"Meaningful yield optimising effort requires in-depth knowledge of the approaching wind field and associated loads acting on the turbine," says Fabio Bertolotti, SSB's research and development director. Modern turbines typically monitor only wind speed and wind direction values taken at a single point, roughly at the rotor centre, but turbine performance is dependent on the distribution of wind speed and direction over the entire rotor-swept area, he explains.
Measuring instruments are usually mounted at the nacelle behind the rotor plane. As a result of this, wind flow is disturbed and slowed down at the point where it is measured and a transfer function is required to compensate for these wind-speed and wind-direction errors, Bertolotti adds.
Transfer-function values are often derived from one stationary meteorological mast somewhere on the wind farm, which by definition lacks sufficient accuracy. Alternative nacelleor rotor-mounted two and three-beam Lidar (light detection and ranging) devices are functionally still single-point devices, he points out.
To describe rotor wind flow adequately, wind data must be sampled at many different points across the swept area at one-second intervals to provide a near-instantaneous data stream, according to Bertolotti. Comprehensive sampling is essential because all wind variables vary with height and width across the rotor plane.
At least seven key variables must be continuously monitored. These are wind speed; wind direction; vertical and horizontal wind shear (wind speed differences in vertical and horizontal plane); veer (change of wind direction with height); vertical velocity; and turbulence level.
BladeVision components are mounted inside the blade structure, where they are easily serviceable and protected against harsh environmental conditions, such as saline air, rain and lightning. The components consist of an industrial in-house developed high-speed camera mounted inside the blade root, and multiple pairs of vibration-proof reflectors fitted at regular intervals along the blade's inner reinforcement structure.
"Blades are the turbine components most impacted by wind-induced forces. That makes them an excellent sensor to measure the impact of key wind variables," says Bertolotti. "Even a minor deviation in one or more variable translates instantly in a blade deflection change recorded by the cameras, which take a fresh picture every 50-100 milliseconds."
A computerised in-house data processing and analysis tool inside the nacelle offers wireless communication with individual cameras. Its main function is to convert deflections into blade loads and wind-related data by inverting the physical cause-effect relationship between wind and blades. The processing unit output produces a precise, continuously refreshed representation of the wind field in front of a turbine. BladeVision, he says, offers owners and operators an improved measure of turbine performance represented by sharply outlined power curves in all wind fields, including complex terrain, and wake shadowing from upstream turbines.
BladeVision is claimed to minimise yield losses by near-eliminating yaw and pitch errors and pitch-angle asymmetries, and by (indirectly) indicating blade surface damage, icing build-up during blade rotation, and faults in turbine control parameters.
Structural blade damage such as delamination is indicated by deviations in system output readings when the damage is sufficiently large to affect blade stiffness.
Bertolotti agrees that swept-beam Lidar devices, mounted inside the rotor hub or spinner, are also capable of effectively monitoring the seven wind variables, but points to the high cost of these devices, which prevents them from being fitted as standard to individual turbines. BladeVision, he claims, offers a pay-back period of less than two years. "This figure is based solely on cost-of-energy benefits linked to the near-elimination of pitch and yaw system and blade-symmetry faults, without simultaneously considering extended turbine operating life, and reduced demand for asset upkeep."
Commercial product launch is expected for later this year, following a six-month field test.