Customer Interview: Kate Geschwind - Forecasting Wind Farm Production

Kate Geschwind is a 10th grade student at Mayo High School in Rochester, Minnesota. She recently participated in the 2011 Intel International Science and Engineering Fair (ISEF) in San Jose, California with thousands of participants from around the world. Kate placed third in the Mathematics category and received two honorable mentions from different mathematical societies for her project on "Developing Analytical Approaches to Forecast Wind Farm Production: Phase II".

Miss Geschwind has been only using NeuroSolutions and NeuroSolutions for Excel for the last 6 to 8-months to build on her 2010 science fair project that analyzed historical hourly output data of a Minnesota wind farm using only regression analysis. That project earned her several awards including an all expenses paid trip to attend the Intel International Science and Engineering Fair. The 2011 project moved from an analytical approach to actually predicting the output and also introduces the use of neural networks for the model development process to determine if they can outperform regression-based models. In addition, this year's project applies the developed forecasting approach to multiple wind turbine farms in Minnesota, Oklahoma and Vermont.

The project used a variety of neural network’s including: the Multi-Layer Perceptron (MLP), Generalized Feed Forward Network (GFNN), the CANFIS (Fuzzy Logic) Network, Time-Lag Recurrent (TLRN) and the Support Vector Machine (SVM). The results of the project were compared to different regression, neural networks and a general persistence model commonly used in wind farm forecasting. As a result Miss Geschwind found that “the neural network models generally perform the best over the different forecast time periods…”

Miss Geschwind believes that everyone stands to benefit from more accurate forecast models for wind farm production. The electrical grid would be more reliable if they had accurate forecast on how much wind generation they would have on the system. Accurate forecast would mean fewer fees, which could translate into lower rates for utility customers. Kate had many great things to say about NeuroSolutions including “because it enabled me [Kate] to do my analysis work in Excel...” We would like to thank Miss Geschwind for sharing her success with NeuroSolutions for Excel and we hope to continue to see great things from her as she completes high school, moves on to college and the professional world.

2011 Intel International Science and Engineering Fair (ISEF) in Los Angeles, California

• 3rd Award, Mathematical Sciences Grand Award
• American Mathematical Society
  • Certificate of Honorable Mention
• American Statistical Association
  • Certificate of Honorable Mention

• 2011 Intel International Science and Engineering Fair (ISEF) Finalist Award (all expenses covered to attend ISEF in Los Angeles, CA)
• State Science Award (trip to state science fair)
• Top high school math project
• Clean Energy Resource Teams of SE MN award for Outstanding Clean Energy Project
• U.S. Army Achievement Award for an Outstanding Science Project
• Fair Director Award for Outstanding Data Book
• Top science project for Mayo High School
• Mu Alpha Theta Award for Most Challenging, Original, Thorough, and Creative Investigation of a Problem Involving Mathematics
• National Oceanic and Atmospheric Administration’s (NOAA) “Taking the Pulse of the Planet” award
• American Meteorological Society’s Certificate of Outstanding Achievement
• High School Paper Presentation
  • 2nd Place High School Paper Presentation
  • State Science Award for Paper Presentation