Solar Modeling and Performance – Solar Capacity Factor Modeling
A typical approach to modeling a solar facility is to essentially derate the unit for the entire year. The map below comes from the National Renewable Energy Laboratory (NREL) and shows the average capacity factor for solar in the US .
I have seen many analysts model solar with a fixed capacity for the whole year, but we know that the sun does not shine 24 hours a day and different seasons have different amount of sunshine. On top of that is the uncertainty of cloudiness on an hour to hour basis. The typical shape of solar can be seen in the figure below.
I decided to create hourly shapes and test the difference between modeling the solar facility more accurately versus the general capacity factor derates. AURORAxmp is an ideal tool to add this complexity. I used the change set feature and modified 4 tables – Fuels, Time Series Weekly, Time Series Monthly, and Resource table.
For the Fuels table I just added a fuel called PV. In the Resource table I added 4 rows – 2 resources in two zones with flat derate, and 2 resources referencing to the more complex hourly shape. I used the ERCOT world and decided to put solar in West TX and Houston. West Texas got the higher capacity factor ~21% vs. Houston ~12%. In each area I added the two units as mentioned above.
To start my hourly shape I used the shape from above and developed 3 shapes for 3 seasons – Winter, Shoulder, Summer. Therefore, in the forced outage factor I needed it to refer to a monthly reference but I needed the monthly to change by month and refer to a weekly table.
The weekly table is the perfect mechanism when things need to change hourly, since it contains 168 columns. Each column represents an hour for each week. As an example of the weekly shape changes, the summer season for the highest capacity factor averaged ~24% versus the winter being closer to 20%.
In the end, the result showed a significant economic performance gain. I ran ERCOT in zonal mode for the entire year 2012. The 8760 run took around 2 minutes. In terms of revenue, modeling solar with an hourly shape shows almost a 200% gain in economic performance versus the standard derate modeling. They both produce the same amount of energy but dispatching during the day gave a significant economic boost.
The setup for this model run took only a few hours, with MS Excel providing a means to create the hourly shapes.
AURORAxmp’s speed and easy and flexible interface allows you to achieve a more accurate valuation of your projects.
Please call me if you want more information 614-356-0484 David.