227 75 18MB
English Pages 226 [227] Year 2023
Francis Xavier Ochieng
Ground-Based Radar in Structural Design, Optimization, and Health Monitoring of Stationary and Rotating Structures
Ground-Based Radar in Structural Design, Optimization, and Health Monitoring of Stationary and Rotating Structures
Francis Xavier Ochieng
Ground-Based Radar in Structural Design, Optimization, and Health Monitoring of Stationary and Rotating Structures
Francis Xavier Ochieng Jomo Kenyatta University of Agriculture and Technology Nairobi, Kenya
ISBN 978-3-031-29007-7 ISBN 978-3-031-29008-4 (eBook) https://doi.org/10.1007/978-3-031-29008-4 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The three-tier Structural Health Monitoring (SHM), particularly of renewable energy systems with rotary movements, presents one of the unique challenges with these energy generators like wind turbines and hydropower systems from the point of non-contact sensors. The need for such non-contact sensors arises due to metrological features like vertical gradients in the wind (or water current) speeds and direction, as well as turbulence intensity. These are becoming more decisive in the design and deployment of wind turbines (WT), ocean and tidal power systems, and hydropower systems. Such needs have subsequently fuelled a need for new measurement techniques to collect the specific parameters for structural health monitoring (SHM) of the mast, turbines 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 blades, turbines, and masts. One such novel approach is the use of ground-based radar (GBR) as a non-contact SHM sensor. GBRs are increasingly being used either as vibration-based or as guided-wave- based SHM sensor for monitoring of blades, turbines, and masts components. This book contributes to the monitoring of these components during design, testing, and in-field operation. The optimal monitored condition parameters (CPs) that the GBR can utilize are identified as the component modal frequencies and deflection at component blade tips and tower/masts nacelle. The book seeks to enhance knowledge and skills as well as address research gaps with the subsequent novelty to be addressed, including: First, an 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, for example, the WT is operating under theoretically normal conditions under international IEC standards 61400-1 Design load conditions (DLC) 1.1.–1.5. Secondly, the apparent lack of studies for utilizing GBR under the SHM framework for components operating in actual conditions. A clear rationale is provided by previous studies (and not under the SHM framework) that focussed on WT in a parked position or under laboratory set-ups but not under actual conditions. v
vi
Preface
This book thus seeks three primary objectives: (a) to provide an understanding of how ground-based radar works (theory of ground-based radar); (b) to entrench the use of the three-tier SHM (structural health monitoring) framework in the context of non-contact sensors; and (c) to provide application or use cases for GBR in areas of civil engineering, energy, and human health. On a secondary level, the book using Wind Turbines (WT) as a specific case study seeks (i) GBR determination of the two SHM condition parameters (CPs), modal frequencies and flapwise deflection, for an operating in-field wind turbine blade and masts, and (ii) integration of GBR into the three-tier framework and its validation thereof. In achieving these objectives, a conceptual framework cognizant 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 is utilised. By employing several methods, including using simplified Sammon mapping, 2D visualization techniques, operational modal analysis (OMA), and three sets of multi-sensor experiments; the objectives of the book were progressively achieved. In addition, verified experiments in the utilization of GBR for non-contact SHM of WT are included in the book as case studies. One such 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 experiment involved use of a custom- made rotating arm to acquire deflection of a beam structure in a rotating motion. The third experiment focussed on the CP acquisitions from an operating in-field WT. The main WT components are the WT blades and tower. The GBR results were in the first experiment validated using a Geographical Positioning System (GPS). The aim of these two experiments was primarily to assess the capability of the GBR to acquire CPs, the accuracy of the acquired values and, most importantly, CP’s acquisitions by GBR within a three-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 results validated using an OMA in the form of a Campbell diagram under a three-tier SHM framework. The key results from the three-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 to those of a GPS, which indicated a maximum divergence of ±3%. 3. Operational in-field WT: The GBR was able to acquired both CPs (modal frequencies and flapwise deflections) for the blade and mast. The validation results were obtained from the OMA Campbell diagram. An accuracy of 3–7% was achieved when the GBR results were compared to WT design parameters as
Preface
vii
provided by the Campbell diagram. This experiment was the focus of this book, 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 and hydropower blades and towers/masts while also enabling its integration as a three-tier SHM framework. Thus, providing significance of this research to the wider wind, hydro, ocean, and tidal energy industry and monitoring for: (a) Flutter through the acquisition of SHM condition parameters (CPs) 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 three-tier SHM framework to address the:
(i) Deficiency of current fatigue damage metrics in blade-tip monitoring (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 book also entrenches the use of GBR as a non-contact sensor in level 1 damage detection for infield WT composite blades within a three-tier SHM framework. This work suggests additional work be done regarding whirling mast movement 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 three-tier SHM framework for off-shore WT, where vertical subsidence of the sea plays a key role. Keywords Modal, Deflection, Radar, Wind, Mast, Blade, Monitoring, Flutter, Wheeling, Aeroelasticity Nairobi, Kenya
Francis Xavier Ochieng
Contents
1
Introduction to 3-Tier SHM Framework���������������������������������������������� 1 1.1 Why the Three-Tier SHM Framework?�������������������������������������������� 1 1.2 Condition Parameters������������������������������������������������������������������������ 2 1.3 Organization of Book������������������������������������������������������������������������ 3
2
The Need for GBR and Coupling It with SHM������������������������������������ 5 2.1 Challenge of Aeroelasticity and Flutter�������������������������������������������� 5 2.2 Operational Gaps in use of Contact Sensors������������������������������������ 7 2.2.1 Deficiency of Current Fatigue Damage Metrics ������������������ 8 2.2.2 Insufficient Understanding of FRPC������������������������������������ 9 2.2.3 Real-Time Operating Verification of Design Results������������ 10 2.3 Defining Scope for GBR Use������������������������������������������������������������ 11 2.4 Broad Methodological Use of GBR�������������������������������������������������� 11 2.5 The Novelty of this Book������������������������������������������������������������������ 12 2.6 The Motivation for Using GBR in Three-Tier SHM������������������������ 12 2.7 Summary ������������������������������������������������������������������������������������������ 12
3
