-
Notifications
You must be signed in to change notification settings - Fork 39
2018 12 07 Task 3 Coordination Meeting
InaDJ edited this page Dec 18, 2018
·
5 revisions
Date: December 7, 2018, 9am Brussels time (UTC+1)
- Progress of Building modelling subgroup
- Presentation of comparison between 5 building models
- Define steps forward (= clearer definition of maximal powers of the ideal heating system for the different models)
- Set deadline for simulation results with different maximal power of the ideal heating system
- Discuss is how to coordinate our DESTEST building models with WP1 building "emulators" that are used in the BOPTEST (proposed by Igor)
- Progress of Network modelling subgroup
- Discuss general framework of DESTEST
- Outline of paper (IBPSA Rome)
- Join Skype Meeting
- Trouble Joining? Try Skype Web App
- Join by phone
- +3216379898 (Heverlee) Dutch (Netherlands)
- +3216379897 (Heverlee) English (United States)
- Find a local number
- Conference ID: 1417905
Present: Dirk Saelens, Michael Mans, Ilaria Vigna, Ina De Jaeger, Felix Bünning, Annelies Vandermeulen, Igor Sartori, Bram van der Heijde (Notes), Enora Garreau, Hanne Kauko, Michael Mans, Alessandro Maccarini
Presentation by Ina: Presentation of comparison between 5 building models
- Difference operative/air temperature
- Different libraries use different standards
- Decision: all ideal heating systems controlled on air temperature
- Peak powers are defined differently
- Belgian/German TABULA data is largely the same, but IDEAS includes reheat factor, AixLib doesn't
- Dirk: models tested with BESTEST?
- DIMOSIM: Yes, was okay
- Buildings: Yes, okay (there's a paper)
- IDA-ICE: To be confirmed, probably yes
- AixLib: To be confirmed, check with Moritz
- Annual values:
- To be confirmed after further checks, but order of magnitude is largely the same
- Temperature profiles:
- (general comment from a colorblind: I can't distinguish between Buildings/IDA-ICE/AixLib colors)
- March
- DIMOSIM has higher temperatures, but should be fixed now
- IDA-ICE cools down very quickly
- But air temperatures, logical
- Multiply air mass by 5 to account for furniture?
- Schedules?
- Some misalignment, some schedules start sooner or later
- To be looked into
- Peaks in IDA-ICE?
- Handling of time steps? Igor will check
- September
- Overheating
- Check thermal masses
- Set up correct building envelope for AixLib
- December
- Time schedule problems are apparent
- AixLib: system is underpowered, high temperatures can't be reached
- Load duration curves
- First align powers
- Air temperature
- Is not per se more realistic, but used for the sake of comparison
- Peak power calculation
- Align better
- Decision to take 99% duration value as peak for every model
- But results are still the same
- Idea to choose one peak power value
- Idea to use average over the first hour per model to compare
- One moment in time is hard to compare
- All to remodel with single maximum power
- Action Ina: set deadline (in January)
- BUT data is needed for Paper, deadline 15 Jan
- Michael: Should we aim for exactly same results? Every lib uses different standards, validations,...
- Should we remove AixLib from results?
- Igor: differences because of human mistakes, different inputs - we aren't ready yet to compare
- Either all rush to get same results in the last weeks, or accept that there are differences and comment on difficulty to implement same situation in different model libraries
- Proposal to try and solve current issues before Christmas, write paper first half of January.
- Conclusion: present conclusions (ongoing work) anyway, hope for even better results.
- E-mail sent by Ina to the whole group on December 18, 2018 for the next round of building simulations
- Dear WP3 colleagues, In our previous Coordination meeting, we discussed how we will run a second round of building simulations, but with a limited peak power. We propose the following: 1 simulation, based on a design power calculation as specified in the standard, including a reheat-factor. This means that the power for the day zone is set at 8024.46 W and the power for the night zone is set at 8548.50 W. In this simulation, all the building should have sufficient power, so not a flat beginning of the load duration curve. 2 simulations, based on the previous simulation, where we do define a limited power, so we end up with a “flat beginning” of the load duration curve (in this case, we can compare the duration of and extent to which the comfort is not met). This means everyone takes their own previous simulation, calculates the load duration curve, takes the average over the first hour of this load duration curve and uses that as a max value in a 2nd simulation. Additionally, everyone takes the average over the first 24 hours of this load duration curve and uses that as a max value in a 3rd simulation. What do you think? If you have any suggestions or comments, just let me know :) As discussed during the Coordination meeting, it would be nice if we could include these results in the BS2019 paper (of which the deadline is on the 15th of January), but I do understand that this is rather late notice. So please, check within your own agenda what is possible. It would be nice if you could let me know when it would be feasible for you to perform the 3 additional simulations (just for my information ;)). All the best, Ina
Presentation by Felix 14-15 people at network modelling tutorial meeting
- Waiting for results
- Alessandro sent results
- Hanne tried, but still weird error messages with data input
- Still waiting for the rest
- Problem presented in Paris
- Not in integrator, problem is in stiff solver
- Euler with small time step provides much more similar results
- Variable time step is chosen so large that it misses the start of the heating season
-
Use fixed time step solver
- Worth mentioning in the paper
- Show results so far
- Compare injected heat
- Results are largely similar (dynamic pipe vs Plug Flow)
- Compare heat losses
- Peaks should be discussed later on
- Compare injected heat
- Felix is writing paper now
- Should be finished tomorrow
- Next steps
- No network meeting so far
- New meeting should be scheduled as soon as new results are there
- Prefer sooner than later
- Outline, tasks, what we want to do
- Building models and results
- Network models
- Felix started
- Bram contributes?
- Outlook of future work
- Send out to review beginning of January
- Building modelling subgroup: tba
- Network modelling subgroup: tba
- Next coordination meeting: tba