Merken

Master class with Heinrich Jaeger

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
I'm finally Mahiga from University Chicago it's really wanted to
do this master class with all of you tonight many of you are quite familiar with training materials and so I I will tell you something that may or may not be familiar but I'll probably give a slightly different twists and so what I would like to start out with it is tell you a little bit about the England matter by design and would like you to interrupt me at any time just yell out if there's something that's unclear aura or where you have some opinion that needs to be discussed so I wanted to be as active as possible it at this late hour of the day on my idea is to give you and overview and that's really what I wanted to an idea of what I'm thinking right now uh when it comes to this whole issue of designing granular matter so there will be some basic questions that just highlight the field as hazardous and then some new concepts and maybe applications right so the idea is to 1st .period you know tell you what you kind of already know why is this class of materials interesting and then talk a lot about what we mean by design but more generally what we mean by materials designed and maybe more specifically what I think design is all about animal focus that on granular materials and that's kind of dead is the 3rd thing here is really 1 of them the big pieces of of leftists lecture tonight all right back so we introduce ourselves already sold go and introduce yourself you when you must believe and rightly understand text on with and have thank you so there the material where we're talking about but something that is seemingly simple idea that could fit in a little flask like this it could be that he carried around it might look like you see over there a bunch of marbles it is all Our universe in fact is really ubiquitous and it involves materials that you know certainly when the states from from and breakfast materials Dad are important subpoena know for food industry they come in larger scales and I were talking about me the iron ore coal pharmaceutical industry all particles large aggregates of particles and it is in they can can be large and as large as cars and making fuel really stretch the idea of what a granular materials individual entities that gets into hopefully not too much contact in this particular case then you see that there is not limited to our planet so these images in the upper right there and then a senator from Mark right to the same material exists there and certain man-made objects like that we can get into trouble so when I think
about Wrangler matter I would really mean by that in the most abstract form I need an aggregate of individually solid grains of particles that are in close proximity may be in contact the interesting thing about granular material is that whatever happens inside materials is not as important as the interaction between so now whether you really want to ignore the internal degrees of freedom are not that depends on the application right but you you downplay them because in the end it is the interfaces that really matters so for me a granular material as material where behavior is dominated by the interface prior that he also notified Jordan is not a nice regular and not with nice regular or spherical shapes but irregularly and the key thing is dead there is behavior that emerges from many of these entities particles coming together that is qualitatively different from what you might have expected from an individual prior to really come concrete individual party of a solid and you know that it's clearly assault maybe even a crystalline followed by the time he put many together it could behave liquid like that when I was a graduate student a grand within physics was a crunch condensed films individual particles were tiny Nanice tropical islands of metal on the substrate Everything I told you is still true writers granular Anderson said diesel entities are individually there are separated by the gaps you can worry about conduction that's what I did actually where supra conduction 2 but now more recently I granular systems led there physics are the focus has shifted a bit and the emphasis on situations where there is no substance you know the particles are just coming together and then they can move actually did not clenched into place all right now you can ask of course what are these interactions and that there are many different ones I'm listing appeared to the typical once that that we are familiar with rights refuse a macroscopic particles now that's focus on um then there will be the report suggests that penetrate and you can worry about the particular form of that maybe hearts and tied form that doesn't really matter the friction and then there could be something that I would call a geometric friction which means that particles can move past each other because they interlock rights with their not spherical but then this is now also with the field is actually and from from the physics perspective pushing a little bit and where engineering for for obvious reasons of course has been for a while is to look at finer particles with electrostatic or sticking interactions become very important and also you can take many of these ideas about talk about and apply them to situations where you put between the particles not just air of action thing but so a didn't suspension in many ways will behave like a granular material and some of you I know from when he just introduce yourself you know that really well I will not spend much time on that today but you can of course have right to just drive that point home again .period there are really 3 scales yes the skill of the individual particles unless scale he really zoomed out of the ball added and what interested in you know what I think is still an important thing for granular material is that the action happens in this intermediate-scale actually work a few particles coming together in exploring the consequences of having his interface In that sense if this is really the micro scale and this is the macro scale then this is the maze of the intermediate-scale and you know it just like a mezzo-soprano 1 mezzanine of the intermediate flora for 2 analogy this it is d Intermediate regime that determines what happens right but it's not affixed to a particular size scale so well that 10 Michael 100 micron or 5 kilometers or or more that doesn't really matter if issue particles coming together and we assume right now that the details of the single grain level are no and you can answer sovereign was the 1st time somebody worried about ideas In goes back a lot this is 1 of my favorite slides and you immediately of course all this is a book that was published in the year 16 11 can see down here and it a 10 year and reported on by an eminent scientists about the Snowflake you know that's what it's this year will be eminent scientists you have kept ideas the following that you have an object in this case he had units call material for tonight a snowflake and there's something that strikes you struck him and the symmetry of the snow and the idea was that from the macro scale behavior of something you would like to deduce something about key building blocks in that .period drives the inherent she wants no flaking something in and the idea is very simple in this case that course is what Keppler says a very simple idea already can explain much of that and that is indicated here take identical objects and say coins or little pebbles and put them together and you get this particular cemetery you get at it if you go to this arrangement which is the closest packed arrangement then you get 6 nearest neighbors :colon that gives rise to this excitement of the snowflakes in this idea this is if you want to have protection that predates the police effect crystallography by 163 here in it is
of course something that we use a lot Miami's is exactly that idea you take now a little bit more complicated situation it takes 2 sizes he said now collards you put them together and by just changing relative propensities of dollars to buy you get a whole variety of different kind of and arrangements made for only with these 2 sigh this work by colleagues of mine Dimitri can and land Shevchenko and you know if you get that many already with 2 we can imagine that this can get really massively complex and go beyond what you could do actually win 100 elements in the periodic table but there is 1 thing that
is important about which adjusts or the nite is that these are equilibrium of work equilibrium configuration these were special configuration those with those where temperature had time to explore the parameters basic configurations space and give you exactly what you saw there this is the same as saying going back to Keppler the the stacks hears you I'm not interested in any arrangement that you are interested in the very particular arrangement namely the wanna give the dentist possible pack that's this 1 the little side remark it is obvious to everyone who's ever stacked oranges it took over 400 years to prove mathematically right I think it the proof was 1998 the started in and knowledge 2011 I think it was fine except now this is the crystal pack this is if you want a ground state but Of course there are other possibilities of throwing these particles together especially if you don't have temperature available so if you go to large enough particles and here you see a random packing there similar objects right in 3 dimensions than others not automatically falling into crystalline arrangement infected person who started to study the 1st was very knowledgeable and all you see the physicist never who thought of this as a proxy for the amorphous liquid like arrangement you haven't things like liquid um argon expected liquefied gasses now notice 1 thing this is particularly begins with the dancers you can this is much left right it is essentially a configuration that is quenched into 1 of many possible states this is just 1 particular configuration and there are a myriad of different ones and they all have to see him density right they all trapped far from equilibrium and if if if no further motion Kirsten day are trapped in what we would call a nun and got regime ready can explore phase space and not surprisingly if you pushed on this now and you could expect nonlinear response to how how can we the idea of what that means getting trapped in 1 of the state's right and not being able to explore blog configuration space so that you get to the dancers possible pack and there was a nice way of
thinking about that it's like that's imagine somebody says he's amazed and you're supposed to find the accident and you have to really run through somebody says alright now fast right you don't have time to explore everything you just thrown amazing and somebody says not even 10 seconds while you probably get stuck in 1 of the that right that's exactly what happened to this package on the other hand if you have time to explore everything you might find that if you get lucky and you get on top of the the bridge you can survey the lay of the land you find the exit right by these other packing is essentially a clenched no
there are many ways to get you to manic Libyan behavior is amended to classic bonds or the 1 that I just mentioned getting stuck right the quenching you could of course try the system also really hard that's another possibility and the interesting thing is that you might say that in his disordered configuration many of the beautiful properties of a crystal obviously lost and therefore the properties of the material might not be as desirable and obviously that is true With respect to many properties but they also have properties that are rather than enhanced so for example if you look at the glass 1 this this is just image from the web that shows you how you might prepare piece of window glass we could look at a metallic we look at Glacier alloys which began in the sort of metal stable trapped states that they have behavior in many um always outperforms which you would have with with typical of the kinds of materials was flooded he the strength and when things break and this is the elastic limit you know how far you can pull on something before it gets lonely so obviously polymers and I have a high elastic limit as good but in a particularly strong steel is better but something like blast hour-long disorder amorphous material actually performs exceedingly well I'm just to remind him down while
is state that may be a little bit more forcefully um yes if you use order as guiding principle you get perfection homogeneity you get certain predictability of course God but you also lose something so for example here the crystal obviously might be really strong but it could also then with smaller defect suddenly break and clean rights also the robustness can be already P built into material that inherently Is this you know whether you have 1 more defect on 2 more defects doesn't really matter suspensions and in particular granular material fall into that category right and then the issue that we're addressing here it really how do you want to now packed these ingredients to get particular performance I don't know that translates into a question about the structure performance of structure function relationship because he never come back and I'll just be
labeled right macroscopic materials are models of them that's why a physicist likes them right the new the simple image here for an avalanche of seeds that is material at the same time can be solid like as well as liquid like of more complex than the ordinary To material and I will go through some of those there are things that are listed there now just to show you that not just in the static state but also indeed um
um moving state is very interesting physics associate with crying materials and that's maybe something you really know what the movies it's interesting after eluding group here from Trenton this is simulation where you start particles now with minimal velocities and they have some average energy and all they do is they collide and lose energy at every collision that's all they do In remarkably by the mere fact that you got collision with dissipation right to to let's say friction on the elastic collisions you get structure of and you can think of business that went to particles colliding with energy slowed down right now there together a little bit longer and 3rd 1 will also pile up as like the traffic jam growing and local leaders blue regions are these enormous traffic jams but they move around and and and changeover but eventually everything "quotation mark coalesced there right let me
