Open jecollins opened 6 years ago
The most principled way to compute regulation capacity seems to require information about model state that's not currently being maintained. For example, curtailing a water heater only works if the unit is currently drawing power, and from the customer's standpoint should only be allowed if the temperature in the water heater is above its minimum value. A simpler way would be to just offer a fixed fraction of current consumption for up-regulation, and zero for down-regulation. One question is whether to include all or part of shifted consumption from previous slots in the calculation for the current slot.
As far as I know, this issue applies only to factored-customer. I have spent way too much time looking at that model while planning for implementation of the evcharger model. My conclusion is that this is a difficult issue that will require significant resources.
Demand-response in the form of up-regulation and down-regulation for balancing is based on customers declaring their maximum regulation capacity to their subscriptions in each timeslot, by calling TariffSubscription.setRegulationCapacity(). Currently, no non-storage customers do this, so regulation using simple curtailment is not possible.