Develop a function to update information

[jd-lara]

No description provided.

[claytonpbarrows]

Currently the design is that a PowerSystem struct contains the timeseries information for relevant devices. This creates an issue in PowerSimulations.jl where creating problems on time selections that are not perfectly aligned with the timeseries included in the PowerSystem requires the re-instantiation of the PowerSystem object.
I believe that the correct resolution is to:

1. Depend only on the timeseries information in the Forecast structs when creating/simulating problems in PowerSystems.jl
2. Remove the population of timeseries fields such as scalingfactor in the PowerSystem struct and the related device structs.
3. Provide the appropriate checks to ensure that the required timeseries information exists in a Forecast struct to enable a PowerSimulation with a PowerSystem struct.

I'll leave this open for comment, then migrate to new issues from 1, 2, and 3 as the discussion dictates.

[jd-lara]

I am for option 2 and then have more time series structs where we would have a device field that we can use to point to the device for which this is relevant.

[claytonpbarrows]

@jd-lara is that to say that you are in support of the proposal that includes all of 1, 2, & 3? Because I think that 2 isn't strictly necessary, but would be a good idea to improve clarity.

[jd-lara]

@claytonpbarrows yes, implement 1 but not only for forecasts also a new struct for just time series.
Implement 2 to simplify how the data looks like and also is parsed and 3 might be needed or not.

[claytonpbarrows]

@jd-lara, in some sense everything that we have enabled in PowerSimulations is optimizing schedules against a 'forecast', so until we do something different, I don't see the need for a new struct.

[jd-lara]

This will be resolved with the implementation of the time-series update in PowerSimulations.

[jd-lara]

I am closing this since now all the structs are mutable.