A number of commercial production cost models are available to utilities, system operators, and planners to evaluate the operation of the grid. These models are used to help plan system expansion, evaluate aspects of system reliability, and estimate fuel costs, emissions, and other factors related to system operation. The models have the primary objective function of committing and dispatching the generator fleet to minimize the total cost of production while maintaining adequate operating reserves to meet contingency events and regulation requirements. Modem production cost models often include transmission power flow simulations to ensure basic transmission adequacy for the generator dispatch. These models are increasingly used to evaluate the impact of incorporating VG sources, such as wind and solar. Integration studies evaluate the impact of VG on power plant ramping and reserve requirements and explore changes to grid operations needed to incorporate increasing amounts of VG (GE 2010).
As the penetration of VG increases, studies have found an important role for grid flexibility techniques and technologies, including new market structures, flexible generators, demand response, and energy storage. One option for flexible renewable generation is CSP with TES. This dispatch – able energy source can provide grid flexibility by shifting energy over time, providing ancillary services, and ramping rapidly on demand, enabling a greater penetration of VG sources, such as wind and solar PV (Denholm and Mehos 2011). Several previous studies have included CSP to various degrees. The Western Wind and Solar Integration Study (WWSIS) (GE 2010) included CSP with TES but assumed CSP was dispatched in fixed schedules. Integration studies by the California Independent System Operator have included CSP but assumed very little storage (CAISO 2011). Two more recent studies, the SunShot Vision Study (U. S. DOE 2012) and the Renewable Electricity Futures Study (Mai et al. 2012) incorporated CSP with TES into a commercial production cost model (Brinkman et al. 2012) and allowed the model to dispatch the TES resource. These studies demonstrate qualitatively the value of dispatchable solar but did not attempt to isolate the value of CSP with TES or compare how the value of CSP changes as a function of storage or other grid components. Alternatively, there have been studies that focused on the value of CSP with TES but were limited in modeling resolution. An example is a study that used a “price-taker” approach to dispatch a CSP plant against historic prices, assuming these prices (and solar availability) are known with varying degrees of certainty (Sioshansi and Denholm 2010). This type of study can identify some of the additional value that TES adds in terms of energy shifting and ancillary services; however, the value of this analysis is limited because it cannot examine the impact of different fuel prices, grid mixes, or the ability of CSP to interact with variable renewable sources, such as wind and PV.
A more comprehensive 2012 study evaluated CSP using a reduced form commitment and dispatch model and quantitatively identified a number of benefits of TES (Mills and Wiser 2012). This study evaluated changes in the long-run benefits of CSP with and without TES in California using an investment model that included a “fleet-based” commitment and dispatch component for conventional generators. The study also isolates the value proposition for CSP with TES, including energy, day-ahead forecast error, ancillary service requirements, and capacity value, although the simplified commitment and dispatch component of their model did not have the fidelity to represent detailed individual unit commitment and dispatch decisions.
Several studies initiated in 2011, including the second phase of WW- SIS (Lew et al. 2012), examine CSP in greater detail. These studies use the PLEXOS production cost model and simulate the operation of the Western Interconnection in the United States. Simulation of the grid over large areas is important because of its interconnected nature and the corresponding ability of utilities to share resources over large areas. The Western Interconnection consists of thousands of generators, each of which must be simulated in detail.
Given the complexity of a large grid, it can be difficult to validate proper operation of a new generator type and isolate the cost impacts of a relatively small change in the system. As a result, to evaluate the performance of CSP, we began with a test system within a subsection of the Western Interconnection. This test system was used to evaluate the performance of CSP and the incremental value of TES under various grid conditions, including penetration of renewable generators, and compare CSP with storage to other generation sources.