results to the European wind industry through:
- a website;
- publication of 25 conference papers, particularly at EWEC2009, Marseille, EWEC2010, Warsaw and EWEA2011, Brussels;
- publication of 12 papers in international journals;
- delivery of 17 presentations at international conferences, seminars or workshops addressing more than 500 wind industry professionals;
- preparation of a RELIAWIND video available on the RELIAWIND website.
The consortium also interacted with the wider industry through its users group, who contributed operational results, and through a reliability panel, consisting of wind industry colleagues outside Europe, where the results of the work to date have been received with great interest and augur well for the development of international standardisation in the areas of WT availability, taxonomy and data standardisation. Finally, the consortium developed the following documents which are available to the industry on the website:
- literature survey on reliability of wind turbines;
- monograph of all publications;
- final publishable set of training materials;
- standardisation recommendation to the industry for IEC 61400.
Dissemination measures (Templates A1 and A2)
This section describes the dissemination measures, including scientific publications relating to foreground. RELIAWIND produced a research monograph, downloadable from its website, which lists all the research publications of the consortium over the project period.
Potential impact and results exploitation
Expected project impact as stated in the description of work:
- Reduction of manufacturing, logistics and maintenance costs combined with increased power-to-weight ratio, reliability and robustness should lead to lower production costs for wind generated electricity.
- Cost reductions through improvements in technology, up-scaling of turbines, large-scale deployment (including offshore) and grid connection should lead to a cost below EUR 0.04/kWh in 2020.
- Develop robust, reliable, cost effective and low maintenance onshore and offshore wind energy systems which are easy to transport and to install.
- Development of individual components and aggregate sub-systems (e.g. rotors, drive trains, controls) using advanced materials, design tools and validation models, and the development of innovative manufacturing procedures.
- Innovative large scale on and off-shore wind power plants based on improved technologies, more robust, reliable and low-maintenance multi-MW turbines, combined with dependable output forecasting tools as well as with standards and certification schemes should bring wind power to higher levels of market penetration.
Potential users of results
The main beneficiaries of the RELIAWIND project outcomes will be WF operators, WT producers, WT designers and other supply chain members of WT production. The results enable the industry to focus on significant aspects of reliability. The outcomes have also pronounced somewhat different view as compared to previous studies, showing that the gearbox problems are not dominating faults at least for relatively young installations.
Results of RELIAWIND project may encourage other manufacturers to share fault statistic information. Results may have impact on how different events are recorded, e.g. not all events leading to downtime are perceived as faults (to be reported further to component supplier for instance). It can give an objective view on the actual reliability and associated downtime of different components. Today, reliability related discussions are dominated by some issues, such as the gearbox, while the actual reliability of some components is overseen. Using these data, a more accurate return on investment calculation can be made for new projects, and O&M can be optimised to further reduce the cost of energy produced by WTs.
Other potential beneficiaries and stakeholders include standardisation organisations, engineers, researchers, designers of asset management tools and systems, IT industry; more generally also environmental organisations and non-governmental organisations (NGOs) who have interest in more reliable renewable energies and energy control experts to achieve more balanced wind energy production.
The project considered fully the requirements of users. However, as in the case of most innovative projects, the results drove to technical or behaviour breakthrough that in some cases indicated significant changes in the users norms and values.
Exploitation of results
There are a number of outcomes that are relevant for exploitation and the following list is not exhaustive:
WP1. Field survey
- Literature review on the reliability and availability of WTs
- Identification of public domain data on failure and downtime rates for WT sub-systems
- Standardised WT taxonomy
- Standardised reliability data collection templates
- Identification of the most critical sub-systems of WTs
- Establish risks levels for the principal components and identify the critical system - Comparable failure and downtime rates for WT sub-systems, assemblies and sub-assemblies.
WP2. Reliability analysis
- A reliability model that was built for the R80 and R100 WTs
- Understanding the usage of reliability calculations for WTs
- Use of the FMECA analysis to identify the key risks in current and future technology WTs
- Identification of the most important Failure Modes in the most critical sub-systems of WTs
- The possibility to simulate the reliability related results of the design changes in the WT structure
- Work done to identify the importance of systematic reliability design methods in WT development, including the use of the FMECA and reliability testing like HALT.
WP3. Algorithm development
- Understanding the structure of SCADA data in WTs and the related technical-IT issues
- Understanding the influence of control on the WT reliability and its potential for providing solutions to mitigate faults
- Analysis and validation of possible advanced WT control methods
- Developments of new algorithms for failure detection, location and prognosis for the critical failure modes of the six most critical sub-systems, including: engineering knowledge based solutions, alarm sequence analysis, data-cleaning, probabilistic methods for the alarm sequences, detection of non-conform situations, SCADA data based failure detection, close to real-time component residual life cycle estimation
- Harmonised and integrated handling of different maintenance activities, e.g. retrofit, failure prognosis
- Structured specification of WF maintenance constraints and expected key performance indicator calculations
- Algorithm development for finding the optimal short term scheduling of WFs and zones
- Definition of SCADA database reports to support maintenance related decision making at operator and at WF owner level
- Establishment and direct ordering SCADA reports to different WT working situations, e.g. failure cases
- A concept for integrating all the above results in one framework system.
