Back to the programme printer.gif Print

Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Advanced operation & maintenance' taking place on Thursday, 13 March 2014 at 11:15-12:45. The meet-the-authors will take place in the poster area.

Christoph Heilmann BerlinWind GmbH, Germany
Christoph Heilmann (1) P Anke Grunwald (1) Michael Melsheimer (1) Prof. Robert Liebich (2) René Kamieth (2) F
(1) BerlinWind GmbH, Berlin, Germany (2) Technical University Berlin, Berlin, Germany

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Dr. Christoph Heilmann has been working in the wind industry for fourteen years.
Since 2009, he is head of R&D at BerlinWind GmbH, Germany.
In 2005, he finished his doctorate in mechanical engineering at Technical University of Berlin, Germany. Then, he was for five years with Deutsche WindGuard Dynamics GmbH, leading R&D projects and wind energy training courses.
At BerlinWind, he develops loads and vibration measuring systems and software for wind turbines and drive trains (e.g. imbalance checks), as well as systems for rotor blade angle checks. Moreover, he is involved in load measurements and consulting projects.


Reconstruction of a wind turbine's endured operational load spectrum using a short-time load measurement and operational data


A wind turbine (WT) is normally designed, tested and certified for a design life of 20 years [1, 2]. The design loads are validated by extensive prototype load measurements for certification. 20 years is as well the typical service life (SL) for the structure (tower and foundation) approved by the German building authorities [3]. Due to low average wind speeds in the past years, a SL extension is discussed. However the remaining SL can only be quantified by comparison of the design load spectrum with the endured load spectrum which is mostly unknown since WT lack a permanent load monitoring.


The real loads in the WT structure depend on several factors, e.g.:
- Excitation by the complex 3D wind field,
- unplanned excess vibration, e.g. due to faulty rotor conditions like blade erosion and damages, aerodynamic and mass imbalance [4],
- additional loads from the control behaviour, e.g. controller-induced tower vibration [5],
- malfunction or lack of the tower dampers,
- amplification of amplitudes by WT structural defects, e.g. pre-damaged structure, loose bolts, etc.
The guidelines on continued WT operation [6] and [3] propose an in-depth visual inspection and/or a complete design analysis with site wind data. However, visual inspection finds fatigue cracks only when there is no reserve anymore and the structure is damaged, i.e. when the design loads have been exceeded. A simulation for fatigue analysis neglects the real WT operational behaviour and its structural condition. Therefore, the only suitable way to assess the loads is to measure them. A complete load measurement [7] is too costly for a single WT. Therefore, an approach has been developed in order to estimate the endured load spectrum using a short-term load measurement and combine the results with recorded operational data, e.g. of the SCADA system to reconstruct the endured load spectrum. The load measurement reveals the real loads e.g. at the tower top and bottom, Fig. 1. Simultaneously recorded operational data allows to classify the loads for the individual WT. Then, the classified operational data of the service period can be used to reconstruct the endured load spectrum and compare it with the design load spectrum to determine the structure’s fatigue and the remaining service life. The corresponding method including a suitable measurement system (BerlinWind) and the analysis software (TU Berlin) need to be developed since the typical WT design software is not able to perform this evaluation based on the operational data but only with the data from the extensive prototype measurement campaign. Moreover, it is necessary to evaluate the suitability of the method by measurements. In addition, a cost-effective calibration method for the strain gauges for the load measurement is needed.

Main body of abstract

In Germany, every year more than 1000 WTs reach the 20 years of SL and operators want to continue operation because wind conditions and energy yield in the past years was in the range of 20% below the siting assumptions. There is a discussion about how to assess the lifetime consumption of the WT’s structure and determine the remaining SL. For WT certification, the validation of the design load spectrum by extensive WT prototype measurements is required. However, the endured operational load spectrum of an individual real WT is mostly unknown since there is no permanent load monitoring. Thus, the main goal of the presented research is the development of a reliable method for the assessment of the remaining SL . The approach is divided into the following steps:
1. Preliminary inspection of the turbine: The WT, especially the structure (i.e. tower and foundation), is examined for obvious damages like visible cracks. If these damages are too heavy for any continued operation, the turbine will have to be dismantled. Otherwise, the load measurement can be planned.
2. Load measurement: Considering the requirements described in [7], a load measurement system is installed, Fig. 1. Besides the accelerations in the nacelle and the strain in the tower, weather and operational data is recorded simultaneously. To safe costs there is no additional wind measuring mast. The duration of the measurement will be approximately one month. It depends on the intensity and variety of the wind speeds since it is necessary to fill enough bins in the capture matrix where the loads are sorted according to the classified 10-min mean wind speed and turbulence intensity. With this short-time measurement, a relatively quick and, for the operator, economical assessment can be achieved. Moreover, investigation of the rotor imbalance will be performed since this is a cause for increased fatigue loads, Fig 2.
3. Reconstruction of the endured loads: With the load measurement data including operational data, a correlation between the operational state, e.g. recorded wind speed and the response of the predamaged structure of the turbine (e.g. stress at the tower bottom) for the individual WT can be made. As a conservative approach, it is assumed that the measured loads after e.g. 15 years of operation have occurred during the entire lifetime despite that they might have been lower in the first years. Then, using the available information about the turbine’s life, Fig. 3, the endured loads from commissioning up to the time of measurement can be reconstructed for comparison with the design load spectrum. For example, known downtimes, where the loads on the turbine are reduced, will extend the remaining service life, while extreme conditions with heavy wind speeds might contribute to a higher lifetime consumption, Fig. 4.
The result of this process will be a realistic estimation of the remaining SL of the structure, reduced by a quantifiable uncertainty due to measurement errors, calculation inaccuracies or missing lifetime data.
A 17 channel load measurement system has been developed which comprises three acceleration sensors, at least four strain gauges applied at the turbine’s tower and additional channels for weather data like wind speed and direction, Fig. 1. The measurement system uses a USB data acquisition board and a small PC. There is the possibility for a periodical status email via UMTS or remote control e.g. via modem. In order to calibrate the strain gauges without the cost-intensive use of a mobile crane for defined excitation, a video-based calibration method has been developed and tested. A video of the nacelle movement is recorded and analysed, providing the nacelle displacement over time. By means of a mechanical model of the tower, the strain and thus the set values for the calibration of the strain gauges is obtained, Fig. 5. The feasibility of this approach for load reconstruction was shown in another research project on the model-based imbalance reconstruction for wind turbines [9].
The measurement system and the video-based calibration method have successfully been tested for several weeks at a small wind turbine in urban wind conditions and extreme weather conditions in winter and summer. The data has been used to extensively test the evaluation software, Fig. 6.


