Author David Clark
David Clark / Condition Monitoring expert / www.bachmann.info
A speaker at a recent AWEA conference remarked that the wind industry was getting a handle on the gearbox problem and its next big challenge would be blades. The speaker is right. What the industry needs is a condition monitoring system capable of detecting and predicting failures and conditions related to the blades on megawatt-class wind turbines. Several technologies have been tried, or adapted, from other markets with limited ability to cost effectively detect all major failure modes. One system judged most appropriate is currently not commercially available to the market and will be described later.
A blade condition-monitoring system will likely be a requirement as wind turbines and blades increase in size, insurance requirements dictate reliability, and offshore wind turbines increase in number.
How blades fail
Six major failure modes and other conditions must be monitored by a blade condition monitoring system, or BCMS. Although several approaches have been tested in the recent decade, no commercially available system is capable of detecting major failure modes.
Many attempts to adapt other technologies to this application have been met with limited success. At least three issues have hobbled current blade-sensing technology.
The first issue is whether or not the technology or product can detect the likely encountered failure modes. The second issue: Can the system retrofit to existing towers. The third issue: Is the system cost effective?
Failure modes
An effective BCMS must detect several conditions.
Crack detection will be critical for the system. Cracks typically occur at four major locations: the root, leading edge, trailing edge, and tip. While there is some variance in the locations due to the variety of blade manufacturers, generally speaking these four locations are consistent.
Delamination predominately occurs at trailing edges and is caused by separating layers of composites and laminations.
The accumulation of ice on blades should be detectable. An ice accumulation is obviously not conducive to safety or operation. The extra loading creates uneven stress on blades, which can be measured under specific frequencies.
Imbalance (either aerodynamic or static)
While blades are balanced from the factory within a tight tolerance, several factors including leading edge erosion, contribute to a difference in static balance. A more frequent aerodynamic imbalance is caused by variations in pitching.
Loading issues come from wind shear, pitch deviation, tip in/out, and yaw deviation. Ideally, all these conditions should be monitored with a BCMS. This would not contribute to condition monitoring as much as performance monitoring for improved reliability and performance production output.
When not monitored correctly, these operational conditions may appear as imbalance. A blade CMS should be able to discern between these differences.
A blade CMS system should also detect lightning strikes, a common condition which contributes to a failure mode.
A few attempts
Historically, the different technologies applied with limited success include:
Fiber optic devices, which have been difficult to install, are retrofitted, expensive, and do not detect all failure modes. This is likely why they have had limited adoption and success by the wind industry. Installing a fiber optic-based sensor usually involves cutting a shallow slot into the perimeter of the blade into which the fiber optic strand is laid and epoxied in place.
Strain gauges are easy to install, but when retrofitted to existing turbines, they do not detect all failure modes. Although costs are relatively low, the sensors are not robust in the field. Some tests and trials have shown sensor life as short as 6 to 9 months.
Acoustics, used by one wind-turbine manufacturer to detect cracks, has worked on a small number of units. It mounts a microphone atop the nacelle and focuses it towards the hub to detect high frequency acoustics common in surface cracks. While easy to install and relatively cost effective, acoustic detection and fiber optics never caught on. The technology has a limited detection ability because it depends on a crack as the failure mode, which is not always the case.
Vibration sensors have been used mounted near the hub, but not on the blades. There is good measurement ability on some failure modes such as icing, imbalance, but vibration sensors have a less than optimal operational condition. As with the others here, vibration sensing has limited detection of the common failure modes. However, ease of installation or retrofit, and cost are good. As a secondary benefit, these sensors are usually mounted on main bearings which are also monitored. Although main bearings have the least frequent failures, the capability is an added benefit.
Laser reference uses a laser and prism system to compare differences and changes between known reference points inside a blade. This is done by aiming the laser at the prism and then redirecting the laser to internal locations. This would be an excellent system for quality control of blades – measuring manufacturing deviations in substrates and composites. Once again, the ability to detect all common failure modes, ease of retrofit, and cost, all factor in to a lack of widespread acceptance. Secondly, the blades expand and contract during operation, so temperature changes make valid reference measurements extremely difficult.