- Specification of a comprehensive software system realizing the results of WP3 in an integrated manner
- Definition of interfaces between WF operator's software systems and the newly established software solution
- Development of user surfaces of the system, like web based usage and mobile equipment based access
- Implementation of failure lifecycle handling and maintenance scheduling related business processes and the database structure of the system
- Set up of a demo application to represent all the steps of an integrated failure lifecycle handling solution
- Verification of the solution based on data from the field
- General introduction to WTs and their reliability issues
- Reliability calculation and modelling
- Demonstration of the power of the FMECA to reveal reliability issues in a WT
- Demonstration of the effectiveness of various algorithms to detect WT faults in advance from SCADA signal data
- Demonstration of the importance of SCADA alarm data in terms of the very high alarm rates being experienced by WTs and the potential for using these alarms to warn of incipient faults
- Demonstration of the impact on WT reliability and availability of systematic reliability design during development using the FMECA and reliability testing like HALT.
- Scientific papers
- Establishment of recommendation of a standard
- Participation and presentation at various EWEC and EWEA events
- Project website
- Project video
- Dissemination of training materials.
The RELIAWIND partners' exploitation plans are as follows.
Some of the know-how and methodology developed can be directly distributed to the users, such as normalised fault rates, fault distribution, downtime distribution, training materials and courses, software model for reliability and reliability calculations, software prototype for failure lifecycle handling and maintenance scheduling. The software developed to demonstrate the principles of the project findings or algorithms could be distributed as software or could be a basis to develop a customised programme. The results could be even more widely used providing specific implementation support to users.
The recommendations from the reliability debate at the RELIAWIND side event at EWEA2011 were:
- that the he RELIAWIND reliability database should be expanded;
- data sharing about WT reliability and availability is to be encouraged in the wind industry but involves major issues of confidentiality;
- in order to aid the development of more reliable WTs pre-normative research is needed into: definitions of wind turbine sub-assembly loadings taking account of the wind's stochastic nature; the influence of wind resource and weather on reliability results; development of knowledge management systems for large wind farms, particularly those offshore.
In order to build user capacity to make continued use of the RELIAWIND results, some of the next steps could be:
- more training of wind industry users in reliability techniques;
- realisation of the changes in WT and WF company software environments to permit greater data exchange;
- continuous improvement of the currently available results;
- continue to gather and record fault information on the original group of turbines to see if the reliability of components will change over time e.g. gearbox faults;
- setting up a uniform and anonymous field intervention database;
- implementation of the methodologies developed in RELIAWIND;
- assessment of the improvements made by the use of RELIAWIND outcomes in availability and in cost of energy;
- standardisation of reliability and downtime related definitions;
- standardisation of the WT taxonomy for the purposes of more accurate data collection;
- standardisation of the reliability data collection methods;
- standardisation of service intervention reporting;
- standardisation of reliability analysis tools.
This section specifies the exploitable foreground and provides partners' plans for exploitation.
Applications for patents, trademarks, registered designs (Template B.1)
The partners of RELIAWIND project have not submitted any applications for patents, trademarks or registered designs (template B1). However, Garrad Hassan and Durham University developed a recommendation for wind turbine taxonomy.
Durham University sent a questionnaire to the standards committee representations of the consortium partners. In their replies, the standardisation of structure, taxonomy and terminology were ranked highest. Consequently, Garrad Hassan and Durham University devised a recommendation on an updated wind turbine taxonomy. The recommendation drew on the taxonomy developed by the RELIAWIND project and other taxonomies such as that derived by Sandia Laboratories, US.
The taxonomy should be adaptable for application to the common reliability analyses needed for WTs, such as failure mode and effects criticality analysis, failure rate Pareto analysis, reliability growth analysis and Weibull analysis.
The intention of adopting such taxonomy would be to overcome current deficiencies of the data collection which can be summarised as follows:
- consistency of naming of the systems, sub-systems, assemblies, sub-assemblies and components of WTs;
- non traceability of the system monitored;
- unspecified WT technology or concept;
- problems of confidentiality between parties when exchanging data.
This recommendation was sent to all the consortium members who are members of standardisation committees of International Electrotechnical Commission (IEC). It is also available on the RELIAWIND website and was advertised at the side event of EWEA2011 in Brussels.
