Topic: Thin glass composites with 3D printed polymer cores
Speaker: Christian Louter & Daniel Pfarr
Oganization: TU Delft & TU Dresden
Thin glass is mainly used for displays on electronic devices, but it also offers interesting characteristics for architectural applications. Due to its high strength and small thickness the glass can easily be bent in architecturally appealing curvatures, while the small glass thickness (≤ 2 mm) offers a significant weight reduction compared to traditional window glazing.
Research at TU Delft and TU Dresden focuses on exploiting these beneficial characteristics for the creation of lightweight composite façade panels. More specifically, composite panels are developed that consist of thin glass outer facings which are adhesively bonded to an inner stiffening 3D-printed open-cell polymer core. Besides the benefits of high strength, high stiffness and low weight, the composite panels also offer the potential to influence daylight entry through customisation of the 3D-printed core pattern.
The presentation provides an overview of the research, from the initial concept until the latest research findings.
Led by scientists of the Technical University of Darmstadt, BE-AM | Built Environment Additive Manufacturing connects researchers and users interested in 3D printing in construction. The platform spans architecture, civil engineering and manufacturing automation.
Annual highlights are the BE-AM | Exhibition and the BE-AM | Symposium at Formnext in Frankfurt/Main. This year’s exhibitors and speakers will showcase and discuss work from 9 companies in the industry, 10 universities and 2 research institutions. The event is hosted by the Formnext fair in Frankfurt, and organized by Prof. Ulrich Knaack, Prof. Oliver Tessmann, Dr.-Ing. Philipp Rosendahl, Dr. Nadja Gaudillière-Jami, Christin Gandyra, Emanuel Nowak and Malcolm Unger from the Technical University of Darmstadt.
Contact: be-am@dg.tu-darmstadt.de
BE-AM: https://be-am.de
DDU: https://www.dg.architektur.tu-darmsta…
ISM+D: https://www.ismd.tu-darmstadt.de
Follow us on instagram @think.beam to stay up to date and don’t miss information about the upcoming events.
Okay good afternoon uh my name is Christian Lao professor of structural design and building engineering at T Del and I’m happy to be here together with my colleague Daniel farar who is a researcher at The Institute of Building Construction at t dden uh and today we want to uh show you
Our research about think glass Composites with a 3D printed uh core pattern uh this is just a small example or a small sample of it U the idea is that we combine very strong thin glass uh to make a very uh lightweight um and strong um composite panel um that we
Intend to use as a facade or perhaps as a roof or even you could think about a floor um system and the benefit is that is that you use much less material than convential glass as you see here in the building today um and that you can also
Optimize or customize your p P that you see in between the glass um just as a short outline we um present a little bit about the concept of it the upscaling uh the performance uh a short Outlook into what will be the next step and then we uh close it
Up uh so first of all the the concept so how did we come to this ID um it started all with uh thinking about how you can use thin glass and then I refer to the glass that you have on your devices your mobile phone or your tablet which is in
The range of .1 to 2 millimet let’s say and it’s very strong because it has been chemically strengthened so 10 times stronger than a regular glass more or less and it’s also very flexible so you can easily uh easily Flex the glass in desired curvatures so this is a piece of thin
Glass and you can see that you can easily bend it um into desired curves so by hand by Cold bending we call it many people think it’s plastic but it’s really uh glass actually um here you can see a small test on 2 mm glass so actually in the
Thicker range I would say uh so we do some column buckling by pressing it downwards and you can see the glass quite bending outwards uh we have a strain gauge on there to measure the The Strain and calculate the stress from that to see how far we can bend it before it
Fractures and then slows down the the video because we arrive at the full fracture uh just one disclaimer it explodes quite uh well explosively because there’s a lot of strain energy in in the glass if I would drop this piece of glass it will fracture in big
Pieces so a bit safer than this uh but now we have so much buildup energy that once it breaks it’s exploding all directions so very strong uh thing glass here you can see a plot of it um we reach um strength vales of well over 400 megapascals and actually in this setup
It all uh it filled from the clamp so we could push it much further but we had to improve the setup um for that and then you can get to 800 or 900 megapascals which is uh let’s say in the range of high strength
Steels um so we had a a series of Master students looking into what can you do with it so first concept was to look into adaptive facad so on the right hand side you see a uh shape memory alloy activated uh thing glass facade uh that
Opens up so by pulling up this bar in the middle you will open up your facade for ventilation when it’s too hot and then when it has cooled down then it closed closes automatically again um and this makes use of the flexibility of the thing glass um but then another other concept
That we would like to talk about today is more to combine the thing glass uh with um a core pattern to stiffen the glass because you can imagine that this uh this vibration is too much for a building it’s causing noises it’s causing too much uh deformation and too