just Ellis III back additionally none experiment he has something dead is obviously the dusty fine powder and all we did is we dropped the uh 1 inch to here into it all right and let me do that again he is the material he sees splash like you'd expect for a dusting of coming up of course but then is this column made jet coming from so again is the fact that he's a disappeared of systems that let stuff I still cling to the knuckling together but let rather prevents stuff from from spreading out right and focuses that momentum sunlight and has beautiful work done and went on exactly
the same thing infected did that before we started and you see some photos here after that that Gerrard that comes out this summer From from videos that were taken little snapshot so the liquid state itself also has a rather interesting properties gas and liquid like state and don't know you had a question of who would like yeah but that doesn't establish simulation what would happen so if you don't fit any more energy and and eventually to the last yet solid nowadays don't stick together but day-to-day slow down because of collections and there's also of course if you several particles with that people guys in between you get a rattling events that don't lose energy very efficient I mean it's almost catastrophic and but many people were in the world yes so doubt it that is the idea that in these macroscopic states due the internal degrees of freedom the phone on most of the heat essentially takes have all the right solves all of course I mean we're not saying the know the energy has to go someplace that's that's sure but we not worrying about it say the band structure of the individual particles typically right we do not worry about their energies Specter would just assume that the warm that's a that's a very good question fell just to make sure right we we use their macroscopic nature of these part of essentially as an energy sink now as to gold that for a
2nd to the static situation while clauses that situation imagine you very very slowly push particles together and you start on the left here and you end up with particle beam squeezed together on the right this is a time and taken year pictures that come from a review from Morton from hackers as you sit down there and did you can imagine that at some point as you push them together just in 2 dimensions shown here for clarity they will start to right and this is a very special .period forward essentially the material that was totally loose on becomes mechanically richer In answers .period Figes packing fraction how much volume is occupied by the particles divided by the total volume as we saw previously in 3 dimensions you you can't really go above roughly 64 per cent for equal size particles until you start squeezing the particles together right in the form of course you can go higher and so on Justice Ruth roughly as that level here now if you squeeze them a little bit more or if you take 1 of the particles in the interior of 1 of these systems and regular you see that forces are then transmitted through contacts in neighboring particles but what I'm showing you hear use veins are essentially the surpassed worked at the fabric of Faustian that is highly headed In fact if you push 1 where pushed other way of blue in red they look statistically the same but it clearly on identity so this heterogeneity it is automatically that's also true if you share so you see is public confidence from the back this dark stuff on Blue Arrow's they're supposed to tell you which way the body of respondents you put you on the box on the outside of Soria compression here sharing the Red powers tell you what happens on the outside and then you see highly hemorrhaging use most and interior is a hallmark of his granular materials and you can't do experiments and a group that is that most known for that is but just group at Duke using for last tests and across polarizes and you can visualize forced changes the computer generated images in there it's a natural experiment so this particular point where prices are just coming together and the whole thing becomes rigid this jamming .period has been called the epitome of the source In a sense is this is not just a lightly disorder visitors gonna the most is ordered situation particular also has faster forced transcriptions concern now the challenge has been found for maybe you 10 or 15 years to find a conceptual framework to explain In and out of this gainers idea of a jamming faced Indian put Noble we can talk about it more I don't want to be labeled right now there is a framework that has been established and many people all over the world here as well as other places have been working on hand In the idea basically as stud dude I have to regimes and jammed an agenda regime this is the point Vijay of course if you jammed I know you sharing this system you can also on German you can push yourself out of a traffic jam but as you can see right there in this has led to a number of outcomes that debt I'm listing right here particularly in connection to the the glass transition that's that's very exciting but what I want to do is just 1 extract 4 points about jamming for you here and then applied right so the 1st thing is that this is the onset of rigidity usually cooperative ship and did jammed state is not a crystal and we saw that we said it's really because of the geometry of the particles as geometrical constraints it that there's no size killing them the principle In and the other thing is very reversible can jump in and you know something that and we keep doing and so pirated the talk here tonight will be you know can we use the right so by writes so well the outline again so I gave you hang on just a 2nd it a little bit idea of why I find that interesting widely meeting should find an interesting this time of material and now we want suspicion that someone there was a question of whether or I'm here and I don't think of it as part of during the day of the you'd certainly can't talk about that she wants the framework that has been developed and you know wrestling skills associate with it with jamming in particular there are certain correlation length and asked jamming at at that point um these correlation length blow up so basically if you're interested that phase transition itself and monitor what happens exactly at that point you need to go to the infinite system size In order to capture so what do we mean by materials design this goal Back to what I talked about before that I mean I think it is a relationship between the properties of the material as a whole and the ingredients president from many if you might .period as simple as calling structural properties with I put 2 errors and are errors in both ways no if you remember what Kaplan did he started with the object itself the Snowflake it decided that they're gonna be something structurally and average looks like so if I but that is no abstract that will bid he starts with the assembled given up there was there was given any sign something out about the building now I don't know would you do every day but what I do and what my colleagues do is rather different we usually start with the given building blocks In assemble object united by synthesizing molecules here I can do stuff like that Indiana some wonderful color a new film that I make that way this was what this was right exactly or you can go to more complex
objects soldiers has said that that capitulation beautiful water out of the trade and comes from the factory and if you look at this process a little bit more meaning engineering perspective you say what's not just structure functions just process there instead he nearly fell out of bed you start with you rob staff you run it through some process whatever that is for example you just heated up to some temperature cooler down to some that gives rise to local structure and from that function in March In usually that is exactly the way the arrow is running and that's what usually I would imagine you do in the lower or in the office on your computer and so that's how an experiment run maybe virtually stimuli now maybe you get something maybe you just start here grinded through and come up with the function that's OK maybe you are interested in the structure you don't care what the OK but you run in that direction know what I'm saying is that true design goes the other way if you ever asked a designer and where they start from the start from the goal they don't start from the ingredients so what that means is for finding fizzes like me when I have to talk about designer matches in England in the World Cup the problem that it is counter to what I usually do and that makes it hard so some people talk about inverse design I think but that doesn't make sense to design is this way for so that's can act in a natural gonna talk about how you that right and so on knowledge but it's too bright and unless you have some questions but that I wanted to get across is kind of the big thing design is going that way you start with what you want to make it happen question all good so let's let's start with some opportunities that the start with something that may be you would like to design or are you see something that you might like to optimize all right let that stop just play around a little bit get excited and then we'll talk about how to actually maybe do some current but the robot come in all types of shapes and had just picked 3 and many many different In a very different already here but they have 1 thing in common they all made out of hot pots all right and they all have identifiable parts and is clearly is a wheel it is clearly is a leg and it's always going to be a real and always kind so now as contrast that
with this object right here in a few concerned about it also moved life is on a smaller scale but if even and the final biologist hard to tell where the front and back again you know that it often can you identify the wheel there was awaiting word on it will moving create stuff as it needs right voted Celia little sack filled with some .period so let's take that idea little step further imagine this is 1 of now move the market take a flexible sacked remembering Newfoundland part you can put some liquid and never does not necessary imagine you put so many particles in that they can move past each at that point they are collectively Jan or you squeeze the bag tighter and tighter so that they can move past each other at that point yet collectively jammed at that point the whole thing is which can you use this transition between rigid and then maybe let some room in again uh liquid like behavior to make a robot in the answer of course it's yes I when show it to otherwise so we got involved in a few years ago company called I Robot and of course you know that's what I think this would look like that at all but here's the idea to take a flexible and closures triangular-shaped silicone balloon-like saying any fill it with Graham tell you pulled they're out and then let them back in will have wide it can so the engineer their animals acre did a great job during this will stop motion animation and to show what jamming is all about and I love it and I think you might enjoy 2 super-Gs when I talked through it there's also some sound if you pick it up on the way right so this is the idea that you can make something that is hard as the break in western please allowing these particles not to move past each other and use of force change their right to and then there's humility system is on GM and it can move it's malleable so let's say we talk now a grapefruit-sized soccer ball With a double How old skin is he is 1 of the other 1 mn we turn this thing into various sections In each section could now be champion and individual To right now they all jammed let's say if you expanded this volume you much would happen because the but if you and Jim Davis set of 2 sectors right there and now you expand the volume you could essentially gold that wasn't the you could shift the center of gravity and to use the very 1st start robot by that I mean obviously assistant little grapefruit likes him but um and you see it still tethered because these
pants and all that they weren't part of it yet but the idea is that right and
as but 2 computer control
I do want thing can do is control they can crunch and of the 1st gymnastics so that was just the beginning of the end it is started a very interesting direction which was to use jamming for soft robotics that that hadn't been done before so let me show you another idea the next