Structural Damage Detection Methods�������������������������������������������������� 13 3.1 Damage Detection Using SHM�������������������������������������������������������� 13 3.1.1 Unbalanced Parameters�������������������������������������������������������� 13 3.1.2 The Significance of SHM in Assessing Wind Turbine Loads ���������������������������������������������������������������������� 18 3.1.3 Damage Detection Using SHM Frameworks������������������������ 20 3.1.4 Tiered SHM Frameworks������������������������������������������������������ 21 3.2 GBR Applicative Area and Problem Characterization���������������������� 22 3.2.1 Blade Flutter������������������������������������������������������������������������� 23 3.2.2 Component and Type Certification �������������������������������������� 27 3.2.3 GRB Niche in Addressing the Research Problem���������������� 41 3.3 A Conceptual Framework for GBR Niche���������������������������������������� 42 3.4 Summary ������������������������������������������������������������������������������������������ 43
ix
x
Contents
4
Overview of GBR ������������������������������������������������������������������������������������ 45 4.1 Three-Tier SHM Framework������������������������������������������������������������ 45 4.2 SHM Sensors������������������������������������������������������������������������������������ 47 4.2.1 Direct and Indirect Contact-Based Sensors�������������������������� 47 4.2.2 Non-contact-Based Sensors�������������������������������������������������� 49 4.3 Radar SHM of WT Blades and Mast������������������������������������������������ 57 4.3.1 Patents for WT SHM Using Radar �������������������������������������� 57 4.3.2 Ku Band SHM Radar Systems Affixed Partly on WT���������� 57 4.3.3 L Band SHM Radar Systems������������������������������������������������ 59 4.3.4 Vector Network Analyser Radar Systems ���������������������������� 60 4.3.5 C Band SHM Radar Systems������������������������������������������������ 61 4.4 Fully Non-contact Ku Band SHM Radar Systems���������������������������� 62 4.4.1 Evolution of GBR ���������������������������������������������������������������� 63 4.4.2 Operating Frequencies���������������������������������������������������������� 67 4.4.3 Continuous and Discontinuous GBR������������������������������������ 67 4.5 Applicative GBR Case Studies �������������������������������������������������������� 68 4.5.1 GBR Monitoring of a Bridge������������������������������������������������ 69 4.5.2 GBR Deflection Monitoring of Masts���������������������������������� 69 4.5.3 GBR Deflection Monitoring of Buildings���������������������������� 69 4.6 Summary ������������������������������������������������������������������������������������������ 71
5
GBR: Working Theory and Signal Processing�������������������������������������� 73 5.1 Quasi-Monostatic GBR�������������������������������������������������������������������� 73 5.2 GBR Planar Waves and RCS������������������������������������������������������������ 76 5.3 GBR Integration into a Three-Tier SHM Framework���������������������� 80 5.4 Mathematical Model’s WT SHM Monitoring���������������������������������� 81 5.4.1 Signal Decomposition Using Short-Time Fourier-Transform (STFT)���������������������������������������������������� 83 5.4.2 Limitations of STFT ������������������������������������������������������������ 84 5.4.3 Signal Decomposition Using Sammon Mapping and 2D Visualization ������������������������������������������������������������ 84 5.4.4 GBR Results in Validation Using OMA-Based Control Charts������������������������������������������������������������������������������������ 86 5.5 GBR Windowing������������������������������������������������������������������������������ 87 5.5.1 Windowing Techniques on GBR Signal Processing ������������ 87 5.5.2 Determination of the Framework for GBR Operation at an Extremely Close Range������������������������������������������������ 89 5.6 Load Determination of Nacelle, Blade and Mast������������������������������ 90 5.6.1 Blade Deformation���������������������������������������������������������������� 91 5.6.2 Mast Deformation ���������������������������������������������������������������� 94 5.6.3 Nacelle Deformation������������������������������������������������������������ 95 5.6.4 WT Blade Testing During Design (IEC 61400-23)�������������� 96 5.6.5 WT Tower Deformation (IEC 61400-1 and Draft IEC 61400-6)������������������������������������������������������������������������ 97 5.7 Summary ������������������������������������������������������������������������������������������ 98
Contents
xi
6
Deflection and Modal Analysis from GBR Experiments���������������������� 101 6.1 Experimental Aims and Justifications���������������������������������������������� 101 6.1.1 Experimental Aims��������������������������������������������������������������� 101 6.1.2 Justification for the Choice of Sensors��������������������������������� 102 6.1.3 Justification Methodology Used and On Non-use of Modelling Techniques������������������������������������������������������ 102 6.1.4 Deformation of Structures���������������������������������������������������� 103 6.2 Experiments Undertaken������������������������������������������������������������������ 103 6.2.1 Static Laboratory-Based������������������������������������������������������� 105 6.2.2 Dynamic Laboratory-Based�������������������������������������������������� 110 6.2.3 In Situ In-Field Experiments������������������������������������������������ 112 6.3 Key Equipment Used������������������������������������������������������������������������ 114 6.3.1 GBR�������������������������������������������������������������������������������������� 114 6.3.2 GNSS/GPS System �������������������������������������������������������������� 115 6.3.3 Accelerometer ���������������������������������������������������������������������� 116 6.3.4 Wind Turbine������������������������������������������������������������������������ 118 6.4 Summary ������������������������������������������������������������������������������������������ 119
7
Case Study Application of GBR for Rotary and Non-rotary Cases���� 121 7.1 Test Cluster 1: Acquisition of Modal Frequencies���������������������������� 121 7.1.1 SHM Step 1: Data Acquisition and Normalization�������������� 121 7.1.2 SHM Step 2: Feature Extraction Using Sammon Mapping�������������������������������������������������������������������������������� 122 7.1.3 SHM Step 3: Results Validation�������������������������������������������� 128 7.1.4 Sensitivity and Error Analysis���������������������������������������������� 128 7.1.5 Analysis of the Non-dominant/Resonant Frequency������������ 129 7.1.6 Summary ������������������������������������������������������������������������������ 130 7.2 Test Cluster 2: Acquisition of Deflection CPs���������������������������������� 131 7.2.1 SHM Step 1: Data Acquisition and Normalization�������������� 131 7.2.2 SHM Step 2: Feature Extraction Using the Sammon Method���������������������������������������������������������������������������������� 134 7.2.3 SHM Step 3: Results Validation�������������������������������������������� 135 7.2.4 Sensitivity Analysis of GBR Deflection ������������������������������ 140 7.2.5 Summary ������������������������������������������������������������������������������ 142 7.3 Test Cluster 3: CP Acquisition for In-Field Operating WT�������������� 142 7.3.1 SHM 1: Data Acquisition and Normalization���������������������� 142 7.3.2 SHM Step 2: Feature Extraction Using Sammon Mapping���������������������������������������������������������������� 145 7.3.3 SHM Step 3: Results Validation�������������������������������������������� 149 7.3.4 Error Analysis and Applicability of Results������������������������� 151 7.4 Summary ������������������������������������������������������������������������������������������ 154
xii
Contents
8
