Abstract
Eight separate energy prediction methods, developed independently
across European Universities and Research Centres, have been compared
with respect to their estimated DC energy generation for five different
photovoltaic (PV) module technologies and 7 different sites distributed
over whole Europe. The analysis of this work is the basis for further
improvements of each of the modelling approaches and thus enables
a reduction of the prediction error in PV yield estimations. The
recently completed first of three planned round-robin inter-comparisons
found that the agreement for all methods and all technologies is
within 5% on an annual basis, provided that the environmental parameters
incident irradiance and the module temperature are well described.
This good accuracy was also found when translating the energy yield
measured at one location in Europe to another for an identical module
utilising shorter time periods (months). Significantly higher errors
were found when using different PV modules of the same manufacturer
and technology to predict the energy yield at other sites. Here the
variation in module power rating dominated the results of the energy
prediction methods and a correction for these differences had to
be applied.
Users
Please
log in to take part in the discussion (add own reviews or comments).