Welcome to Task 33 Summary Page

Reliability Data: Standardization of
Data Collection for Wind Turbine Reliability
and Maintenance Analyses

Division Energy Economy and Grid Operation
Fraunhofer Institute for Wind Energy and Energy System Technology (IWES)

Operating Agent Representative
Berthold Hahn
Fraunhofer Institute for Wind Energy and Energy System Technology (IWES) Königstor 59, 34119 Kassel, Germany Phone: +49 561-7294 229
E-Mail: berthold.hahn@iwes.fraunhofer.de

Task 33 Reliability Data: Standardization of Data Collection for Wind Turbine
Reliability and Maintenance Analyses, approved in October 2011 will build upon the discussions and work already performed in regards to failure statistics during the 65th IEA Topical Expert Meeting “International statistical analysis on wind turbine failures,” March 2011, Kassel, Germany.

• Provide an open forum on failure and maintenance statistics on wind turbines for exchange of experience from individual research projects
• Develop an IEA Wind recommended practice for collecting and reporting reliability data
• Identify research, development, and standardization needs for collecting and reporting reliability data.

Three Subtasks apply the experience of reliability analyses and failure statistics to determine common terminologies, prepare formats and guidelines for data collection (inventory, maintenance, failure, and possibly condition data), and set up procedures for analysis and reporting. The expected outcome is the formulation of guidelines for data collection, data structure, and data analyses for overall wind turbine failure statistics. The topics are complimentary work being performed to standardize data management.


High reliability guarantees a high degree of operating and personal safety, high system availability, and low necessary maintenance. These characteristics are important to reducing the cost of energy from wind plants. Modern onshore wind turbines attain high technical availability of up to 98%. Evaluation of maintenance work in previous projects shows, however, that high wind turbine availability requires additional maintenance work, which can be costly. Moreover, offshore wind farms with their increased difficulty of access for maintenance stimulates the demand for improved reliability while keeping maintenance requirements low.

Maintenance of wind turbines is currently planned and carried out according to statutory requirements and rough guidelines from manufacturers. Unplanned maintenance measures due to sudden malfunction of components can cause serious economic losses, especially offshore. Experience has shown that reliability and maintenance procedures can be improved when maintenance strategies are based on sound statistical approaches.

Statistical analyses of operation and maintenance data of turbines and their components can be used to identify weak points and to define maintenance services at an early stage. Maintenance strategies should be shifted from unplanned and corrective measures to more preventive measures based on experience acquired at many locations.

To take full advantage of historical data on reliability, a semi-automated and highly simplified data management approach is needed. Effective analysis to improve reliability and maintenance requires more parameters, data, and additional information than we see collected today. Electronically supported reporting by service teams will be necessary to supply this increased detail.

The Data

One of the most renowned wind turbine failure statistics has been established in the scientific measurement and evaluation programme “WMEP” (“Wissenschaftliches Mess- und Evaluierungsprogramm”), included in the German subsidy measure 250 MW Wind [1].The WMEP database contains a large quantity of operational and maintenance data and detailed information about both the reliability and availability of wind turbines. It provides the most comprehensive study of the long-term reliability behavior of wind turbines and the most trustworthy characteristic reliability parameters, Mean Time Between Failure (MTBF) and Mean Time To Repair (MTTR) published to date.

Besides the WMEP database there exist more publicly available sources of experience. However, these databases differ from each other in monitoring period, number, size and type of wind turbines, in the definition of subassemblies and failures, in the level of detail, and in the overall structure. Nevertheless, surveys have been compared with one another, including WMEP, Windstats Germany and Denmark, Landwirtschaftskammer Schleswig-Holstein (LWK) Germany, Elforsk Sweden and VTT Finland. It has been found that, despite their differences, there is agreement to a certain degree. However, the loads on wind turbine components differ due to the technical concepts and site conditions, which lead to a dispersion of results.

None of the publicly available failure databases are detailed and large enough for appropriate reliability analyses. Even with a broad database, like the WMEP, the breakdown in concept groups, power classes, site conditions, etc. lead to a point where the statistical basis is insufficient. The different strains of wind turbine components, e. g. because of the technical concepts in use or different site characteristics, but also the use of identical and similar components from different manufacturers will lead to different lifetime expectations of the components and thus to a spread of results. This shows the need for broader databases and for appropriate standard data structures.

Today, the wind energy industry lacks cooperation among parties: operators, manufacturers, component suppliers, designers, service providers, and researchers. Such cooperation is common in other industries such as aerospace and brings great benefits. For this reason necessary steps have to be introduced for operation and maintenance of wind turbines to bring available knowledge together and to use experience for improvements.
In addition to that, although wind energy use has been established well during the recent years, still common standards for the documentation of operation and maintenance measures as well as for a uniform structure of databases are missing.

So far, data collection is at least for operators limited, mostly as written reports from manufacturers or service companies briefly described, with encoded description. The reports are usually not detailed enough and without any failure analysis. A consequence of this lack of structure is the difficulty to carry out an effective and optimized maintenance. It is therefore essential to pursue a standardized form of data. The biggest challenges in this connection are systematic collection of data, uniform description of sub-assemblies and description of operating conditions, malfunctions and failures equally.

Currently there are several national initiatives aiming at collecting failure information for reliability analyses, e. g. Offshore~WMEP [4], Reliawind [5], EVW [6], Sandia CREW database [7] and OREDA database [8]. They all (except of [8]) intend to establish a database for wind turbine failure statistics and therefore they should somehow be coordinated to assure that the results may be combined in order to increase the statistical basis available. All initiatives share the following crucial issues:

1. Which data is to be collected?
2. What data are needed for the different analyses?
3. How to implement a system to collect information in an appropriate, structured, detailed and strongly automated way?

The uniform labeling of components, operating systems and the systematic storage of errors and data will initially enable the management a largely standardized and electronically supported logging. As a result, the monitoring process can be simplified, the financial and technical reporting improved and cooperation with similarly oriented businesses enabled. This detailed documentation of all maintenance measures of a large population of plants and a purposeful structured database are necessary to extract sound conclusions out of the operational experience. This way of documenting and collecting data and information provides a number of possibilities for optimizing availability of wind turbines both in design and construction and in operation and maintenance, which result in higher turbine efficiency.

By using detailed, systematically recorded operation and maintenance data that has been processed with standardized and electronically aided protocols, validity can be achieved. However, only through a large amount of information, weak points can be identified clearly and statements on the failure probability of certain components get meaningful and only such a large database allows improving and optimizing maintenance strategies. For this reason, defined and standardized structures are an indispensable basis for comparing or merging different databases. Analyses of captured information from collaborative databases provide resilient figures for detecting weak points and cost drivers as a basis for decision-making processes. As a result weak points can be identified, components can be qualified in cooperation with manufactures and suppliers and statements about the probability of failure behavior can be made.

For realizing standardized data collection and the comparison of different data sources with the overall aim to optimize reliability and availability of wind power plants, it has to be investigated how to possibly establish collaborations between the different initiatives with respect to sharing or grouping some data.


(1) WMEP - Wissenschaftliches Mess- und Evaluierungsprogramm/scientific measurement and evaluation program funded by the German Federal Ministry for the Environment, Nature Conversation and Nuclear Safety.