The Future for GBR Nexus with Three-Tier SHM������������������������������ 155 8.1 GBR’s Contribution to DLC ������������������������������������������������������������ 155 8.1.1 Research Significance in the Achievement of Research Objectives������������������������������������������������������������������������������ 156 8.1.2 Use of GBR as a Non-contact SHM Sensor ������������������������ 156 8.1.3 Type Certification������������������������������������������������������������������ 157 8.2 Main Conclusions ���������������������������������������������������������������������������� 159 8.2.1 Comparing Results with Research Objectives���������������������� 159 8.2.2 Knowledge Gaps Addressed ������������������������������������������������ 159 8.2.3 Current Limitations of GBR Applications���������������������������� 161 8.3 Recommendations for Future Research�������������������������������������������� 162
Annexes ������������������������������������������������������������������������������������������������������������ 163 References �������������������������������������������������������������������������������������������������������� 197 Index������������������������������������������������������������������������������������������������������������������ 209
List of Figures
Fig 1.1
Failure rates for different-sized wind turbines���������������������������������������� 2
Fig. 2.1 S-N curves for various structures������������������������������������������������������������ 6 Fig. 2.2 Aerodynamic loadings variation with height������������������������������������������ 7 Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 3.5 Fig. 3.6 Fig. 3.7 Fig. 3.8 Fig. 3.9
Component reliability and failure rate�������������������������������������������������� 14 Main design and operating loads on a wind turbine ���������������������������� 19 Three-tier SHM framework for GBR use �������������������������������������������� 21 Wind turbine within the atmospheric layers ���������������������������������������� 23 Potential contribution of GBR in aeroelastic modelling���������������������� 24 Turbine loads discretization into local-component loads��������������������� 40 Subcomponent failure rates of wind turbine���������������������������������������� 41 GBR role in three-tiered SHM of wind turbines���������������������������������� 42 GBR conceptual framework used in wind turbine SHM���������������������� 43
Fig. 4.1 Fig. 4.2 Fig. 4.3 Fig. 4.4 Fig. 4.5 Fig. 4.6 Fig. 4.7 Fig. 4.8 Fig. 4.9 Fig. 4.10 Fig. 4.11 Fig. 4.12 Fig. 4.13 Fig. 4.14 Fig. 4.15 Fig. 4.16
Infrared thermography challenge for SHM of rotating WT������������������ 50 Fraunhofer tracking laser vibrometer���������������������������������������������������� 52 Laser scanner WT blade monitoring [58] �������������������������������������������� 53 Mast deflections using laser scanners �������������������������������������������������� 54 Deflection monitoring using contact radar�������������������������������������������� 59 LISA – Fore-runner of GBSAR������������������������������������������������������������ 64 IBIS-FS in RAR mode�������������������������������������������������������������������������� 64 IBIS-FL in SAR mode�������������������������������������������������������������������������� 65 FastGBSAR in SAR mode�������������������������������������������������������������������� 65 FastGBSAR in RAR mode ������������������������������������������������������������������ 66 Modal shape at 0.054 Hz���������������������������������������������������������������������� 69 GBSAR for mast deflections monitoring���������������������������������������������� 70 GBR acquisition set-up ������������������������������������������������������������������������ 71 GBR time series displacement�������������������������������������������������������������� 71 Spectral analysis������������������������������������������������������������������������������������ 72 Modal shape structure �������������������������������������������������������������������������� 72
xiii
xiv
List of Figures
Fig. 5.1 GBR processing techniques������������������������������������������������������������������ 74 Fig. 5.2 GBR acquisition of unbalanced parameters using micro-doppler effects������������������������������������������������������������������ 75 Fig. 5.3 GBR range fields���������������������������������������������������������������������������������� 77 Fig. 5.4 GBR planar waves at the far field �������������������������������������������������������� 79 Fig. 5.5 Campbell diagram for windey WD110 used in this research �������������� 82 Fig. 5.6 TFD windows at various L for Doppler signal analysis ���������������������� 88 Fig. 5.7 Kaiser window at various L������������������������������������������������������������������ 88 Fig. 5.8 Feasible ranges for different square-type antennas������������������������������ 89 Fig. 5.9 Feasible ranges for different square-type antennas������������������������������ 90 Fig. 5.10 Deflection of the rotor blade tip������������������������������������������������������������ 93 Fig. 5.11 Deformation movements of mast, nacelle and blades�������������������������� 94 Fig. 5.12 Static blade testing�������������������������������������������������������������������������������� 96 Fig. 5.13 Fatigue blade testing ���������������������������������������������������������������������������� 97 Fig. 6.1 Fig. 6.2 Fig. 6.3 Fig. 6.4 Fig. 6.5 Fig. 6.6 Fig. 6.7 Fig. 6.8 Fig. 6.9 Fig. 6.10 Fig. 6.11 Fig. 6.12 Fig. 6.13
Multi-sensor approach damage detection ������������������������������������������ 104 Deflection and frequency extraction using GBR�������������������������������� 105 Relating I-beam to WT blades for GBR feature extraction���������������� 106 Static beam impact hammer test �������������������������������������������������������� 107 Schematic drawing of the beam set-up ���������������������������������������������� 108 Deflection and frequency extraction using GBR�������������������������������� 108 GBR positions during the experiment������������������������������������������������ 109 Deflection of choke ring antenna�������������������������������������������������������� 111 Deflection characterization of a rotary system������������������������������������ 111 Deflection monitoring using remote radar������������������������������������������ 113 Orthogonal placement of the GBR ���������������������������������������������������� 114 Leica GPS 10 receiver and choke antenna������������������������������������������ 116 Schematic drawing of the Windey WD-110 turbine �������������������������� 118
Fig. 7.1 Fig. 7.2 Fig. 7.3 Fig. 7.4 Fig. 7.5 Fig. 7.6 Fig. 7.7 Fig. 7.8 Fig. 7.9 Fig. 7.10 Fig. 7.11 Fig. 7.12 Fig. 7.13 Fig. 7.14 Fig. 7.15 Fig. 7.16 Fig. 7.17
GBR acquisition of modal frequencies on a beam������������������������������ 122 Relationship between main and Sammon frequencies������������������������ 123 I-Beam Sammon-mapped frequencies������������������������������������������������ 124 Case 4 frequencies comparisons for different bins ���������������������������� 125 Frequency acquired from accelerometer�������������������������������������������� 128 Accelerometer results for frequencies below 100 Hz ������������������������ 129 Correlating the SNR to the rotating arm �������������������������������������������� 131 Rotating arm slant range, range bin, and geometric sketch���������������� 133 Deflection of the choke ring antenna�������������������������������������������������� 134 Deflection excerpt of step 1���������������������������������������������������������������� 135 GPS deflection for the complete experiment�������������������������������������� 136 GPS results related to GBR using time of day������������������������������������ 136 Cases 6 and 7 of GPS-GBR case relationship������������������������������������ 137 Case 8–12 GPS-GBR case relationship���������������������������������������������� 137 10-s Sammon map of GPS choke antenna deflection ������������������������ 138 Alternative 10-s mapping of GPS choke antenna ������������������������������ 138 2D visualizations of the GBR and GPS Sammon maps �������������������� 140
