 So thanks for coming, everybody. I'm just going to say a few words about our partnership with TREMCO and the TREMCO Sustainable Practices series that we have. Steve Hughes is here from TREMCO. And over the past three years, they've invested in us all. And we've invested in them a little bit to support lectures related to sustainability and other activities related to sustainable practices in general, and that includes us providing opportunities for students, or them providing opportunities, but us providing a little bit of support for students to take up internship trainings at TREMCO's locations, either in Toronto or Cleveland, or kind of both, which are their North American headquarters. And TREMCO is part of a kind of global company, sort of headquartered, I think, in Germany called RPM. There's a host of very interesting companies, if you want to look that up sometime. And what's been great, again, there's always been, in this three-year time, terrific support from TREMCO and some very interesting things that have come up. TREMCO started playing with architecture schools at the University of Toronto and established a thing called the best series. And Christoph Reinhardt, who's our lecturer here, has actually given a lecture in that series at Toronto. We were just talking about that. And one of the things that's been evolving a little further, that's something hopefully to look forward to here, is we've been reaching out a little bit to set up a bit of a network around the US and Canada, starting with the University of Toronto as a bit of our home base, and maybe us as a bit of an American rather than North American home base, including schools like Clemson and the University of Washington and Florida International. And I'm sure I'm forgetting some other ones, but there's about six or so of us at the University of Maryland to try to see whether spreading some related sustainability education might set up a network of opportunity for you guys as students and for faculty members to share ideas. And we'll keep cooking that along, and maybe we can get Christoph to play in that and MIT and other schools coming forward. So I just wanted to start out by thanking TREMCO for that and giving you a little bit of that background. And now I'll have Nate Fash come up and introduce tonight's lecturer. Thanks very much. Thanks everybody. Thanks, Steve. Thanks everybody for coming out tonight. Let the record show that despite the power out as a couple of weeks ago, we persevered. And we're really grateful to Christoph for being accommodating and rescheduling on short notice. So let me say a few things about our guests tonight. Christoph Reinhardt is a building scientist and architectural educator working in the field of sustainable building design and environmental modeling. At MIT, he leads the Sustainable Design Lab, which is an interdisciplinary group with a grounding in architecture that develops design workflows, planning tools, and metrics to help evaluate the environmental performance of buildings and neighborhoods with respect to operational embodied energy use, walkability, daylining potential, some really interesting stuff. Products originating from the group, such as Diva, Mapdwell, Desim, and Umi are used in practice and education in over 90 countries. And I'm sure we'll hear a lot more about them tonight. Before joining MIT in 2012, Christoph led the Sustainable Design Concentration at Harvard Graduate School of Design, where the student forum voted him the 2009 teacher of the year out of 77 instructors in the department. And having been a student of his during that period, I can tell you that was entirely well-deserved. Prior to joining the faculty at Harvard, Christoph had worked as a staff scientist at the National Research Council of Canada and the Feinhofer Institute for Solar Energy Systems in Germany. He currently serves on the editorial board of several journals and has authored over a hundred peer-reviewed scientific articles, including a textbook on day lighting and four book chapters. His work has been supported by a variety of organizations, including the National Science Foundation, Autodesk, United Technology Corporation, Sage Electrochromics, Trans-Solar Climate Engineering, and the governments of Canada, Kuwait, and Saudi Arabia. Christoph's work has been recognized with various awards, among them the Arab Best Paper Prize at the Building Simulation 2009 and the 2010 Leon Gaster Prize from the Society of Light and Lighting. In 2013, Buildings for Change magazine voted him one of its inaugural stars of building and science, and the Sustainable Design Lab Spinoff MAPDWELL has been recognized with fast companies designed by Innovation 2015 Award for Data Visualization, as well as the Sustainia 100 Award. So with that, please join me in welcoming Christoph Reinhardt. Thanks, Nate. Good evening. Thanks, Nate. That was very nice. I think I have to update the website and see how the fix that we have on here. Thank you all for coming and joining us tonight for this lecture series. I think it's fantastic when schools of architecture engage always and have dedicated series related to sustainability as well, and hopefully you'll learn something new here since this is a school of architecture, even though the title speaks of Sustainable Urban Architecture, I effectively broke the talk into two components. In the first component, I mostly speak about our work related to the individual building level, and then we are spanning out and also including work at the neighborhood and whole city level in terms of an analysis. And as always, the research that I'm showing has been the work of many people, a lot of them listed at the bottom that worked with us in the Sustainable Design Lab or at the various companies that we spun out. Here you see us last year on our annual lab hike. To give you a sense of the profile of the individuals working on here, they are all architects by training but also have degrees in computer science and building science. So I try to work with people that have all three fields in one so that we can basically design the type of tools that we do design. And our common interest is that we want to change current practice whether it comes to architecture or urban planning. And what we mean by that, we're trying to observe being in a school of architecture and urban planning how currently the profession is working and then we are trying to introduce new workflows that are cognizant of current practice so that they easily can be implemented and then hopefully lead to better informed decision during building design. So who's Brett, I eat who song I sing? Usually we provide some background of the various companies that have supported us over the years. I guess there's a space for Tramco here. Yeah, just saying. No, so typically the way we are operating we've done a lot of work for basic organizations such as National Science Foundation but we find our work increasingly being working with energy companies such as Exelon and that is really a phenomenon not just in our world but in general that you can still design individual buildings but as buildings become larger and we are nearing a low carbon economy the relationship between the individual building and the surrounding grid becomes more and more integrated and that's just a reflection of us trying to basically understand this integration better and help designers to move productively forward there. So the background of what we are doing is a field called building performance simulation that's really the bread and butter so it's the computer-based attempt to model the various heat and mass flows around buildings. So what does this mean? It basically means that we want to use this awesome tool to predict our everyday experience, everything about us so the heat going into and out of the buildings the air, the sunlight that shines onto you and once we are able to basically mimic in a computer our everyday life experience then we can not only predict energy use in buildings but we can also derive new types of metrics such as the return on investment of a building or the comfort conditions in the building even walkability and that's the goal behind our work. So why do you wanna do that? Do you live in the virtual world in Sim City? Well, effectively there are two approaches for that. One is when you design a more complicated building that before the building's being built you wanna test