idea and in particular imagine now you with a have 1 linear actresses green the actuators essentially something that can move forward about my arm right now with forward backward at factory but knowledgeable supported jamming cell Over here and 1 underneath it right of 1 of them is jammed let's say this 1 and I pull this thing is it can do for much Pollack and was part if you put rounded you can turn 1 degree of emotion 1 degree of freedom forward backward into more degrees of freedom depending on which it is German solitary activities you can read that they move left right up down and of course you can over also in out right so with that you can make a leg and let's see how that works so this is not a robot has 6 of these legs In the income all right so that's pretty nice is 1 that's totally soft still tethered but but even the body announced off except for that 1 camera on top it and now it's being jammed in noticed 1 thing by the way that this January right at the transition involves a very very small change in volume typically need less than about a cent only since it's not really a perceptible and you can do all sorts of little tricks with so I I don't wanna go through that that movies through essentially waving at you various fashion I want to show you some other things yes no only it was just something that I was going as for the question is whether and there's a lot of energy can consume right then and there or whether there he generated right so I think you have to question both it's good that you that you pointed out on a couple of things the 1st of all the heat demand a few days is negligible that was that but there were some of companies the picture is that's so all you have to offer the over 1st of all I want you all to create a new study OK so when we find that the question is what is the energy balance right there then we can talk about that I actually would throw in a few other considerations in the equation right so you can ask OK that is is heeded your major concern you can throw in another thing which is it is the difference which actuation activation it's not a simple natural policies are exceedingly worried about and may be 1 way of life showing that is also right here each degree of freedom say with that parent which is a masterpiece chemical engineering needs them to begin actuator often not always but often Asamoah per degree of freedom OK which is hard to control and and also costly I so there 1 way of working an imbalance of considerations has to say I would like to and have as many degrees of freedom have activated with as few actuators that that that cost energy is part of few motorists and different example in this particular case you notice that every time the 2 like moved there some element headed to work In their vitamin a jamming transition by itself doesn't work so there's obviously some energy costs for that but the number of degrees of freedom that activated right and actuated is such that you can get more degrees of freedom activated clearer natural energy costing factory is something that I want to drive home with this 1 all 70 cities are fabulous machines in announcing that we won outperform any of us lot of good and maybe get away with a simple 1 maybe you want really hold that ping-pong ball just so you definitely need something like this but the design principle often has been to start their anthropomorphic gleaming gold with what we but evolution has that right maybe take a opposable thumb and 2 fingers something like that if the task is to secure and pick up familiar objects there's no question that you can specifically make something work actually fantastically and every if the is to pick up unfamiliar objects that you don't know a priori then a lot of issues come you indeed many degrees of freedom to grasp objects that you don't know what they look like you need some compliance knows a lot of sensory feedback to crush something and you need a break a computer that deals with all of that and then make sure that use fingers and hands out processes rights are now years the idea supposed but I turned off all the lights in this room and asks you to pick up whatever you have in front the 10 or the the bottom line could you do it in the hands of Corsica and way you do it as not like I pick up this thing right now which is bye hand feedback I come here I open this up on the sale wanted grip described a gage how close I am open this up I get my thumb today a lot of processing but rather with the lights off I can use ceasing all I do is I kind of feel like as open wide notice that just good down and grass Hi now that has the advantage I can do it in dark it also the grandest at all I need to stress close so let me show you know about a proper dead works by this
principle so this is at the end of a robotic arm it's assessed the bear little balloon filled with grain material and and as 1 degree of freedom which is just pull something president insisted more rock hard hit the that the air back in and it's soft right now what but the material doesn't essentially computes exactly how it has to deformed right by itself without any
processor so you can run over the shock absorber like that it figures out the shape very closely picks it up right and an open look without any programming nothing except if to know where the object is you stick it over and technology only have to go all the way around when you pick up a raw egg and just to make sure that wasn't out of wood or something like that this was really a rod the mandate conforms to all sorts of possibly brittle object gasoline and electricity in fact all it needs is 1 line of air right in Novell so I know it is based on his idea of jamming right now this is the 1st generation and when you this air back in there is 1 GMT can pull it off the next-generation actively put some areas that just a 2nd it can go bowling with the so now you can push out the object and you don't need any rest movement or normal body down would have a role like you were just kicks it out and has pretty good accuracy with that the nice thing is that these are totally passive system riders no active feedback we can put that but a simple form doesn't have any no alright as wasn't particularly far away from the goal I understand but if you look at that it gets said 3 guys 350 starts in there pretty good I think tomorrow show you another 1 where it's over longer distances so the nice thing is out here all the a part that usually would involve a lot of processing is done by the material itself as it comes in contact right it's adaptive and conform but noticed that the ingredients in there that we haven't even talked about
it right away olive involves a jamming transaction there is something nice that takes the same idea but now would you want pick up essentially is a limit so you moved that thing around 11 and now perhaps in that there is product project that comes out of humanity and the issue that's being addressed is that 1 of the biggest biggest problems with the opera's deceased is 50 especially if you're dealing with situations where are you can go to a hospital and have the thing custom fit which she would have to 4 properties or if it's for children integral just think automatically but that was the idea and that's something they're pushing with particular project has another idea that's being worked on out of Stanford use especially for making their tactile displays right there where if you can moved duties compartments up and down but also you can change a stiffness a gives you 1 more we have reading something get a question here the last part of the question the services of all this things it ended this year with the most beautiful Black said that as that the users of the of the of of the work of art was present at the meeting of the not quite so the the question is does it depend on a on a surface and they have to undergo the material itself and that is true of course because it works and then we didn't tell you exactly this he has 3 some effects said it uses potentially and linear combination depending on what object pitting picking up 1 it's simple friction so as I said this about the depending on how close you start out near the jamming transition . 5 to 1 per cent in volume change involved to that music very slightly grips the option not hard but just a little bit that's enough to exude friction so fractional gripping happens right To the extent that does for example Robert the lunar surface that is not totally frictionless that works the 2nd thing is section so this was a "quotation mark rejected 1 of the things right yes I was reading really think now sold the same as if you had um temperature that suppose you took a piece of metal you drill a hole in the new cool it down Will the whole expand or shrink OK so that's the same idea of game a lot of things you around I hope that you have to worry for 2nd about friction and in a certain angle dependence you know it's 45 degrees with respect to that does that's a minimum angle I mean you can't just put it at the apex but you do not have to go around so that is due of course depends on the friction coefficient right now nearly hit if you have something that is not chorus then you get additional suction effect and finally In his account of the advantage of this thing the more complex the surfaces the easier it is of course want to pick right which is exactly where other grippers would have the difficulty now this is not gonna outperform or make obsolete any Gripper that's optimized for picking up something very specific but this has advantages where you're confronted with situations we don't mockery or if it's brittle or not but whether it complex or not right we need something can shove against now Ituri object pick it up and for example that happens a lot in the automotive and industrial now for for picking apart and and and by the way is now a small company that got founded by 1 of our co-workers there from Cornell and discolored empire robotics and duty make this thing so you know right now we just use the fact that there is a gem transition but we don't really designing the right uh that let's step back for 1 2nd and their asked more generally mean so why would you want to design these things and and more generally why would you design anything with training materials and 1 of the reasons is that granular materials are of course Houston all Over industry agriculture like assured 1 of the 1st slides pharmaceutical the companies and if you think about outfits that package for that handled these types of materials you can compare their performance with 1 the comparative outfits I would do if they dealt with liquids on that by itself these are companies that are a little more production lines that have built for certain output let's say performance of 100 per cent and in the case of a liquid bottling plant for example the thing opens on day 1 of the 1st month it comes back with roughly 85 person of the performance was planned for the interesting thing is if you think about the solid processing plants the performances come so that means we we just had a really hard time understanding How these materials behave when they flowing and that has to do with these issues that we talked about instructions and mountaineering right there's a high payoff if we could make that behaved little bit better so what are we talking about food as bed reactors and that's where most of the time university of plastic and in all of these products come from and certain types of power plants of the summer particularly common but today did do exist but where the nuclear material is part of a temple that gets fed through a final like this and then he just taken out that way
here maybe in Holland you worry about the particles that individually weigh about 25 to 30 times from wave energy dissipation and you want to fall into place and don't move that much even if they get agitated with with lots of ways a little bit smaller you'd like them to take of energy in terms of these packing peanuts maybe and then you go into smaller knew that some water and you get to suspensions that say of colloidal particles size range 10 migrants also and they are particularly good in terms of stab resistance when you when you combine with with and then there's some advanced techniques that came out just the last 3 years the really rely on the granular material metal 3 printing for example layers of grains get packed and infused with the laser In the packing properties obviously determined but if you assistant the behavior of the material as a whole right an issue that look at the
dynamic response I hinted that already with with the stabbing and stab resistance impact and here you see a two-dimensional experiment again by Pub interest group where the an object of this top is pushed into the layer photo lasted beads and distressed passes and his wife airlines right into you see this lightning like parent get most around in space that means that the material is presented genius right and stays 100 years and transmits the stresses in a way that it adapts to a slight shift in the part of that automatically computers and obviously centers already amorphous it doesn't matter if you have additional defects and on and you know you can slow that down more you see really now holiday individual stressed pulses burst really propagate through this that has to do of course with elasticity of the particles OK media skipped this next 1 year you have to talk about it looked suspensions let me
moved Back to static systems just tell you that there is independent movement within architecture certain parts of architecture work people have been exploring whether you could build large-scale structures maybe even structures that you could walk through like here without using any fasteners cities by by just poor In the idea that this would be good for rapid deployment it would be like Wade configurable maybe an also each and every 1 of these parts would be really but you could potentially take it on .period and that the lightweight aspect has
been used in a few years ago Delft now pretty impressive right how all this bridge tolls these people they used the Crestline arrangement right and what we are basically saying right now as you don't have to that's 1 option but it boy took a lot of work To get that just so under jammed amorphous system what want this all right so we have to go
