Metrological features like vertical gradients in wind speeds, wind direction, and turbulence intensity are becoming more decisive in design and deployment of wind turbines (WT). This has fueled a need for new measurements techniques to collect the specific parameters for structural health monitoring (SHM) of the wind mast and blades. Importantly it has led to an increase in non-contact measurement techniques as well as the need for better-validated modelling techniques that address problems of (a) flutter and (b) type and component certification of WT blades and masts. One such novel approach is the use of Ground-Based Radar (GBR) as a non-contact SHM sensor.
GBR are increasingly being used either as a vibration‐based or as guided‐wave‐based SHM sensors for monitoring of WT blades and masts. The study contributes to the monitoring of blades and masts during design, testing, and in-field operation. The optimal monitored condition parameters (CP’s) that the GBR can utilise are identified as the component modal frequencies and deflection at blade tips and tower nacelle.
The research gap and subsequently novelty being addressed include: First, the apparent lack of singular studies that consider a portable GBR for SHM of in-field blades and masts of level 1 damage detection while the WT is operating normally under international IEC standards 61400-1 Design load conditions (DLC) 1.1. to 1.5. Secondly, the apparent lack of studies for utilising GBR under SHM framework for WT operating in actual conditions. Previous studies (and not under SHM framework) focussed on WT in a parked position or under laboratory set-ups.
This research thus sought two main objectives (a) GBR determination of the two SHM Condition Parameters (CP), modal frequencies and flapwise deflection, for an operating in-field wind turbine blade and masts and (b) Integration of GBR into the 3-tier framework and its validation thereof. In achieving these research objectives and answering the research questions, a conceptual framework cognisant with type and component certification for the blade and mast design (IEC 61400-1, IEC 61400-2) and full blade SHM under IEC 61400-23.
Using simplified Sammon mapping, 2D visualization techniques, Operational Modal Analysis (OMA’s) and three set of multi-sensor experiments the objectives of the research were progressively achieved. The first experiment involved GBR determination of the frequency on a Steel I-beam hit by an impact hammer with an accelerometer as the ground-truth. The second involved use of a custom-made rotating arm to acquire deflection of a beam structure in a rotating motion. Third experiments focussed on the CP acquisitions from an operating in-field WT. The main WT components being the WT blades and tower.
The GBR results were in the first experiment validated using a Geographical Positioning System (GPS). The aim of these 2 experiments was primarily to assess the capability of the GBR to acquire CP’s, the accuracy of the acquired values and most importantly CP’s acquisitions by GBR within a 3-tier SHM framework. Correction of the GBR results was also undertaken using Welch Power Spectrum estimation and Group delay response for the frequencies results. The third experiment had its resulted validated using an OMA’s in form of a Campbell diagram under a 3-tier SHM framework.
The key results from the 3 cluster experiments were as follows:
1.I-Beam experiment: - Comparisons between the GBR results with those of an accelerometer indicated a divergence of ±0.1% from accelerometer results when a correction was applied, and ±3% without correction.
2.Rotating arm experiment: - Employing a rotating beam structure with GPS Leica AR 25 choke antenna attached at the tip, the system was used to model a rotary structure. The deflection characterization was done using a portable GBR and the output from the GPS. Using Sammons mapping, GBR results were processed and thereafter compared those of a GPS, which indicated a maximum divergence of ± 3%
3.Operational in-field WT: - The GBR was able to acquire both the both CP’s (modal frequencies and flap wise deflections) for the blade and mast. The validation results were obtained from OMA Campbell diagram. An accuracy of 3-7% was achieved when the GBR results were compared WT design parameters as provided by the Campbell diagram. This experiment, was the focus of this thesis, by demonstrating the actual deployment of a GBR for WT blade and mast monitoring
The above three sets of experiments were thus able to demonstrate the ability of GBR for SHM of WT blades and masts, while also enabling its integration as a 3-tier SHM framework. Thus, providing significance of this research to the wider wind energy industry and monitoring the world for
(a)Flutter through the acquisition of SHM condition parameters (CP’s) that can be used to verify the results of the flutter design process for optimization of WT blades.
(b)Easing the Type and component certification process for WT blades and masts by the implementation of a 3-tier SHM framework to address the
(i).Deficiency of current fatigue damage metrics in blade-tip monitoring as well as,
(ii).Insufficient understanding of the structural behaviour of FRPC materials under long-term real operating conditions.
In conclusion, using GBR for onshore infield WT during real-time operations provides data to enable validation and improvement of current aeroelastic models for flutter analysis. Thus, providing significant information towards flutter analysis and improvement of future flutter models. The thesis also entrenches the use of GBR as a non-contact sensor in level 1 damage detection for infield WT composite blades within a 3‐tier SHM framework.
This work suggests additional work be done regarding whirling mast movements monitoring using GBR. This will particularly support the development of a flutter analysis model. Further work may also need to be done in the application of GBR monitoring within a 3-tier SHM framework for off-shore WT where vertical subsidence of the sea plays a key role.
|Date of Award||16 Nov 2019|
- Univerisity of Nottingham
|Supervisor||C.M. Hancock (Supervisor), G Roberts (Supervisor), J. Kernec (Supervisor) & Xu Tang (Supervisor)|
- mode shape
- Principal component analysis
- finite element analysis
- operational modal analysis