whether it's actually working and it's code compliant. You may think of green building rating systems such as LEED from our end that's actually the really boring part that's just basically ticking off boxes so we really wanna use it in situ during design. And whether it's a neighborhood or just a facade in all cases the idea is really the same, right? That you come up with a concept, with a design and then you analyze it with one way or another. May, is it comfortable? Is there light? Is it energy efficient? And then, and that's the hard part you have to adapt your design if you love it how it looks or not, right? Or if you do so then at least these type of analysis has been productive and then you go around in circles. So how often does this actually happen? This is some work from us a couple of years ago where we asked couple of larger architecture firms that have in-house consultants and that do a lot of sustainable design work. So how often do simulation actually change your building, change your design? And you see two bars here on the right hand side this is the firms that outsourced the consulting and basically have firms that they pay for that at the end of the design process and check off the boxes. Not surprising there in many cases the simulations don't change the design which is just built in the relationship between the simulationist and the design firm that makes a lot of sense. But even the leading firms that have in-house consultants even there you have pretty high levels of times when somebody does an analysis and it doesn't do anything to the design. And that's really the number that we wanna change and the reason for this is it's not very easy to do this even if you do an analysis of a building very often the designer might not understand what it means doesn't know how to react to that and this productive discussion making it better is really what we try to do within our lab. So within today's presentation I have really three components. The first one is relatively short it's just to review and to convince you that the tools that we are talking about here actually work. So then the second part is about using and applying these tools at the individual building level for daylighting and for sustainable facade design. Sorry here, I have something popping up on my screen. And then finally we're spreading out to our work on all modeling itself. So the question first is to today's simulation performance simulation programs actually work. And a lot of this work has been done during the 90s already here you see this really nowadays we would find them cheesy boxes that people put in out and measure temperature and put in thermal slabs is the Hades of the solar architecture movement from the 70s and onwards. And there we developed effectively, well not we those before us developed the tools and methods that really allow us to very reliably predict if we understand everything about a building how the temperature in the buildings for example gonna change. And nowadays this has become pretty much an accepted industry. If you think of ASHRAE, ASHRAE has basically standards for both the simulation tools that you can use and you can even become a certified energy modeler. And that is nice but what is really nice is that the IRS also supports this. So when you have a building and you want to implement certain energy efficiency measures then obviously you're not gonna build this building twice once without the measure to show how much savings you have. So somewhat for your building you have to make these predictions. And the IRS accepts the use of these programs effectively when you say this is building is gonna save 20% of its energy use then you can write this difference of your taxes the energy savings. And this is ultimately the financial engine that drives a lot of this industry. This is all nice now if we look at the moment of truth and we look at the first generation of LEED certified buildings then what you see here is a plot that compares the simulated energy use and the measured energy use during design and first year operation. And if these tools were all nice and perfect then you wouldn't see any dots because everything would lie on the identity line. Which is obviously not the case. In fact the lab buildings are the ones you can basically keep it as a rule of thumb. The larger and more complex the building is the more energy it uses compared to what the prediction was. And that's not really surprising if you think about it in the computer the world is perfect. Every window is clean. Every firm starts to get set the way we wish and reality is not as clean. So there's some kind of discrepancy there which is why commissioning a building is very, very important. Well you could of course also say well maybe the tools don't work so in order just to address that we worked with a company called NRModal in Canada which has been the lead consultant for 30% of I think all lead certified buildings in Canada. And here you see effectively doing the design process the predictions of the simulation model and doing a commissioning process. So you see that this is slightly moving over towards the identity line. And that means on the one hand side that these tools work. It's also a new evolving industry or activity that when you have a complex building you actually have your building facilities manager run the simulations to see if your building actually does what it should be doing or not. So this is the energy part. When it comes to lighting that's of course something that's very dear to the heart of many designers. So when we talk about a lighting simulation then there can be a lot of confusion because in a way modes of simulating light there are many of them. So here you see a little study for example that's the top floor of Renzo Piano's Harvard Museum's extension in Cambridge. So this is the space that's being built. This is actually a sketch from the design team very early on in the process. And there they already tried to capture something of how the space is gonna feel. Then on the top right you see a typical architectural rendering of how the architects would communicate to the client how the space is gonna look. And now the physically based simulation looks similar but only here you can see every pixel in the image is a lighting measurement. So this is really the only one of the four that tries to communicate how bright in absolute terms is it actually going to be whereas the other ones are equally important, right? Obviously very often this lighting simulations are not very successful in communicating the mood of the space for example. But when it comes to us trying to predict really how bright this is gonna be this is what we're focusing on. And how good are we at doing so? There's been a whole series of studies over the years, over the decades and you can remember that these tools when you use them right can effectively model reality with a very high degree of accuracy. You have an error of only 20%. So if you take a light meter in a building and I model it right and I measure it then you're within 20% discrepancy which is really excellent because our eye acts as a logarithmic sensor so you couldn't say the difference within your eye as well. So we could say we can model reality if we do everything right good enough. And now we are moving towards not just modeling in luminances but modeling the whole luminance distribution in the field of view. And one of the recent efforts from our group is to make these tools faster. So those of you that are doing lighting architectural rendering probably know that these tools take forever and they're really far away from being real time. We developed for radiance in new version which is graphics card based and that's about a hundred times faster than radiance itself. But where we really wanna be is real time renderings because there's something interesting happening either you design with a building you do a lighting analysis and then you wait 10 minutes or an hour that interrupts your flow. If you really wanna be in the building and understand how the building works you want your results back in half a second. So this is our current prototype that's from Nathaniel Jones from our lab who just graduated, he works now at AROP. This is a real time glare rendering of the gun tall at GSD. So effectively you can walk through that space and whenever your meter here says it's too bright or not too bright then