back to their immediate what I proposes that we take a little break and there some then bring talk a little bit more about the challenges now actually doing that for granular materials and how we can get around all right so I think that what what will do is I have some more stuff keep interrupting me keep asking questions and then at the end hopefully did there will be some time left also for some more discussion all right OK so it back to our major theme here that we really want to think of designed as a starting from a target and then trying to figure out what the individual building blocks and ingredients are to be and now the interesting point for tonight is 1 of the challenges that you find in doing that when you're dealing with granular systems right and then of course we have looked at many of those before but yeah list them 1 more time to the typically situations that are not Crosland far from equilibrium but many jammed In this is sort that today at what you do at the outset of the material really will make a difference when with the jamming for example the simplest thing he you can remember of course that a pack of vacuum-packed espresso coffee right the reason it is hard as a break there's not that something about the material suddenly got changed but rather that you can find it on the outside he left that confined Resnick open a quality level but they're into the bag expense just a little bit nothing changed to the individual particles In the flow right at the boundary conditions and the preparation history how much you tap something how much you knew it what you do to it that matters and then this is something else that we haven't really talked about and that is I think 1 of the you know interesting new directions to to worry about most of the work it has been done for year we were in terms of thinking about terms of modeling it and you know then variants of that so maybe we now I can think about Polly he drove but did not particles often more complex In the last thing is something that was that it is often forgotten so let's suppose you do some design process he do exactly what I indicated you go from the target to Europe ingredients and you find they ought to be some very complicated little thing while you make it and certainly 1 the 1st 6 level only very recently Have there been ways of making couple arbitrarily shaped particles quickly I did you know 10 years ago you couldn't just do that said so far so let's look at some very simple problem just to get a story just look at a situation where you poor some materials into a container lots make that container national but more interesting it's given to heart walls top and bottom but some of the nearly let it be an elastic sleek so what I would like to we have is a situation where the pressure on the outside of the sting of the pressure across the membrane gift rights so that's a confining pressure not just hot walls on all sides it is she now applies confining pressure identically everywhere right but then squeeze in addition from the top tier of some union Axel compression it can ask yourself Hollister material not responding so essentially the ideas you forward the coffee grounds into bed aluminum container of tinfoil you pull the vacuum now you got a break now you push on the way and what you get is a stress strain relations like this ecology Dettori just because we're starting essentially from isotropic confides to the right and typically essentially for all granular materials on nonpareil he's still just poured in like appetites you get a curved that looks roughly like this In other words it comes up and then it turns and that really means the behavior is such that the Michigan's weaker right this slope decrease now maybe it turns flat almost flat it could even come back down a bit and at that point clearly Matero has failed I did a lot of strain a lot of motion for essentially no additional stress so that I can ask questions such as What is the stiffness of theirs mitterrand honest if this means this lawbreaking what does it depend on you can also ask I should mention packing strength maybe for the strength we we we take some failing criteria now you can pick 1 for example which runs nonlinear or maybe the plateau stress but it's clear at some point in Mitchell feels that's that's a strength so you can ask how does that relate to what you know what ingredients and this is obviously a multidimensional Our parameter space I'm just showing too and I'm asking a very simple question to take that the stiffness of modular the slope at the beginning and answers of I'd change the particles in the what would I expect right so for example you can already imagine shapes that are convex don't have any on sticking out of date might pack more densely right and shapes that have lots of orange sticking out did pack less densely now what is denser parking shape give you a stronger response would that be stiff right or uh well but these guys they can interlock like years maybe they are particularly or maybe there is some optimum somewhere in between and again this this hat could have many parameters this case basis could in principle be very bright complex right so let's play without it's what I wanna do is I would wanna tell you really a story that goes along the lines of what we did not have I mean the state of the art right now it is that's you start with the particles and you explore what various ships would actually do so I'll show you that we did that but in the end I wanna go beyond that and 1 really find the ship's right so was still just doing what everybody else does and so 1 thing that is little nifty is to say OK what if we go to shapes that are not just a single sphere but they consist of connected to the so-called now 1 thing that's really interesting is already depending on the length of his change flexible the packing density very sold very very simply by putting in more and more and more chains you can dial in patent it's so if you want a particular packing density you can read about how long you change should be impeached couple solves adjacent randomly packed right and also this is the packing density is given after we have captives thing out you let it settle 10 thousand tap that's just finest you don't start with particles he disclosed to 64 per cent as we showed early on and then the longer you make these things the loosen the package keep in mind Moussa packing doesn't mean notices worst packing right we we should we should think about that now we can go in there to X-rays we went to the synchrotron you can find everything goes here you can find out which 1 is connected to which you can look at the statistics and if you know you go to town and that's what we did should pull out every little bit of statistic you can imagine and we don't have to be labor that this is the so-called pair college unchecked tells you if I had industry over here where is the next right so this is a serious Annex 1 that's the distance here and it is essentially is the diameter of 2 but pigs could also Split riots Owens random packing so you have been to a distances when a square with 3 other loans to aid you can ask at what happens when you have rings and they can ask about angles all of these things and you can try to go from structured to function of in this case that means to go to stiffness right and it's devilishly hard it is not clear how to do that how this knowledge translates known to knowledge about what it would do now no 1 1 thing that we
did it's obviously going and identify where individual particles it would be so this is not a computer simulations real data you're just pulled out of a couple thousand chains these for 5 days and then you can ask well when teams come together in contrast to particles they can wrap around each other so does that help In the answer is that it will definitely
does work by an undergrad imports but Brown then you can define when something might be coming together in a way that that they can entangle so they have to cut through the so-called manifold right there it turns out he got enough of these clusters sorry is Chrysler polymers his chains tender sets said the entangle chain span all the way across the sample you get different behavior right and so that happens here the win and you have about change that at least that's in this case 5 6 7 1 8 0 9 and that means for this particular chains actually and because the beads they are space not far apart but close they can bend only so far and about 8 or 9 you may correct writes another was once you can make rings once you can wrap around things that's when the behavior change that's wonderful and you can now look at the stress-strain curves and took 2 different data sets 1 for uh a flexible a set of change that is little bit has a different rates of courage and the other but what you sees in both cases the following there's a set of course right here at the bottom where very little James untangle these are shortchanged they they didn't do not able to really wrapped once you make and longer then they start to wrap and an essentially Everything is intact the whole thing 1 big entanglements and what happens this year that behavior is rather unusual I just told you that tropical materials always strain we can know these materials used stress here's the strange the strange stuff very very unusual material that's a hard push it too hard pushes back but polymer stood at real poles and described polymers to that too so this rather unusual and remember just to make sure we're not putting something into a cup with hot waltz fried the walls can move constant pressure this will be the setting up of a piece of the real writing brings the mean restraints on so the question is is there a characteristic length scale on number of beads were you switch from softening to stress different and debt so here yet so hard to to see because on this scale you can see the usual softening really I mean that this kind of buried in here and that is what we call this this correlation length to switch over to strain stiffening once they the polymers good long enough that they conform ring it's pretty much exactly at the point where taken from rings that you see this behavior by itself in this but for this particular type of changes happens between this this between the yellow 1 and green 9 people for this particular that that's stenciling still associate with international that here I'm
so for example and to to be precise that's just look at
a strain level out units at 20 per cent of the end just look at this level right here as a function flanks and you notice that this gap here right in this another gap humanizes pronounced In I'm plodding now what happened is a function length In a single month so that there's a particular number where were things cross over that is that links now you might have all at that point may be the packing is just dancer and that's why it's different right so you look at that and you say OK look here around 10 and here it is again a number of chain the and a change and years the pack a fraction is save anything pack fraction went down right so interesting thing the shortest strains stiffening but they actually could loose so here is now you're you're answer to 1 of these questions wedges that you can get this situation were actually I didn't lose I mean the good chains that that are not particularly dense give you if you want the better performance strange different performance what you see here is that there you know that packing fractions that lowered so let's say below he actually this dress is particularly high so low packet for rather opposite to what you normally would have expected there's a lot of scatter of course and then use of statistical considerations you would agree with me I think that In this range but up to here and that if you look at the color that means up to about this correlation length of 10 the packing density as is but beyond that coalition life is lower and their distress is high so and knowledge switch over to particle shape of individual particles and that this is not really a smooth transition but I needed to show you slide look at this time I mean this is not great vegetables static this is 1 of the most wonderful titles in the scientific literature and death this little story behind that this gentleman here Stephen Hales the United right and you know that the person who actually assaulted Kapler conjecture conjecture His name is also hails not this 1 thing noted that is all that the question is what is the Keppler conjecture is rich in all can you prove that the the greengrocers stack of equal size is indeed the densest packing in 3 dimensions but that's to Keppler conjecture that that's there has hailed injustice gentleman also but but here actually what he did he put a piece in the job and a soak them in water and then if you created essentially a jammed situation right little bit compressed and fact and deformed right and then because the swelled in fact they did deformed against each other and many could look at the the facets and found that there were a lot of 12 sites which is actually a modern way of thinking about it indication To have a random type of packing you got these see a lot of TI's Pentagon him face so this observation he didn't know what it meant by now we know that such a packing it Jan because it expanded and champ intuitive storage to last year's out this peace rather odd circle what if you went to other ships so as I said you you can do that with a 3 D printer so that's another integrated in our and then we have a pretty nice printer so we went ahead made Adonal for 5 thousand initiate wicket-taker right and try to and what you see here are deeds stress-strain curves From that at a particular confining pressure which happens to be in the killer pass constitutes a little bit less then the atmospheric conditions were just what you can do with a simple little pump just Pulliam indeed decide