Much uh NOS for building um occupants so we thought about uh combining it with more conventional on the left hand side Clos cell or um honeycom kind of patterns but uh also perhaps with open cell configurations and there the ID come came also to um print those patterns and the first student looked
Into that to make or to print a trust structure that’s what you see here so that you have a pattern that is stiffening the um the thing glass for that and that you create this very lightweight um D glass facade panel uh of course we tested that in
Very small scale really A4 scale um and luckily also it um it worked with our analytical model so it combines very well and then we had next steps in a student looking into uh can you use foro optimization there or destillation there uh to create perhaps a bit more
Structure actually optimized pattern um and at the same time you can imagine that you have now viewing zones in your class that you can look through and you also have zones which are blocking the Sun and provide a bit of shading there uh so he made quite some nice symbols
There um some of these are also passing around in the audience um and he also proposed a nice design of how to implement that in a building or a roof or a facade and then optimizing your your core pattern for structural but also for daylighting purposes that you can really
Regulate how much light uh you can get into your building and the next step was to look into okay um since the glass is easily bendable can we also exploit that to make uh curved uh 3D pinted patterns and then glue the glass uh easily on top of
That and you don’t need any heat or something to to uh to bend the glass so you can easily bend it let’s say by hand um and make this small scale uh um curved composite panels um then for a while I moved to t d for three years as a professor of
Building Construction where I met Danielle first as a master student and later as a researcher PS2 student um and one of the focus there was to really upscale this concept so we looked into how can we make it bigger um and the first idea was to uh well Stitch a
Smaller prints together uh which of course worked well but it was still quite labor intensive uh and Danielle devised a very nice interlocking scheme that you can uh Stitch um such smaller panels together um at the same time of course there’s it’s not shifting yeah uh at the
Same time of course there was the challenge of uh gluing uh so we are gluing the thing glass to the to the pattern uh which was done manually at that time and of course there we also wanted to uh see if we can automate automate that process and be more efficient
There uh so the whole question was about how do we get from a small scale A4 sized panel to a big scale panel and we saw no other option than to apply for um significant Grant or significant fund to to buy a um large robotic 3D printer um
And luckily we got the grant um so here you see a 4×2 M uh print B which can be heated and the robotic um extruder there or the robotic operated extruder there sorry um to be able to to print our uh well facade skill kind of
Patterns and so how do we do it so we can print um different kinds of polymers um either with or without glass fiber reinforcement in there um or um Opa or also bit more translucent and we prefer to print on a a flat piece so here you
Be printing on aluminum which is heated them from the Bottom by the by the heated bet uh but later you also see that we print on glass directly to have a very smooth um surface to be able to glue to and then in Next Step um we grind the
Top surface so the top surfaces M off also robotically uh just to be able to get a perfectly smooth um surface again to be able to glue on it and that gluing is then done in the next step uh by this also robotically operated uh Reservoir let’s say applying the glue with
Specific um pressure and specific speed and this glue is UV curing so it’s only getting rigid let’s say if you apply UV light from the sun or from an artificial light uh so um until that point that you apply the the light you can still maneuver and you can still um take your
Time to apply the ades let’s say another thing is that this adhesive is very thin normally less than 1 mm so you don’t have that much tolerances to compensate or not that much capacity to to compensate for tolerances um that’s why we really want to Mill it down to have a smooth surface
For the gluing and later you will see also Danel will show a bit more about Hot Melt uh adhesive which are a bit more tolerant in terms of sizing and then we come to the assembly so here see the robot still applying the glue and on the right hand side you see
Uh the thing glass being laid down by suction cups already ready to be moved on top of the adhesive there now it’s there and then the next step would be to illuminate uh the adhesive with a UV light so following these four steps of uh additive manufacturing of the core
Then to Mill down the the core to facilitate adhesive bonding and then adesive bonding and the assembly um that’s our process to come to really large scale uh or full skill uh facade uh prototypes that we can then also test in the lab um and there comes in the
Research of Daniel uh who is looking into the design and the performance of uh such panels yes um thank you very much for having me here as well yeah when it comes to Performance we can of course talk about different objectives or different categories uh today I want to
Give a brief introduction into some mechanical um uh properties at first and because we want to achieve a facade element so so uh it the idea is very close to test it in a windload test therefore we have a big facade testing wall at our laboratory in Dron you can
See here already in the right to the right hand side the um the test specimen which is built into the facade test wall on the left uh picture that’s the technical drawing that was already uh only a preliminary test at first uh Dimensions from roundabout almost 2 m by 1 meter in size
