Feasibility of monitoring large wind turbines using photogrammetry
Introduction
It is state of the art to use accelerometers and/or strain gauges placed inside the blade or tower for dynamic measurements performed on wind turbines [1], [2], [3], [4], [5], [6], [7], [8], [9], [10]. However, these measurement systems are sensitive to lightning and electro-magnetic fields. Besides, some extra installations inside the blades such as placement of cables for power supply and data transfer are required for these applications. The signals from rotating sensors on the blades are transferred to stationary computer via slip rings or by radio/wireless transmission. For large commercial turbines the required installations and preparations (sensor calibration) might be very expensive and time consuming [11].
Moreover, the frequency range of the vibrations to be measured also limits the use of these sensors. Accelerometers cannot provide very accurate measurements for low frequency vibrations (0.3–1 Hz) that are expected to dominate the response of large wind turbines. Therefore only the higher frequency (greater than 1 Hz) vibrations can be captured accurately. The complicated nature of wind loads also makes the efficient use of these sensors on these specific structures very difficult. Since the deflections under the action of wind loading can be considered as the sum of a slowly changing static part and a rapidly changing dynamic part, identification of low frequency vibrations plays a crucial role in predicting the wind response of structures [12]. Several researchers reported that in wind response measurements, accelerometers should be used together with other systems such as GPS (Global Positioning System) which are able to detect these low frequency motions accurately [13], [14], [15]. However, it is also not practical to place the GPS sensors in the blade structure.
Fiber optic strain gauges are proposed to be a promising alternative to accelerometers and conventional strain gauges since optical sensors are not prone to electro-magnetic fields or lightning. However, it is reported that some additional feasibility tests are still needed to ensure the effective and cost efficient use of this measurement system. The factors affecting the performance of the fiber optic sensors such as sensitivity to humidity and temperature variations and the required error compensation methods should also be investigated further [16], [17].
This work aims at investigating the feasibility of applying photogrammetry to large wind turbines and the accuracy that one can expect with current state-of-the-art software and hardware. The final goal is to use the measurements for model verification and health monitoring of wind turbines. Hence, as a following study after the estimation of the measurement accuracy, the applicability of Operational Modal Analysis on the measurements to identify the modal behavior of wind turbines will be discussed as the next research step.
Section snippets
Photogrammetric measurement techniques
Photogrammetry is a measurement technique where 3D coordinates or displacements of an object can be obtained by using the 2D images taken from different locations and orientations. Although each picture provides 2D information only, very accurate 3D information related to the coordinates and/or displacements of the object can be obtained by simultaneous processing of these images as displayed in Fig. 1.
Several applications of photogrammetric measurements are currently in use and proven to
The test turbine
Our tests were conducted on a pitch controlled, variable speed Nordex N80 wind turbine with a rated power of 2.5 MW. The turbine has a rotor diameter and tower height of 80 m and can be considered as one of the largest wind turbines that are commercially available at the time the tests were conducted. Detailed information about the technical properties of the wind turbine can be obtained through the website of the manufacturer [38].
The measurements were performed by GOM mbH [39] (GOM Optical
Acquiring vibration data and post-processing
Since the turbine was not kept at a fixed yawing angle during the measurements, the relative angle between the normal of the rotation plane and the 4 camera axes changed during the measurements. In order to prevent possible sources of error, rotation plane and the corresponding axes attached to this plane were continuously updated throughout the measurements. The deformations shown below were defined with respect to this continuously updated rotation plane and coordinate axes.
The first step in
Estimation of the measurement accuracy
In photogrammetry, the measurement accuracy is usually described either in terms of pixel or the ratio of the absolute measurement error to the field of view. The accuracy is mainly related to the type and resolution of the cameras, illumination intensity, the size and visibility of the targets, and camera calibration.
Very high accuracies such as 1/50,000 (1 part in 50,000) of the observed field of view can be reached in controllable laboratory environments by using 3D calibration tools, 3D
Order analysis of the elongation error
The frequency domain analysis of the elongation data provides important information to be used in determining the source of the error. Fig. 10 aims at showing the important frequencies identified in the corresponding length changes. The frequencies, normalized with respect to the rotational frequency, are displayed along the X-axis. The Y-axis represents the marker number. The numbering of the markers was depicted in Fig. 6. The Z-axis represents the amplitude of the Fourier transform. As can
The possibility of correcting the systematic measurement errors
From the discussion in the previous section, it is clear that the causes of systematic errors can be traced back to geometry and optical effects varying along the circular path of the markers in the field of view of the cameras. Hence one could be tempted to use this knowledge to correct the measurements for those errors, such that the only remaining errors would be the random errors (arising from unavoidable pixel quantization and fitting errors).
