When planning the operation of power systems with high shares of generation from
solar PV (e.g. solar panels on rooftops), up to a few days ahead, it is vital to know
the minimum level of power, which can be guaranteed with a very high certainty
from the PVs. Of course nothing can be guaranteed with 100% certainty, however it
is possible to apply statistical extreme value models to the predict a minimum level,
which can be guaranteed with a very high cetainty e.g. with a 99.999% certainty (e.g.
for hourly resolution the realized power will only be below this level in average one
hour out of 10000 hours).
In this project a case, either wind or solar, will be chosen, possibly together with a
company, and the necessary data collected. Then the aim is to build a suitable sta-
tistical model for the particular case. It can both be the output from a single wind
farm or single PV system, which is modeled, or it can be a larger area. The project
can include an analysis of the effect of spatial averaging on the uncertainty, i.e. if the
average output from a larger area is predicted, this should lead to lower uncertainty.
Advanced courses in statistics and experience with R (or similar scripting language).