Evaluation and Optimization of Heliostat Calibration Methods

Abstract

In solar tower power plants, the sun's energy is focused onto a receiver at the top of a solar tower by means of many biaxially movable mirrors (heliostats) and the high energy density generated is used, for example, to generate electricity via a steam power process or to trigger chemical processes for the production of new types of fuels. It is very important that the heliostats track the movement of the sun correctly throughout the day, so that the incoming solar rays are reflected as accurately as possible onto a defined point on the tower. To achieve this, motor positions of the two linear actuators, which change the orientation of the heliostats, are calculated from an error-prone kinematic geometry model of the individual heliostats. To minimize alignment errors, selected parameters of the geometry model are adjusted by means of a regression to calibration data recorded by the camera-target method. Within the scope of this work, on the one hand, the regression method used is investigated and improvements and further algorithms are tested and developed. On the other hand, the selection of optimization parameters from the geometry model is analyzed. It was found that changing the regression methodology does not improve the alignment accuracy. However, a selection of optimization parameters and their optimization order could be found that indicate a significant reduction of the alignment error.