Open mostaphaRoudsari opened 10 years ago
@chriswmackey have you ever had the chance to export an idf model from designBuilder and see what is their approach in applying natural ventilation? I found it pretty tricky to trust their single check mark.
Hey Mostapha,
Yes, this is going to be really tough. I know the basic principles of an AFN thanks to a class I took with Les. I think the buoyancy driven flow stuff is brand new to E+ 8. I'll try to send a design builder IDF shortly of a situation with and without nat vent. Designbuilder does include a lot of other things other than the check mark (they have inputs for schedules and a buttload of other stuff, I remember).
All of this said, I remember us talking with Colin and thinking that the natural ventilation was "really optimistic". The more I have been delving into the comfort stuff, the more I am believing that there really is a buttload of energy to be saved with this and we really need to have this in our workflow to be worthwhile.
-Chris
On Sun, Jul 20, 2014 at 5:23 PM, Mostapha Sadeghipour Roudsari < notifications@github.com> wrote:
@chriswmackey https://github.com/chriswmackey have you ever had the chance to export an idf model from designBuilder and see what is their approach in applying natural ventilation? I found it pretty tricky to trust their single check mark.
— Reply to this email directly or view it on GitHub https://github.com/mostaphaRoudsari/Honeybee/issues/134#issuecomment-49560026 .
Thanks. I shouldn't have said that about DB. The last (and the only) time that I tested it was 3 years ago. They should have made so much progress from that time.
He Mostapha. No worries. I seems that you were still pretty correct about DB.
Here's a DB idf without natural ventilation: https://www.dropbox.com/s/6gpndg0c6llpr7w/Without%20Nat%20Vent.idf
and the same idf with natural ventilation: https://www.dropbox.com/s/cz28p6uxkwhvvww/With%20Nat%20Vent.idf
Hi Guys
My 2 cents here is that we should more or less copy the Designbuilder scheduled and calculated natural ventilation approach (Not sure if I should be using the world "copy") as the core for HB+LB. It worked very well for most of the energy modelling work I have ever done so I suppose it would work well for other people as well. We could then add other functions once this core part is finished e,g Buoyancy driven flow.
If you both think this approach is good I'll put it on my to do list. I'll then write the basic puesdocode in the next two months and once I am back in Australia get started on it (Not working full time) unless someone else whats to start it in the meantime.
Let me know your thoughts
Thank you, Anton. I agree that a first step is more-or-less implementing what DesignBuilder does for the airflow network. I know that E+ v 8.1 has the buoyancy stuff and that would be awesome if you could look into implementing it. If you have any questions about EnergyPlus that you cannot easily find the answer to in the E+ input/output reference, I would post your question on this wonderful site: http://unmethours.com/question/42/what-is-the-best-practice-to-model-an-atrium-in-energyplus/
While you work on that, I am going to investigate a fairly large bug in the source code of E+ airflow network that one of my peers has recently enlightened me to. It is a bug that does not affect about 90% of the simulations that you would run with E+ but, in cases where you have a particularly low amount of thermal mass in the zone, the airflow results that E+ gives can be off by an error of 40%. I know that my peers found the bug a few months ago and sent an email to E+ team but never received a response. More specifically, the bug results from them using a constant pressure of air instead of a constant air volume in the zone energy balance. Presumably, this bug will be fixed soon but I would like to keep track of it. I will keep everyone posted here on what I find.
Thanks you @antonszilasi. Right now @chiensiTB is leading the effort on this. Let's talk over the weekend and I will put you guys in touch. It is very nice to have you on the team.
Hi guys
I've started to consider how to put this together by looking at Designbuilder.
In Designbuilder after setting the Nat vent to calculated the other settings that must be set when using calc nat vent in Designbuilder are...
All this information is then combined in Designbuilder to produce the following in E+ IDF
What I propose to start calc nat vent in Honeybee to copy Designbuilder's approach is as follows
This will cover most of the functions that Designbuilder has to offer and we can integrate more over time
@mostaphaRoudsari let me know your thoughts on this, we also need to have a skype date so you can walk me through the code of all the E+ HB components Ive read most of the components but I can't really understand whats going on. Once I can understand the code you've written I'll understand how to modify it to achieve what I've suggested above.