you get real time feedback as you design. And obviously we have to make everything now virtual reality. So with this you can actually put on your glasses and walk through the building and see right away where the glare source is within the building and then in real time you can change and see what's happening next. So the tools are available and they're becoming reaching a point where you can just use them while you're having your rhino or other cat environment open basically see how the actual building is gonna behave as well. What's coming next, this is something that we've spent the last years of working on this spectral rendering. So you might have heard about the ongoing life and health debate. Actually thinking about the Nobel Prize in medicine was just given out to the three researchers that found out the effects of light on our circadian rhythm in food flies. Well we are not food flies but we also have a clock in our brain. And there's a lot of interesting work going that basically when we trigger blue lights we can advance or delay our circadian clock. And so to be able to measure and predict this in buildings we've been working with Harvard Medical School on this method that allows us to model the interior of a building everywhere in the world any direction you look spectrally. So then we can overlay different action spectrum on top of that and see how alert is this place going to be for you. So if you wait a couple of years we'll be able to effectively model for a whole year and a space tell you how awake you will be right now in this building. Especially if we know how much you slept last night and so forth. So this kind of concludes part one. This was really just more of an overview and assure you that the tools that are gonna use in the following actually can mimic reality. So now when you have this awesome ability to model everything anyway where you want to what can you actually do with that? And when it comes to lighting and daylighting we're using this schema here. So when we wanna evaluate a space and say well is this a well-delayed comfortable space then we think of these three different concerns. So to call it daylight obviously there has to be daylight in the space so the daylight availability is key. Then depending on the lighting condition you encounter in a space there is we are doing this for the occupants. So we wanna understand visual comfort conditions within the space how people are going to react to that. And then there is another reaction going back because if the occupants don't like it they might leave or they might close the blinds turn on the lights. So you have this interesting relationship you're going on between the buildings that you design and the people in the buildings and then all of this falls into a new energy use. And this is a very interesting topic to understand. So the worst case you can encounter is this oh this is really dark here. So this is a typical European building where the occupants got beware can do everything they can open the window they can close the blinds and they can turn on the lighting. And so we somehow have to predict and that's a messy business what people are gonna do in buildings. And then all you pick the other extreme this is the Rolex center by Sana in Switzerland where the occupants can't do anything effectively. The clockwork is basically setting the blinds at every moment in time to create adequate lighting conditions. Or if that's too complicated you can just think of more architectural control. You find that in a lot of libraries where there are no moving parts here in New York City Public Library the form and the shape of the building just you have to get it right so that by itself it can accommodate all the daylight conditions over the year and not create any glare within the building. And maybe as a third alternative, if you go back to the Sana building that's actually a building without any real furniture. If you've been there it's a really great building. It's a university workspace of sorts but there are only beam backs in the space. So if you feel glare anywhere pick up your beam back and leave, right? So I mean this is basically you're flipping it around rather than having very constrained conditions where everybody sits at their workspace and we have to control lighting there. We can also just open it up and leave it up to the users to do whatever they want, right? So now quickly, how do you actually evaluate the daylight availability part? So we have models for this. Obviously we can model controls and we have even some models that predict how people interact with the space. So how do I mimic daylighting? And this is some work that has been done by now eight, nine, 10 years ago. It's now in the lead green building rating system. It's our effort to say when I look at a building where's the daylight and where's it not daylight? That's of course not really a binary thing because when you go in a lit space it's not that anywhere within your building the lighting completely disappears. But you can say at one point it's just not as useful anymore. So we're gonna say, well this is where we need the daylight area ends. And what has been done a couple of years ago is basically that we develop computer methods that allow us to model for the whole year for every hour in the year how much lighting levels there are. And then if we say we need 300 lux we wanna know how often we reach that level. And that's called daylight autonomy. So for example here so you see for south facing space or I think I have to stay here. The daylight autonomy levels away from the window. So easy to remember when you're right sitting near a window then the designer cannot mess it up. You basically over 95% of the time will get lighting levels. It becomes more interesting if you move further and further away. And at the 50% mark this effectively means half of the time when people are in the building there's plenty of light. And that we consider a daylight space. Well the question then of course is just because we say there's a soul how do people in buildings perceive this? In order to understand that we first did a study with architecture students it's the Carpenter Center at Harvard by Le Corbusier. So here this is the second for studio space. It's actually more daylight than the image suggests. And we gave everybody a floor plan. Here's the window, here's the floor plan. Just set draw the line where you think the lighting ends. And that we've done with a lot of people because everybody thinks differently and this is the kind of results that you get. So everybody draws their own line because this is obviously a subjective evaluation. But what you see is when you take groups from different years together we did that once in the spring and once in the winter. The mean result is surprisingly constant. And in this case we were really lucky because when we ran our simulations on top of it we nearly had the same result. So there's one space in the world where this works. And in order to understand whether there's more than one spaces we worked at that time with I think 15 universities and asked them to do the same thing. And I don't want to go into lots of details but effectively what we found this really strong correlation between the predicted percentage of the daylight area and what these students were saying about a variety of spaces. So this paper I think ultimately got daylight autonomy into the lead green building rating system. It shouldn't necessarily work. It's in a way surprising that it works so well because when you look at a space the lighting is really more a matter of contrast. When the window is very bright there's this constant fall of weight becomes so dark that comparatively if you look at the front wall and the back wall you feel here it ends. So it's surprising that an absolute simulation such as this one works but so far so good. And if you wanna apply it it's actually relatively easy to do this type of simulation. So if in your next studio project you would just wanna say this is a daylight space and you just have to run this annual calculation and then you see where the daylighting ends. And it doesn't mean it's gonna be great daylighting but if you don't have in your model when you set it up bright any light then there's no light in the space. So then probably anything else that you deduct from there on is not really gonna work. So that's relatively easy. One could need