Estimating wind speed and direction with optimal sparse sensor placement on a cylinder

For a wide variety of applications, specifically those where automation is an integral part of the operation, there is a need to determine the magnitude and direction of an oncoming flow. Recently, there has been a growing body of work that has looked at flow field estimation, under the implicit assumption that one already knows the uniform flow field direction. In this work, we consider the simplified scenario of estimating the wind speed and direction of uniform flow over a 2D cylinder. Specifically, we are interested in determining the minimum number and location of sensors on the obstacle with the aim of developing a framework for other geometries, such that they can be used for subsequent flow field estimation methodologies. We consider Reynolds numbers in the range of 25,000 < Re < 125,000, where Re = UD/ν, with U being the oncoming flow velocity, D the cylinder diameter, and ν the kinematic viscosity, and use surface pressure readings to determine the wind speed and direction. Our results are compared to other established sparse sensing techniques, such as POD with QR-pivoting, with the results suggesting that the same level of accuracy can be obtained using fewer sensors for the proposed technique.