Representative virtual heliostat: A reduced-order approach to accelerate flux estimation for a central receiver tower system

Flux estimation over the receiver surface is critical in design and analysis of a concentrating solar power (CSP) system based on heliostat field and central receiver tower. Monte Carlo ray tracing (MCRT) provides high accuracy in estimating flux distribution. However, it is computationally too expensive to be employed for year-round estimation for various sun-positions, or for different receiver geometries, for a large heliostat field of practical use. To address this limitation, a novel method is developed in the current study where a reduced-order framework based on Representative Virtual Heliostats (RVH) is introduced. In the present approach, particle swarm optimization (PSO) is employed to identify one or more RVHs which reproduce the optical signature of the full heliostat field on the receiver-vertical plane via utilization of only a fraction of the computational resource. Optimization minimises the deviation between the flux distribution generated by the RVHs and that obtained from the full-field MCRT. For a heliostat field of small width, a single RVH suffices. On the other hand, for large heliostat fields, multiple RVHs are introduced to preserve directional variation in the incident flux distribution. A scaling strategy is implemented to maintain parity in the input energy between the full-field MCRT and the RVH ray tracing. The proposed methodology is tested against the full-field MCRT across multiple sun positions spread through a year. It is demonstrated that the RVH-based novel method can effectively expedite feasibility estimation for the flux distribution on the receiver vertical plane with minimal loss of accuracy.