WP1.2 Coordination Meeting April 19, 2021
Meeting Report WP1.2 ‘Modelica library for MPC’
Subject: WP1.2
Date: 19-04-2021
Location: Skype for Business
Minutes taken by: Lieve Helsen (KU Leuven)
April 19
- Emulators
- BOPTEST trials
- BOPTEST journal paper
- Breakout sessions for next on-line IBPSA Project 1 Expert Meeting
- Further issues
1-year simulation in JModelica succeeded (Tao). Documentation is ready, add unit testing for 2 week-periods (copy and adjust scripts from other emulators – action Dave), scenarios are defined. Dave points to documentation on the scenarios: scripts to find the day where the system need for heating is highest and where the system need for cooling is highest, also typical heating and cooling days, and a mixed day.
Dave progressed, controllability problem is solved (issue with priority of valves), documentation has been updated. Final revision and ready!
READY, will be updated when bug in weather file in IDEAS has been solved.
Iago started working on his emulator again, updated documentation, to be checked whether there are problems with JModelica in yearly simulation.
Dave made progress, ready for review (by Filip) on BOPTEST branch
Yeonjin made progress, ready for review (by Konstantin), sent by email to google account
Call to test BOPTEST as much as possible
Authors have been contacted by Dave and outline can be found here.
Good progress, 90% of content is in. Javier continued testing on the BESTEST hydronic case with heat pump focusing on showcasing BOPTEST, not on the performance of the MPC used. Prediction horizon and control time step have been varied to compare control performances. Reviewed by Dave and Jan, working on a revised version. Dave wrote text on the BOPTEST, Sen reviewed and made good suggestions. Another iteration and then the draft is ready for careful review.
May 7 & 12, 2021
here.
- Adoption of Modelica
a. Roel: not a good timing to contribute and more related to MPC (instead of modelling).
b. Feedback from practitioners is indeed very useful.
- Discussion about continuation
a. SDU has Modelica modelling as a core strategy, support continuation, mainly as user.
b. Filip stresses that a critical mass is needed. Could we make the difference between developers and users more formal? Can be easily done by assigning a role to each person (developer/user), but keep them both in the same WP since their interaction is important.
c. Not easy for new people to take up a developer role since we progressed already to an advanced stage. So, having Modelica used by practitioners is important.
d. LBNL wants to continue.
e. More time needed to have real impact, otherwise big risk that everything ends.
f. Training material needed, extra WP in new project for translation towards practitioners?
- Progress needed in:
a. BOPTEST trials
- Outreach
a. VOLTRON (operating system to help implementation of building control): implements controller as an agent, they have agents for EnergyPlus simulation, they work on an agent to be coupled to BOPTEST
b. Igor Sartori (SINTEF) is engaged in IEA Annex 81 and is preparing an ERANET proposal on a competition focused on ML and data-driven analysis and control (2021-2023), competition in 2023, wants to use BOPTEST for this competition. A separate emulator will be developed for this competition (funded by industry). Prize: winning-controller will be implemented.