Advanced image processing for turbulence wedge detection in thermographic flow visualization on wind turbines in operation

Abstract

Abstract Environmental conditions like the presence of rainfall or insects can disturb the rotor blade surface of wind turbines in operation, triggering a premature laminar-turbulent flow transition in the boundary layer flow. The local contaminations develop a wedge-shaped surface area of turbulent flow in the area that would otherwise be laminar if the surface would be undisturbed, decreasing the size of the laminar flow regime. This change in the ratio between overall laminar and turbulent flow regime sizes has a negative impact on the aerodynamic performance of the profile, decreasing the efficiency of the wind turbine. While the spatial distribution of the flow regimes can be visualized with thermographic flow visualization, the state-of-the-art image processing method for applications on wind turbines in operation is not robust against localizing the position of the flow transition along these turbulence wedges. Therefore, this work introduces an advancement of the image processing method for localizing the flow transition in thermographic images with a focus on decreasing the localization uncertainty along the turbulence wedges. The state-of-the-art one-dimensional evaluation method is enhanced by a two-dimensional image processing method in order to increase the directional gradients at the turbulence wedges’ flanks. Six out of six previously undetected turbulence wedges are successfully detected in a flow visualization image of a rotor blade of a GE 1.5 sl wind turbine in operation. The new approach yields an improved application of the thermographic flow visualization for locating the flow transition and quantifying the reduction of the laminar flow area on disturbed rotor blade surfaces of wind turbines in operation.