Wind Turbine Active Control: Sensor Input System

Wind gusts, turbulence and wake interactions create a complex velocity field over wind turbine rotor blades, particularly in wind farms with large turbines. Even if complete details of this complex, unsteady, three-dimensional velocity field were available, the needed local force distribution over the blade surface (“wall”) could not be accurately determined, as our knowledge of turbulence lacks an accurate model relating wall shear stress, wall pressure and the turbulent velocity field very close to the wall. Sampling the two components of local wall shear stress and the local pressure would directly provide samples of wind-induced blade surface-force fluctuationdata in a few selected locations on blades. Active control based on direct measurement of the three components of measured wind-inducedblade-surface forces should help combat stall, reduce fatigue and extreme load failures, reduce power fluctuations, and could help extend the wind speed envelope for power extraction. Suchdata may be reasonably proposed as the best input for active control of wind blade force fluctuations, using devices such as flaps, microactuators, suction or blowing –however, wall shear stress fluctuations have proved notoriously difficultto measure accurately. We present development of a thermal microsensor for measuring fluctuations of turbulent wall shear stress. Flush-mounted, single-hot-film microsensors are non-intrusive and can be made in a variety of sizes using microfabrication techniques. However, such probes suffer from large errors caused by significant heat conduction through the solid substrate. Multi-element guard-heated microsensorshave been designed to remove these errors. Prototypes of guard heated sensors have been microfabricated. Numerical and analytical studies of response to shear stress, showthat errors from substrate heat conduction can be reduced to near-zero, with proper geometric design. We show that a strong nonlinearity in the standard single-element sensor can be corrected by the guard-heated design, indicating the new sensors may also contribute more accurate data towards a better fundamental understanding of near-wall turbulence.