Handling (custom) weather data

[TeodorChiaburu]

Hello!
Thank you for your great contribution! We are trying to use Teaser for public buildings in Berlin. I have some questions regarding the integration of weather data:

1. Where exactly is the model making use of the weather information. I have tried digging for the responsible script/function starting with calc_all_buildings() from project.py but had no success...
2. The data we collected from the weather stations in Berlin has been delivered as csv-files with the following features: Mean Temperature (Celsius), Day Light Duration (Hours), Cloud Amount (%), Wind Speed (km/h) and Humidity (%). Of course, I noticed your default weather file from Mannheim has other attributes and also a different format, namely .mos. However, when running the model with our csv instead of your mos, I did not get any error message. Can the csv be translated into a mos? And are we missing key weather attributes necessary for the model to run as expected?

Thank you!

[PGorzalka]

@TeodorChiaburu as far as I know, you need to convert the weather files to the exact same table format that AixLib expects in its weather data reader. This reader basically expects you to deliver EnergyPlus epw files.

@DaJansenGit I'm currently working on my issue #679 / AixLib issue 1080 and I have a similar issue there: The ReaderTMY3 with files as a source is fixed in the AixLib Multizone template. With a custom multizone template (which I am currently implementing), it would at least be possible to either change the reader or use the possibility to fill the reader's value without a file. My favourite way, actually, would be, to implement a reader that can directly read DWD's test reference years which are available for all Germany, but have another data structure than EnergyPlus and AFAIK cannot be converted without data loss.

However, I think Mean Temperature (Celsius), Day Light Duration (Hours), Cloud Amount (%), Wind Speed (km/h) and Humidity (%) are not sufficient for simulating the models created by TEASER. You will at least need hourly temperatures and solar radiation (diffuse and global and/or direct). Solar radiation is not commonly covered in public weather station data.

[DaJansenGit]

Hey @TeodorChiaburu and @PGorzalka. Thanks for reaching out regarding the weather files. Actually the current implementation of the TMY3 weather reader of IBPSA can be used with DWD test reference years. They just needed to be converted correctly. Sadly, our own converter is currently not available outside our institute, but we are discussing to make them available.

But as @PGorzalka already pointed out: Mean Temperature (Celsius), Day Light Duration (Hours), Cloud Amount (%), Wind Speed (km/h) and Humidity (%) will not be enough to run a dynamic simulation. Please have a look at the ongoing issue regarding the TMY3 reader and possible improvements in AixLib: RWTH-EBC/AixLib#1222. There you can find a list of needed data for the simulation.

[PGorzalka]

Actually the current implementation of the TMY3 weather reader of IBPSA can be used with DWD test reference years. They just needed to be converted correctly. Sadly, our own converter is currently not available outside our institute, but we are discussing to make them available.

That would be great :) Until then, the custom multizone template option from #679 will at least allow to directly write into self-created Modelica models that fill the weather bus in a user-defined way.

By "data loss", I meant the missing upwelled infrared radiation by the ground, which we already discussed about in AixLib issue 922 - and I will post another comment there as the discussion does not really relate to this issue here.