Summary for an ideal system
So what would work? An ideal blade sensing systems should:
- Detect all six major or common failure modes
- Use robust sensors
- Identify blade position
- Identify the blade
- Identify the sensor
- Provide a cost-effective fit into blades in the field and those on the manufacturing floor
- Install easily
- Use technology that is wireless, self-powered, and self-contained
At this writing, a recent patent shows promise for combining all the features mentioned here. It is under development at California-based Bajou Engineering. Company president Dr. Shaw Makerami says the system should be commercially available in about 18 months. This should be good news to those developing offshore turbines or those facing blade issues. WPE
Crack detection will be critical for the system. Cracks typically occur at four major locations: the root, leading edge, trailing edge, and tip. While there is some variance in the locations due to the variety of blade manufacturers, generally speaking these four locations are consistent.
Delamination predominately occurs at trailing edges and is caused by separating layers of composites and laminations.
The accumulation of ice on blades should be detectable. An ice accumulation is obviously not conducive to safety or operation. The extra loading creates uneven stress on blades which can be measured under specific frequencies.
Imbalance (either aerodynamic or static) While blades are balanced from the factory within a tight tolerance, several factors including leading edge erosion, contribute to a difference in static balance. A more frequent aerodynamic imbalance is caused by variations in pitching.
Loading issues come from wind shear, pitch deviation, tip in/out, and yaw deviation. Ideally, all these conditions should be monitored with a BCMS. This would not contribute to condition monitoring as much as performance monitoring for improved reliability and performance production output.
When not monitored correctly, these operational conditions may appear as imbalance. A blade CMS should be able to discern between these differences.
A blade CMS system should also detect lightning strikes, a common condition which contributes to a failure mode.
A few attempts
Historically, the different technologies applied with limited success include:
Fiber optic devices, which have been difficult to install, retrofit, expensive, and do not detect all failure modes. This is likely why it has had limited adoption and success by the wind industry. Installing a fiber optic-based sensor usually involves cutting a shallow slot into the perimeter of the blade into which the fiber optic strand is laid and epoxied in place.
Strain gauges are easy to install, but when retrofit to existing turbines, they do not detect all failure modes. Although costs are relatively low, the sensors are not robust in the field. Some tests and trials have shown sensor life as short as 6 to 9 months.
Acoustics, used by one wind-turbine manufacturer to detect cracks, has worked on a small number of units. It mounts a microphone atop the nacelle and focuses it towards the hub to detect high frequency acoustics common in surface cracks. While easy to install and relatively cost effective, acoustic detection and fiber optics never caught on. The technology has a limited detection ability because it depends on a crack as the failure mode, which is not always the case.
Vibration sensors have been used mounted near the hub, but not on the blades. There is good measurement ability on some failure modes such as icing, imbalance, but have a less than optimal operational condition. As with the others here, vibration sensing has limited detection of the common failure modes. However, ease of installation or retrofit, and cost are good. As a secondary benefit, these sensors are usually mounted on main bearings which are also monitored. Although main bearings have the least frequent failures, the capability is an added benefit.
Laser reference uses a laser and prism system to compare differences and changes between known reference points inside a blade. This is done by aiming the laser at the prism and then redirecting the laser to internal locations. This would be an excellent system for quality control of blades – measuring manufacturing deviations in substrates and composites. Once again, the ability to detect all common failure modes, ease of retrofit, and cost, all factor in to a lack of widespread acceptance. Secondly, the blades expand and contract during operation, so temperature changes make valid reference measurements extremely difficult.
Summary for an ideal system
So what would work? An ideal blade sensing systems should
· Detect all six major or common failure modes
· Use robust sensors
· Identify blade position
· Identify the blade
· Identify the sensor
· Provide a cost-effective fit into blades in the field and those on the manufacturing floor
· Install easily
· Use technology that is wireless, self-powered, and self-contained
At this writing, a recent patent shows promise for combining all the features mentioned here. It is under development at California-based Bajou Engineering. Company president Dr. Shaw Makerami says the system should be commercially available in about 18 months. This should be good news to those developing offshore turbines or those facing blade issues.
Shaw Markaremi 805 452 3014
Filed Under: Blades, News
