Simulation and Optimization of Heliostat Fields using Radiance and MOO fot the Design of a 300 MWth Central Receiver

This work discusses solar power towers and especially the simulation and optimization of heliostat fields. The technological stateof the art is given through typical characteristic values and dimensions of some existing central tower power plants such as th so-called Solar Two in California and the PS10 in Spain. Leaving cycle losses aside, the field losses involved are described and roughly estimated thanks to commonly measured and computed values, typically coming to an 80% overall field efficiency. On the other side, the coats that must be taken into consideration when designing a heliostat field are listed and approached through correlations, giving for instance a total investment cost of $15 million for relatively large fields ( approximately 700 [m]). The raytracing software Radiance is then chosen for modeling an entire heliostat field with a central receiver, and thus for running simulations over time for different sun positions. Some given field layouts such as those obtained in published cases can be visualized and heat flux at the receiver‘s surface is computed thanks to Radiance tools before being displayed through a MATLAB post-processing. For validating Radiance calculations, the results are first compared with published cases: the receiver‘s total thermal energy produced over a year is generally between 5 and 10 [%] lower than these published results. Secondly, the Radiance heat flux profile has to be compared with some published experimental results: based on measurements led at the PSA Solar de Almeria). the heat flux profile (gradient) is close to measurements, but computed and measured total thermal powers do not match when computed and measured peak values do match. Therefore the Radiance heat flux still requires further validation probably involving some model adjustments. Nevertheless, this does not prevent from implementing some very large fields (more than 3000 heliostats) optimized with the MOO multi-objective optimizer after having modified routines for generating surrounding fields. and thus resulting in a first overview of what the heat flux profile on a cylindrical receiver may look like for a total thermal power of 300 [M Wth].