averages ensemble averages over let's say 5 to 10 runs in day color around alliance gives you an idea of the the variation with within the ensemble now you can look at that and there 14 different ships it considered itself can I discover what shape that is the same experiment every single time put it in a rubber sleeve put 80 killer Pascal's on then compressed it's all mechanics that that's called attracted to start with the simplest 1 right there and I have applauded here when you see a sphere right that's that's the lowest that's what we have been talking about right comes up and flattens 1 of which now if you put facets on this year you notice a couple things 1 of these curves to rise up to overall we would see the strength the yielding went up the strength went up to nearly became hot you can also see that maybe this slope moved out a little bit then let's go to another group of different particles eyes because in some sense they are spheres central spheres with sphere with arms coming up soon I asking what happens when I put on something and something very interesting happens here can see so let me moved out of the way so you should compare this curve with 1 for this year which is actually kind of right around here so putting tiny little arms on didn't do much putting LA a longer arms on that's the 2nd 1 here which would be uh this yellow 1 actually did did a lot right and then you make the arms even longer and what happens is gets softer is kind of sense right very long arms they can obstruct and finally make the arms even longer that's in this curve up here and of course it gets even softer but now things get so messed up that if you keep pushing eventually outperforms it so we don't know what happened with the other 1 maybe the force commander so arms do 2 things so it's a complicated situation what's really interesting is that and then you can go to other ships right you can call for example to to these things that called Dole knows that in real life meaning out here in Holland on South Africa use jetties in their 30 times of concrete here and there they don't behave particularly wild In fact they are outperformed all sorts of other ships and 1 thing that looks look particularly interesting at all seems to be really interesting is behaving so there is a complexity that we really don't understand unless we would have to think about it too much much more this it was actually also on my question because this looks very much like my book the charity's growers of surrendered physical explanation waters this was right this no and and so on and so do did the question is could there be buckling and and other explanation indeed the honest answer is right now this is there fairly complete experimental baseline at baseline effects from Algeria and I could give you know humanism hand-waving ideas but but there's no explanation of the details of the current right a particularly notice of course that there is a lot of data to each curve I will only focus on 2 parameters essentially 1 of the initial slope and then 1 maybe the yield stress properly defined but after that the 2nd defined for yesterday with the when the question was Is that preparation protocol look at very very good .period and at least me directly to the next slide to keep that in mind but I'll I'll get back to yes you that you would think that contact area that was a question should matter tremendously and of course that's what we believe but shouldn't becomes can you use it as a design principle which even know what to do without knowledge could you have predicted that all could you have designed something so that the particular outcome would happen right in Seoul at this stage you and I right now we're in we take that as this additional knowledge experimental baseline knowledge but now we have to deal with right so
let's go through that 1st thing is this is how this experiment works right the materials and sleeve and it's compressed and what we do is that in this case it is um poured in and in the cyclically compressed 10 times so so we got to 3 percent strain end and it's run up and down 10 times and the 11th time Seattle expect I think and then the whole thing gets repeated whatever 10 times in ensemble averages take this is the particular protocol for for these experiments and then the other thing we should talk a lot about how do we define things OK as get some cover from the bench you just 4 here so 1 is you take the the slope here that in the limit vanishing at home so we have fitted to travel on to the near term that's that's or slope here and then to define something like he was just see sometimes these his things still have a slope from Anatoly flat we decided just operationally to uh take the intercept here and that gives you strain if you want to win now a definitional you run this would give the what strength this is just operationally doesn't really matter that much but that many many ways of defining yielding and some people take the deviation from linear so there would be the only it now dozens of just focusing on these 2 parameters that we just extract parameters from all the states for you right and from surely now is essentially every single data point for his 100 90 or so try actual tests the 440 different ships total rights 100 and he is all the day married all right so it's is 1 big mess you might say but the music on all but it also falls in it particularly rise we can start thinking about what might be going on so let's go through that little bit more careful this is the slope right here there are plotting so couple orders to here change here is that this yield stress that I'm plotting this direction in these colors therefore the amount of confined OK so the very 1st thing you notice staring at that is that the colors they seem to be clustered together so does as well as by itself means most bites of blue barely overlaps with France that means if I change the confinement I really moved through the streets of confinement is the big now that changes everything so this is this is just barely over ambiance so 1 killer come AT killer calipers essentially pulling roughly 1 atmosphere so I can I can have a material by the way which I can move over at least an order of magnitude may be romantic in half in terms of staff just by doing something to the outside but by itself is kind of interesting try that with steel now the other thing is Egypt is symbols is a particular the particular shape and so you see that as some shapes like these here then on top of each other right which means and this 1 here this is a statistical uncertainty and separate measurements so as some spread in there but generically you see that Sears for example the symbols near the end up at the end of meeting both low here and they along stretch so yield early and not really struck back packing slide and you also notice that there's a certain spread here which which changes as you go to higher confinement but remember a la Blocks K spread so you have to be more careful in comparison so this distance roughly the same as you move over but in this direction things could crunch and so now you can ask yourself so other issues that stand out because a particularly good a particularly bad so I would have thought I don't know what you but I would have thought Wow something complicated like this interlocking what about that so let's look at the December stars and you look at the stars and you old there they are so they relatively steady but only if you don't confine him much and I find these things and his the stars and up at the end they want to him much good when you can find things now they happen to be designed for just using the ambient conditions so it's OK right they were designed for a lot of confinement but if you want the parking that's really really stiff really wanted to throw in these jacks tech-support well that only is OK here but they didn't pretty pretty good right here at the front but certainly not when you can find it mn conversely about these guys daylight is outliers yeah they always low but they can interplanetary he's a whole phrase right to their certainly not particularly good with respect to stiffness however they don't yield so easily so that's why they found at least you get distorts the front of the green which but then they lose it for high conform to that all this complicated situation and you see that certain things that I just said there were front here on top belong on top there it's a nonlinear relationship now you could say what what about density if I only knew which was partly dense then I would know what should happen right what we already know from the change that's not so easy you can't use that so you can look at that now they are this is now not frictionless pack on a computers as this is actual packing right so the densities on on as high as friction obviously in there and there was a rather large because he said somewhat small systems only a few thousand particles um too but this is the order you get and is at the bottom you because you have arms or the hollow anyway OK that makes sense accused of course potentially ticket Pakula events so cubes you might have thought nice plant faces they should really give you this strong on that series stiff packing right now that's look at that and going back to where show AccuWeather cubes are there once the look like you here and die the cubes are here for example when cubes here they don't stand out at all prime it's not that easy on the other hand just 1 more thing because is kind of exciting these guys look at the this trying to trial by they can .period indeed pack really low 48 per cent compared to only no 1 will know but look within up so by pyramids are look like to feel triangles end out front and out front so given that there is a long time and they are merely good performers there would be lovely to know what happened the right and and I can give you 1 more bit of insight 1 thing that you can see is you take a particular shape you pick nation and ask yourself How does that say it's stiffness move as you change the pressure right away how how does it perform
as you squeeze these things together more and more and if you do that you find we found actually rather remarkably it's a power fall of teachers now this is not to prove its Apollo saying it's compatible with the powers and I'm shifted shifted his power losses to connect and see what happens in the successive of rather subtle changes you moved to the shapes but what a straight line would indicated here is that the it was the stiffness does vary with the amount of pressure put on this is the sick were confined to home much confining pressure we put on and if that seems to vary with an exponent and here which goes for about . 8 To Bollard . 4 now and we all know that individual particles the interact in some way when they hit each other right maybe with hearts along whatever it is locally but this is for the aggregate as a whole and if you I wanted to assume that the force between individual particles skills as some um power lower Texas the overlap as they push into each other with some export Alfonse then you can show a couple lines that in fact and exponent of files related to that exports so for example there was words you hear 3 half right you would get out one-third that would be an effective medium approximation but one-third doesn't even show up here right all much so that it's not that simple so if I plot now this exponent of the function of something about the shape the minister obviously changes is exported but what about the shape know what I could say how spherical are you part of wider mean if if really satirical maybe that means something about how easily it responds to pressure changes that you really if you arms maybe there doesn't really matter how much pressure and so this is what I'm plotting here sought assistance exponent again and for a city is essentially the measure of how trickled in object so it's it's the ratio of the and close volume to the volume of CIA actually the other way around and sold the sphere here as for a city of 1 side and and things that have a lot of warm so whatever they used use jacks they have a much lower 1st sitting right here he has the In all these guys you shouldn't take too serious because there these frames so I don't even know how to exactly count them in here but you see the very interestingly the moment you put facets of it seems like the stockings then can and will respond to pressure changes much more sensitive there is if you had on the this suggestion that I would give to you and said that maybe because the fastest they have usually been random packing stalled lineup immediately but they might line up almost if you have a higher confinement pressure you might just line 1 more and so that a little bit change in alignment can give you potentially very strong changing stiffness or stronger changes stiffness and store changes stiffness of pressure means that you have a high school but that's as far as we know right now it would be lovely to have some more both simulation have this but also experiment and possibly the OK I and so is the last time we get back to this you must be familiar with this by now and dead now OK so why don't we just go backwards like inversion which is doing well know we can't write there is no principle that we know that says here is the way you do that and given all these complexities was we just talked about it it's even hard to simulate the things so and we give you 1 more example of how 1 can be the approach that but then also the he issues with it let's suppose I wanted to allow for oratory shapes but I want to create shapes not with Polly he drove right now because the part I don't really know how to think about facets hitting facets versus edges sitting facets of 2 corners or not the interaction is kind of complicated but suppose stick with fears any contact but I made more complicated particles by gluing years together that's up to them so here's stress-strain curve in general alright and now what I'm showing you hear it again this is initial slope but I'm doing that for shapes that I make out of 3 particles and the only thing I do is I changed the opening this I can actually numerator straight destined no nothing fancy just put on the computer and run it through for all these different angles make the packing suggests measure would still been now I don't know what you what I would have guessed that maybe it should come down all right but I would certainly not have guessed and it is very much yes the bars sold that's true but it is a rugged landscape even if you take care of I'd again this is something they are not intuitive and you couldn't have known that you have to guess find it out so if we now go back to and say Well really want to go the other way I what do we do about right and in principle what we could do is just make many many many more experiments right and and lay out all the results and fine Our favorite outcome about their only just by the sheer exhaustion is probably not the best approach given what I just told you so what can we do things that the looks left right Unisys this is this is word that probably many possible ways of doing that and there we have now I felt 1 that seems to work rather nicely Tallulah but about that and that is essentially taking their cue from biology where you also have a really complicated landscape that organism environment let's say has to cope with and somehow performed right and did the idea is that we need so if I go back 1 that we don't just run 1 virtual experiment right off and will experiment with some parameters but we run many experiments and see what happens and based on what happens you could just stop if you hit the jackpot and the result is what you wanted or the of these outcomes you have to can ever imagine how you wanna not tweak the input parameters so that you get closer to so you need some sort of a feedback loop and that's where we take this cue from evolution so we built a cult population here the whole