Three different approaches can be potentially
Frequency domain investigation of measured vibration data
Fig. 11 and Fig. 12 show typical examples of PSD (Power Spectral Density) graphs of edgewise and flapwise vibrations measured by photogrammetry. It should be noted that all the frequencies displayed below are normalized with respect to the rotational frequency. In order to protect the manufacturer’s interests, the real frequencies are not given explicitly in this article. Fig. 11, Fig. 12 are presented to provide a 3D frequency distribution which also includes the information related to the
Conclusion
Based on the results of the analyses reported in this paper, it can be concluded that the random component of the coordinate measurement error is in the range of ±5 mm or 1/16,000 of field of view. This amount is consistent with the random measurement errors reported in literature. If compared to the amplitudes of the deformations (summarized in Table 1) that are expected, this accuracy can be considered to be high and even be improved further by using higher resolution cameras.
Initial results
Acknowledgements
This research project was partly funded by the We@Sea research program, financed by the Dutch Ministry of Economical Affairs.
The authors would like to thank ECN (Energy Research Center of the Netherlands) for providing the tests turbine and other technical equipment.
The authors also acknowledge the extensive contribution of Pieter Schuer (GOM mbH), Wim Cuypers (GOM mbH), Theo W. Verbruggen (ECN) and Hans J. P. Verhoef (ECN) in organizing and performing the field tests.
References (43)
Dynamic response of a Darrieus rotor wind turbine subject to turbulent flow
Engineering Structures
(1988)- et al.
Structural investigation of composite wind turbine blade considering various load cases and fatigue life
Energy
(2005) - et al.
Measurement of wind-induced response of buildings using RTK-GPS
Journal of Wind Engineering and Industrial Aerodynamics
(2002) - et al.
Application of GPS technology to measurements of displacements of high-rise structures due to weak winds
Journal of Wind Engineering and Industrial Aerodynamics
(2002) - et al.
Monitoring dynamic and quasi-static deformations of large flexible engineering structures with GPS: accuracy, limitations and promises
Engineering Structures
(2006) - et al.
FEM modeling of structures based on close range digital photogrammetry
Automation in Construction
(2009) - et al.
Out-of-plane deformation measurements of an aluminium plate during quasi-static perforation using structured light and close-range photogrammetry
International Journal of Solids and Structures
(2007) - et al.
Along-wind response of a wind turbine tower with blade coupling subjected to rotationally sampled wind loading
Engineering Structures
(2005) - et al.
Modal testing of a rotating wind turbine
(1983) - et al.
The natural excitation technique (NExT) for modal parameter extraction from operating wind turbines
(1993)
Structural response of 34-m Darrieus rotor to turbulent winds
Journal of Aerospace Engineering
Damping measurements using operational data
ASME Journal of Solar Energy Engineering
A method for determination of damping for edgewise blade vibrations
Wind Energy
Two methods for estimating aeroelastic damping of operational wind turbine modes from experiments
Wind Energy
Optical motion analysis of wind turbines
GPS measurement of wind-induced suspension bridge girder displacements
ASCE Journal of Structural Engineering
A fiber Bragg grating sensor system monitors operational load in a wind turbine rotor blade
Measurement Science and Technology
Introduction to modern photogrammetry
Cited by (138)
Non-contact damage detection in stiffened composite laminates using surface strain-field and robust regression analysis
2023, Materials Today CommunicationsUnmanned aerial vehicle-based computer vision for structural vibration measurement and condition assessment: A concise survey
2023, Journal of Infrastructure Intelligence and ResilienceCoupled static deformation of rotating blades measurement based on tip timing and validation using laser displacement sensors
2023, Mechanical Systems and Signal ProcessingAE health monitoring technique for composite wind turbine blade: a state-of-art review
2023, Acoustic Emission Signal Analysis and Damage Mode Identification of Composite Wind Turbine Blades