And you can see the honeycomb structure which is 14 mm thick it was a PG based um polymer with 20% glass fiber and on the top on each side we glued a still 3 mm thick I would say um glass sheet on it and uh yes the whole test specimen
Was uh supported on each side and we can blow in and suck out air into the cavity behind the facade element and yeah we test the element until 4,000 Pascal which is still quite a lot so if you wouldn’t want to calculate the facade panel you would estimate in Germany as
Far as I know 1 and a half thousand pascals so it is a really uh strong wind load uh on this panel and uh we tried to measure everything with the DIC system with the digital image correlation system so that we get uh some deformation and displacements from the
Panel when it is uh loaded and when we look into evaluation we compared two different systems so in the top uh or in the bottom part of the picture you see the DIC picture so that is the actual um deformation of the panel during the loading and on the top
There’s an fined element simulation which uh represents the physical tests quite well and uh we measured at the maximum uh load of 4,000 pascals uh only 2.3 mm um deformation and just to compare it a little bit to other structures or glass structures that would be the same or roughly the same
Deformation uh which you would get when you have a 12 mm monolithic glass pan so one solid uh glass pane and that already means that we save 50% uh of the glass used and we only have or yeah we only have 30 40% less weight because we are
Adding a little bit of weight during uh the additive manufacturing process uh another example to compare it with is an igu so an insulating glass unit that it’s probably more common like you see here in the facade as well there you have a 10 mm glass sheet outside a 16 mm
Cavity filled with air or the different gas and an 8 mm um glass sheet in the the inner side and that would mean already 70% less glass and 65% less weight and uh in our current project we are investigating okay what what impact can we uh expect um to CO2 emissions or
Other environmental uh aspects and therefore we have a um a test facade or um an example which is a building in Dron that is kind of a conference center or our students have their lectures there and uh only this facade what you can see here in the picture is 300
Square met of glass which means already 30 tons of uh glass and only the production of this glass uh would um would uh make 20 tons of CO2 equivalence and so our idea is basically to not only reduce the mass in the building but also reducing the material from this energy
Intensive um glass and uh so switching from the left side from the traditional igu to a right side to a composite panel with only 2 mm glass on each side and there we can see we can expect a lot of reduction in weight and CO2 emissions yeah so far we have mostly
Looked at the structural performance of the of the panels but also thermal performance is an issue of course uh that we need to have a Clos look into uh on these images you can see some models of the thermal model let’s say and you can see that we are losing Heat at some
Points and that’s uh where you you um expect where you where the core is touching the glass let’s say um so there we um see some thermal uh bridging happening of course uh because normally an igu or insulating glass panel doesn’t have this core inside uh but now with
The blue core that I drafted there you see that there’s a Thal Bridge coming from the left to the right hand side of the glass so bringing your heat or your cold uh to the other side of the panel which we uh would like to prevent of
Course um I cannot point with this computer but um the IDE is then to perhaps to apply a second layer as you also see now with triple igu triple insulating glass uh and make sure that the cores in these two indid individual Chambers doesn’t touch or they don’t
Touch so that you don’t have direct conduction there or simply apply an extra layer with a cavity which is empty um which can still benefit actually from the stiffness of the other panel um but in this case you avoid this thermal Bridge from inside to outside and the
Other option is to um to separate the core let’s say to make two cores to print them on either side of the glass panel um so that you avoid the this core is bridging from the one to the other glass um and another issue that we would
Like to look into bit more later is also daylight control um of course this panel can be optimized for structureal purposes but also for daylight control purposes uh so having more pattern where you would like to block the light incoming light and having a more open
Pattern when you uh would like to have the view or the light in and perhaps also to use the curvature or the angle of the pattern to reflect light inwards or to ceiling As You Wish uh so here you can see some render on the left hand
Side but also you can see that the the shading already indicates a little bit of what you can achieve in this respect yes then uh let’s come back a little bit to 3D printing again um what we are looking right now is actually to go to curved panels uh we have the
Robotic arm we have the thin glass which is very flexible so we want to make use of it we have the potential of computational design to uh to design in a a variety of geometries and so the next idea is to go for uh curv panels and therefore we need um design
Procedures production procedures which can be automatic which can be automatically and so here on the left side you can see uh rendering uh how we could design a panel like that so we are directly um designing the print path which is expected later on with the
Robot and then we are projecting it to some kind of free form which represents the facade and on the right hand you can directly see it doesn’t matter which panel you are choosing the print path is directly generated and we can transport it or we can give it to the robot and it