List of Figures
xv
Fig. 7.18 SNR at different turbine locations������������������������������������������������������ 143 Fig. 7.19 The GBR horizontal beam plane. (a) GBR focused on WT3 nacelle. (b) GBR focused on WT3 mid-mast ���������������������� 143 Fig. 7.20 Telescopic focussing of GBR to particular WT component �������������� 144 Fig. 7.21 GBR results from stationary and operating WT. (a) Stationary WT deflection. (b) Operating WT deflection. (c) Operating WT frequency ������������������������������������������������������������������������������������ 146 Fig. 7.22 CP extraction of modal frequencies���������������������������������������������������� 147 Fig. 7.23 Mast whirling movements along the x-axis���������������������������������������� 148 Fig. 7.24 Campbell diagram used a damage selection criterion������������������������ 149 Fig. 7.25 Bladed simulation of mast whirling movement along the x-axis�������� 151 Fig. 8.1 Role of GBR in emerging SHM needs of wind-turbine blades and mast���������������������������������������������������������������������������������������������� 159 Fig. A.1 Different chord lengths at different wind speeds�������������������������������� 164 Fig. A.2 Different linear chord and twist distributions ������������������������������������ 165 Fig. A.3 I-Beam slant range SNR profile���������������������������������������������������������� 167 Fig. A.4 Experimental Case 1 – 10:09:27 �������������������������������������������������������� 170 Fig. A.5 Experimental Case 2 – 10:10:37 �������������������������������������������������������� 171 Fig. A.6 Experimental Case 3 – 10:11:04 �������������������������������������������������������� 172 Fig. A.7 Experimental Case 4 – 10:11:29 �������������������������������������������������������� 173 Fig. A.8 Experimental Case 5 – 10:11:55 �������������������������������������������������������� 174 Fig. A.9 Experimental Case 6 – 10:12:35 �������������������������������������������������������� 175 Fig. A.10 Experimental Case 7 – 10:13:01 �������������������������������������������������������� 176 Fig. A.11 Experimental Case 8 – 10:15:37 �������������������������������������������������������� 177 Fig. A.12 Experimental Case 9 – 10:16:02 �������������������������������������������������������� 178 Fig. A.13 Experimental Case 10 – 10:16:29 ������������������������������������������������������ 179 Fig. A.14 Experimental Case 11 – 10:16:59 ������������������������������������������������������ 180 Fig. A.15 Experimental Case 12 – 10:17:31 ������������������������������������������������������ 181 Fig. A.16 Experimental Case 13 – 10:18:22 ������������������������������������������������������ 182 Fig. A.17 Experimental Case 14 – 10:18:48 ������������������������������������������������������ 183 Fig. A.18 Experimental Case 15 – 10:19:08 ������������������������������������������������������ 184 Fig. A.19 Sammon maps and 2D visualization�������������������������������������������������� 186 Fig. A.20 WT structural damage due to cyclic loads������������������������������������������ 193
List of Tables
Table 3.1 Table 3.2 Table 3.3 Table 3.4 Table 3.5 Table 3.6 Table 3.7 Table 3.8 Table 3.9
Non-contact SHM sensor characteristics������������������������������������������� 15 Blades and mast sensors to monitor unbalanced parameters������������� 16 Review of aeroelastic modelling techniques�������������������������������������� 25 Review of aeroelastic modelling techniques�������������������������������������� 26 Certification systems variations���������������������������������������������������������� 30 Load cases assessed by IEC DLC������������������������������������������������������ 31 DLCs for wind turbines [82, 88]�������������������������������������������������������� 32 Partial safety factors for loads γf�������������������������������������������������������� 33 Basic parameters for defining wind classes���������������������������������������� 38
Table 4.1 EM spectrum-based non-contact SHM of WT ���������������������������������� 58 Table 4.2 Portable radar systems for movement detection�������������������������������� 63 Table 4.3 Applications of GBR�������������������������������������������������������������������������� 68 Table 5.1 Determined GBR range fields������������������������������������������������������������ 78 Table 5.2 Wind turbine blade testing regime������������������������������������������������������ 98 Table 6.1 Table 6.2 Table 6.3 Table 6.4
Characteristics of GBR used������������������������������������������������������������ 114 Characteristics of GBR antenna ������������������������������������������������������ 115 Characteristics of AR25 choke ring antenna [246]�������������������������� 116 Technical specifications for the wind turbine [222] ������������������������ 119
Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 7.7 Table 7.6 Table 7.8 Table 7.9 Table 7.10 Table 7.11
Selected range bins and ranges�������������������������������������������������������� 123 Error analysis of GBR���������������������������������������������������������������������� 129 SNR scan results for range bin 25���������������������������������������������������� 132 Average deflection between GBR and GNSS/GPS�������������������������� 141 GBR sensitivity toward arm rotation velocity���������������������������������� 141 WT3 out-of-plane blade tip deflection �������������������������������������������� 144 Experimental case 1 WT frequency�������������������������������������������������� 147 WT3 deflection calculations ������������������������������������������������������������ 148 WT3 resonant frequencies compared to design������������������������������� 150 WT3 validation between GBR and OMA���������������������������������������� 150 Comparison of modelled, averaged OMA and measured data�������� 153 xvii
xviii
List of Tables
Table 8.1 Research gap addressed by the use of GBR ������������������������������������ 156 Table 8.2 WT3 resonant frequencies���������������������������������������������������������������� 157 Table 8.3 WT3 resonant frequencies���������������������������������������������������������������� 160 Table A.1 Table A.2 Table A.3 Table A.4
Relationship between antenna diameter and field type�������������������� 166 Transition boundaries of near and far fields ������������������������������������ 166 Binning of data from different experimental cases�������������������������� 168 GBR feature extraction using Sammon mapping���������������������������� 169
Abbreviations and Notations
* Fd
Partial safety for fatigue (only in reference to Sect. 3.2.2.5) Design value for the aggregated internal load or load response to multiple simultaneous load components from various sources for the given DLC fd Design values for materials fk Characteristic values of material properties γf Partial safety factor for loads γm Partial safety factors for materials γn A partial safety factor for consequences of failure A Abnormal A Designates the category for higher turbulence characteristics (only in reference to Table 3.9) A2LA American Association for Laboratory Accreditation B Designates the category for medium turbulence characteristics (only in reference to Table 3.9) C Designates the category for lower turbulence characteristics (only in reference to Table 3.9) CCD Charge coupled device CP Condition parameters (also called unbalanced parameters) CSLDV Continuous scan laser Doppler vibrometer DLC Design load case DNV-GL Det Norske Veritas and Germanischer Lloyd ECD Extreme coherent gust with direction change EDC Extreme direction change EOCs Environmental and operational conditions EOG Extreme operating gust ETM Extreme turbulence model EWM Extreme wind speed model EWS Extreme wind shear F Fatigue Fk Characteristic value for the load xix