set the simulations that all run parallel to the population of simulation 10 20 30 40 100 however many ones we let these things do a task so forgive input parameters to give output each 1 of then we compared the output with the goal and we have to kind of select which of them got close enough to the goal that we want come at me keep them and which ones we wanted card and we need a way I have to take it now whatever happened so that we get hopefully closer to that it's a lot of physics sits in the green box To the extent that we have a simulation on the computer now that when you give it some input you trust the . is that output is a physically reasonable out but if you don't have that just doesn't work but if you do have it he just 1 of many of them and then you need this other thing and that's called now in evolutionary strategy that's the interesting thing is all you need really In addition to the red card is tools things 1 way it is to evaluate which of the many outputs from the population was best still better you need a metric amid some some minor criteria something worked or didn't and it can be really brutal right intersect at the very best closest distance to Michael In the other thing is now that you have found that good performer you need some way of changing it to make it hopefully even better perform so that this mutation part but you need to allow somehow in the structure of the algorithm that something can be mutated because of its couldn't you could make it better but that's all you need analysis 1 other thing these evolutionary strategies they you might have heard about genetic algorithms by which are similar in flavor but it turns out that these evolutionary strategies are essentially the next generation of approaches it as very very clever ways of dealing with this and they come out of computer science there will be 12 years sold by announcing a relatively recent now they're starting to be applied for all sorts of things for example crystal structure predictions about memory we talk and disordered systems that non equilibriums systems you can also use for equilibrium audit system when you have many many parameters and so the free energy landscape that he had at the could look terribly are complex and the question is how do you find the minimum right and what these evolutionary arguments are really good at is not getting stuck so if you have starts skate that is high-dimensional very rugged maybe this is the real landscape right there for gold palladium some sort of With with a local manual hidden somewhere in there these things up particularly good at finding or if she had really flat area with a hole somewhere in the middle of the golf course right front that 1 that's hard for creating this approach these things have very good and they also being applied for 4 other things for example discovered Hambletonian we have applied and that with a colleague over an incident from the Clinger nearing Pablo too turn justice random fingerprint into some letters that certain journals like and you know these are far from equilibrium situations where you wanna bias the outcome in a particularly clever way and you can use them to teach robots 2 what is a complex problem again many promises that just give you for a fine example of justice robotic gait evolution not this is a robot that only exists in silico right this is not a real problem in the physics engine isn't quite calibrated for real walking back to the ideas the following right in the computer I want a robot that gets as fast as possible from over here 2 over there I reward you for speech freight anybody who is not fast by Darwinian principle gets eliminated and I have 4 kinds of tissues that can be put together in a random combinations computer find what is the right combination right and here is the number of generations this thing ran and this is the distance had gotten a particular time and the further against the more gets rewarded and reward means that a particular configuration is apparently better not sometimes it trips right sometimes it doesn't get quite there but the computer does know anything about walking or anything in fact all also has to get from a to B. fast so it can set with unorthodox solution right which we couldn't have guessed maybe because we typically awarded in you know going the other way from the ingredients the final things we use things that help us let's asymmetries stuff like what this doesn't have happen right so if you can look it up on the Web is confined here's another thing just terribly complicated for an ordinary soldiers church algorithm in matters of forest of spikes and inside little parabola and new starting out there and you will find that 1 of maybe force of spikes with holes in between you won't find that 1 place right to what a genetic soaring evolutionary with really important evolutionary strategy will do is generate a whole Swarbrick trials solutions Anderson very clever about and mutating right in the particularly um 1 we use is called the covariance matrix adaptation method and I'm sure you hear what that thinking do 1 of these a test cases right swarms on explores a wide parameter space in it's very good at using prior data to estimate where to go next and it really zoom in and so does this combination of exploration exploitation None of course there is no such thing as a perfect optimizer right Everything is only good for certain kinds of problems but for these types of problems we found that this is remarkably remarkably good news to the at the bottom of this the citations this comes out paper originally by Hansen from 2003 right now the 1 reason 1 other reason was as good as that which you rotate stretch here parameter the doesn't care subsidiary into that and so you see that you duration or thing is not simply a problem and stretched and you'd 1 and this is the distance to the minimum in the middle and for these 2 cases since essentially indistinguishable right so it's robust with another example turns out for humans it's really easy to answer the following question I give you 4 spheres was the most compact way to assemble the Will move on to the streets .period appear not try with a computer it is really hard and especially if you go to 5 or 6 it is essentially near-impossible this a beautiful test case and he's saying what we usually in physics to write Monte Carlo similar kneeling and thing In this is again the distance to the known minimum duties have been tabulated and an award a minister member function calls are answered you see after while they 81 similar but then this method really goes to town and takes over prices is rather remarkable so we gonna do now
is take this particular "quotation mark as matrix adaptation he has evolutionary strategy and use it together with a simulation which is a molecular-dynamics simulation of what we think should happen this week for particles in the back and squeeze that and then this thing spits out a particular solution and we test by printing it and running the actual express prices that would closer look that way and again the particles
are going to be arbitrary shape Of ends years stuck together like this examples of 380 printed articles where a few millimeters each year and then the colors the only anything and the end 10 that means tens years that computer could put together in any which way it will rain weather and that the mutation all
we don't have time to maybe go through that in detail but you can ask me later this is something that marked the most integrated student developers a very nice way of dealing with are they hear what he calls a blueprint of telling the computer where to place spheres for each component particles that actually cuts down on the overall such dimensions a nice way to control you can think of this essentially is going from a genotypes that's what the computer knows that these numbers to tell the computer based in what direction to put another particle to the phenotype all right so now we can ask to computer to do a few things so we can say find us Aguayo molecule that under these given conditions right paw in there but confinement on inscrutable but that gives you the stiffest or the softest respond and so it starts with some they generated particle Kompong particle and here we said that actually limited to forces was early work we did and we don't have enough computer power to let it run with higher number in and just get stiffer and stiffer other generation number so it is a generation numbers 20 30 40 70 and these are the actual values for stiffness Pfizer comes up once in a while gets stock and searches around and find something better and eventually thousand told 1st we can tell at this stage With the 60's and similarly for this looser once they come down at the end of the day and you can see that the smaller serotonin wiggles around tries to see if it's better nephew compare these where what the Monte Carlo I would have done so just look at it just randomly generated ships get the histogram like that and you see what the sting founders actually significantly outside the bell writes I mean to to find that just by random searching the Unisys 6 6 Mark Six Sigma offices for signoff are you'd have to be very very lucky but this is disempowering 1 thing you can of course ask not also know I found right not no 1 seen our science starts I mean I've just biased myself into the part of such things that I was interested and non-selective figure out why it does what it does this not the computer doesn't tell me that and for the stiffest in particular scanner interesting this is not flat it's a little bit Boettcher actually it turns out if you look at this same question for different numbers of it and you can find the particle shape that will give me the stiffest packing accorded randomly for different by taking the party that is the most compact shape with plus 1 resilience canceled the flight just 1 year has 4 goes vote was monitored and if I wanted to have guessed what it looks like it should have taken the most compact particle for 5 and take 1 day off the networks actually quite well up to share and of course and cause large eventually doesn't doesn't work and what is battling what take 1 but he his for and it was strange you take the 1 with any Eagles for taekwondo and goes for it at the 5 year this would want that's the principal that seems to be underneath that behind and it is interesting because by taking 1 cutting off you essentially have a little hollow that means that almost makes it easy for another particle coming in there and it's like a socket in of all types but this is a conjecture right now we we haven't proved it yet but that's what OK so um computer find me the shape after going a molecule that packs to the highest density when according to contain under gravity so now I changed the conditions Knoll confinement just poor but you know which 1 that is is going to be serious I mean all kids all spherical particles tried but here is saying you can take up to 10 Sears gloom together any which way you want and you can't change the radio In fact if you let the Raiders go to 0 is effectively remove received by so you have up to 10 go ahead and the computer says OK and went ahead and what I'm sure you years be recycled through this and give me the worst to test the media can result and the colors just didn't mean if there were different sizes and dead but this is what another with Mickey Mouse is not interesting and here's what happened to the density start with some random thing here complicated thing and had a lot of particles and it got rid of most of the particles ended up with its 3 it turns out that as a reason that jibes with that and that is that in the structure of voids there actually holes between neighboring particles that are such that they have to have these years are about 70 degree apart and have there really is no bigger than one-third of the big 1 day statistically the chance of finding does what's best the fact that the question is is it is not like water now would whenever you're interested in all it's not to Henkel's offer to host as seriously about it yet so that would that would be nice but it would be a current students ride I read is this is a consequence of of pouring on the gravity and in this particular case so if you want and no 1 look at for example what happens Back to the whole the whole ensemble you know I am not plotting essentially everything that the computer went through and in the end we remove the best 1 up there and and the worst 1 down there and each of Jesus 1 generation of the grinds through right and it and ran all of these things in parallel rights essentially everything a computer seeking action look back with that evolution and find out what it did if you want to analysis that they got rid of most of these extra guys and shrink some of them to tiny values right and eventually all goes away and this happy right now I'm realizing this is almost 5th time here but I used still with the program tried to do 1 more thing because this is our kind of interesting a coming up so you you see how this convergence essentially to the making of right I can run a little bit longer but it's up almost yet so so that you look at for example the 2 target them and at different distances to target value and you say how close and my right so for example is a very good question how how do I know it's
converged here I run this all too few hundred simulations maybe a thousand and I just say how flat so I'm asking for them in this case the densest packing I'm really interested in this snow I could ask for other things in there that would compare that number with my time and do you it you acute can sort of this of excellent question how sure are we that finding the global optimal you cannot be sure the stings that's the thing that's really not looking for the grounds they were not identifying any ground state we are identifying 1 of many states and see if other states and so therefore I weekend not 100 percent be sure and the only way around that is in this game to run it over and over again different scenes and testicle if so so that there is no that that is kind flip side of the coin and the cost that you incur said it was a there is no link mention to the spirit of the results of the regions it yes however and that is so the question is is it when the grounds to be the crest but that wasn't the question the question was pouring under gravity I will guarantee you that you will this is measure 0 the 1st phase 2 poor staff and have find a Crystal Heights so this is the beauty of this you take the conditions into account and that has been Florida's is critically important in this business this approach does take into account automatically pass to right but it's not that the bad thing is in fact part of the beauty of
