I am currently doing research on sustainability through smart parametric design of a 3d printed wall, focusing on optimized thermal conductivity.
The idea is that purely through better design and with the freedom that 3dp construction gives, a wall with a certain interior printed design could be more thermally efficient than a hollow wall or one with traditional insulation.
In PHASE 2 of the project I will look into inserting different types of insulation into the cavities, as well as how different cement or mortar mixes influence the thermal performance.
For PHASE 1, I will remain material agnostic, hoping to acertain an efficiency ratio, of Design 1 vs Design 2 VS Design 3 (assuming this ratio could than be applied to any material that is later used)
This is particularly important because currently, I only have access to PLA type 3d printers. Next semester I will hopefully have access to a concrete 3d printer. (Parametric design also helps me scale from model to life-size).
So, I have some experience using DIVA, Climate Studio, EnergyPlus, Ladybug for daylighting and thermal analysis, but have never used honeybee.
What would you recommend I do?
Is ladybug and/or honeybee the best option?
Here in Taiwan they use COMSOL or ANSYS, but I lack experience with these programs, and my design is in Grasshopper, so I would like to stay within the Rino environment if possible.
I am in contact with researchers that have done thermal analysis on a 3d printed clar/cob wall segmet, but they used Therma.
Any recommendations/tips/suggestions would be appreciated.
I agree with your choice of THERM. Assuming you aren’t doing anything too exotic with the insulation, then the majority of your heat transfer will be through thermal bridging in your construction. THERM has certain limitations in that it only simulates steady-state conditions[1] and only models 2d planes, but I think both these constraints are good simplifications to make given the overwhelming impact of insulation on the wall thermal performance relative to the more marginal impact of transient processes or 3D details.
However, if your wall doesn’t have insulation, then you may need a more representative simulation engine. I would suggest one, but I am a little unclear on your thermal performance optimization strategy in relation to a typical insulated wall. In particular, I am skeptical of this claim:
a wall with a certain interior printed design could be more thermally efficient than a hollow wall or one with traditional insulation.
Maybe I’m misunderstanding your claim, but I don’t see how an uninsulated wall construction can outperform the thermal performance of a typically insulated wall through geometry optimization alone. I can imagine some exceptions:
An extremely deep, thermally massive wall (i.e. multi-layer masonry) that could buffer heat flow enough to compete with a minimally insulated wall, but that isn’t more efficient in terms of cost or material quantity.
A (very unconventional?) hollow construction with sealed air or thermally inert gas, essentially acting like a window. However if you don’t have that seal, the insulating effect of the hollow should be marginal. The air temperature of the hollow construction should approximate the exterior ambient air on average, and the insulating effect of the air film on the surface relatively small, relative to actual insulation.
Moderate climates where the exterior air temperature is already within thermal comfort setpoints. In this case I can see how a 3d printed wall can outperform a typical insulated assembly by optimizing just radiant and convective heat transfer, independant of the conduction across the envelope. In this case, the optimization would have to address something other then average energy impact, since the baseline climate doesn’t have a large thermal energy need. So the focus here would shift to thermal performance during extreme cold/hot periods, or thermal comfort.
Outperforming a typical hollow construction with geometry seems reasonable, although I’m not sure if the thermal improvement would be significant against dominating impact of insulation layer.
Am I missing something? How can your 3d-printed wall compete with a typically insulated wall with just an interior print?
[1] Last I heard, THERM was in the process of releasing a transient-state simulation, but I’m not sure if the old Honeybee-legacy components will work with that release.
So, I have gotten caught up with trying to model the width and girth of each printed bead/layer, with the idea of precision, but this has proven very PC demanding and time consuming. I have since reduced my model size, but might just drop this idea altogether in the interest of focusing on interior geometries alone.
About the interior geometries- I have found two (conflciting) results in experimental literature. One states that very closely spaced, thin webs (running parallel to the wall length) are optimal, while the other states that closed hex or quadrahedrals (but not triangles) are the most efficient. Both types have air locked inside.
A different work from the UAE, focusing on masonry brick design arrives at a similar conclusion that more, thinner webs parralel to the wall create a better barrier.
The idea of using TUNY, Galapagos or Optimus to help me optimize the interior geometries is attractive, but I will focus first on completing a working model for thermal testing, and once I have confirmed that my workflow produces (desired?) results, I will test other geometries.
I have recently also found works on surface (facade) designs that optimize thermal mass to dissipate heat, optimzing use of material soley by varying their geometry.
The end goal of my study is interior thermal comfort. I believe that natural ventilation could help greatly with this (termite mounds, passive breathing buildings), but that aspect will not be considered in this study.
Thank you very much for your advise!
If my design proves to still be too heavy to comfortably manipulate, I might import a more basic one from CURA5, and try to do an analysis on it instead. (Every minor alteration currently takes about 800s to compute…)
Here is a screenshot to give a general idea of what I am working with. I have since oriented my gaps along my wall width, but somehow broke my containment barrier…
I figure as long as its solid, it should work for proof of concept.
I have managed to get my model up to a workable height, and if I dont manipulate it too much, my computer doesnt crash.
Unintsalling ANSYS-Fluent to make room for ABAQUS. Seems its much easier to port from GS to ABAQUS, and multi-step from GS to Fluent, which would break my feedback loop when looking to have the wall optimize itself with OPTIMUS or Galapagos.
I will try THERM one more time tomorrow. I was unsuccessful getting it to work today.
It seems by far the simplest solution, so it is very apealing. I cant use it for natural ventilation, but I dont need to test that at this phase.
Hi I think I am working on a project that has similar concept as you, also interested in how geometry and morphology can alter the thermal conductivity. May I ask which project were you referring to when you mentioned “I have recently also found works on surface (facade) designs that optimize thermal mass to dissipate heat, optimzing use of material soley by varying their geometry.”
I am also looking for simulation scripts for uniform geometries. Perhaps we can share information if you are still working on the same project!
Hi, I am doing a similar project for my undergraduate thesis. I am working on modifying brick morphology and stacking for better internal thermal comfort. I just wanted to know which software you ended up using for thermal performance analysis? Also, did you perform the analysis for just one brick or for the internal thermal comfort of the room?
Any guidance with this would be immense, Thanks in advance.