@antonszilasi It all sounds doable. Let's talk tomorrow about this.
Hey Guys, Peter Ellis gave me some good advice about this yesterday. There are some basic air mixing objects that we should incorporate before going for the full AFN thing I will try to post more info here soon.
Cool. Let's also make sure to keep @chiensiTB in the loop.
Alejandra and I had a very productive meeting yesterday and we got down to the heart of what DesignBuilder is outputting in the IDF. Notably, we found that the two methods that DB uses to account for natural ventilation (scheduled and calculated) correspond to natural ventilation set up with simple Aiflow objects and Natural Ventilation done with an Airflow Network. The attached picture shows the difference between these two types of E+ objects.
I have decided that, for the purposes of my thesis, I am going to implement most of the first class of objects (the simple AirFlow objects or the objects used in the "Scheduled" DB option). Both Alejandra and I seem to be aficionados of this type of calculation because it gives us a clearer sense of what is being simulated and, in the E+ reference, some of these objects use formulas that we understand from classes that we have taken at MIT. My implementation will include ZoneVentilation:WindandStackOpenArea and ZoneMixing, the latter of which which will effectively bring air walls into HB. Note that this implementation will be different from DB's use of ZoneVentilation:DesignFlowRate, which to me an Alejandra, seems to use a bogus equation that we have yet to verify as accurate. I am going to verify with Les whether this equation has any precedent before pushing full steam ahead with the implementation but my gut is telling me now that we probably do not want to use it.
Note that this implementation of the first class of airflow objects is what Peter Ellis had suggested to me instead of the AFN. I checked the DesignBuilder AFN file and there is a crazy amount of info that has to be written in for it to work. Anton, this work will still be there if you ever want to tackle it but be warned that it will take a lot.
Alejandra, I think that my current thesis plan is shifting from using your excel file to pre-determine the airflow to implementing these IDF objects. I will then feed the results of the E+ simulation into your formula for air stratification for the final leg of my comfort maps. I see this as better than trying to pre-determine an airflow since, at the end of the day, both the E+ method and your excel sheet are using the same formulas but the E+ method can change the results as the simulation happens.
I will include the option of either having ventilation triggered by the indoor temperature rise or inputting a schedule that specifies the fraction of window area that the windows will be open. The latter will avoid the issue that you ran across, Alejandra, where the windows were being opened and closed every minute of the simulation as the indoor temperature fluctuated around the desired temperature.
I attached IDFs of DesignBuilder Natural Ventilation with and without the AFN for reference. Let me know if this all sounds good and I look forward to implementing this soon! https://www.dropbox.com/s/onwnxwvsy6zt3s8/NatVent_No_AFN.idf?dl=0 https://www.dropbox.com/s/u17k44ireixx9ds/NatVent_AFN.idf?dl=0
Hello All, Given that I have implemented a lot of the natural ventilation stuff so far, I am going to change myself to be the asignee here. All that I have to implement to close out the issue now is the airflow network and the basic natural ventilation objects on the OpenStudio workflow. -Chris
Basic natural ventilation objects have been implemented in the OpenStudio workflow using a work-around for the fact that OS currently doesn't support these basic nat vent objects. @antonszilasi is also working towards the airflow network. I would suggest @antonszilasi check the status of OpenStudio's work towards the airflow network on the SDK as he continued work.
@mostaphaRoudsari @chriswmackey thanks as you can see here: https://github.com/mostaphaRoudsari/Honeybee/issues/506
I've written most of the code I'll need to check on the status of the open studio AFN at the moment, please do give me your thoughts on whether the work flow that I have proposed is ok.
@mostaphaRoudsari and @antonszilasi ,
Given that this issue has changed so much, it seemed to me that it should be renamed. We currently have the ability to model natural ventilation with E+/OpenStudio and this discussion is specifically about the airflow network for complex natural ventilation cases.
It also seems that this is not necessary to clear this issue in order to claim that we have full support of OpenStudio. This is especially because OpenStudio currently has no means of modeling natural ventilation and we had to write a hack around OS to be able to do so.
-Chris
Model natural ventilation
This will be such a challenge! I can help with items b, and c.