An improved complex ICA based method for wind farm harmonic emission levels evaluation

Abstract Wind farms, as typical renewable energy infrastructure, provide clean energy to the power grid but also inevitably generate harmonic pollution. In the case of a multi-wind farms system, evaluating the harmonic emissions for each farm is essential for the division of harmonic responsibility and the design of the harmonic mitigation scheme. However, this evaluation process faces two challenges. First, the background harmonics of a single wind farm might be unstable. In addition, the harmonic impedance of wind farm side is reduced by the installation of filters. As a result, traditional evaluation methods are no longer applicable. Complex independent component analysis is a well-known method for evaluating harmonic emissions. To make this method feasible in the above situations, we analyze its shortcomings. Furthermore, an improved method is proposed by introducing two screening mechanisms based on sparse component analysis and the maximization of hybrid negentropy. Additionally, the results of simulations and field case verification indicate that the traditional methods have large calculation errors when the background harmonics fluctuate greatly or the harmonic impedance of the wind farm side is not far larger than that of the utility side. In contrast, our novel evaluation method has high precision in the above cases.