Session: 05-13 Heliostat Consortium 3

Paper Number: 116619

116619 - High-Speed Assessment of Heliostat Fields Without Disrupting Operations 

Heliostat fields are concentrating solar power (CSP) collectors comprised of many individual mirrors that can be independently steered to reflect and concentrate light onto a target receiver.  Heliostat fields are an attractive technology for industrial decarbonization, because they can achieve both high temperatures (>1000 °C) and high power (>100 MW_th).  High temperature is accomplished by a high concentration ratio, and high power is accomplished by employing a large number of heliostats.  Example industrial heliostat fields are vast, with >10,000 heliostats, total mirror area >1,000,000 m², and land area >2.8 km in diameter [1].

The high concentration ratio required for high temperature requires accurate heliostat optics.  Optical error targets can be as low as 1.0 mrad (0.06°), and heliostat optical error exceeding 1.5 mrad (0.17°) can lead to a 47% increase in the levelized cost of heat (LCOH) [2].  Thus highly accurate metrology systems are needed to assure that heliostats meet these tolerances, both at installation and over the heliostats’ 30-year life cycle outdoors.  A widely-used system for heliostat calibration is the beam characterization system (BCS), where heliostats individually direct reflected sunlight onto a target, typically on the tower, and the accuracy of the resulting spot position is assessed [3][4].  However, this approach is slow and does not provide detailed information on heliostat shape errors.  Further, BCS performance is reduced for heliostats far from the tower, both due to large spot size and reduced spot intensity.  But these distant heliostats are the most critical to measure, due to their very long optical path length to the tower (up to 1.6 km).  For these distant heliostats other tower-based metrology approaches also are susceptible to distortion [5], optical aberrations due to thermally-induced atmospheric turbulence, and occlusion for targets lower than the receiver.  Making things even more challenging, heliostat field metrology systems must be very productive – calibrating a large heliostat field may require >100,000 full-heliostat measurements.  To accomplish this in, for example, a month, would require a measurement rate faster than 7 heliostats/minute, or an average measurement time less than 9 seconds per heliostat – all over a land area up to 12 km².

This presentation will report progress toward the development of a system designed to measure optical errors and mirror status at high speed across large heliostat fields.  This approach uses an uncrewed aerial vehicle (UAV) programmed to fly constant-heading, constant-speed passes over the heliostat field, capturing a video of heliostats passing under the UAV.  The resulting flight paths can scan heliostats at a rate of approximately 4 seconds per heliostat, although overall average rates are lower due to pass stepover time and turnaround time between flights.   Flight paths are computed to cause viewed heliostats to have reflections of adjacent heliostats sweep across each viewed heliostat, providing a rich time sequence of images containing reflected optical targets and avoiding long optical path lengths.  The resulting video is then post-processed to identify heliostat boundaries throughout the video sequence, yielding a time-sequence of identified heliostat corners that can then be used to estimate the 3-d shape and position of each heliostat, and also the UAV position relative to the heliostat.  The resulting ensemble of UAV-to-heliostat relative positions is then used to estimate both the UAV trajectory and the orientation of the heliostat within the field.  In preliminary work so far, post-processing calculations estimated intra-heliostat facet position deviations of 2-4 mm, and the measurement technique is sensitive enough to detect heliostat tracking motion between scan passes.   Future work is planned to process reflections seen in the resulting heliostat images to obtain measurements of optical slope for many points on the mirror, thus producing an estimate of heliostat optical shape.

References

[1]  https://solarpaces.nrel.gov/.

[2]  G. Zhu, et al. 2022.  NREL Report NREL/TP-5700-83041, 2022.

[3]  J. Strachan.  Sandia Report SAND92-2789C, 1992.

[4]  M. Ayers, et al.  AIP Conference Proceedings 2303, 030004 (2019).

[5]  R. C. Brost.  SolarPACES 2021.

Presenting Author: Randy Brost Sandia National Laboratories

Presenting Author Biography: Dr. Randy Brost is a technical staff member at Sandia National Laboratories in the Concentrating Solar Power Technology group. He currently leads projects related to concentrating solar optics. He received his Ph.D. in Computer Science from Carnegie-Mellon University in 1991, and performed robotics research at Sandia National Laboratories until 1997. He then served at Eastman Kodak Company until 2007, implementing a variety of custom software tools supporting advanced manufacturing, metrology, and physics analysis. He then joined SkyFuel, a concentrating solar power company, where he helped develop the SkyTrough, a utility-scale parabolic trough collector, and applied computational methods to optimize new solar collector designs. He returned to Sandia in 2011, and pursued a variety of computer science research topics before joining the Concentrating Solar Technology group in early 2020.