to say this is solved in terms of an analysis. The tricky part is really how do we deal with visual comfort and comfort in general in spaces. And typically what we nowadays say is kind of nihilistic where we say you are comfortable if you don't complain and you maybe have a view to the outside. So as long as you don't have experienced severe glare and we have some connections to the outside we are already happy. You can liken that for those of you that have studied or learned about thermal comfort in the 60s and 70s. We also said thermal comfort is the absence of people complaining. Nowadays we actually try to get more towards people actively expressing a liking for the environment and in lighting we are just not there yet. So if you follow this what can you actually do? There's some interesting work from Denmark and Germany where they did comparisons of daily spaces, did hydamic range photography and had people in spaces and effectively they came up with an algorithm that you can apply which is called daylight glare probability. And what you do with that you can effectively take a photograph in a certain way of a space and then run this algorithm on top of it and it's gonna tell you the likelihood that somebody's gonna experience discomfort glare or not. So the nice thing about this is and this is how we use it for teaching that we just ask our students or other designers to take pictures of spaces that they really like and that they don't like just with a hydodynamic range photography and then they get a sense of these glare levels. And the nice thing is that within our simulation world we create the exact same images. So we basically have a way now to go back and forth between reality and simulations. And what is basically not rewarded you don't wanna have too much contrast in the field of view and you don't have too much to watch bright lights in the field of view. Now the tricky thing is of course when you talk about visual comfort is that changes all the time. So again we are in the situation where we can say you for one moment in time you'll be fine but what about all the 4,000 other hours of the year when it's daylight? And what about if we look in different directions? So here for example you see a side lit space this is south facing and just when somebody looks towards the window you experience glare when you look away you don't. That makes total sense. So what one would practically say well don't look to the window, right? And that's of course what people do in spaces there if you give them the opportunity think of the sauna space they adjust themselves to not experience glare. So we somewhat have to basically not over be too afraid of glare we just have to provide opportunities to avoid it effectively. So here you see something like an analysis like this looks so here we basically tell the occupant you have to look straight the whole year and then here this is an annual map and whenever it's orange or yellow that's in the middle of the day this is January 1st till the end of December it's glaring. So this is a terrible space unless you let people adjust but the moment when they adjust this is not so bad at all. So when it comes to doing an analysis of the visual comfort conditions we effectively provide or give people the ability to adjust themselves and move and then when we look back at our concept we can basically look at how much light there is when do they feel comfortable and how do they control their energy. And here's a little example study for four different facades how you could apply something like this and we're using a facade effectively one with exterior Venetian blinds which is more of our default setting one where you can actually adjust the blinds in the upper part and the lower part independently one with a nano gel which is effectively a system that has very high R values very high insulating properties and lets to diffuse light and a switchable electrochromic blazing and with a shading system and this work is a few years ago old but you might like it a lot of designers are into static shading system because that's a really strong statement and I see you are into this as well looking at your school you have this east facing shading system. So how do we design an east facing shading system properly? If it's a south facing facade and you want to do a very clean simple solution then we know since the 1957 when Orge and Orge published their book on static shading system how to design just a static overhead. But of course if you look at a building like here that's Genie Gang's Aqua Tower there's this desire for more complexity so in a way how do you design something like this that actually works, right? So in order to do so this is work by from our group John Sargent and Jeff Niemers that was their thesis at the GSD they developed a method that we call shader rate and effectively how that works is you have a space and you can put whatever you wanted shaped shading system that's standing in front of it and then what the analysis does it meshes your shading and then colors it for you so it's purely visual feedback and it basically says when it's blue it's good, when it's red it's bad meaning that this is actually a liability which is not surprising when you have this funny hatch in front of the building and it's so big then you take away too much light and where it's white it doesn't matter. And that's of course key this is the kind of thing where we try to provide a feedback that a designer can work with, right? So in this case if you really wanted to do something you just cut away the red and maybe consider the funny form here. So how can you apply something like this in actually a whole building? So what you see here is the top floor of the Aqua Tower in Chicago and here this part of the animation just goes through the notion again. So we always wanna know for all of these different patterns these different blocks is it good to have a balcony there for the floor above or not? And basically as to running through this you need in order to do this analysis you have to run a full thermal simulation for the floor with and without this individual shading system and then you can decide is it good or bad for the building. And then you can color the whole thing in and basically wherever ever it's blue you want a balcony and where it's red you don't want a balcony. And it becomes interesting because now you can use that for form finding. So in a case we say the inflection point is the point where the balcony ends. So now I basically have my solar optimized balcony and then I just do that for the whole building. So you basically see this whole thing coming down and now there comes their trademark move which is the time that we move up the city. So when Chicago is coming up you basically see the building reacting and the balcony reacting to the building and then you can really design back and forth so you can move your windows up and down and you see the whole thing reacting to that effectively. So this is really then circular sustainable optimized design. Yeah and when you go to Buenos Aires the building obviously becomes fat because you're closer to the equator and you want to avoid shading and when you go to anchorage that doesn't make a lot of sense. So this is basically an example where you can use computational design and environmental performance and link them together. So if we go back to our little dinky self facing office then here you see the type of analysis if you were to do that with a simple shading system in Boston if we have split lines we have static shading, a shader rate system a translucent aerogel or you have a switchable glazing and effectively really what works best is a split blind. So what does this mean really in the case of our climate here being able to have a shading system that adjusts in the upper part independently from the lower part is really the best you can do for your system because otherwise our regular lines go down all the way and by the time you block your glare you block all the daylight. So having some cleverer systems really worthwhile the static shading unfortunately even though it was so cool in the way it looks doesn't really do that much in this climate why is that because this is a cold climate right? So you don't have that much benefit from a static shading system even worse we talked about that in the car actually coming Boston is the worst climate for a static shading system