OK so but now look uh we we found dead dead loses Packer here even for 10 it's a lot like the Bird's Nest stinks like meals that they packed loosely and signed pact with impact and way that soft with push we found this particular guide this little Mickey Mouse being the 1 that this is the highest packing and then of course while this thing right and we generated a lot of data it started somewhere and then ended up someplace and that's all good data just isn't used right to know if you could fill the space with this I mean if you wanted a particular density this density you you could pick a point here and you have the answer so this gives you answer for questions of the sort and if there's a hole in and you just run simulation that another now hard what is this thing him this is an axis that leaves I mean that many ways of doing that we wanted to have some way of that separating different shapes up and the way we did that it's just look at this so courtship and justice captain the distances to started 1 and let's say that the distance to all the others In this at this stage is still not really satisfied right because you know have a way to go backwards in reverse that's good but for every question every target you principle have to run this thing to get 1 that's the not so satisfied would not be nice if you could find more general answer right and particularly if you look at this there is no guarantee that neighboring points have shapes that don't look anywhere Simone they could look totally different and so you have no way of predicting anything unless you run however and is the interesting thing a clever if you use the same Hargrove now not to identify anything in between but you say finally the densest Lewis's packing or the strongest are the weakest issue 1 look for extreme we ready showed you the existing identified extreme and if you have found extreme then he actually gains with images showing that here so suppose we found extreme on say that 1 and that allow me particularly access and promises and of course also Associates in between you notice as possible that the same um aggregate property has several solutions there is no guarantee there's only 1 but as long as we have identified the extreme on weekend now something their particular what we just say we focus on the extreme only right now so I know already held a get to those and I know now what shapes correspond to the and if I'm not contain any transformation gets me from 1 extreme to the other and shapes then I get all of these things in between under Freedom of condition right because I know these are extremely and I know that any change in shape would give me something in between I can pick now my path according to completely different principles for example was the easiest to fabricate or the lowest cost or whatever I don't need to follow the line that their comparatives in fact I don't have to go through all the shades I could for example just trappers from this into that by shrinking along with the new taxes if we did that by by direct calculation right and so you see here you go from down there and take 1 off or shrink it down and you can for example starting and trade 1 and the next and the next he also Maja right Indian principle would march to the to the corner although it with the word access and then you can invent another rule that says grow between it went to a simple rules to cover the without having to worry about all this ships In between OK so and then emitters to drive to hold down at point in home this this is so I think important so we use the optimizer now to find extreme under this this really have something but then you can also ask the computer to find you things that you don't have any idea about but suppose you as a computer to find the ingredients for a property that you did know said now it was a good simulator it would look for if it found it you had discovered something new right so you can use the same process for that of course is no guarantee that you will discover but that's true for any discovered right you just have to try to let pass a discovery goal Ryan you already know there is a solution the computer didn't know that the action did not want to so we said computers this is what everybody knows happens defined as something qualitatively different I mean really different slope that way so maximizes the the 2nd something like that In the computer went to town and we said OK you have and make it also the smallest molecules we already knew for and was wanted that that's the core of an 11 minute doesn't work right and he put 2 we have done already knew it and it works really well I showed you the graphite why should you slope but we took the whole Graf didn't work so it had to be a little bit more so we gave it I don't know 6 0 7 to play with and this is what I came back with so he is I compare that with the ordinary Daimler and this is this thing none of this is the smallest obviously not the strongest strains differ about it does as advertised and that's what it looks like so I have here for you the the big version or have to hand and you get to play with them don't break it at I was sorry I only not handed over you guys and there who would have thought right I mean this is highly asymmetric that has the beginnings of a miss not playing at the the right said it has the beginnings of something that you really call a star-shaped polymers connection with the 3 arms coming out all of that but again it's 1 of those cases where I couldn't have predicted that a priori writing needed just discovery tool to find OK so now we said OK you need to stick his years together just prior touching you can relax at the end and so currently out the wrath of the student and she allows no overlap so you can make smoother shapes and then look at the role of geometrical friction that way too there are other extensions you can imagine that we just talked about packing many many things 1 can thank you to optimize you can have in particular faster computer to optimize computing goal right maybe something that is really still yet highly permeable a lot of water should flow through but but I would also released against waste are you could imagine during coevolution here very interesting direction and of course the interactions because we could have anything Beyond going beyond just repulsive interactions when they touch right friction so you could imagine doing something that you do in chemistry right where you go really beyond just their interactions and have functional groups and so certain parts of molecule touch another certain part of another molecule could get very interesting and you could that temperature and you know that's step 1 is starting to play well and then some finally the same logarithm you know you you can also be used to really to America processing path and and this is maybe a little bit too complex at the end but just look at that picture over here this is a highly corrugated landscape that's a rather sticking out of the blue is going down at the end you wanna go from this point on the left To that other point over there as fast as possible and what you can do is you can tilt the potential well into directions and you can change temperature Bucher said if you play with this game with with the marble that rolls and you turn denouncing have to kind of move around holes it similar to that so much of this thing do should just carefully Neil things and then let it slide around or should tilt like crazy run over some of the hills and stopped Turns out the computer like that right did later generations of this .period curve which means a a lot of tilt here and in some braking action in both directions while the temperature actually turns out to be not so important particularly case but you can really model processing passed so that's that's an interesting New Directions OK a job ballots so many close up so I told you love bit about this as a model system for approximately behavior and down the key thing really is that we should embrace the question we should use it write to you jamming as a transition between amorphous state liquid like an amorphous did that's all at once no structural transition necessary that's what makes a powerful and then I think finally there are some special opportunities right now that are coming into view because looking a bit over our offense and with it junior in computer science and you said to to try to get a handle on this design processes is universe process party with that I think you very much but if you don't and I know what wrote later than plan but I'm ready and willing to answer any questions you might have I just don't want to talk about places program is that you don't already know I think you all yeah yeah so the question is why is this particular strategy let's CMA strategy sold good and it's really a combination of maybe handful different things but but 1 thing is obviously not gradient breeze OK so that doesn't care what gradients so therefore doesn't get stuck in minimal right it basically that generates solutions for swarms of the trial solutions and that's the point number 2 picks out of dollars with this by looking at the core of areas which ones have a chance to be more successful which direction have a chance to be most successful at it and then the 3rd thing is that it is invariant to various types of scaling so if you buy a chance didn't come basically and idolized Joe parameter matrix it can in the best possible way just this thing is it's rather insensitive to them and then there are the couple more but it's really the combination of and the other thing 1 has to I always say right there is no such thing as 1 perfect unworthy for every problem is that there may be problems were some of the algorithm probably is is more useful it just turned out that for these type of far-off Mecklenburg equilibrium situations and we haven't found anything that is as good or possibly better justice really outperforms everything we have seen and that includes all Sanader's work that we did on the polymer so even in cases where the temperature available and it still seems to work exceedingly well yes and those who stand in the way of saying that we have a real means that no 1 know this that there haven't yet and then 1 OK is sold the dirt I think if I understood the question is if I have the yield stress fluid then if on they're not dealing I and jam let's say end once the deals that would be Unger that well so for example if you have a suspension that behaves like such an injustice food with dense suspension with with many particles in the typically that I would say typically behaves just similar to regret material and you can think of the jammed state as the 1 that is not yielding right In other words that the 1 that is essentially static where's the relative motion between particles in the 100 m status state where this village of more here this year profile something like that so in that sense of yielding can be compared to a and jamming transition it was that kind of what you're asking yes I I I would think that status that's a possible way of looking at that a mail in many ways suspension then suspensions I have characteristics that make them look similar to dry granular materials to the extent that you have frictional interactions of the dense enough that the frictional interactions I don't often indeed the fact that you also liquid there is it's not longer the central role that it can play a secondary role and and really the thing can be dominated by frictional interactions in which case behaves very much like would you expect from the try curriculum materials including yielding and behavior like that this I'm going to be a lot of planning the people in the world to has it was the end of the war started by saying that you used the word of the 1 thing that actually realized it was the correct thing to do in the next 2 years using the word really only yes absolutely no line OK so they did the great question the 2 questions and I want to have poured answers will be different for the to select 1 question is did we actually put this shares started the strain stiffening molecule that did we tested it's really yes so not with these big ones but we we made a few thousand smaller ones and it was tested and that was the that was a Blue Cross there was action experimental data I guess also a computational answer but I sure to the actual expense for both the diamond also this this 5 and the other question is is there there deeper ones do we understand why this particular molecule now does what it does right and we can we tell and and the the answer there is um that I have some punches but I can't give you the you know the proof of anything in particular for 4 sample what we would love to say is disabled and arrange in a particular local structure which then gives rise to this global function right and as that has eluded us so far some as 1 of the interesting facts about this game right it as such complex structure function relationship that it's not clear what about the local connectivity right now b beyond some rather general statements obviously this thing will pack initially you poured rather loosely right and in order to strains different must have a chance they had to to compress and so you can think of to it's also estranged of Nazir lol plus all racial material some sense but it will probably behave they're not isotropic under them in this industry compression clearly is not expanding outward this much as an ordinary isotropic material would do but still that's something where I think there would be great now to go in and To some more experiments and more simulation so what this does is it it finds things for you that you'd like to find but it doesn't absolve you from what you do as a scientist now which is to worry about you know now why do you want to it's also designed properly but there still a science problems and that's also 1 reason why the designs the problems typically goes in the opposite direction from the design and we want an answer he added this is the smallest strains different I know I know and you're right I can prove it is there's nothing else out there that there does pass that we were 3 particles but doubts records of OK this don't all right I am ready for some lovely drinking this is not going to sit and you
