Model for a data-driven demand response client that predicts the future load and allows to apply a load shedding factor.
This model takes as a parameter the number of samples in a day, which is generally
24 for one hour sampling or 96 for 15 minute sampling.
Input to the model are the consumed energy up to the current time instant,
the current temperature,
the type of the day, which may be a working day, non-working day or holiday
as defined in
Buildings.Controls.Types.Day,
a boolean signal that indicates whether it is an event day,
and a signal that if true, causes the load to be shed.
The input signal yShed determines how much of the load
will be shed if shed=true. If shed=false, then
this signal is ignored.
Output of the model is the prediction of the power that will be consumed
in the current sampling interval, i.e., generally in the next 1 hour or the
next 15 minutes.
If the parameter nPre > 1, then the prediction is done
for multiple time intervals. All of these predictions can be obtained from
the output PPreNoShe. This output does not take into account
yShed.
The output PPre is
PPre = yShed * PPreNoShe[1] if shed=true,
otherwise it is
PPre = PPreNoShe[1].
The baseline prediction is computed in Buildings.Controls.Predictors.ElectricalLoad.