Validation of SAM Modeling of Concentrated Solar Power Plants

Abstract

The paper proposes the validation of the latest System Advisor Model (SAM) vs. the experimental data for concentrated solar power energy facilities. Both parabolic trough, and solar tower, are considered, with and without thermal energy storage. The 250 MW parabolic trough facilities of Genesis, Mojave, and Solana, and the 110 MW solar tower facility of Crescent Dunes, all in the United States South-West, are modeled. The computed monthly average capacity factors for the average weather year are compared with the experimental data measured since the start of the operation of the facilities. While much higher sampling frequencies are needed for proper validation, as monthly averaging dramatically filters out differences between experiments and simulations, computational results are relatively close to measured values for the parabolic trough, and very far from for solar tower systems. The thermal energy storage is also introducing additional inaccuracies. It is concluded that the code needs further development, especially for the solar field and receiver of the solar tower modules, and the thermal energy storage. Validation of models and sub-models vs. high-frequency data collected on existing facilities, for both energy production, power plant parameters, and weather conditions, is a necessary step before using the code for designing novel facilities. National Electricity Market data tell us that even installed capacity of 18.1 GW sometimes is not enough to deliver 0.2 GW to the grid in cases, for example, of low wind after sunset), the coupling of concentrating solar power with molten salt thermal energy storage is extremely attractive, as it may produce fully dispatchable energy.While the concentrating solar power solar tower coupled with thermal energy storage has so far performed badly in the real world, the concentrating solar power parabolic trough with thermal energy storage is already delivering good performances, [8][9][10][11][12], even if still less than the predictions. Solar photovoltaic works with annual average ε about 0.27-0.29, but with much larger than the capacity factor high-frequency standard deviations, for coefficients of variability in excess of unity [15]. While in the virtual reality of model computations concentrating solar power with thermal energy storage may achieve much larger and more uniform ε, with average ε well in excess of 0.5, also addressing the issues of lack of production during night times, and drastically reduced production with clouds, with dramatically reduced standard deviations, the best real-world experience has so far delivered an annual average ε of about 0.36 (Solana). By contrast with photovoltaic, concentrating solar power, especially the solar tower, suffers from cloud coverage, and this may considerably affect the accuracy of the models.As concluded in [8][9][10][11][12], the concentrating solar power parabolic trough has the potential, once a satisfactory design will be industrialized, to deliver the same or better than photovoltaic costs of electricity with...

A 140 MW Solar Thermal Plant in Jordan

“…In the much simpler and more reliable PT technology, CSP is certainly performing much better and the reduction in the capacity factors of existing plants in comparison to the design ones is significantly less. For example, Solana is operating on capacity factors of 36.4% in comparison to the design value of 43.11% [27]. Surprisingly, Genesis and Mojave Solar Project, which are based on the more established PT technology but without any TES, are not lagging the design values at all.…”

“…And finally, the technology Crescent Dunes operates with is ST, with a TES of 10 h and no NG boost. It is worth mentioning here that all the presented capacity factors discussed in [27] are calculated based on the year 2018. Upon investigation of the latest electricity production data of ISEGS, capacity factors are found to be in the low 20% [27].…”

“…It is worth mentioning here that all the presented capacity factors discussed in [27] are calculated based on the year 2018. Upon investigation of the latest electricity production data of ISEGS, capacity factors are found to be in the low 20% [27]. These undesired numbers are even not accounting the burning of significant NG amounts, which represents additional generating costs and pollution.…”

“…These undesired numbers are even not accounting the burning of significant NG amounts, which represents additional generating costs and pollution. As a matter of fact, the actual capacity factors are drastically reduced by one-third to less than 15%, once the consumption of NG is accounted for at the fuel energy conversion efficiency of a reference combined cycle gas turbine plant [27]. The low capacity factors are significantly alarming in light of the fact that the designed capacity factor for such a plant was 32.68%, with a much-limited boost by the burning of NG.…”

“…The low capacity factors are significantly alarming in light of the fact that the designed capacity factor for such a plant was 32.68%, with a much-limited boost by the burning of NG. Moreover, the Crescent Dunes plant, which is now operating with capacity factors of around 10%, had a designed capacity factor of 51.89% [27]. Not to mention it is continuously subjected to periods of shutdowns.…”