Dual-UAV nighttime detection of misaligned heliostats and cooperative inspection scheduling for dual-tower CSP plants

Heliostat-field optical performance directly affects solar-to-electric conversion and sustainable operation in tower concentrating solar power (CSP) plants. Conventional inspection methods rely on natural sunlight, are conducted during daytime, and may interrupt power generation. This study proposes a nighttime inspection framework using two unmanned aerial vehicles (UAVs) and artificial light sources to detect misaligned heliostats without occupying generating hours. Real-time kinematic (RTK) positioning, solar-trajectory simulation, and coordinate transformation are integrated to reproduce solar incidence conditions with a UAV-mounted light source, while an imaging UAV hovers near the receiver aperture to capture reflected images. A U-Net network segments heliostat regions under low-light conditions, and spot-centroid deviation and grayscale-distribution features are extracted for multi-temporal tracking-state assessment. For dual-tower CSP plants, a segmented cooperative UAV scheduling strategy is developed to optimize inspection routes and reduce idle flight time. Case studies demonstrate reliable nighttime identification of misaligned heliostats and improved inspection efficiency under representative nighttime operating conditions in the experimental and simulation analyses. By integrating solar-direction simulation, receiver-consistent imaging, optical diagnosis, and cooperative scheduling, the proposed framework provides a practical method for intelligent heliostat-field inspection, helping maintain optical performance, improve solar energy utilization, and support efficient and sustainable operation and maintenance of large-scale CSP plants.