xx
Abbreviations and Notations
FMCW Frequency-modulated continuous wave FRPC Fibre-reinforced polymer composite GBNW-SAR Ground-based noise waveform SAR GBR Ground-based real aperture radar GNSS Global navigation satellite system HT Hypobook testing IDIC International Doctoral Innovation Centre IEC International Electrotechnical Commission Iref The expected value of the turbulence intensity at 15 m/s JTF Joint time-frequency LDV Laser Doppler vibrometer N Normal (only in reference to Sect. 3.2.2.5) NTM Normal turbulence model NWP The normal wind profile model OMA Operational mode analysis PR Rayleigh distribution RO Research objective RoC Receiver operating curves SAR Synthetic aperture radar SCADA Supervisory control and data acquisition SFCW Stepped frequency continuous wave SHM Structural health monitoring SL-FMCW Step linear frequency-modulated continuous wave SWCC Small Wind Certification Council T Transport and erection U Ultimate strength UKAS United Kingdom Accreditation Service UWB Ultra-wideband Vhub Wind speed at hub height (m/s) Vin Cut in wind speed Vmaint The upper limit for wind speed during transport, assembly, maintenance and repair VNA Vector network analyser Vout Cut-out wind speed Vr ± 2m/s Sensitivity to all wind speeds in the range Vref The reference wind speed average of over 10 min Ω Rotor RPM 𝐴 The amplitude of the reaction force 𝐹𝑥𝑁𝑅 Non-rotating reaction force on the mast in forwarding and backward direction 𝐹𝑦𝑁𝑅 Non-rotating reaction force on the mast in the left and right (sideways) direction w Edgewise frequency of the blade
Chapter 1
Introduction to 3-Tier SHM Framework
1.1 Why the Three-Tier SHM Framework? Recent advances in radar technologies have made it possible and practical to use ground-based real aperture radar (GBR) for monitoring displacements for beam- like structures such as bridges and masts [1–3]. However, using radar for structural health monitoring (SHM) of WT components (blades and masts), during design, testing and operation, may not have been fully explored within the framework of a three-tier SHM framework [4, 5]. A three-tier SHM framework is a monitoring three-stage monitoring framework that includes (a) data normalization, (b) feature extraction and (c) hypothesis testing (HT) [6, 7]. The data normalization as the first tier involves the acquisition and clustering of the data into bins based on the environmental and operating conditions (EoCs). Tier 2 – feature extraction involves the identification of condition parameters (CPs) like modal parameters (modal frequency and damping) and time series (eigenvectors) among others. The CPs to be acquired in this study are the unbalanced parameters – blade resonant frequency and deflection. Subsequently, the three-tier employs analysis of the CPs through HT. The HT tier ascertains damage to the WT component by either (a) comparison of the data in the unknown state with that of a known state, e.g., healthy state, or (b) application of statistical models with established thresholds and decision boundaries that can be tracked over time using methods like x – control charts [6, 8]. Generally smaller WTs have smaller failure rates compared to WTs rated 1MW and above [9]. It is noted by [9] that in the first 8 years of operation, a turbine rated 1 MW and above has almost 4 times the failure rates of smaller WT (Fig 1.1 from [9]). Previous studies using three-tier SHM of WT blades and mast may not be numerous [10]; they do however indicate that to detect damage, four main approaches exist. These include the transmittance function, resonant comparison, operational © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 F. X. Ochieng, Ground-Based Radar in Structural Design, Optimization, and Health Monitoring of Stationary and Rotating Structures, https://doi.org/10.1007/978-3-031-29008-4_1
1
2
1 Introduction to 3-Tier SHM Framework
Fig 1.1 Failure rates for different-sized wind turbines
deflection shape and wave propagation methods [11]. By providing this information to WT manufacturers, owners and operators, they can save on costs and time during the 10–30 years life cycle of the turbine. Most importantly SHM provides significant early fault detection to prevent catastrophic WT failures. However, the emphasis and application of SHM are entirely different from condition monitoring (CM), as explained below. CM monitors for machinery parts that are in motion such as bearings, gearboxes, shafts, converters (mechanical to electrical) and fluid machines. SHM is geared toward monitoring load-bearing components like blades and masts [12, 13]. The focus of SHM is the load-bearing capacity, stability of the structure and remaining lifetime and the end of the design life. SHM does not usually consider the ultimate limit state of the structure [12]. In [13] CM is concerned with the mechanical imbalance and aerodynamic imbalance as pertains to rotors while for masts its fatigue and corrosion, thus validating the definitional focus of SHM.
1.2 Condition Parameters While it is arguable as to what CP to use for CM or SHM, [12] points out that the WT has features that lend credence to both. For instance, the root blade is connected to a shaft that is linked to the gearbox and the generator. The latter components, except the blade itself, will be monitored using CM. The blade is however monitored using SHM since the interest is on the load-bearing capacity and the stiffness of the blade and mast. The CP for such a case would then be natural frequency and deflections. SHM is commonly used in the laboratory for blade lifetime testing
1.3 Organization of Book
3
through deflection cycles as well as with natural frequency [9]. The main purpose of this is to provide data for the validation of the simulation and theoretical models. With increasing sizes, costs and fabrication sophistication, the need for continuous SHM is becoming essential, especially considering that WTs in remote areas may require them to be monitored remotely [9]. Thus, this work builds on the previous works and improves upon them, by exploring the use of a non-contact, quasi- monostatic ground-based interferometric radar system (GBR) to undertake non-contact three-tier SHM of a WT blade and mast. In this research, the author used the GBR and contact sensors to acquire three representative tests using three different configurations: a rotating arm, a stationary beam and a large-scale operating WT (>1 MW). By comparing the results of each representative test with results from a contact sensor or known receiver operating curves (RoC) like a Campbell diagram, the GBR can then be assessed whether it is accurate and fit-for-purpose as a non-contact sensor in three-tier SHM This enabled the achievement of two key objectives: (i) RO 1: Determination of the modal frequencies and flapwise deflection parameters of an operating WT blade and masts using ground-based radar (GBR). The process which can be applied to other rotary systems in Ocean - tidal and hydropower systems. (ii) RO 2: Integration of GBR into the three-tier framework and its validation thereof within the three-tier SHM framework for WT blades and masts. The integration of a GBR into SHM has the potential to improve the next generation of supervisory control and data acquisition (SCADA) under Industry 4.0 [13], while also opening an entirely new frontier in non-contact/remote sensors for a rotor blade and mast SHM. It may also contribute to the understanding of mechanical and aerodynamic imbalance under ISO 16079-1:2017 – condition monitoring and diagnostics of WT [13].