because in March and September the sun is at the same position but if you've been out in March and September then the needs are very different right? You are still cooling in September and you're heating in March so having anything there that doesn't move creates by per se it's a conflict that you can't solve you can optimize it but the optimization doesn't mean that you're really saving a lot whereas when you do that in a hotter or in a colder climate you get way more clearer solutions there. So this is the framework that we're working on just in case you wanna try some of these things we have a plugin that's called Diva for Rhino unfortunately if who's using Rhino here? Yeah well not unfortunately awesome so if you use Rhino you can just use the plugin it works for Rhino and for Grasshopper and this is how we do all our simulations here which gets me to the next topic when you wanna do something like this how actually do you learn to do a proper simulation and that's not that easy to teach as well so the first time I taught this we used a tool called EcoTech that probably many of you don't even know or remember which was really phenomenally elegant at the time and I just gave it McGill an assignment model your crit room and I thought no one in the world could possibly get that wrong and I was wrong so here you see basically from the two years when I taught this this is the good solution and these are all the assignments that were submitted so completely and utterly all over the place and that is Mary's reasons that is software compatibility between the environment you work in that means that you don't really know how the simulation environment works but it also means that we have to when you wanna use this to have a basic understanding of the algorithms behind them and what they do so we got a lot better results nowadays we have these two books out well the first one will be out in the spring and this one has been out for a while but they basically take you step by step we have a lot of tutorials six or seven hours on our website how you actually learn to do a simulation and here you see a comparison from the MIT students when they did the same assignment and we got way better results but they only did it after going through all these different steps so just giving you a sense how do we teach that this is the exercise in the first week and this is really trust building so you have to get an object and model it in Rhino and you put it in the sun and take a picture twice in the day and then hopefully you can reproduce your shadow in Rhino right away and obviously that should work and if it doesn't work then you know that you're building probably you don't get it right then the orientation is wrong or so but this is a really strong first step and it's also helping in terms of trust building and then we go through a series of other things at the very end there's a student assignment where you take a picture of a space and you do a high dynamic range version of that picture you then do a simulation and hopefully the two are very similar so for lighting we can really get the lighting the digital world and the visual world the real world really close together so for energy that's a lot harder because people can't see energy you can't walk in a building and if somebody asks you is that a good building you can look at certain indications such as when the windows are leaking or it's a single pane glazing it's probably not awesome but even if it's a regular put together building it's very hard to judge so how do we basically help you get an intuition for energy and for that we started a game the first time we did that was six years ago and there that was the original game so we had this human cluster of experts that were sitting there and then groups of students had a simulation order form did you play that game mate? I think you missed it, right? Yeah it was really fun so basically you have this it's like an ordering paper that you get and you have a budget so you can basically pay for window upgrades all kinds of things we could spend at the time you were allowed eight GSD dollars to spend over all on your upgrades and then you get an energy use so basically here you see the students with their shopping list and then somebody ran the simulation and effectively the idea was to say how low can you get energy use and that was my favorite I got an email in that evening I think it worked well that today the class was definitely not boring just kind of backfiring compliments so what about all the other classes but usually I take what I get so we are still playing nowadays if you want to do this we call it the Diva simulation game now actually we let everybody do the simulations themselves so the way this works is you work in Grasshopper and in Rhino you can do a massing study the Grasshopper component calculates the price of the building for you and again this time you can spend 50 Diva dollars effectively out of 100 and it's about how do you design what is the best combination so when you play the game you get a different climate and a different size of the building and you have to work based on that and the way it works every time you run a model it keeps track of your model so you can basically run over 90 minutes maybe 30, 40 variants and then you can always go back and click on the one that works best and this is I think generally how you want to start designing that as you try things you keep a track of what you've done so that you go back and you actually learn from it so this, sorry I keep on having windows popping on my side so this works pretty well it's easy to learn again it comes with some tutorials so if you want to try it in teaching or individually it's a lot of fun and finally if you want to join us once a year we do Diva Day somewhere the next one is at UC Berkeley and we have student competition and the winning student group that successfully applies these tools in their studio project we usually fly them wherever we are and let the students present so try it out the next one is probably in Rome the next one is the one in a year as a motivation for all of us to go to Rome so now I showed you a bit we have some tools that can be reliably applied and we can apply them in sustainable building design and facade design so now the next part is really what our lab has mostly been working on for the last three, four years which is the question of urban modeling so what does it mean, urban modeling? why are we interested in that? well obviously humanity is moving into cities more than 50% of the world's population is living in cities if you look at this in numbers we effectively have to house net two million new city dwellers every week and if we were to house them all in Boston then we need around 400,000 new buildings every week and there are about 100,000 lead buildings in the world so if we just put these two numbers together to keep up with all the building activity that's going on we have to look at more urban level performing and of course cities become denser and denser because we don't just want to spread out we have to densify so we have to really understand how cities work and here just looking at some examples for urban living there are of course multiple variants informal settlements in Rio very planned high density buildings in Shanghai you have plus zero developments in central Europe back bay in Boston the endless suburbia of Las Vegas and the single family home somewhere in the US so if we were only looking at energy use in these buildings then of course these two buildings were the winners but of course it doesn't make sense to use this as an example for future urban living so our goal has been since then to develop this tool that we call UMI the Urban Modeling Interface which basically allows you it's kind of a diva for the city it allows you to model whole cities looking at operational energy use, mobility, comfort, daylight finance, embodied energy and urban agriculture and I want to show you a few of these things if you are going to use the tool probably better send me an email because our internal version is a lot more powerful than what we have publicly online so our first effort in this realm was to combine big data with building performance simulation and for that we started with Cambridge and the background was really Boston had a solar map and Cambridge wanted one and I said we can do a better map than Boston very important