Will additive manufacture the core structure and there’s also included that the the gluing after the printing or the Milling process if needed and so we want to have a file to factory process uh which is uh almost digitally uh until you come to the to the real printing how
We start with that is that we um yeah just uh M out this kind of formwork so that’s the wooden form uh formwork which is as well comput or parametric designed so we can easily change it and it it’s quite low cost to produce it um so that
Is uh for example the Milling process which is needed then to uh get the shapes and then we uh just simply um put the the wooden blades together when we once we have ch uh changed the shape of them yes and then we have a curved
Surface which we can print on uh you can here see a very thin sheet of glass so this is only is 0.7 mm thick we can heat it from below and then we can directly print on the curved surface and uh we have a curved uh core structure and the
Next step is to glue it somehow and therefore we already heard about the acrylic which is UV hardening but it’s very uh like honey so when you have a curved surface it will flow down to a certain point so we are investigating new methods to glue it uh one promising
Example could be the Hot Melt uh which we are using here so we just mounted the Hot Melt glue gun on the left side to our robot and then we uh took the same way like the last layer of printing and so we are um yeah applying the Hot Melt
That is totally fine when it’s uh when it’s co uh cooled down because then we go to a second lamination Pro process where we stack up the glass with the uh with the core structure and then we put it into the OV into the oven and and
Then we are laminating the core and uh the print together yes and some closing words then again yeah so we come to the end of presentation um yeah still a lot of work to do in the in the next years I think but um well we would like to convey the
Message that we’re working on um strong and stiff yet very lightweight uh facade panels or floor panels or roof panels uh which uh have less embodied energy than conventional facades and you can also optimize the pattern according to your well performances or also do your architectural um intent let’s
Say um this presentation is short you can read more in the beam booklet um or in the meantime you can also uh look at one of our recent Journal papers um published with Springer in the architecture structure and construction material uh Journal sorry and we would
Like to uh thank very much the fund that we have um obtained for for the printer itself on the on the left hand side but also for continuing in a project with several Partners at theen um that is looking more into the real implementation of such um components in
Facades and we would like to thank you very much for your [Applause] attention yeah fantastic thank you very much for the wonderful presentation and the projects do we have any questions from the audience thank you for the presentation uh I he a question uh am I understand
Right that the usage of uh glass with ultraviolet protection is not possible because of glue glue process ah yeah you mean that if you have any uh UV blocking mechanism that would be difficult to use the UV uh curing glue I think to a certain extent that’s that’s valid especially if you
Come to laminated glass uh but at the Hot Melt that Daniel showed in the end doesn’t have this UV uh activation need so that’s heated or that’s cured by heat in the end okay and uh for the second uh possibility to for adhesive uh assembly
So it means last layer is not hot all the time so it will glass will be put on top and then it will be made so you bring heat again uh yes so we uh first we 3D print the core structure then we take off the the 3D printed core and
Then we we can put it back to apply the um the adhesive then we um flip it around to put the adhesive on the other side of the core element and then we have a core structure with two-sided um Hot Melt which is cooled down then we
Can stack the um the glass on it we put it in a vacuum B because we want to achieve the the curvature as well so it is pressed down to the um to the adhesive and then it comes to an oven at 120° for example and then the Hot Melt
Is melting and it’s reconnecting uh and then yeah it’s connecting the core structure with the glass so it’s a second step after the printing welome thank you very much have any further questions sorry thanks for the presentation uh I got a question if you have looked into
The full life cycle because I guess you can’t get around this topic these days anymore and if you have looked into this how to recycle or how to what to do with the product once it’s not needed anymore because glass only I guess you could recycle and then there’s plastic glue on
Yes a very good question that is a big part of our current research project which is now um running for one year and so we still have two year two years left so at first we looked into the a module of the building cycle so only the
Production of the materials and the CO2 emissions during production the next one will be the living um phase so now we have to care about u values and performance in terms of thermal insulation and things like that and uh we go the full circle until the d level
And we want to or we make Concepts how to recycle it the UV um hardening acrylic is very challenging to recycle because that is like yeah the bond is very strong and we can’t really um uh disassemble the the whole product the Hot Melt is a little bit more beneficial
Because we can just heat it up to 90 100° and then we can take it off so that’s the idea um but yes we still working on Concepts how to completely disassemble the the whole product but uh that’s a little bit way to go still