Optimization of stored energy dispatch for concentrating solar power systems

Abstract

Concentrating solar power (CSP) is an emerging technology capable of generating renewable, dispatchable power using cost-effective thermal energy storage. Dispatchability imparts significant value to a power generation technology, both expanding the applications in which it is useful and -- perhaps more critically -- enabling the proliferation of other non-dispatchable renewable technologies by supplementing their performance during transient periods of low or variable production. Current CSP technology is functional but sub-optimal, especially with regard to solar field design and to strategies for optimally dispatching power. A comprehensive, immediate, and systematic optimization approach is needed to drive down technology costs, thus improving the competitive position of CSP in the marketplace. Several tools are a necessary part of this process, including those for solar field design and optical characterization, plant productivity simulation, optimal operations scheduling (i.e., dispatch optimization), cost assessment, and financial performance evaluation. However, advancement in each critical area has been uneven to-date, with solar field design and optical characterization and dispatch optimization less established than other areas. Herein lies the challenge undertaken in this thesis -- namely, to identify mathematical models and computational techniques that advance these two aspects of CSP technology development. We present methods for improved solar field design, characterization, and optimization that advance previous work by extending an existing technique for analytical flux modeling to individual heliostat reflections, enabling rapid computational prototyping of a broader range of field designs. Once defined, a plant's dispatch schedule can be optimized to maximize revenue from electricity sales, minimize costs due to subsystem start-up or change in production, and enforce contractual or technological constraints. We undertake this task by formulating a mixed-integer linear program that more realistically and accurately accounts for a variety of operational modes and subsystem performance characteristics. The advances in field design and characterization and in dispatch optimization are adopted by an industry-leading project developer who demonstrates through simulation the viability of CSP molten salt tower technology as a dispatchable resource, where previous work had left in question the impact of optimized dispatch.