that actually works so this is the MIT campus this is LIDAR data who is using LIDAR and GIS data here? some of you these are data sets that if you are in urban planning are very common you get them for free from many parts of the world so as you see here LIDAR data is basically our point clouds that can automatically from an airplane be generated out of which you can create a model of a city and nowadays LIDAR data is such high resolution that you see this looks pretty full compared to the level of details that you get nowadays so we can then basically run for a whole city for every building for every hour of a year solar analysis until you really exactly for how each building is going to do for solar photovoltaic so this is this map tool where that works in eight cities right now you can click on an individual address and then you see your rooftop you can then draw on your rooftop and gives you with all state incentives exactly how much payback you get for your PV system how much it's going to cost how many trees we're saving and so forth so the idea is really here that we again start changing this time not the minds of architects but of the general population living in a city so that in five minutes they decide how much they pay when they get their money back and just to test how well something like this work we went to a well fleet on Cape Cod that probably some of you know so it's a small town on the Cape where we did the solar map and we worked with the Solarize Code and the Solarize Code reached out and 80% of everybody living in that town looked at the map, contacted the installers in 40% of the time the installers looked at the map and said your roof is bad we're not even coming so what doesn't make the roof bad? it faces the wrong way it's shaded by trees there are lots of reasons why a roof isn't good when the installers came in 94% of the time they made an offer and incredibly 54% of the time the owner said yes so that really means that the trends act and so then after four months 10% of the households bought a solar system in this case which is about seven times higher than any other Solarize program in New England that was run at the time so the reason is really that we want to get to a level where when you look at your building everybody gets the same information and it's reliable should I get a new furnace? should I get anything new? we should do our best to predict the performance and then hopefully people are gonna do so or not that's really the goal behind this this is our urban modeling interface it's again, it's Rhino-based links Rhino to a database so you can just draw your city in Rhino and now we want to do an urban energy model so how do you do that? we call that open the urban building energy model well you need various data sets if you want to model the whole of Bristol Providence then you need weather that's pretty easy to get you need the geometry and you need to describe the physical properties of all of your buildings and that's the tricky part but when you have that then you can build a model of a city so this is the GIS model of Boston and then for every model we extended it we guessed how many floors there are where the windows are what the type of building is and then we ended up with an energy model for every building in the city and that allowed us then to do this more blanket predictions in Boston for every building 80,000 buildings how much energy is being used and here you see how you can apply this so here we use if you know Boston this is Back Bay Prudential Center MIT is on the other side of the river so this is our prediction for electricity use hourly for the hottest day in the year and this is if 30% of all Bostonians there would put PV on their rooftop and what you see is this mismatch this is called the duck kerf even though I have a hard time seeing the duck but what this is, this dark gray this is the resulting load profile and the reason why it looks like a duck of sorts is because our electricity use peaks late in the afternoon because this is when you are home have the air conditioning on, you're at work have the air conditioning on and this is just where we use most but the PV peaks in the middle of the day so if you are utility and this really goes back to my earlier comments you are not just doing your utility a favor by putting PV on your roof because the utility always has to match supply and demand and this is incredibly painful that if in half an hour you have to ramp up and double your output that's not great if you are utility so you are not necessarily creating value here for society if we have too much PV and we don't orchestra that in meaningful ways and these are tools to work with that so an example of what you can actually do is in this case we assume that everybody has a nest thermostat and we are big brother or big sister as a utility we change the thermostat settings of everybody you want by up to three degrees Celsius and when you do that you can basically use all buildings as a thermal battery because it really doesn't matter if you're in a building just for 20 minutes somebody relaxes your thermostat setting nothing is happening but when you're doing so you turn off the air conditioning piece by piece in different buildings and you basically you tunnel through that peak so doing this type of things where you have basically a fully integrated system you can create enormous value as you reduce the peak by as much as all this photovoltaic at really no extra cost and Comet in New York actually started a program right now where they give certain clients rebates if they let them control even in reason obviously the thermostat if you're critical which I hope you are and you're taught to be you might say well how reliable are these models could be complete baloney given that an individual building energy model cost $15,000 for somebody to build so we want to do that same thing for 80,000 buildings so in order to understand how good these things actually work we went into the desert in Kuwait where our funders work and we wanted to model these builders here so these are all very similar single-family residential homes in Kuwait and this is the energy use intensity so the normalized energy use that the range is enormous but it also shows you which is generally two in residential construction occupant behavior really matters so you can double triple the energy use of a building and what are the reasons for that well some of the reasons might be you have more kids you have more people living in the building that obviously adds to the energy use or everybody has a TV everybody sets the thermostat to 63 degree Fahrenheit in winter and in the summer so usage behavior matters so how do we want to basically get this whole variety and understand it so when we first model this here you see the total energy use intensity distribution here you see when you have one building type one occupant type we effectively model a city of robots and then you don't get the spread that you want we then went into the city and we looked at different types so which were the buildings that were new old and renovated and then that got us a little closer and finally we started something that's called the Bayesian calibration where we throw 800 different stories at each building so we say you have 12 kids you have a family, a Dinkley family you have a lot of TVs, you have no TVs so we throw in this calibration process that at each building and we only keep the stories that fit the measured energy use for a few buildings from that we learn so you can think about it we create basically behavior profiles for a neighborhood and once you have this then you can model the whole neighborhood very well we did that for the neighborhood that we modeled first and then for two neighboring neighborhoods and that works very well then we went into Cambridge because you could say well this is in Kuwait what about here so then in Cambridge, Massachusetts we got energy use for two and a half thousand buildings and again we could model it very very well if we have that so what does this mean we cannot I cannot tell you when I see a building on the street how much energy it uses because I don't know it's living there but when I look at hundreds of buildings like this then the distribution the statistics work in your favor and I can tell you the mean for a hundred buildings what they use very well and then of course when your utility