Wind Field Reconstruction Using Inverse Process With Optimal Sensor Placement

Physical and data-based approach for wind forecast has been under rapid development recently. Yet challenges remain when high spatial resolutions of the velocity distributions are desired for a wind farm. One of the major causes is the huge computation load that cannot be accomplished in time for a large area. In this study, a method for rapid wind field reconstruction is developed that can provide refined spatial resolution. The method first generates a large number of numerical wind fields using computational fluid dynamics, and extracts the basis vectors of dramatically reduced dimensions by means of proper orthogonal decomposition. With the basis vectors, a wind field can be rapidly reconstructed in an inverse process based on limited measurement data by the sensors. A procedure is proposed to optimize the solution process, which effectively leads to optimal sensor placement. Both simulations and experiments are carried out to evaluate the methods. The results are compared and good agreements are observed.