Compensation of heliostat drift by seasonal sampling

Abstract Heliostat image drift is a common phenomenon in central receiver solar power plants. Several geometrical errors produce drift of the heliostat solar spot at receiver surface, increasing radiation spillage. A heuristic drift compensation method is proposed, based on a polynomial approximation to the drift trajectories. Results of the practical implementation of the proposed method for the control of 10 heliostats in a solar tower facility are presented. A substantial improvement of heliostat tracking is observed on the experimental tests. Because heliostat drift experimental monitoring is a time consuming task, a numerical analysis of the yearly behavior of the compensation method, based on simulations of heliostat drift, was carried out. In these simulations, the behavior of the daily RMS deviation of the concentrated solar spot centroid is evaluated for a whole year, as the polynomial correction is applied. The simulations serve also to test the effectiveness of the proposal polynomial method in a wider range of conditions. Thus, heliostats with a variety of primary error values are simulated. Random wind induced vibrations are introduced in the simulation to evaluate the effectiveness of the calibration method under noise conditions. It is found that a very effective calibration can be achieved with a few sampling events of the heliostat behavior during the year, taking only a few minutes. The RMS deviation can be reduced to values of the order of the wind induced noise level. The proposed polynomial compensation looks like a promising alternative to be implemented in heliostat fields.