In this chapter, vulnerable offshore wind turbine components, which could lead to long downtimes, are discussed based on the reliability data observed. As the failures of these components can potentially lead to long downtimes and significant revenue loss in the offshore environment, their operation and health condition should be properly monitored. In this chapter, the significance of monitoring these vulnerable components is further emphasized through discussing the consequent economic impacts of their failures. Following this, the relevant wind turbine condition monitoring techniques and commercially available wind turbine condition monitoring systems are discussed. Finally, the chapter ends with a summary of the existing challenges and issues as well as the future research tendencies in the field of offshore wind turbine condition monitoring.
Table 18.1
Non-destructive techniques applicable to wind turbine condition monitoring
No | CM techniques | Cost | Online CM | Fault diagnosis | Deployment | WT components |
1 | Thermocouple | Low | Y | N | Already used | Bearings Generator Converter Nacelle Transformer |
2 | Oil particle counter | Low | Y | N | Already used | Gearbox Bearing |
3 | Vibration analysis | Low | Y | Y | Already used | Main shaft Main bearing Gearbox Generator Nacelle Tower Foundation |
4 | Ultrasonic testing | Low to medium | Y | N | Being tested | Tower Blades |
5 | Electric effects (eg discharge measurement) | Low | Y | N | Already used | Generator |
6 | Vibroacoustic measurement | Medium | Y | N | N | Blade Main bearing Gearbox Generator |
7 | Oil quality analysis | Medium to high | N | Y | N | Gearbox Bearing |
Table Continued |
No | CM techniques | Cost | Online CM | Fault diagnosis | Deployment | WT components |
8 | Acoustic emission transducers | High | Y | N | N | Blade Main bearing Gearbox Generator Tower |
9 | Torsional vibration (encoder-based) | Low | Y | N | Being tested | Main shaft Gearbox |
10 | Fibre optic strain gauges | Very high | Y | N | Already used | Blade |
11 | Thermography | Very high | Y | N | N | Blade Main shaft Main bearing Gearbox Generator Converter Nacelle Transformer |
12 | Shaft torque measurement | Very high | Y | N | Being tested | Blades Main shaft Main bearing |
13 | Shock pulse method (SPM) [22] | Low | Y | N | N | Bearing Gearbox |
Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673–693.
Table 18.2
No | Name | Product information | ||
Product | Company | Major functions | Notes | |
1 | WindCon3.0 | SKF (Sweden) | Collect, analyse and compile condition-monitoring data that can be configured to suit management, operators and maintenance engineers | The system focuses on the condition monitoring of wind turbine blades, main bearing, shaft, gearbox, generator and tower by the combined use of vibration transducers and a lube oil debris counter |
2 | TCM | Gram & Juhl (Denmark) | Advanced signal analyses on vibration, vibroacoustic and strain, combined with automation rules and algorithms for generating references and alarms | The WT blades, main bearing, shaft, gearbox, generator, nacelle and tower are monitored by using spectral analysis methods |
3 | WP4086 | Mita-Teknik (Denmark) | Integrated with WT SCADA, the system provides real-time frequency and time domain analyses of turbine operational signals and gives off alarms based on predefined thresholds | With the aid of eight accelerometers, the WT main bearing, gearbox and generator are monitored by using both time and frequency domain analysis techniques |
4 | Brüel & Kjær Vibro | Brüel & Kjær (Denmark) | Collect and process data at fixed intervals and remotely send results to diagnostic server. The time-waveform of the data at any time is accessible for further analysis | The WT main bearing, gearbox, generator, and nacelle (temperature and noise) are monitored by the approach of vibration analysis in combination with temperature and acoustic analyses |
5 | CBM | GE Bently Nevada (USA) | The system gives monitoring and diagnosis of drive-train parameters. Correlate CM signals with WT operational information (eg wind and shaft rotating speeds), and give off alarms via SCADA | The vibrations of WT main bearing, gearbox, generator and nacelle as well as bearing and oil temperatures are monitored |
Table Continued |
No | Name | Product information | ||
Product | Company | Major functions | Notes | |
6 | CMS | Nordex (Germany) | Actual vibration values during WT start-up period are compared with the reference values. Some Nordex turbines also use the Moog Insensys fibre optic measurement system | The system focuses on the monitoring of main bearing, gearbox and generator. The WT blades are also monitored if the WTs also install Insensys' fibreoptic system |
7 | SMP-8C | Gamesa Eolica (Spain) | Continuous online analysis of WT main shaft, gearbox and generator and compare their spectral trends. Warnings and alarms are given through wind farm management system | WT main shaft, gearbox and generator are online-monitored through the spectral analyses of their vibrations |
8 | PCM200 | Pall Europe Ltd (UK) | This is a real-time system for testing and assessing fluid cleanliness | The cleanliness of gearbox lubrication oil is monitored |
9 | TechAlert 10/20 | MACOM (UK) | TechAlert 10 is an inductive sensor to count and size the ferrous and non-ferrous debris, while TechAlert 20 is a sensor only for counting ferrous particles | Both systems are designed for monitoring the debris contained in lubrication or other circulating oils |
10 | Rotor Monitoring System (RMS) | Moog Insensys Ltd (UK) | RMS is in fact a blade-monitoring system, conducting the condition monitoring of wind turbine blades and rotor by measuring the strains in blade–root sections using fibreoptic technology | The load measurement by RMS is also helpful for the load control of pitch regulated wind turbines |
11 | MDSWind MDSWind-T | VULKΛN SEΛCOM (Germany) | MDSWind measures the vibrations of main bearing, gearbox, generator, and tower of the wind turbine, calculates and displays their statistic indices (eg RMS and Crest factor) online | MDSWind-T is a four-channel portable system developed based on MDSWind |
Table Continued |
No | Name | Product information | ||
Product | Company | Major functions | Notes | |