For safe continued operation of WT beyond the design life of e.g. 20 years, an assessment of the WT’s structural condition is required by the civil building authorities. However, the relevant guidelines for continued operation [3, 6] neglect that
a) the structural loads are influenced not only the site wind regime but by many factors, e.g. and
b) load measurements show directly increased loads long before the accelerated fatigue is visible due to cracks.
Since e.g. a rotor imbalance causes increased fatigue loads, only a method based on load measurements at the individual WT is suitable to quantify the endured load spectrum. A corresponding method has been developed including a short-time load measurement with parallel recording of operational data. By combination with operational data and the WT’s lifetime recordings the endured WT load spectrum can be reconstructed. The comparison with the design load spectrum then gives the remaining SL. The uncertainty will highly depend on the amount of lifetime data and recordings and as well on the wind conditions during the load measurement. Therefore, sensitivity studies will be required to estimate the impacts. Moreover, the cost-effectiveness of the method will as well vary with the amount of data available.
The measurement system tests have been successful, and a load measurement at a thirteen year old 600 kW WT will start in a few weeks. Since the operational data of the WT are available for nearly the entire operational time, the reconstruction of the endured load spectrum can be performed. It was necessary to develop the evaluation software since commercial WT design software is not able to perform the load reconstruction but works with simulated wind data. Most software modules are successfully programmed and tested, so the feasibility of the method can be demonstrated in the next months.

The doctoral dissertation grant of Mr. René Kamieth at Technical University Berlin for the research project is gratefully funded by Reiner Lemoine Stiftung. The small wind turbine was provided by the Reiner Lemoine Institute and Hochschule für Technik Berlin, University of Applied Sciences.

Learning objectives
In Germany, for continued operation of WTs beyond the planned service life proofs concerning the structural safety are required.
A realistic determination of the remaining service life requires investigation of the WT’s endured loads.
The innovative approach reconstructs the endured load spectrum using a short-time load measurement and lifetime recordings to quantify the remaining service life by comparison with the design load spectrum.

[1] International Electrotechnical Commission: IEC 61400-1 – Wind turbines, Part 1: Design requirements, Third Edition, 2005
[2] Germanischer Lloyd Industrial Services GmbH: Guideline for the Certification of Wind Turbines, Edition 2010
[3] Deutsches Institut für Bautechnik: Richtlinie für Windenergieanlagen (Guideline for Wind Turbines), Edition October 2012, Schriften des Deutschen Instituts für Bautechnik, Reihe B, Heft 8, 2012
[4] Grunwald, A., Heilmann, C., Donth, A., Melsheimer, M.: Improving Performance of Wind Turbines through Blade Angle Optimisation and Rotor Balancing, Proceedings of EWEA 2011, March 2011, Brussels, Belgium
[5] Gasch, R.; Twele, J.: Wind Power Plants, Springer Publishers, Second Edition, 2012
[6] Germanischer Lloyd Industrial Services GmbH: Guideline for the Continued Operation of Wind Turbines, Edition 2009
[7] International Electrotechnical Commission: IEC 61400-13 – Wind turbines generator systems, Part 13: Measurement of mechanical loads, First Edition, 2001
[8] Gasch, R.; Knothe, K.: Strukturdynamik 2; Springer: Berlin, 1989
[9]. Niebsch J., Melsheimer M.: Rotor imbalance determination fit for Condition monitoring, Proceedings of EWEA 2013, Februar 2013, Vienna, Austria