A comparison of price-taker and production cost models for determining system value, revenue, and scheduling of concentrating solar power plants. (1st December 2018)

• Record Type: Journal Article
• Authors: Martinek, Janna; Jorgenson, Jennie; Mehos, Mark; Denholm, Paul
• Is Part Of: Applied energy. Volume 231(2018)
• Page Start: 854
• Page End: 865
• Publication Date: 2018-12-01
• Subjects: Concentrating solar power -- Thermal energy storage -- Dispatch optimization -- Grid integration; CSP concentrating solar power -- LMP locational marginal price -- MILP mixed-integer linear program -- MSPT molten-salt power tower -- PCM production cost model -- PG&E Bay Pacific Gas and Electric Bay area -- PT price-taker -- SAM System Advisor Model -- SM solar multiple -- TES thermal energy storage -- VO&M variable operations and maintenance; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2018.09.136
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: Comparison of operation from self-scheduled vs. system-scheduled perspective. System-scheduled optimization requires two orders of magnitude longer solution time. Optimal dispatch profiles exhibit strongly similar characteristics. Main operational discrepancies occur during low-value/low-cost periods. Operational value decreases only 1–5% when operated per self-scheduled solution. Abstract: Concentrating solar power (CSP) plants convert solar energy into electricity and can be coupled with low-cost thermal energy storage to provide a dispatchable renewable resource. Maximizing the economic benefits of CSP systems requires deliberate selection of the timing of thermal energy dispatch to coincide with high-price or high-value periods. The performance and value of CSP systems have traditionally been evaluated using two distinct approaches: (1) production cost models (PCMs), which optimize commitment and dispatch schedules for an entire fleet of generators to minimize the cost of satisfying electricity demand; and (2) price-taker (PT) models, which optimize dispatch of the CSP system against historical or forecasted electricity prices to maximize revenue available to the CSP operator. The PCM approach is, in principle, superior because of the ability to capture system-level details such as transmission pathways and associated constraints, and feedback between CSP dispatch schedules and operation of other generators. The computationally simpler PT technique cannot explicitly address these concerns, but the comparative reduction in solution time is attractive for sensitivity analysis and rigorous optimization of CSP plant design and operation. In this work, we directly compare the dispatch profiles, net revenue, and operational value of CSP systems using implementations of each technique to assess how closely the solutions from the PT technique can approach those from the PCM for a specific market scenario. The PT implementation produced characteristically similar dispatch profiles as the PCM for both a large solar-multiple configuration and a small solar-multiple peaking configuration, while requiring only 1% of the computational time of the PCM. In addition, dispatching the CSP plant according to the self-scheduled PT solution—as opposed to the system-optimized PCM solution—only reduced the total CSP operational value by 1–5%, with consistent trends across CSP configurations and locations. This analysis validates the ability of the PT approach to substantially capture the economic benefits of CSP systems when pricing profiles are available, and indicates that the PT approach is suitable for initial design and optimization of CSP technologies.