-
Notifications
You must be signed in to change notification settings - Fork 39
2019 05 28 library development
Date: May 28, 2019, 8am-9am Pacific Times The purpose of this meeting is to coordinate the development https://github.com/ibpsa/modelica-ibpsa
Join from PC, Mac, Linux, iOS or Android: https://lbnl.zoom.us/my/mwetter
Or iPhone one-tap (US Toll): +14086380968,6614042296# or +16465588656,6614042296#
- Or Telephone:
- Dial: +1 408 638 0968 (US Toll) or +1 646 558 8656 (US Toll) Meeting ID: 661 404 2296 International numbers available: https://lbnl.zoom.us/zoomconference?m=_h5BuJ686mPy3rWEyKf4NROxLaeOV5J8
- Update to Dymola 2020
- Block to overwrite and to send control signals (https://github.com/ibpsa/modelica-ibpsa/issues/1059).
- Regression testing
- Actuator model revision
- Update about refactoring of regression tests.
- New media R290 from Aachen (Christian).
- Weather data reader (https://github.com/ibpsa/modelica-ibpsa/issues/842)
- Library restructuring: https://github.com/ibpsa/modelica-ibpsa/issues/778
- Update regression tests to Dymola 2019FD01
- Open pull requests and issues at https://github.com/ibpsa/modelica-ibpsa/issues
Once final version is released, update CI testing and make pull request.
Michael make last review and then PR to merge to master branch.
Funnel is now integrated in BuildingsPy. All unit tests and plots are reported in html as a stream to html. Pending is last review by Michael and Filip.
Need to design how/where to store results on a server.
PR is done for damper model, but numerical problems exist in one of the regression tests that computes slow. More iteration steps are taken inside the central flow domain where the flow rate is computed as a fraction of the nominal value plus a regularization term. The issue seems related to the regularization coefficient which scales with the inverse of the pressure drop difference between fully closed and fully open position for the required flow rate. When the required flow rate is close to nominal this leads to a very low order of magnitude.
No update from Aachen.
Massimo fixed problem with integration bounds. Massimo implemented validation test for full borefield to ensure that heat transfers are properly divided between boreholes. Soon ready to make PR. Will also add average borefield temperature in the model.
In September will have a student working on borefield modeling.
PR is pending, Filip waits for feedback on how to parameterize model (https://github.com/ibpsa/modelica-ibpsa/pull/1113). Michael to provide feedback on issue.
See https://github.com/ibpsa/modelica-ibpsa/issues
Michael to update to Dymola 2020 when it is released.
Michael to provide feedback on https://github.com/ibpsa/modelica-ibpsa/pull/1113 (cooling coil parameterization) and on funnel integration in BuildingsPy.
Massimo to provide PR for borefield validation.
None.