Kraftfahrzeugexport
Maßstab <Messtechnik>
Material
Weltall
Schiffstechnik
Computeranimation
Teilchen
Steinbildhauer
Elementarteilchenphysik
Speckle-Interferometrie
Stunde
Personenzuglokomotive
Brennpunkt <Optik>
Tag
Schaltelement
Weltall
Übungsmunition
Angeregtes Atom
Schiffsklassifikation
Gruppenlaufzeit
Flüssiger Brennstoff
Halo <Atmosphärische Optik>
Material
Planet
Feinkohle
Münztechnik
Stoff <Textilien>
Gesteinsabbau
Kugelblitz
Konfektionsgröße
Dschunke
Rutsche
Big Crunch
Material
Energieniveau
Kondensor
Oberfläche
Quadrophonie
Passfeder
Proof <Graphische Technik>
Druckkraft
Kaltumformen
Faraday-Effekt
Schalttafel
Schaltelement
Jeep Wrangler
Übungsmunition
Angeregtes Atom
Satzspiegel
Motor
Atmosphäre
Jahr
Oberfläche
Elektrostatik
Exosphäre
Analogsignal
Schaltelement
Kristallgitter
Gasdichte
Schneeflocke
Direkte Messung
Maßstab <Messtechnik>
Festkörper
Zwangsbedingung
Schnee
Phasengesteuerte Antennengruppe
Antiteilchen
Federung
Teilchen
Metallschicht
Umlaufzeit
Zylinderblock
Elektrolytische Leitfähigkeit
Energielücke
Bahnelement
Tagesanbruch
Klangeffekt
Zwischengitteratom
Innere Reibung
Teilchen
Schutz <Elektrotechnik>
Brennpunkt <Optik>
Metallschicht
Flüssigkeit
Munition
Flüssigkeit
Sturmgewehr
Umlaufzeit
Innere Reibung
Material
Lithium-Ionen-Akkumulator
Gasdichte
Atomphysiker
Thermalisierung
Mikroskopobjektiv
Grundzustand
Nacht
Teilchen
Stack <Brennstoffzelle>
Kristallwachstum
Fertigpackung
Kommandobrücke
Kaltumformen
Gasdichte
Explorer <Satellit>
Kraft-Wärme-Kopplung
Beschusszeichen
Elektronenschale
Gas
Kristallwachstum
Angeregtes Atom
Thermalisierung
Flüssigkeit
Bestrahlungsstärke
Jahr
Rucksack
Exosphäre
Temperatur
Amorpher Festkörper
Niederspannungsnetz
Potenzialausgleich
Thermalisierung
Toleranzanalyse
Atmosphärische Störung
Material
Federung
Kristallwachstum
Angeregtes Atom
Luftstrom
Glasherstellung
Kristallgitter
Herbst
Speckle-Interferometrie
Brennpunkt <Optik>
Federung
Abwrackwerft
Prozessleittechnik
Feldstärke
Metallschicht
Atmosphärische Störung
Eisenbahnbetrieb
Glasherstellung
Strukturelle Fehlordnung
Angeregtes Atom
Schmalspurlokomotive
Material
Weiß
Gitterbaufehler
Rundstahl
Astronomisches Fenster
Kristallgitter
Amorpher Festkörper
Modellbauer
Atomphysiker
Kaltumformen
Thermalisierung
Gagat
Sonnenstrahlung
Gruppenlaufzeit
SEED
Fahrgeschwindigkeit
Impulsübertragung
Schwache Lokalisation
Teilchen
Angeregtes Atom
Römischer Kalender
Innere Reibung
Lawinenschutz
Modellbauer
Fahrgeschwindigkeit
Material
Experiment innen
Speckle-Interferometrie
Störsender
Dissoziation
Schaltelement
Summer
Konfektionsgröße
Fehlprägung
Leisten
Schiffstechnik
Material
Energieniveau
Computeranimation
Edelgasatom
Maßstab <Messtechnik>
Kühlkörper
Experiment innen
Straßenbahn
Kümpeln
Prozessleittechnik
Stromschiene
Kaltumformen
Ladungsdichte
Quecksilberschalter
Schaltelement
Angeregtes Atom
Konfektionsgröße
Schiffsrumpf
Schiff
Jahr
Direkte Messung
Elastische Spannung
Festkörper
Störsender
Wärmequelle
Gasdichte
Lunker
Atmosphärische Störung
Zwangsbedingung
Impakt
Teilchen
Kristallwachstum
Umlaufzeit
Zylinderblock
Halbleitertechnologie
Phototechnik
Bandstruktur
Speckle-Interferometrie
Fließfertigung
Schreibkreide
Gruppenlaufzeit
Gauß-Bündel
Tag
Flüssigkeit
Störgröße
AM-Herculis-Stern
Telefon
Gas
Druckkraft
Videotechnik
Flüssigkeit
Source <Elektronik>
Buntheit
Material
Ersatzteil
Kolbenverdichter
Störsender
Dissoziation
Zwangsbedingung
Schwingungsphase
Rucksack
Gasdichte
Kugelblitz
Nyquist-Kriterium
Schiffstechnik
Material
Behälter
Lenkrad
Satz <Drucktechnik>
Teilchen
Avro Arrow
Luftstrom
Prozessleittechnik
Aktives Bauelement
Pfadfinder <Flugzeug>
Kristallgitter
Experiment innen
Passat
Messbecher
Zyklotron
Tagesanbruch
Zwischengitteratom
Schall
Teilchen
Potentiometer
Mark <Maßeinheit>
Front <Meteorologie>
Zementation <Metallurgie>
Bett
Irreguläre Galaxie
Druckkraft
Nivellierlatte
Flüssigkeit
Nassdampfturbine
Motor
Gleichstrom
Januar
Koffer
Kommunikationssatellit
Jahr
Uhrwerk
Ersatzteil
Störsender
Kosmische Strahlung
Interstellares Gas
Kristallgitter
Leistungsanpassung
Temperatur
Akustische Rückkopplung
Myon
Greiffinger
Regler
Kugelblitz
Aktor
Griff
Leitungstheorie
Schlauchkupplung
Zelle <Mikroelektronik>
Prozessleittechnik
Maschine
Experiment innen
Neutronenaktivierung
Proof <Graphische Technik>
Bahnelement
Logger
Prozessleittechnik
Abwrackwerft
Theodolit
Kaltumformen
Waffentechnik
Licht
Unwucht
Tag
Proof <Graphische Technik>
Akustische Rückkopplung
Energielücke
Übungsmunition
Schiffsrumpf
Gleichstrom
Jahr
Ersatzteil
Direkte Messung
Störsender
Wärmequelle
Kaltumformen
Waffentechnik
Längenmessung
Elektrizität
Regelstab
Schaltelement
Akustische Rückkopplung
Stoßdämpfer
Leitungstheorie
Uhrwerk
Küchenmaschine
Weltzeit
Holz
Prozessleittechnik
Atmosphäre
Kopfstütze
Ersatzteil
Material
Experiment innen
Gasballon
Familie <Elementarteilchenphysik>
Greiffinger
Mikrowellentechnik
Kleiderstoff
Thermalisierung
Großkraftwerk
Summer
Anzeige <Technik>
Closed Loop Identification
Wellenenergie
Konfektionsgröße
Rutsche
Material
Satz <Drucktechnik>
Kernreaktor
Leitungstheorie
Griff
Zentralstern
SEED
Regelstrecke
Buchdruck
Prozessleittechnik
Passung
Kernenergie
Monat
Sonnengranulation
Temperaturabhängiger Widerstand
Federung
Prozessleittechnik
Buckelschweißen
Theodolit
Kaltumformen
Personenzuglokomotive
Bleibergbau
Buchdruck
Störgrößenaufschaltung
Übungsmunition
Greiffinger
Kometenkern
Edelsteinindustrie
Halbleitertechnologie
Jahr
Festkörper
Störsender
Anstellwinkel
Temperatur
Reibantrieb
Festkörper
Stoßdämpfer
Begrenzerschaltung
Teilchen
Metallschicht
Spiel <Technik>
Pfadfinder <Flugzeug>
Urkilogramm
Laser
Kiesabbau
Vakuumphysik
Klangeffekt
Tastverhältnis
Teilchen
Anzeige <Technik>
Kombinationskraftwerk
Metallschicht
Absorption
Bett
Tag
Flüssigkeit
Eisenbahnbetrieb
Proof <Graphische Technik>
Weltall
Band <Textilien>
Taschenmesser
Flüssigkeit
Nassdampfturbine
Schwarzes Loch
Innere Reibung
Kernreaktor
Rucksack
Panzerung
Material
Ersatzteil
Pneumatik
Material
Schiffstechnik
Teilchen
Uhrwerk
Funksender
Luftstrom
Phototechnik
Kapazität
Kristallgitter
Experiment innen
Erdblitz
Federung
Buckelschweißen
Kraft-Wärme-Kopplung
Verkehrsflugzeug
Gruppenlaufzeit
Impakt
Untergrundbahn
Rotverschiebung
Schiffstechnik
Elektronische Medien
Jahr
Ersatzteil
Material
Gitterbaufehler
Erder
Elastische Spannung
Thermalisierung
Feldstärke
Closed Loop Identification
Synchrotron
Stellwerk
Energieniveau
Geokorona
Material
Schwingungsmembran
Brennspiegel
Leitungstheorie
Schlauchkupplung
Buchdruck
Prozessleittechnik
Biegen
Sicherheitsbehälter
Proof <Graphische Technik>
Randspannung
Druckkraft
Abwrackwerft
Kaltumformen
Feldstärke
Ladungsdichte
Längenmessung
Zugangsnetz
Übungsmunition
Angeregtes Atom
Schiff
Druckfeld
Jahr
Vakuumphysik
Steckverbinder
Absolute Datierung
Direkte Messung
Anstellwinkel
Elastische Spannung
Ersatzteil
Schaltelement
Anstellwinkel
Gasdichte
Tomographie
Target
Bildqualität
Förderleistung
Teilchen
Druck
Mitchell, Giurgola Architects
Zylinderblock
Fertigpackung
Kommandobrücke
Teilchen
Gasdichte
Kraft-Wärme-Kopplung
Basis <Elektrotechnik>
Röntgenstrahlung
Atmosphärische Störung
Bett
Reif
Parkuhr
Gleichstrom
Drahtbonden
Koffer
Rucksack
Modellbauer
Material
Elastische Spannung
Maßstab <Messtechnik>
Erdefunkstelle
Satz <Drucktechnik>
Teilchen
Schlauchkupplung
Biegen
Messbecher
Clusterphysik
Kontrast
Randspannung
Tonnenleger
Tag
Braun
Reif
Übungsmunition
Konfektionsgröße
Rundstahlkette
Druckfeld
Jahr
Steckverbinder
Material
Elastische Spannung
Dipol <Nachrichtentechnik>
Unterwasserfahrzeug
Trenntechnik
Rucksack
Theoretische Mechanik
Zelle <Mikroelektronik>
Direkte Messung
Konfektionsgröße
Erdefunkstelle
Energieniveau
Satz <Drucktechnik>
Dezember
Flüssigkeitspumpe
Computeranimation
Zentralstern
Teilchen
Druck
Monat
Brechzahl
Comte AC-4 Gentleman
Farbe
Energielücke
Randspannung
CHAMP <Satellitenmission>
Randspannung
Anthropisches Prinzip
Rundstahlkette
Theodolit
Mikrocluster
Gasdichte
Lichtstreuung
Bleibergbau
Behälter
Ladungsdichte
Halo <Atmosphärische Optik>
Strangeness
Konfektionsgröße
Nassdampfturbine
Rundstahlkette
Schiff
Januar
Jahr
Buntheit
Rucksack
Druckerei
Kolbenverdichter
Elastische Spannung
Initiator <Steuerungstechnik>
Reibahle
Schalter
Unterwasserfahrzeug
Messung
Parallelschaltung
Mechanikerin
Sternatmosphäre
Rutsche
Big Crunch
Geokorona
Schwächung
Computeranimation
Schlauchkupplung
Regelstrecke
Trenntechnik
Messschieber
Biegen
Einbandmaterial
Randspannung
Feldstärke
Längenmessung
Faraday-Effekt
Optischer Verstärker
Schaltelement
Ausreißer <Messtechnik>
Übungsmunition
Angeregtes Atom
Satzspiegel
Schiff
Uplink
Druckfeld
Jahr
Elastische Spannung
Rundstahl
Natürliche Radioaktivität
Exosphäre
Postkutsche
Begrenzerschaltung
Teilchen
Dezember
Zylinderblock
Wagenheber
Randspannung
Radiergummi
Klangeffekt
Gasdichte
Kraft-Wärme-Kopplung
Front <Meteorologie>
Waffentechnik
Gruppenlaufzeit
Tag
Druckkraft
Band <Textilien>
Nivellierlatte
Tageslicht
Parkuhr
Nassdampfturbine
Gleichstrom
Innere Reibung
Rucksack
Buntheit
Material
Tuner
Kolbenverdichter
Schiebersteuerung
Frequenzsprungverfahren
Zentralstern
Tischler
Unterwasserfahrzeug
Myon
Schaft <Waffe>
Messung
Magnetisches Dipolmoment
Parallelschaltung
Maschine
Interstellares Gas
Material
Geokorona
Satz <Drucktechnik>
Leitungstheorie
Computeranimation
Weichzeichner
Schlauchkupplung
Matrize <Drucktechnik>
Buchdruck
Bildfrequenz
Biegen
Kristallgitter
Druck
Feile
Prozessleittechnik
Familie <Elementarteilchenphysik>
Explorer <Satellit>
Längenmessung
Schaltelement
Elektromagnetische Verträglichkeit
Kardieren
Elektronenschale
Übungsmunition
Werkzeug
Closed Loop Identification
Römischer Kalender
Druckfeld
Jahr
Flugsimulator
Gießen <Urformen>
Exosphäre
Kristallgitter
Anstellwinkel
Stromausfall
Niederspannungsnetz
Tandem-Fahrrad
Irrlicht
Ruhestrom
Teilchen
Kristallstruktur
Druck
Wagenheber
Flavour <Elementarteilchen>
Herbst
VW-Golf GTI
Randspannung
Klangeffekt
Tastverhältnis
Teilchen
Behälter
Waffentechnik
Kombinationskraftwerk
Inversionsschicht
Tissue
Adapter
Telefon
Druckkraft
Weiche Materie
Kristallwachstum
Thermalisierung
Atomhülle
Schwarzes Loch
Schreibkugel
Rucksack
Ersatzteil
Isostatisches Heißpressen
Initiator <Steuerungstechnik>
Papier
Unterwasserfahrzeug
Rucksack
Schaft <Waffe>
Target
Parallelschaltung
Durchführung <Elektrotechnik>
Diwan <Möbel>
Geokorona
Postkutsche
Dreidimensionale Integration
Familie <Elementarteilchenphysik>
Satz <Drucktechnik>
Nanometerbereich
Teilchen
Woche
Kristallgitter
Experiment innen
Randspannung
Familie <Elementarteilchenphysik>
Teilchen
Kaltumformen
Gasdichte
Ladungsdichte
Elektronisches Bauelement
Längenmessung
Adapter
Scanner
Tag
Proof <Graphische Technik>
Rauschzahl
Türglocke
Übungsmunition
Band <Textilien>
Druckkraft
Elektronische Medien
Nassdampfturbine
Schwarzes Loch
Gleichstrom
Schiff
Jahr
Matrize <Drucktechnik>
Rucksack
Buntheit
Sprechfunkgerät
Ersatzteil
Exosphäre
Grosspackmittel
Fliegen
Rucksack
Münztechnik
Erder
Modellbauer
Greiffinger
Messung
Thermalisierung
Erwärmung <Meteorologie>
Lineal
Grundzustand
Familie <Elementarteilchenphysik>
Satz <Drucktechnik>
Minute
Seewind
Leitungstheorie
Computeranimation
Schlauchkupplung
Schwingungsphase
Prozessleittechnik
Steinbildhauer
Kristallgitter
Federung
Randspannung
Prozessleittechnik
Abwrackwerft
Amorpher Zustand
Familie <Elementarteilchenphysik>
Theodolit
Eisenkern
Kaltumformen
Optische Dichte
Beschusszeichen
Ladungsdichte
Längenmessung
Impakt
Körner <Metallbearbeitung>
Werkzeug
Übungsmunition
Werkzeug
Angeregtes Atom
Tonbandgerät
Römischer Kalender
Matrize <Drucktechnik>
Jahr
Lineal
Flugsimulator
Ringgeflecht
Elastische Spannung
Festkörper
Temperatur
Druckgradient
Target
B-2
Schnee
Diamant <Rakete>
Erdefunkstelle
Postkutsche
Impakt
Teilchen
Brechzahl
Luftstrom
Discovery <Raumtransporter>
Spiel <Technik>
Quecksilberschalter
Passat
Speckle-Interferometrie
Randspannung
Rundstahlkette
Tagesanbruch
Teilchen
Gasdichte
Zugangsnetz
Waffentechnik
Kombinationskraftwerk
Flüssigkeit
Weltall
Nivellierlatte
Thermalisierung
Flüssigkeit
Nassdampfturbine
Schwarzes Loch
Gleichstrom
Großtransformator
Profil <Bauelement>
Innere Reibung
Rucksack
Modellbauer
Material
Ersatzteil
Störsender
Grosspackmittel
Unterwasserfahrzeug

Metadaten

Formale Metadaten

Titel Master class with Heinrich Jaeger
Untertitel Granular materials by design
Serientitel Physics@FOM Veldhoven 2015
Autor Jaeger, Heinrich
Lizenz CC-Namensnennung - keine kommerzielle Nutzung - keine Bearbeitung 3.0 Deutschland:
Sie dürfen das Werk bzw. den Inhalt in unveränderter Form zu jedem legalen und nicht-kommerziellen Zweck nutzen, vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen.
DOI 10.5446/18038
Herausgeber Foundation for Fundamental Research on Matter (FOM)
Erscheinungsjahr 2015
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Physik
Abstract Granular materials are large amorphous aggregates of discrete, individually solid particles. Despite seemingly simple ingredients, such aggregates exhibit a wide range of complex behaviours that defy categorization as ordinary solids or liquids. This includes non-Newtonian flow behaviour and collective 'jamming' transitions. One of the key issues has long been how to link particle-level properties in a predictive manner to the behaviour of the aggregate as a whole. However, for actually designing a granular material, an inverse problem needs to be solved: for a given desired overall response, the task becomes finding the appropriate particle-level properties. This master class discusses new approaches to tackle the inverse problem by bringing concepts from artificial evolution to materials design. These results have general applicability and open up wide-ranging opportunities for materials optimization and discovery.
Schlagwörter Granular
Materials

Ähnliche Filme

Loading...