is all that matters so now we spend all this time and create a calibrated model what do you do with the model? well you can use that for energy policy decisions again we went back to Kuwait we assembled the twenty more influential people in the country that are responsible for energy use from industry and from government and we said well we could reduce the energy use of your whole city slash country by eighty percent and then the discussion became really animated and interesting because then would this pay for itself which in the case of Kuwait it doesn't because it cost them sixty cents a kilowatt hour to produce and they sell it for a cent that's in many parts of the Middle East the case so that's not a winning financial proposition so in that case it doesn't work but they really got into tiered pricing system and in many parts of the world that's now a way to channel basically the transition to paying real energy costs which means when you pay when you use a lot of energy as a household then you probably can afford it to pay more and when we this is our tiered pricing system that they proposed in Kuwait and here we could effectively show that the payback times are pathetic longer than the lifetime of some people and you really need the tiered pricing system to make any dent there and so at least they proposed it let's see what's gonna happen getting closer to home what else can you do with this models here we looked at a neighborhood that some of you might know that's the Dudley Square Triangle in Southern Boston that's a neighborhood with a rich history a lot of local neighborhood groups engagement and we wanted to understand how could a city reduce in this neighborhood carbon emissions on a large scale so for that we use census data and we basically know how many people it live in the neighborhoods per block that's public data that own the buildings and how many elderly couples, families, young professionals so we model these behaviors and through that at all the buildings and we could effectively show these are current energy users this is by how much energy use could be reduced if we either implement energy saving measures that you can do anytime so ECM-1 is weather stripping buying a thermostat things, getting new LED systems things where you just go to Home Depot once and do and then ECM-2 is when you open your pocketbook and you don't do that that often that's when you get new double glazings and you do insulation to your buildings and so forth and this is the typical numbers that you would expect of course it becomes really interesting when you see how does a government support this and right now in Massachusetts there are 93 incentive programs to get homeowners to invest in this the uptake is pathetic it's three or four percent because we know it's always the well-educated homeowners that do it and here you basically see these are the goals for the city for 2020 and 2080 if only we keep on doing what we're doing right now we don't get anywhere, right? So we basically have to completely change our energy policy and incentive programs to reach the other demographics as well so this I think becomes really interesting as a tool for cities if they really want to engage with it so a few examples now what else can we do going beyond operational energy use one key component is of course at the urban level also daylight and light to health so we have a module that we call Urban Daylight that was developed by Timo Dogen from our group who is now at Cornell so this allows us to do basically this availability analysis for our whole city very quickly and last year New York City's zoning laws turned a hundred and years old you might have heard that in your history classes in New York City was really the first city that provided zoning laws for step backs because you want access to fresh air and light to people in the city and if you think of the Rockefeller Center that's basically built according to the maximum volume buildable at the time and the laws haven't really changed in New York in a hundred years so we wanted to see if we could do that with something better so here you see basically this is a typical Manhattan block and we had five different archetypes and then we grew the block bigger and bigger and bigger to FARs over 30 which is enormous and we just wanted to see can you really hyper densify without killing the daylight throughout the whole city and so these are the three lines the different density that I'm now gonna show you here so these are the three different types the top line are actually this very thin towers that you actually see south of Central Park popping up right now everywhere so commercial base at the Bordenbury thin tower coming up and what you see here, they perform best this is the maximum density that you can build in New York and this is the lead requirement so with these thin towers you can basically keep going to an FAR of 28 and you would still not really impediment the daylighting around you so interestingly if you convert that into money you would basically add an extra billion dollars a block in Manhattan if you build something like this so I mean in this case this is really a win-win situation because we keep people in cities, we densify and if you build this thing you can make a lot of profit here as well but this is basically another example where we can just use this type of analysis to see what makes sense from an environmental performance standpoint the last project, the second to last that I wanted to show is really we call it the daylighting the millennials more looking at how do people use buildings nowadays and you guys are probably the same if I want to meet people from my group I have a better chance seeing them at Starbucks than in the lab because everybody is working at various coffee bars so we try to better understand especially when the weather is nice so you have nice buildings here the Koch Center, beautiful dated spaces but nobody is in the building so how do we actually predict at an urban level where people are when and for that I mean in a case like this this is the MIT Northcourt our parks and benches are very popular so we wanted to get a better understanding between these relationships and design and plant cities and areas around buildings to keep people comfortable so for that we basically zoom out here and this is the Stator Center this is the Koch building and in there are two big public cafeterias and here are some benches so people really have the choice to sit inside or outside so it's an ideal spot to just see when do people choose to be where and what we then did is we used Wi-Fi scanners so what a Wi-Fi scanner is it's a device that listens and your cell phone keeps on screaming all the time I'm a phone and my name is and it gives a number your phone does that nonstop and with a Wi-Fi scanner you effectively can capture that and that's legal what is also legal which is even more disconcerting is you can have a device that pretends that it's a router and then it says, hey phone I'm a router and when your phone connects to that basically the router can see everywhere where you locked in all the systems so that's great if you are a company because you can see where you are where your clients shop but that's really something that has to be changed so the point of this project was really and we had to go through a lot of through a lot of committees at MIT to get the permission really to understand how can we maintain the privacy of individuals how can we create value as city governments without having too much influence on the privacy of individuals so that was the idea behind it with this encrypted ID so we never got even the number we got an encrypted MAC ID for everybody and here are our devices we work with a company called Sufa and we collected this this is what we got between July and May last year to this year so we connected in this court 30 million Wi-Fi popes after cleaning them up we had 14 million we threw away the weekends because we were interested in workdays so 10 million popes then we could throw out then it becomes interesting because from the profile you can see what are people and what is stuff stuff such as just outside routers other devices sprinkler systems so we were left with 2.1 billion signals from over 600,000 visitors because a lot of people visit MIT every day and then the rest these were 16,000 regulars so you can effectively based on this say