Static optimal control: Real-time optimization within closed-loop aim point control for solar power towers. (1st May 2023)

• Record Type: Journal Article
• Authors: Oberkirsch, Laurin; Zanger, David Amos Vincent; Maldonado Quinto, Daniel; Schwarzbözl, Peter; Hoffschmidt, Bernhard
• Is Part Of: Solar energy. Volume 255(2023)
• Page Start: 327
• Page End: 338
• Publication Date: 2023-05-01
• Subjects: Concentrating solar power -- Solar power tower -- Heliostat aiming -- Aim point optimization -- Optimal control -- Cloud disturbance; Solar energy -- Periodicals; Solar engines -- Periodicals; 621.47
• DOI: 10.1016/j.solener.2023.03.051
• Languages: English
• ISSNs: 0038-092X

Abstract

Abstract: Many aim point optimization techniques exist to control Solar Power Towers (SPTs). However, SPTs exhibit optical losses that cannot be exactly modeled. Moreover, cloud passages cause transient incident flux distributions. Due to these modeling errors and disturbances, aim point optimization may exceed the Allowable Flux Density (AFD); consequently, these efficient aiming strategies are seldom applied at commercial plants. In this paper, an innovative closed-loop aim point control technique, the Static Optimal Control, is proposed. Flux density measurements close the open control loop of aim point optimization. Based on this feedback, the Static Optimal Control estimates weights that are embedded in the cost function of the aim point optimization. This GPU-based optimizer finds good aim point configurations in a few seconds even for large plants. Thus, the Static Optimal Control compensates for modeling errors and rejects disturbances to observe the AFD while maximizing the intercept. The performance of the Static Optimal Controller is evaluated for inaccurately modeled mirror errors and under a real cloud scenario. Aim of this control is not to exceed the AFD by more than 5% i.e. the accuracy of the flux density measurements. The aim is achieved for static modeling errors while improving the intercept by 1.7–8.6% compared to a heuristic control. In the cloud scenario, the Static Optimal Control reaches its limits. Even mapping all-sky-imager-based nowcasts in a feed forward manner on the heliostat field does not improve the control quality due to high prediction errors. Graphical abstract: Highlights: Static Optimal Control as closed-loop aim point control technique for solar towers. Suited for large plants due to embedding of GPU-based aim point optimization. Closing the open control loop of aim point optimization using flux density feedback. Compensating for modeling errors while improving the intercept by 1.7–8.6%. Feed forward scaling of the heliostats reflected power by real nowcasting data.