Example Configuration 1

Limited data available for a portfolio of frequently changing assets

Background:

• Frequent changes in farm population (farms removed or introduced on an ongoing basis)
• Limited information about the farms available (no online data)
• Off-line production data available with a time delay

 Suggested forecasting setup

• Farms are grouped into a number of areas
• Grouping is a good way to provide a robust forecast by only adjusting a few key properties of the groups (total capacity), but otherwise assume other properties are stable over time
• Avoids the complexity of continuously having to configure individual farms, which will not yield a significant increase in accuracy anyway
• The setup is mostly suited for day-ahead forecast