water as thermal mass

[macpavee]

hi

i'm simply trying to simulate the passive effect of water as thermal mass. has anyone had success with this?

i have 910mm diameter eight sided 'barrels' in a pillar with openings permanently 'on' between each, hoping to create a single mass leading to a small roof pond above. i have then adjusted the barrel walls construction type to the following: 1mm steel, 450mm 'steel' but with specific heat 4184 J/kg/K, density 999 kg/m3, thermal conductivity 0.58 w/mK, i.e that of water.

the simulation however doesnt show the expected lagging temperature curve of subtler peaks and troughs in the pillar (let alone the room) to that of the outside dry bulb.

does the fact that the thick walls will overlap inside and towards the centre confuse the system somehow?

am i missing something?

thanks

[Complex Potential]

Hello again Mac

The closest thing I have seen to this was when a colleague attempted to simulate ground conditions by creating several layers of sublevels of earth and allowing them to precondition for a year. As I recall, the results looked quite plausible.

In your case my first suggestion would be that the way the 450mm "steel" construction has been set up is perhaps not quite right. Since the pillar is internal, the "walls" of the pillar will be internal partitions rather than external surfaces which means that the 450mm thickness is spread equally on each side of the construction line in plan view. Geometrically , this will create a 450mm pocket of air in the middle of your "barrels" which could well be messing up the calculations. If you switch on internal volumes you may well be able to see this pocket in plan view.

If you redraw each barrel with a diamater of 450mm instead of 900mm it should create the 900mm of actual mass you want without any air gaps in the middle.

CP

[macpavee]

Thanks for the reply CP
much appreciated

i increased the thickness of the 'water' element to the total diameter thickness of the tanks (sort of what you advised) and this did seem to improve things....at least inside one barrel (the lower one) but i suspect there are now other reasons for the difference.

not much change in the room yet....and a million other frustrations.

[Complex Potential]

I suspect you may be hitting the limits of what IES can actually model in detail. Conductive transfers between adjacent constructions are very bespoke and probably best represented by a finite element analysis in a program such as ANSYS.

If it is the room condition you are attempting to model then you could approach from a different angle. Instead of trying to model the barrels themselves explicitely you could set up sensible and latent heat losses/gains within the space controlled by equations (profiles) that reacts to the internal and external temperatures based on the physics of thermal storage.

Of course, this would take a bit of delving into Physics books and CIBSE guides (as well as a bit of practical experimentation possibly?) but it may be the best method.

[macpavee]

Thanks CP i appreciate the input
&
intriguing thoughts

In the long run i can look into your suggestion, i have a year to play with IES in this way. Have you any experience with TAS?
but i wont pretend to yet know how to implement your method practically on IES...i'm very 'newbie'

unfortunately i need to get data together asap (next few weeks) for my current study. the limitation you mention will be useful nonetheless.

however
IES seems to have no problem demonstrating the benefits of a concrete floor added to my model, although with thermal mass surface area is important -(a further limitation of the barrel idea). but still, IES is giving results.

Unfortunately i am researching temporary structures so a thick concrete floor is no go...practically speaking

certainly experimentation is on the cards later. but for now i need to go with some sort of valid demo of water as thermal mass in a temporary structure.

hmmmm
i will run some simulations and compare the effects of a concrete and a 'water' floor setting, each at different thicknesses.
see is there is any difference to the room temperature (degree hours/discomfort coefficient) between the materials

yeah the practicalities of a water floor! i know. off the top of my head i am thinking thick coils of pipe or water beds as tanks....somehow.

anyhoot
lets see what IES throws up and i will post up some results later.

[Complex Potential]

Tas is similar to IES in it generals capabilities although they do work on different calculation engines. I'm not sure it would be much better at modelling this in all honesty. Also, I think that attempting to model a fluid like water as an adapted solid construction may mean components like convective heat transfer are not picked up correctly.

One final thought for the morning.

Going back to your original model concept, a thermal store will need some way of rejecting heat in the summer, otherwise it is simply going to get warm over the course of a few days and then stay that way. Presumably you are modelling the room as purging at night to allow the barrels to re-cool?

If the water is being actually being replaced from an external source then you would need to find a way of modelling that, perhaps by setting the temperature at the centre of the barrels artificially using setpoints. I think this could be fundamentally where your model is struggling since it is not possible to set the temperature of a construction thickness directly, only the temperature of the "room" is it surrounding.

[macpavee]

Thanks for all the thoughts CP

i did set a night cooling schedule using an insulated door on top of what was effectively a roof pond sitting on the pillar of barrels.

however
(ahem) after some head scratching and hours of messing about with macroflo opening types i have realized (a typical novice mistake) that i was sometimes failing to hit the 'macroflo link?' tab in the sim, at least when it mattered

since then i have rebuilt the model (as you know) and am fooling with room grouping (as you know) as well as this and a passive ventilation set up. i'll hopefully be retesting things a bit more properly soon. but have some foundational sensitivity analysis to do first.

the pillar option is more suitable/practical option to the construction than the floor so hopefully this will show some positive relief.

but (fingers crossed) things are slowly coming together now (at least to the point where i can begin to analyse some useful data) and i can only thank you for the patient assistance and this forum for it's existence.

[Complex Potential]

We do our best and don't worry about the missed checkbox, it's very easy to do.

Once ticked and run once that checkbox should remain checked unless you are resetting macroflow licences. Just seems a little odd but I'm glad you're starting to get it together now.

CP

[macpavee]

so initial results are not that encouraging.

water and concrete at a depth of 50mm with 50mm insulation beneath both performed in a similar manner, with the water sim slightly besting concrete but no where as much as it should do.

interestingly i also tried a comparative run using the thermal properties of cob. although the worse dampening performance as regard to temperature could be expected. what was noticeable was the simulations report that relative humidity in the room went rocketing.

this of course is counter to the way cob would respond which is by buffering the RH extremes.


in the hope of assessing the uncertainty I have tried altering the absorptance, first to 0.94 but this has changed nothing. i presume a new radiance simulation needs to be done to recalculate illuminace???. however radiance now keeps on crashing

is this necessary as interior surface illuminance wont be to significant? but i would like to know by how much the simulation responds to this parameter.

i also havent performed the obligatory adjustments to glazing yet to maximise the direct solar gain.

[Complex Potential]

Hi Mac

Radiance should not be needed at all for what you are describing. Generally the only time radiance has a bearing on a dynamic thermal simulation is when you are using dimmable lights and daylight sensors. I would simply uncheck it.

As for the rest, it sounds like it's going into some quite deep questions about the calculation engine and exactly what it can and cannot pick up. Unfortunately the only people who can answer those are the developers and they tend not to frequent these forums.

If you send your model to IES support (archive it and send as a cab) describing what you are attempting to achieve and why you feel the current results are wrong, they may be able to give a definative answer as to whether you are flogging a dead horse.