Airborne real-time solar concentrator orientation estimation for heliostat coarse calibration

Abstract

We present a novel method for the airborne real-time solar concentrator orientation estimation for heliostat coarse calibration, termed ARTSCORE-C. The input to the method is the set of image coordinates of identified concentrator corner points within a single image. Additionally, it requires the known heliostat positions, the kinematic model and the concentrator geometry, quantities that are typically well known in heliostat fields. The method is well-suited for real-time applications and lays the foundation for autonomous condition monitoring systems for heliostat fields. It was developed with the aid of a dedicated simulation environment, which enabled a thorough analysis of noise propagation throughout the algorithm. Insights from simulation were successfully transferred to real-world data, demonstrating that simulation tools can effectively support the development of airborne image-based systems. Validation experiments at the solar tower test facility in Jülich, Germany confirmed the method’s practical utility: Using 20 observations per concentrator, the method achieved an average angular RMSE of 3.5 mrad across 146 validated concentrators with a range of typical orientations. The achieved accuracy qualifies the method as a coarse calibration system, suitable to be used as an upstream-component for fine calibration systems. Moreover, its ability to optically estimate the camera pose for a given image makes it a promising alternative to hardware-based localization systems, particularly in remote environments, where correction signals may be unavailable. This capability enhances its practical relevance for large-scale, low-maintenance solar tower power facilities.