This competition and all ideas within it are archived
Check out the new way Marblar is realizing the promise of science.
The competition has now ended, but the conversation continues. Browse past entries, comment on your favourites, and join the discussion below. If you have any ideas for a new entry, please leave it in the discussion below.
Uniform appearance, varied internal structure: Using 3D printing to make ultra high efficiency passive solar buildings.
29 May '13

This is one of the more speculative ideas since 3D printing of an entire home is at least a decade away, but here's the idea.  

Right now, passive solar buildings are buildings designed to minimize heating and cooling costs by using strategies to take advantage of the natural environment.  It's a collection of techniques like southward facing windows, ultra insulating materials, angled overhanging roofs to shade during the summer when the sun is high but let in light in the winter when the sun is low, and putting thermal mass where it will radiate heat when it's cold and take in heat when it's warm.  

There are companies that do this right now and it tends to cut about 80% of household energy consumption for heating and cooling when done right.  A problem with the process is that sometimes the aesthetics of the house and the demands of the physics don't coincide. Furthermore, the process is done using straight lines and sharp angles while the sun moves in curves. Also, the entire wall has to be built the same way even though the part of the wall two feet off the ground receives a differing amount of sunlight then the part 5 feet off the ground.  Finally, the building has to be very meticulously planned to make sure that the sun is hitting the right materials at the right time of year.  

  My idea is using a program to automatically vary the internal structure of the walls, floor, and roof of a 3d printed house to make it maximally thermally efficient. It uses two unique properties of 3d printing to make this process a lot more precise and, at the same time, a lot less complex to plan.

It relies on two aspects of 3d printing. 

1)  3d printers can vary the internal structure of a material while keeping the external appearance constant. This means that It could make a straight or curved wall that from the outside looked to be a single uniform material, but would be able to vary the internal structure to make it a better or worse insulator.    

2) 3d printers are computer based and directly computer controlled so what would be very complex calculation and fabrication task if set up by a person can be handled automatically by a program.  

Here's how it would work.  The architect would design the house in a normal CAD program. Once that was done, the architect would put the CAD file into the 3D printing software. The architect would then add some basic information, like the exact location of the house, locations of adjacent buildings, the kind of ground it will be on, what the desired internal temperature is, what color it will be, data on local weather etc.  The 3d printing program would then use that information and the CAD blueprints to make instructions for the 3d printer that's going to build the house.      

The 3d printer software would automatically vary the internal structure of the housing materials as it was making the house with the resulting house being thermally optimal for its location.  Parts of the walls and floor that would need to absorb, store, and radiate heat would be far denser (and thus have more thermal mass) than parts that didn't. The important innovation here is that, while the external building would look as square or symmetrical as any previous house, an x ray of the walls would show that they varied in density in irregular shapes designed to absorb or repel heat based on where the sun is. 

(This is a house built the current way, every flat surface has the same thermal conductivity across its entire face.)

The architect could rotate the building or place it another place, and the program would automatically make a new unique internal structure to that would be thermally optimized for the new location.  

The program would also automatically put in things like thermal vents and build around windows, doors, pipes, and electrical outlets and such.  The program could even be designed to to place these automatically based on simple parameters (3 outlets in each room, two windows in the living room etc.)  

Also, 3d printed houses are built so that they don't have the same amount of joins which allow heat to escape  that normal houses do.  

Juan Pablo Garcia on May 30, 2013
Can't a closer in time approach be to build modular walls or covers that have different densities (or the factors that affect the thermal function) so you can mix them and install where matters most? If a full house is a long way now, maybe parts can be made first with your technology and criteria?
Matthew Reinert on May 31, 2013
That's a great suggestion. It would work in modular construction. The basic idea is just that the computer controlling the 3d printer has to know exactly where a given piece is going to be in relation to the sun and the ground and can alter the thermal conductance (either by altering the density or the composition of the material) accordingly.

There is no reason why it couldn't be printed in several big pieces, (North Wall, South Wall, East Wall etc.) or even in the form of thousands of individual bricks which are precisely placed.

The only downside for the modular approach is that it leaves more joins and increases the possibility that energy losing gaps would appear when it was built.
David Bovill on May 31, 2013
We were discussing this last week! I have a friend who is coming to the UK from Italy to work and wants to buy a small house. I showed him the wiki house project at the Westminster Hub where I work - and discussed sustainable issues. He actually wants to buy some land a print a small house designed to maximise energy use exactly along the lines you say!

Great idea - got my vote :)
Will Ray (Judge) on Jun 02, 2013
One of the great innnovation ideas in construction to "grow" a building. The trick is what material to use for printing? You are going to use a lot of it, because houses are quite big, so it need to be cheap. The unitary material has to fufill the different properties that the layers of conventional construction require i.e. weatherproof on the outside (or at least able to clad), structural stability, fireproof, thermally insulative (to quite a high level these days, think U-value approx 0.1-0.14 in Europe), decoratable on the inside with spaces for cables, pipes etc. Also it can't be too thick because internal floor area is the key asset. I'll let everyone see what they think might be good materials to try. Concrete and polymers could both be starting points, but I'd think more about blends of materials.
Ralph Evins on Jun 03, 2013
Regarding shape/form, there are huge gains available. Tools that aid architects in achieving these gains are certainly needed (though there are opposing demands of ease of use / speed and detail in thermal analysis). These tools / design approaches can be combined with novel fabrication techniques to get better building forms. This is already happening for large, iconic buildings, where large design teams of architects, engineers and contractors work together (and charge very high fees).
For the mass domestic market, perhaps there are too many other factors (e.g. plot space, need for useful internal space) for very curvy shapes? This turns it into a highly constrained optimisation problem, which is much harder to solve.
Regarding the variation in internal structure, it would be interesting to see how much difference this makes. The parameters you could control are insulation and thermal mass. Both should ideally be as high as possible. You could trade off between them, but having high thermal mass with low insulation doesn't help much since the heat stored is lost externally. In terms of spatial distribution, insulation needs to be evenly distributed around the envelope to avoid cold-bridges (weak links in the chain). You could have slightly less in floors and more in roofs, but the change is small. For thermal mass, you're right having it intercept solar gains works well, but even if the solar energy is intercepted by a low-mass surface it will be transferred to the zone air and so to all other surfaces. Also in general solar from external windows hits floors and internal walls: this means you can have all external walls highly insulated and all interal walls and floors high thermal mass. The architectural idea that aims to maximise this is the Trombe wall [1].