1.3 Organization of Book The rest of the book is organized as follows: In Chap. 2 the context and objectives in how to relate the GBR-SHM nexus of the book are provided. The contextual aspect addressed in this chapter is the three research gaps – deficiency of current fatigue damage metrics, insufficient understanding of fibre-reinforced plastic composites and gaps in real-time operating verification of designs results. The gaps enable the identification of two research questions, two research objectives and subsequent identification of the novelty of the research, which is reinforced in Chap. 3. Chapter 3 provides the theoretical background of SHM frameworks and potential problems that GBR may address within the three-tier SHM framework. Under SHM frameworks the main aspects covered are the three- and four-tier SHM frameworks, unbalanced parameters and the significance of SHM in assessing WT loads. The
4
1 Introduction to 3-Tier SHM Framework
chapter also addresses two kinds of problems, namely, blade flutter and component and type certification. Based on these aspects, a conceptual framework is developed. To address the GBR-SHM literature review perspective, Chap. 4 presents the state of the art concerning direct, indirect and non-contact SHM sensors. It then provides a detailed description of the evolution and types of GBR. Lastly, it presents case studies of different applicative use of GBR in the monitoring of bridges, masts, buildings and in-field WT. It is demonstrating, particularly in the latter case, the research gap that the book seeks to address. Chapter 5 thus addresses the methodology employed and the experimental setups to demonstrate and validate GBR use. The chapter presents the mathematical framework for blade and mast SHM monitoring which is a signal decomposition using two approaches (a) non-metric multidimensional scaling (neural networks, fuzzy networks, evidential and Bayesian approaches) and (b) Sammon mapping and 2D visualization. As discussed in Sect. 5.4.1, the lack of prior distributions and contact sensors limits the use of non-metric multi-dimensional scaling in the particular use of GBR in our case, hence leading to the use of approach (b) in the research. In conclusion, the chapter addresses the key equipment used as well as the three clusters of experiments undertaken, namely, static laboratory tests and dynamic laboratory tests to assess and validate the GBR accuracy and capabilities and lastly the in situ in-field tests. In Chap. 6, the results and discussions of the arising from the experiments are presented. The section is divided into four parts: general results from windowing techniques and near-field GBR operation to help assess GBR placement during the experiments. The second part addresses the static laboratory tests; the third, dynamic tests; and the fourth, the actual in-field WT tests. The study results are discussed and analysed in each sub-chapter section. Chapter 7 then concludes the book by comparing results with overall book objectives and providing recommendations for further research. It also addresses the limitations faced in doing GBR research work in addition to the main conclusion to be drawn from this research work. The remaining parts of the book are the References and annexes.
Chapter 2
The Need for GBR and Coupling It with SHM
2.1 Challenge of Aeroelasticity and Flutter Wind turbine blades, unlike commercial aircraft blades, are fatigue-critical structures. Their design is dictated not only by fatigue but also increasingly by aeroelastic considerations. The design of aircraft blades will mainly consider ultimate strength, while for WT blades, fatigue and aeroelastic aspects are the major considerations. When being designed, the WT blades will experience approximately 100 million design cycle loads over their lifetime, which is several orders of magnitude more cycles in testing than for aircraft blades (one million lifetime design cycles) (Fig. 2.1). This makes wind turbine blade testing and designs extremely expensive. The approach normally used is for physical testing for some orders of magnitudes, and the residual is achieved by simulations to give as one of the outputs an S-N curve (Fig. 2.1 adapted from [14]). A study by [14] demonstrated that wind turbines have a higher S-N cycle range (7.95–9.21) than that of other rotary and beam-like dynamic systems such as bridges and aircraft wings (Fig. 2.1). This is because wind turbines are fatigue-critical structures (i.e. the design of many components are determined by fatigue conditions like wind loads). In addition, the use of fibre composites increases the number of fatigue cycles to failure for WT blades (Fig. 2.1). It is further noted in [1] that, unlike the sub-megawatt wind turbines with a rotor diameter of 40 m that were prevalent 20 years ago, new models have increasingly longer mast and rotor lengths. For instance, an 8-megawatt (MW) WT averages a rotor length of 154 m. Possibilities of having a 20 MW WT by the year 2020 has also been demonstrated by the European “Upwind” project [14, 15]. The reason for such increased lengths is to capture higher wind speeds at increasing heights above ground (Fig. 2.2) by leveraging on the atmospheric boundary layer phenomena [14– 16] – the higher you go, the higher the wind speeds. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 F. X. Ochieng, Ground-Based Radar in Structural Design, Optimization, and Health Monitoring of Stationary and Rotating Structures, https://doi.org/10.1007/978-3-031-29008-4_2
5
6
2 The Need for GBR and Coupling It with SHM
Fig. 2.1 S-N curves for various structures
Longer blades in tandem with leveraging the atmospheric boundary layer practically increase the effective area available for the extraction of wind power. The optimization of longer blades is focused on weight reduction, thickness reduction and flexibility increase. Consequently, monitoring the blade, especially during operation is thus paramount; new sensing modalities to address the emerging problems of aeroelasticity, composite materials and blade failures are required. Aeroelasticity (particularly flutter), fatigue and the use of composite materials complicate the monitoring of in-field rotating wind turbines with multiple deflections in multiple directions. Consequently, the importance of system health monitoring (SHM) to monitor deflection cannot be overstated. It requires an array of sensors for monitoring not only the generator, gear systems and rotating hub but more importantly the masts and blades as they are increasingly made bigger and employ composites materials (for blades). Until now the design of blades to deal with flutter and composite materials has mainly relied on numerical simulations, but simulations could significantly be improved by real-time data of in-field wind turbines using the non-contact sensor. This book evaluates a ground-based radar (GBR) as one such sensor. Subsequently, this book will; first, provide and formalize a methodological way in which GBR can be used within the SHM framework. Secondly, it will identify, test and validate an analytical method for the analysis of GBR signals to facilitate
2.2 Operational Gaps in use of Contact Sensors
7
Fig. 2.2 Aerodynamic loadings variation with height
the characterization of modal frequencies and deflection parameters for non- stationary rotational GBR signals for SHM.
