SAM against CSP plants in operation

We have been working with SAM doing some production estimations in CSP plants...As a result, we have obtained an annual NET energy value from SAM which could be our guaranteed value in a contract.

Our question the following: which is your feedback about the values obtained by SAM and the real energy generation of plants in operation? Talking about the energy production, are annual results from SAM more or less optimistic than real values of plants? Is there any report talking about these issues?

We would be very grateful if you could give some information about the leeway between the estimation data from SAM and your experience in energy generation from CSP plants.

SAM's performance models represent the general performance of a given technology rather than the performance of a particular plant. SAM is designed for pre-feasibility analysis, not for the design phase of a project that requires engineering tools to predict the performance of a specific design. 

SAM's default input values are convenient starting point for a pre-feasibility analysis. They are not intended to represent a particular plant. SAM is not intended to be used as a design or contract guarantee tool. It is useful as a third-party validation or verification tool, or as a method for different partners involved in contract negotiations to exchange information without divulging proprietary data or the details of in-house models.

NREL can not and does not assume any kind of liability or provide guarantees on SAM's results.

That said, the CSP models are not designed to over- or under-predict system performance. We have designed the models to be as accurate as possible within the context described above, and to provide realistic, "middle of the road" performance estimates.

Regarding specific CSP models:


We have benchmarked the Physical Trough model and found that it estimates performance at the mean of other models.
The empirical trough model was based on performance data from the SEGS plants.
The molten salt power tower model uses DELSOL, which has been validated using Solar II data, and the performance models have compared favorably to limited Solar II data.
The DSG linear Fresnel model has been reviewed by industry, though we have not validated against data ourselves.
We have marginal confidence in the accuracy of the dish-Stirling model at input values away from the default values.



We are interested in seeing studies comparing SAM results to measured data from installed systems, and have created the Case Studies page to share the results of such studies. If anyone has a copy of or web link to other studies to add to the page, please let me know.

Best regards,
Paul.