Partners exploitation plans (Template B.2)
The following partners have prepared exploitation plans: Alstom Wind, LM Windpower, SKF, Garrad Hassan and UDUR. For the question of confidentiality LM is sending its exploitation plan directly to the EC Project Officer. The other mentioned partners' plans are included below.
The purpose is to develop a new platform of Wind turbine generator to gain access to the offshore market. The RELIAWIND foreground is currently being exploited by the Design Department of Alstom Wind introducing RAMS in the design of the new platform.
IPR exploitable measures: Several patents are being under study.
Further research: future new platforms
Potential / expected impact: Gain access to the offshore market where in the next three years several WFs will be located.
Optimise resources of O&M
The purpose is to implement new software. This software will be interactive help to the Manpower to identify the best action when a fault in the wind turbine generator (WTG) appears. Also the software will be fed with the actions taken in the WTG by the Manpower and will help for further actuations. The foreground is currently being exploited by the O&M department of Alstom.
IPR exploitable measures: No measures are so far taken.
Further research: Make prognosis to anticipate the failure of the WTG and implement a predictive maintenance.
Potential / expected impact: Optimise the O&M resources reducing the CoE.
Impact at European policy level
A wider impact at the European level lies in promoting renewable energy and in the proof that efficient green energy solutions are available. It equally encourages further investment in related R&D. There is better return on investment calculations for new wind projects, representing the improved efficiency of WFs operations. It can open up new opportunities for creating new expertise and jobs in the wind energy sector, favouring also rural areas. In addition, RELIAWIND contributed a lot of material for possible standardisation of WTs. RELIAWIND consortium members participating IEC standardisation will strive for common turbine taxonomy and fault event classification.
Report on societal impacts
 Arabian, H., Feng, Y., Tavner, P.J. Deliverable D 1.1 - Report on Previous Publications Dealing With Wind Turbine Reliability, Technical Report (Project Deliverable), EU FP7 project RELIAWIND 212966
 Ribrant, J. and Bertling, L. M., Survey of Failures in Wind Power Systems with Focus on Swedish Wind Power Plants During 1997-2005, IEEE Transactions on Energy Conversion, 2007, Vol. 22, Iss. 1.
 Various Authors, Estimation of Turbine Reliability Figures within the DOWEC Project, Technical Report 10048/4, Dutch Offshore Wind Energy Converter (DOWEC), 2002.
 van Bussel, G. J. W. and Schontag, Chr., Operation and Maintenance Aspects of Large Offshore Wind Farms. In proceedings of the 1997 European Wind Energy Conference, Dublin, Ireland, 1997. pp. 272-279.
 van Bussel, G. J. W. and Zaaijer, M.B. DOWEC Concept Study, Reliability, Availability and Maintenance Aspects. In Proc. of the Marine Renewable Energy Conference (MAREC), Newcastle, United Kingdom, 2001.
 Various Authors, Wind Energy Report Deutchland, Technical Report, Insitut fur Solare Energieversorgungstechnik (ISET), Kassel, Germany, 2006.
 Harman, K., Walker, E. R. and Wilkinson, M. R., Availability Trends Observed at Operational Wind Farms, European Wind Energy Conference, April 2008, Brussels.
 Tavner, P. J., Xiang, J., Spinato, F., Reliability Analysis for Wind Turbines, Wind Energy 2007 Vol. 10, pp. 1-18.
 Spinato, F., The Reliability of Wind Turbines, PhD Thesis, Durham University, United Kingdom, December 2008.
 Spinato, F., Tavner, P. J., van Bussel, G. J .W and Koutoulakos, E., Reliability of Wind Turbine Subassemblies, IET Renewable Power Generation 2009 Vol. 3, Iss. 4, pp. 387-401.
 Landwitschaftskammer (LWK). Schleswig-Holstein, Germany: http://www.lwksh.de/cms/index.php?id=2875
 Hahn, B., Durstewitz, M., Rohrig, K., Reliability of Wind Turbines, Experience of 15 years with 1,500 WTs, in 'Wind Energy', Proc. Euromech Colloquium, Springer, Berlin, 2007, ISBN 10 3-540-33865-9.
Wilkinson, M. R., Gomez, E., Bulacio, H., Spinato, F., Hendriks, B., Reliability Profiles (Methods), Technical Report (project deliverable), RELIAWIND Deliverable D.1.2.
Personal communication: Peter Tavner to Michael Wilkinson by email 20 January 2010.
 RELIAWIND Document: Work Breakdown Structure (WBS), dated 10/06/2008.
Barbati, L., Functional Block Diagrams Specifications. RELIAWIND Deliverable D.2.0.2 January 2008.
 Barbati, L., Reliability Results. RELIAWIND Deliverable D.2.0.4 January 2010.