Evaluation and Sensitivity Analysis of Cost Calculations in the Thermo-Economic Modeling of CSP Plants

Abstract: Thermo-economic modeling refers to the process of estimating the cost and performance of a power plant using cost oriented equations and reference data. In this thesis the fundamentals of cost and performance modeling as well as sensitivity analysis is researched and applied to an existing model in the field of concentrated solar power. The thesis aims to isolate the sources of possible errors and presents comprehensible methods of minimizing the sensitivity these give rise to. The extensive literature study provides the knowledge and methodologies necessary to perform an evaluation of a computer model and these methodologies are applied to the tool DYESOPT developed at the Royal Institute of Technology.   The evaluation highlights the importance of reliable references of operational solar power plants and also the current lack of such data. A particular area suffering from this is the cost estimation, which includes assumptions and requires future revisions. The sensitivity analysis methodologies one-at-a-time and the sensitivity index are used to locate the areas where extra care must be taken in order to minimize error as well as provide an understanding of the internal correlation of critical inputs.   The results show that the accuracy of the model is dominated by three inputs: solar multiple, tower height and storage time, and that certain intervals and combinations of these decide the overall error of the model. By isolating the intervals in which the sensitivity is at its minimum the model error can be roughly quantified with a class system using standard error intervals. For a model such as DYESOPT a minimum error of 20 to 30 percent is a reasonable assumption.