India with heavy drop out live data

[peterdudfield]

Detailed Description

Ive been noticing that the live data from RUVNL sometimes is not up to date. It therefore is probably worth creating a model that can still give a fairly good prediction if there is no live data. Perhaps a model that can handel live and no live data would be ideal.

Context

• large gaps in live generation data
• https://github.com/openclimatefix/india-forecast-app/issues/11

Possible Implementation

• add large amount of drop out to the live generation data (both wind and pv) https://github.com/openclimatefix/ruvnl-consumer-app/issues/8
• perhaps the model already can handle this situation?