Back to Profiles

 

Description: "The German Weather Service (DWD) gives IWES wind speed forecasts for a given time span. From these, IWES would like to predict power output from a given wind farm for this time span. My group worked on optimizing the model which takes the DWD’s wind forecast and creates an IWES power forecast. First, I read a master’s thesis (in German) about the optimization process to familiarize myself with the model. I was then given the MATLAB code that went along with the master’s thesis and spent a while running and modifying the code, looking at the results when we tweaked certain variables and trying new optimizations, such as correcting for wind direction, to see if they improved the model.

Later I worked on trying to find the optimal training period length for the model. I learned about several different methods for testing training period lengths, including cross validation and bootstrapping. I was also given new data in the form of Excel spreadsheets and created a convenient MATLAB structure to hold it all for future use."

Comments: "It was my first engineering-related internship and gave me valuable work experience. I learned a lot about statistics and model validation techniques, and I greatly improved my MATLAB skills."

For further information, please write to enevill6@stanford.edu