Robert Trezona (Judge) on Jun 03, 2013
This is a very exciting idea, though I think you do need to consider how it could meet the criterion of being deployed in some manner within a year.
However, in the meantime, following on from Will's comment, you might want to look at recent developments in 3-D printing of powders that would allow low-cost, powdered building materials to be used. An example is this patent from the University of Southampton
This allows precise dosing of dry powders from a nozzle that can raster over the part being constructed. Various different powders can be combined, so composition can be varied. If this was being used to construct structural elements of the home then a separate firing / sintering process will probably be required, so it might make sense to pursue Juan's idea of off-site pre-fabrication
Christoph Buchner on Jun 03, 2013
You could also add phase change materials into the mix, which have a high latent heat of fusion, and thus can store large amounts of heat.
Cf. also
Nick Goddard (Judge) on Jun 10, 2013
When I first read the title of this post I thought “pretty crazy” and then when I read how well argued the case was I thought “pretty neat”. I would separate two different ideas underlying what you posted. The first is the use of graded structures or ‘functional gradient materials’ for house structural elements. This could have a lot of potential, even if the graded components (panels, beams, tiles etc) were of a standard shape and made (perhaps by 3-D printing) in a factory for assembly by more of less conventional building techniques. The second idea is to build a tailored house by some sort of computer controlled assembly working from a CAD drawing of the house. Again, this could go via an interim stage where each piece was made in a factory and then numbered and assembled on site like a huge jigsaw rather than a building made from undifferentiated standard blocks. I think this just restates what several of the other commentators have already said.
Matthew Reinert on Jun 11, 2013
Wow, the earlier comment about modular construction has had me kicking myself for not thinking of it earlier.

I didn't see it earlier, but you're right about it being essentially two ideas; using a computer to plan according to the sun, and then using strategically graded materials for better efficiency.

The software I was thinking about is one that is used by Legoland when they want to design large life size models. They put in a 3d model of the object (a dinosaur, an X wing etc.) and then the computer prints out detailed instructions of what legos to put where. Legos are not 3d printed, but conceivably something like this software could be used which would have a library of graded materials manufactured by whatever method, a CAD drawing of the house, and the location of the sun and create detailed instructions about how to put these materials together.
Robert Ireland on Jun 17, 2013
Maybe it was in a TED talk I saw recently, but 3D printing of shelters and small houses is completely possible, today. Great ideas for passive thermal though.
Matthew Reinert on Jul 31, 2013

So I did some more work on this using Sketchup and I hope I can use this to explain the idea more and show that it's very close to doable with existing materials/techniques/software.  

Take for example this "house" I made in Sketchup.

One of the cool things that Sketchup already allows me to do is put in my location, which in my case is Taipei, Taiwan. 

Ok, now the program knows exactly where my house is and it can thus give me accurate shadows.


This is how the shading would be on the Winter Solstice at 11 AM. 

The program can show which parts of the house receive sunlight at any time on any day of the year.  Now, my idea would be to get another program which could then calculate which parts of the house are getting the most sunlight and when for the entire year and then alter the density accordingly.  Taking a cue from what people said in this thread about calculating it modularly, I numbered the individual bricks from 1 - 18 on each side and took snapshots every hour on the solstices and compared when different parts got sunlight.

This process took a long time and was very inefficient for me to do by hand, but a computer could do it for all 365 days of the year for every minute that the sun was up, as well as include things like windows, the likely average ambient temperature for that day, likelihood of clouds, etc.   


Now, let's say I had several of these houses on a curvy street, like this. 

It's apparent that corresponding bricks in each house gets different amounts of sunlight at different times and it would be a nightmare to optimize each house individually by hand. However, in this simplified example, a 3d printer attached to a program to calculate the sun's position and the optimal internal structure, could print out 4 bags of 72 bricks each.  Each component would look identical from the outside and be numbered.  It then could be assembled without any specialist training in eco home construction.  Though it's built in straight lines and right angles, the 3d printer would vary the internal structure of the each block in such a way that a density map of of the building after it's finished would show something like this.  

There would curved density gradients to match the path of the sun at different times of year.  The only thing that someone would need to know how to build it is putting brick As1 next to As2 etc. I did this with 72 bricks because it was simpler to work with, but there's no reason why something similar couldn't be done with thousands of bricks using existing 3d printers printing with concrete, wood fibers, plastic etc.      

I want to say a big thank you to everyone who took an interest in this idea!    

Robert Trezona (Judge) on Aug 21, 2013

The new images below look great and add significantly to the idea.  There are two additional areas that could be explored to develop it further.  

The first is to respond to Ralph's post on June 3. I think he makes a good point that you might always want to be going for maximum  insulation and thermal mass. If this is the case, then it is difficult to see the value of a variable internal structure.

The other question is how one could cost-effectively "print" using low-cost building materials (as per the technical question posed above).