2.2 Operational Gaps in use of Contact Sensors Although encouraging results have been obtained in both analytical and small-scale experiments for SHM of wind-turbine blades and masts, several challenges have emerged that need to be addressed before the SHM of in-field rotating wind-turbine structures becomes a reality. Among those, a well-known challenge is an uncertainty in the measured modal parameters, affected by external and uncontrollable environmental and operating conditions (EoCs) such as temperature, temperature gradients and humidity. To meet this operational sensoring gap, this book demonstrates the use of a non-contact ground-based radar for SHM of WT components (Fig. 3.3). Three interrelated reasons underpin the need for the utilization of GBR in wind, ocean, tidal and hydropower structures: (i) The deficiency of current fatigue damage metrics in tower/blade-tip monitoring (ii) The insufficient understanding of the structural behaviour of FRPC materials under long-term real operating conditions (iii) The need to verify the results of the design process for the optimization of wind-turbine blades
8
2 The Need for GBR and Coupling It with SHM
2.2.1 Deficiency of Current Fatigue Damage Metrics Fatigue is considered in this research as the material weakening due to time-varying loads or time-dependent phenomena leading to failure of the structure [17, 18]. It can be in the form of low cycle fatigues (LCF) (i.e. below 1000 Hz) or high cycle (i.e. one above 1000 HZ). Either type can have a loading (force bearing on the structure to cause fatigue) that is cyclic or spectrum in nature. Cyclic loading has its amplitude in constant steps, while spectrum fatigues the cycles may not be clearly defined. Metrics are used to quantify fatigue as shown below. 2.2.1.1 Fatigue Damage Metrics for Wind Turbines and Similar Structures Generally, fatigue damage (FDam) can be measured using damage metrics like decreasing residual stiffness or residual strength [19, 20]. These residual metrics change very slowly until the component is about to fail when they decrease rapidly until it reaches the maximum value for cyclic stress and then failure occurs (sudden death phenomena) [20]. Before the final failure, composites are known to suffer many micro-damages [21]. For wind-turbine blades, this may not be desirable. One would prefer to know the deviation of the damage metric from the normal and take corrective actions beforehand. Fortunately for a wind-turbine mast, they are normally built with rolled steel and/ or concrete and thus are generally isotropic and homogenous and can be modelled accurately. Therefore, wind-turbine mast failures are comparatively much lower than blade failures. Predicting micro-damage and loss of stiffness for wind-turbine blades is, however, extremely difficult due to the computational efforts required [21]. Additionally, though considered insensitive to fatigue, FRPC does suffer from it. It is mainly attributed to the fact they are made from a “composite” of materials (heterogeneous materials). The fatigue behaviour of composite materials is different from that of metallic ones (due to the material having different properties, e.g. Young’s modulus in different directions is anisotropic), and consequently the classical or traditional fatigue life modelling and prediction cannot be applied here directly. It will be further pointed out that current design standards (IEC 61400-1) acknowledge that composites are materials whose design codes are unknown and thus general partial safety factors are attributed to them (see Sect. 3.2.2.7). The impacts of such attribution are: (i) An overdesigning of the system for safety purposes – hence increased costs (ii) Higher computational loads (iii) Expensive blade design process and products (iv) Difficulty in ascertaining accurate FDam
2.2 Operational Gaps in use of Contact Sensors
9
2.2.1.2 Main Fatigue Damage Metrics General deficiency of fatigue metrics arises from S-N curves and Palmgren-Miner rule. The two major limitations of the S-N curve are that they are empirical and usually come from constant amplitude cyclic loading at one given load ratio. Further, they cannot be used under 1 kHz cycles (where fracture mechanics would work better) nor at very high cycles since the generated S-N curves become horizontal, with increased scatter in the results – making it hard to predict the probability of failure. The other deficiency is that the Palmgren-Miner rules assume that the structure’s fatigue is independent of frequency and that there is a linear accumulation of damage. Incidentally, in cases of visco-elastic materials like high-temperature metals, polymers or composites like FRPC, frequency effects become important [17]. Further, the linear accumulation of damage assumption may fail due partly to local plastic strains or visco-elasticity [18]. One approach to relate fatigue damage accumulation with frequency is by monitoring modal frequencies of the first seven modes [22]; another approach is explored in a fatigue damage spectrum (FDS). The FDS relates the total damage a structure experiences at a particular time period and is plotted for a specific range of frequencies. The plotting is done using approaches like Henderson-Piersol’s fatigue calculation method [23], while the total damage is calculated from the energy of the vibration test (also called power spectral density [PSD]). To acquire the vibrational data, Fibre Bragg strain gauges are normally inserted inside the WT blade. These will reflect a narrow portion of the light spectra that is shifted due to a change in stress or temperature. With a strain sensitivity of ≈1.2 pm/ με and temperature sensitivity, 10.2 pm/°C – a 1 °C change in temperature is equivalent to 8.5 με. As noted by [24] the Fibre Bragg gratings are affected by the non- homogeneity of the blade composites; thus these sensors measure over small lengths only – a delimiting factor for overall blade length monitoring especially when in operation.
2.2.2 Insufficient Understanding of FRPC Another operational gap arises in the increased use of composite materials. Such composites include glass fibre-epoxy, glass fibre polyester, wood-epoxy and carbon fibre-epoxy materials in the making of WT blades. Others include fibre-reinforced plastic composites (FRPC) and carbon fibre-reinforced plastics (CFRP). The latter tends to be quite expensive and hence is used on very specific parts, while the other composites especially FRPC are widely used [25]. Being anisotropic and heterogeneous, FRPC is mechanically different from metals and other isotropic materials making their fatigue analysis difficult [21]. Thus, most FRPC WT blade’s fatigue damage assessment is done under the simplified load spectrum giving rise to an M-N curve (applied moment against allowable
10
2 The Need for GBR and Coupling It with SHM
cycles to failure) instead of the S-N curve. This is because composite materials behave differently under different loading conditions, e.g. tensile, uniaxial and compressive. The use of a simplified load spectrum means that WT blades are thus over- dimensioned to compensate for the uncertainty during their design. This increases the cost of the WT. It also creates uncertainty during the SHM of WT blades when it is rotating. The fatigue damage on the blade can occur at different locations by different modes such as buckling, delamination and bolt failure among others [26]. Thus, the damage associated with each M-N curve is determined using commonly accepted fatigue behaviour assessment methods such as the Palmgren-Miner equation to calculate fatigue damage (FDam). Equation (2.1) is used on a machine component under variable stress with fluctuating amplitude of the blades over its lifetime for fatigue and extreme loads [20, 26, 27]. Failure will occur when FDam = 1. i 1
FDam j
ni (2.1) Ni
where i is the load case number, j the total number of load cases, ni the counted number of fatigue load cycles for each case in the bin I of the characteristic load spectrum and Ni the number of load cycles to failure for case i.
2.2.3 Real-Time Operating Verification of Design Results When stochastically fluctuating loads occur due to varying wind shear profiles, pressure can lead to structural failure of the FRPC. A further challenge of FRPC WT components is where the temperature is involved. In [20] it is noted that no method exists in the literature for predicting fatigue behaviour under thermomechanical loading patterns as would be experienced with S-class wind turbines working in very hot (as in the equator and desert conditions), very cold or fluctuating temperature regimes like those experienced offshore, thus underscoring the need for real- time operational monitoring [28], using remote sensors potentially. In determining long-term fatigue damage during blade manufacture, an extrapolation of experimental data using various approaches and modelling including load multipliers/safety factors [29] is applied to Eq. (2.1). Subsequently, GBR monitoring of rotating wind turbines can provide abundant information to verify or correct the deficiencies arising from modelling and consequently tackle the problem of over-dimensioning in the design phase of wind-turbine blades to reduce production costs, reduce weight and ultimately improve power efficiency. This, in turn, would provide insights into the main causes of FRPC failures including cumulative and combined fatigue-related damages causing delamination, matrix failure and fibre pull-out [20].