12 | Ascent | Commtest (New Zealand) | Ascent is a vibration analysis system for monitoring the main shaft, gearbox and generator of the turbine by the approach of spectral analyses and time domain statistics | System available in three complexity levels. Level 3 includes frequency band alarms, machine template creation, and statistical alarming |
13 | Condition Diagnostics System | Winergy (Germany) | The system analyses vibrations, load and oil to give diagnosis, predict and recommend for corrective action. Automatic fault identification is provided. Pitch, yaw and inverter monitoring can also be integrated into the system | It mainly focuses on the health monitoring of wind turbine main shaft, gearbox and generator through vibration analysis and oil debris counter |
14 | Condition Management System | Moventas (Finland) | This is a compact system initially designed for monitoring wind turbine gearbox by measuring temperature, vibration, load, pressure, speed, oil aging and oil particles | The system can be extended to monitor the generator and rotor as well as the controller of the wind turbine |
15 | OneProd Wind | Areva (France) | This system monitors the main bearing, gearbox and generator of the wind turbine by measuring oil debris, structure and shaft displacement, and electrical signals | It consists of operating condition channels to trigger data acquisitions, measurement channels for surveillance and diagnosis, optional additional channels for extended monitoring |
Table Continued |
No | Name | Product information | ||
Product | Company | Major functions | Notes | |
16 | WinTControl | Flender Services GmbH (Germany) | This is a vibration-monitoring system for assessing the health condition of wind turbine main bearing, gearbox and generator. Both time and frequency domain analyses are adopted | Vibration measurements are taken when load and speed trigger are realized |
17 | WiPro | FAG Industrial Services GmbH (Germany) | Temperature and vibration measurements are taken for monitoring the main bearing, main shaft, gearbox and generator of the wind turbine | Time frequency analysis used in the system allows speed-dependent frequency band tracking and speed-variable alarm level |
18 | HYDACLab | HYDAC Filtertechnik GmbH (Germany) | This is a system for monitoring the particles (including air bubbles) in hydraulic and lubrication systems. | It is used mainly for monitoring the gearbox of the wind turbine |
19 | BLADEcontrol | IGUS ITS GmbH (Germany) | BLADEcontrol is a system specifically designed for monitoring wind turbine blades by comparing spectra with historic spectra obtained from normal blades | Accelerometers are bonded directly to the blades and a hub measurement unit transfers data wirelessly to the nacelle |
20 | FS2500 | FiberSensing (Portugal) | This is also a fibreoptic system designed for monitoring wind turbine blades with the aid of Fibre Bragg grating sensors | This system can be potentially applied to wind turbine blade monitoring, but at the moment it has not been extensively deployed |
21 | Oil Condition Monitoring System | Rexroth Bosch Group (Germany) | This is a system for the early detection of gearbox damage and the monitoring of oil cleanliness. High-dissolving sensors for the measurement of particles and water content in the lubricating oil are available. Both permit an estimate of the remaining life time of the lubricating oil | This system not only improves the reliability of wind turbines but also their efficient operation by predictable maintenance |
Table Continued |
No | Name | Product information | ||
Product | Company | Major functions | Notes | |
22 | Gearbox Oil Condition Monitoring | Intertek (UK) | Intertek oil condition monitoring services include testing of gearbox oils and lubricants, helping clients extend runtimes for expensive turbines, windmills and other equipment, while minimizing downtime and costly repairs | This is an offline oil analysis system |
23 | Icount system and IcountPD Particle Detector | Parker (Finland) | Parker's system is an all-in-one system to determine whether or not system oil is contaminated and the best way to detect particles online or offline | IcountPD is a particle detector; while the Icount system provides early warning of any unwanted changes in hydraulic or lubrication oil quality. Thus increasing the availability of the machinery by reducing the need for unnecessary downtime |
Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673–693.
Table 18.3
Technologies being researched for WT CM
No | Technique | Advantages | Disadvantages | Online CM | Fault diagnosis |
1 | High-order spectrum | Able to detect the nonlinear relationships between different orders of the harmonics contained in the signal | It is still a tool for processing linear signals, not ideal for analysing WT CM signals | N | N |
2 | Continuous wavelet transform | Able to analyse non-stationary signals satisfactorily | It involves intensive calculations and is still a tool for processing linear signals. However, WT CM signals are often nonlinear | N | Y |
3 | Discrete wavelet transform | Able to analyse non-stationary signals efficiently | Unable to analyse nonlinear WT CM signals correctly, and unable to locate a desired frequency range flexibly | N | Y |
4 | Empirical mode decomposition | An ideal tool for processing non-stationary and nonlinear signals, like WT CM signals | Unable to locate a desired frequency range flexibly | N | Y |
5 | Energy tracking | An efficient tool for analysing WT CM signals | It inherits the disadvantages of wavelet transforms and the accuracy of its results is highly dependent on the correctness of WT speed | Y | Y |
6 | Wigner–Ville distribution | Able to analyse non-stationary signals satisfactorily | Unable to analyse nonlinear WT CM signals correctly | N | Y |
7 | Neural network | An ideal tool for developing real-time CMS. It takes all CM information into account, however it is able to process them efficiently | Difficult to train the neural network | Y | Y |
8 | Data-driven technique | Attributed to ‘natural’ decomposition of original signal and the use of phase information, it is ideal to detect incipient mechanical and electrical defects occurring in WTs | Complex computation and manual selection of interested intrinsic mode functions make it difficult to use online | N | Y |
9 | Genetic programming | Able to simulate complex problems mathematically | The physical mean of the obtained mathematical model is unknown | Y | Y |
Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673–693.