Identification of shape errors in a parabolic solar collector : An improvement to the analysing algorithms examining the solar collector from optical measurements

Sammanfattning: Increasing global warming with droughts, forest fires and melting polar ice has forced the world to speed up the transition from fossil fuel to fossil free energy. A big part of it is the use of solar energy. To obtain solar thermal energy, different sorts of solar collectors have been developed. One of these are the parabolic solar collector, a trough which concentrates the sun rays onto a receiver tube, which in turn absorbs the thermal heat. Absolicon Solar Collector AB is a Swedish company developing a parabolic solar collector named T160. These collectors are optically verified in the end of the Absolicon production line to decide if they meet the expected criteria. The verification of the parabolic shape is of utmost importance for the performance of the trough while a shape error can cause the beam to hit the receiver tube in a sub-optimal angle or miss it completely. If this were to happen, all of the energy can not be extracted. Earlier research have developed different methods for finding slope errors, deviations in the normal angles, in the trough but does not investigate the connection between slope errors and the trough shape errors that might have caused the deviations. This report aims to develop an algorithm based on slope errors in a parabolic trough collector which identifies four predetermined common shape errors in the trough. Identifying shape errors help to quickly identify and correct systematic deviations. In addition, this work aims to implement a new acceptance criteria based on slope errors for the solar collectors to make sure they hold up to their standard. The algorithm should also be compatible with a new camera system being implemented in the optical verification at Absolicon. This is done by deriving mathematical expressions for the normal angles in the trough with respect to the shape errors. By using the Pinhole Camera Model, the Law of Reflection and geometric properties of the solar collector, it is possible to convert pixel coordinates of receiver tube edges in images to normal angles. The resulting deviation in normal angle compared to the ideal ones are analysed and fitted to the mathematically derived expression for the normal angles by a build in minimization method in the tool lmfit in Python which uses non-linear least squares to detect type shape errors. The acceptance criteria and compatibility with the new system is implemented and taken into account. The results show that the calculation of slope errors from the data is valid with an uncertainty of 0.82 mrad and expected differences in the acceptance criteria quality value is seen when dealing with solar collectors with different type shape errors. The type shape error algorithm finds the correct shape errors for noisy self-created data which shows that the method works. The results when testing on real collectors with forced shape errors show potential but is in need of further adjustments and more clean precise data to produce certain accurate results. The algorithm is a good start to create a tool for finding typical shape errors in parabolic solar collectors.