2.4 Broad Methodological Use of GBR
11
2.3 Defining Scope for GBR Use From the foregoing Sect. 2.2, the main problem rationalising the use of a GBR lies within the first and second tier of the problem schematic specifically this lies in the acquisition and analysis of the CP. Subsequently, two research areas need to be addressed for the efficient use of GBR: 1. Determine the optimal SHM condition parameters for integration of GBR into a three-tier SHM framework1 for in-field onshore WT blades, masts and other structures. 2. How advantageous, relevant and accurate is GBR as a vibration-based damage detection SHM sensor for WT blades and mast components for normal WT operating at low-frequency ranges of 0–5 Hz2?
2.4 Broad Methodological Use of GBR To use the GBR and apply it to different structures, it is necessary to formulate the objectives and the vibrational condition parameters (CPs) as key solutions toward employing the GBR. These vibrational CPs include: • Fatigue loads – Resonances/vibrations, transient sources from mast shadow and ice accretion. • Ultimate strength loads – Transients that could be time-varying, braking, start and stop. Additional sources include cyclic sources like resonance, rotor rotation, wind shear and yaw motion. • Aeroelastic loads – Mainly arising from the aerodynamic aspects of wind shear and wind variations both in speed and direction. Flutter has not been previously considered until now. From the CPs, unbalanced parameters (deflection as well modal parameters like deflection and natural frequencies) may provide the best approach to the test regime used in this book and allow the formulation of two GBR applicative objectives (RO). 1. RO 1: Determination of the modal frequencies and flapwise deflection parameters of an operating wind turbine blade and masts using ground-based radar (GBR) – (proof of principle objective) 2. RO 2: Integration of GBR into the three-tier structural health monitoring framework and its validation thereof for WT blades and mast – (a proof of concept objective)
Referenced to Sects. 3.2.2.7 and 3.2.1 Referenced to Sect. 3.2.2.2
1 2
12
2 The Need for GBR and Coupling It with SHM
2.5 The Novelty of this Book The novelty of this book is thus twofold: 1. It proves the integration of a non-contact sensor, the GBR, within an SHM framework, for masts and heterogenous blades, to support type and component certification3 under DLC 1.1–1.5 2. It introduces the simplified version of the Sammon mapping to extract the GBR results for validation with actual deflection and/or natural frequencies from WT components
2.6 The Motivation for Using GBR in Three-Tier SHM 1. Reduced costs in the design phases of composite blades due to the inadequacy of current approaches and hence the need for enhanced monitoring techniques. 2. Reduced development, operational, monitoring and maintenance costs, since the GBR can be used to monitor several WT at the same time, provided proper benchmarking and space between wind turbines is accounted for. 3. Demonstrate veracity of non-contact GBR use in a three-tier SHM framework for WT.
2.7 Summary This section has covered the contextual purpose for which this book is written. Additionally, it has listed two main research objectives and two applicative areas that will be dealt with in the remainder of the book. Key in this section is the novelty that the proposed research seeks to bring. In the following chapters will be found an in-depth analysis of the aforementioned and proposed contributary novelty in terms of structural damage detection methods as well as deflection and modal analysis respectively.
Type and component certification refer to the international WT standards IEC 61400-23 that certifies WT components for different types of operations. For WT blades this is IEC 61400-23 under design load (DL) 1.1. to 1.5. IEC 61400-1 covers both towers and WT blades. 3
Chapter 3
Structural Damage Detection Methods
3.1 Damage Detection Using SHM 3.1.1 Unbalanced Parameters Unbalanced parameters normally lead to excessive structural vibration and ultimately damage to bearings, rotor blades, mast, and other electro-mechanical component destruction. For wind energy, the key unbalanced parameters would be radial velocity (velocity at which the structure is deflecting/displaced) and the modal parameters [30]. In [30], the modal parameters refers to the dynamic linear and time-varying parameters as a result of resonant vibration. The modal parameter estimation considers the displacement (deflection) and structure model (in terms of modal frequency, modal damping and mode shape) to prevent component/structure failure. Figure 3.1, adapted from [10], shows that the component failure rate per hour is 67.28% higher for blade-tip breaks than for any other part. The rest of the blades, bolts, hub and yaw are on average equal (~7.5%). Thus, the major structural damage in wind turbine (WT) is blade damage (>67.28%). Hence, undertaking a structural health monitoring (SHM) of blades’ Condition Parameters (CP) provides the first early warning system for turbine failure or underperformance [30]. This is critical especially as wind turbines are getting bigger (with blades longer than 45 m) and increasingly being situated in remote regions and open seas, making their inspection costly, time-consuming and requiring the use of non-destructive monitoring approaches, especially in countries with poor infrastructure and inadequate lifting equipment like cranes. Consequential deficiencies in SHM, however, tend to result in damage to the blades and other WT components [10]. Such damage includes layer debonding (skin/adhesive layer, main spar/adhesive layer), panel face/core debonding, adhesive joint failure between skins, tensional/buckling load delamination and gel coat © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 F. X. Ochieng, Ground-Based Radar in Structural Design, Optimization, and Health Monitoring of Stationary and Rotating Structures, https://doi.org/10.1007/978-3-031-29008-4_3
13
14
3 Structural Damage Detection Methods
Fig. 3.1 Component reliability and failure rate
cracking/debonding [31–33]. For these reasons, the rise of remote sensing data acquired from the system’s, SHM such as Supervisory Control and Data Acquisition (SCADA), is becoming more prevalent. While structural vibrations [34, 35] of gearboxes, blades and mast positioning [36] provide a simple basis to measure the unbalanced parameters, this can be achieved by use of contact sensors, additionally various non-contact sensors can be used (Table 3.1). Determining unbalanced parameters (CPs acquired from a rotating structure) is generally achieved by a combination of contact-based static or dynamic [19], sensing technology, signal analysis, and interpretation algorithms [10] for in-field blade and mast tests. The sensors assess the reaction of the blades and their materials (or masts respectively) in terms of strains, accelerations, deformations, and deflections. Typically, strain gauges, accelerometers and displacement sensors are utilized as sensors (Table 3.2), of which the results so obtained are referenced as CP within the SHM framework. Unbalanced parameters assess excessive vibrations indicating blade imbalance. They normally lead to excessive structural vibrations and ultimately damage bearings, rotor blades, masts, and electromechanical component destruction. The main CPs considered are the radial velocity (velocity at which structure is deflecting) and the dynamic linear and time-variant modal parameters. These modal parameters [30] include the radial displacement (deflection) and modal structure characteristics (mode shape, natural/modal frequency and modal damping) (Fig. 3.9). Typical causes [37, 38] for undesirable CP include unbalanced blades caused by high wind shear, errors in control mechanism, manufacturing tolerances, material
Table 3.1 Non-contact SHM sensor characteristics
3.1 Damage Detection Using SHM 15
1 Hz up to 50 kHz for lidar
Optical, photogrammetry, drones and lidar
Embedded or contact
Remote
Acoustic emission
Sensor Type Ground-based radar Remote, portable X-ray ~1 Hz Remote, not portable