search

quintel / merit

A system for calculating hourly electricity and heat loads with a merit order
MIT License
3 stars 2 forks source link

Where should merit read load profiles from? #113

Closed jorisberkhout closed 10 years ago

jorisberkhout commented 10 years ago

I've discussed this topic with @antw, @AlexanderWirtz , @dennisschoenmakers and @ChaelKruip before, but would like to gather their opinions in this issue and come to a conclusion.

All load profiles (i.e. chp's, river, solar, wind and total_demand), are stored in the merit repository (in load_profiles/<county>/). Most of them are only used in merit, except for total_demand, that is also used in the calculation of the Loss of Load Expectation. To facilitate this, the load profiles were copied to ETSource (see https://github.com/quintel/etsource/commit/17dc2f61332ac76616619790ade8dcc6e08327c1) and are used by ETEngine.

While adding the total_demand load profiles to ETSource for the new countries in the Energy Productivity project, I realised that it is very easy to forget to add/update the load profiles on merit or vice versa. In addition, the load profiles should be part of the dataset generation process such that we can document their sources and keep track of the changes (see https://github.com/quintel/etdataset/issues/504). To solve the latter, I am going to create an additional analysis that generates the load profiles and use rake import to copy them to ETSource.

Two open questions remain:

Let me know how you feel about this, @antw, @AlexanderWirtz , @dennisschoenmakers and @ChaelKruip.

jorisberkhout commented 10 years ago

I would prefer merit to read the load profiles form ETSource, to avoid storing the same data in two different places. If this is a problem since merit has to run as a standalone module, we could consider creating a rake import functionality on merit that copies the load profiles from ETDataset in addition to the rake import that copies the load profiles from ETDataset to ETSource.

I prefer .csv over .yml as these are easier to generate with Excel.

antw commented 10 years ago

Where does merit read the load profiles from (ETSource of merit)?

Currently, Merit reads the files stored in the merit repository, not those in ETSource. It is my intention to end this practice soon (once my part in the price convergence project is finished) and move the files to ETSource.

How should the load profiles be formatted (.csv on merit, .yml on ETSource)?

I have absolutely no preference. Merit caches the load profiles after the first read; since performance is therefore not a limiting factor, you can choose whichever file format is most convenient for you.

jorisberkhout commented 10 years ago

It is my intention to end this practice soon (once my part in the price convergence project is finished) and move the files to ETSource.

Thanks, @antw ! I wasn't aware that you were planning to do this. But moving them to ETSource sounds perfect.

I have absolutely no preference. Merit caches the load profiles after the first read; since performance is therefore not a limiting factor, you can choose whichever file format is most convenient for you.

.csv has my preference as it is easier to edit with an Excel analysis on ETDataset.

antw commented 10 years ago

All done!

jorisberkhout commented 10 years ago

Great! Thanks @antw .

AlexanderWirtz commented 10 years ago

I like!