who comes there and sits on that bench all the time or is on campus all the time and now with something like this you can basically link these decisions to outdoor comfort conditions because when you're a visitor you just go if it rains it's not that you're turning around and leaving you're just visiting there for the day but if you work there you make a free choice so here then we did an analysis with a metric called the universal thermal climate index this has been developed over the last 10 years in various mostly European organizations that's a metric that predicts how comfortable people are feeling outside and they had done this type of analysis with maybe 100, 150 people so basically we do this now with 16,000 people so here you see effectively there's an example calculation when you run the simulation you see when it's green it's good and when it's wet we think you won't like it it's not comfortable there and when it gets blue which in this case it doesn't go then you don't like it because you're cold and this is what we got for 16,000 people basically this really nice curve this is how many people are out and this is how the UCCI conditions are so this really helps us to say we now have a means to predicting where outside people will be comfortable and that you can use for all kinds of things we can for example just create little micro areas within a court to make it more comfortable so we have the summer corner we have the winter corner to basically help us keep people outside for a longer time we can also use it for more mundane purposes so here this shows you it's hard for you to see basically this tells you the medium time how long my MIT colleagues and I have lunch outside so when the weather is really lousy the mean lunch break is eight minutes and when it's really great it realizes to 12 minutes fortunately there are some same people that have lunch then in the summer up to 20 minutes and I'm dying to do the same analysis in another city with another population but of course more seriously with this you can effectively also see when people are dwelling where so if you want to have for retail purposes we say we're at the best cafes and so forth you can do that as well so now how do we put that together how do we house these guys or do we provide lighting for them well effectively we are promoting now that you do a full day lighting you do an outside analysis and then you merge these two ways and find the times when it's best to be in the building outside the building so you basically try to august the lighting and lighting condition for people and this is just an example from Rohan Bayoumi from our lab that worked on this in Kuwait and tried to create basically navels that try to mimic that so you see it's very open urban fabric that tries to create a dialogue between the inside and the outside overall what we want to do is for all navels we are creating knowledge scorecards so when we do a class in the spring that's called modeling urban energy flows different aspects of a neighborhood and then it's like your Pokemon card every navel proposal have different strengths and weaknesses effectively this was one that I really like this was a project that we worked on actually for this in downtown Lisbon and this was a group that was really interested in hydroponics urban agriculture how much food can you grow in cities and at the time I thought it's a cute idea but I thought I mean the famous tomatoes in the double skin facade that doesn't give you a lot of food but they've effectively looked at hydroponic systems on rooftops and PV systems so what's better for a city should I put photovoltaic on rooftops or should I grow food and even there we have no models that tell you how much jobs you're creating and it's really stunning so here's an example this is Gotham Greens in Brooklyn, New York these systems here which are hydroponics have incredible efficiencies you 70 times less water than regular system it's local food that's sold down there for Lisbon we effectively showed that with unused land areas in downtown Lisbon you could cover more than 100% of the food needs of the main vegetable groups for all residents and we checked for Lisbon, New York Paris and Singapore it's even better in terms of carbon emissions to grow our tomatoes in a basement with an LED than what we currently do and the reason for that is that the food miles traveled and that includes you for all the tomatoes that you eat over the year the average tomato travels 3000 miles to New York because in the winter we keep on eating tomatoes and then they come from Chile so in a way I'm not sure yet how the finances are panning out but there are more and more people really interested in that and rightfully so it really changes I think the perception of how much food we can grow in cities of course if you think back a few hundred years cities were largely feeding themselves but with these technologies we can actually bring that back and as I said you can see a lot of those from a carbon emission standpoint it even makes sense brings a lot of jobs into cities but it's obviously also really disrupt our whole food production in the world how it's going to work so this ends really this overview of what we're doing I hope I could convince you that there are some great tools and that are usable enough that you can use them doing design be it for an individual building I think the larger question is really that our goal is to use this tool for evidence based design so when you make a claim in your studio project or otherwise that you can somewhat back that up and show does this really work not the famous lines this is my natural ventilation building and this is how the air is flowing but really does this work or not with these new urban methods we can really start developing long term strategies and I think the times are over when I just say I built my green building and I have PV and it's all fine we really have to think of the building as part of a larger urban infrastructure energy infrastructure and that is happening and you will see that reflected in incentive programs to make these tools usable we try to educate stakeholders, practitioners and you and hopefully you're going to pick it up when it comes to daylighting I think it's important also to think about it's not just your building you also want to basically create comfortable outside conditions right outside of your buildings and this is actually where the big challenge comes in because it's very easy for us to know who's going to use it by the building level tools there's really no protagonist right now in cities that does this because municipalities don't have the money necessary to do that but really we're talking here about this meta-scale where we know in pork about 100 buildings we really talk about the urban canyon and how do you make the urban canyon liveable and that's very hard to do I think European architecture tends to do this more I think this is something where we have to catch up a little bit here and hopefully we see that more and more so where do I see all of this going well I think from our end we will have the building level tools and the main part is really that we're linking all these tools to our smartphone apps we are doing this a lot of people doing this so that effectively the energy system goes all the way to you and tells you to change your thermostat settings or maybe it tells my watch tells me that my favorite park bench is currently free and it's comfortable so you could sit there and I'm sure that's going to happen more and more that it basically becomes that's exciting because designers are good at designing GUI so hopefully some of you are picking that up as well and with that I thank you I think that you would be using the tools to simulate that and to see the daylight that was in the interior spaces when those were used on exterior for example because sometimes you see designs where the courtyard portions are so high that maybe an additional bed the tools can be used anywhere and we've done analysis of outdoor courts indoor courts the tools are completely agnostic so you can just if you model it in CAD these ray tracers that lie in the background don't care if you're inside or outside it's actually going to be easier when you're outside because it's less complex there are less bounces in the world of outside lighting whereas inside you have to wait longer but yeah that's possible I'm kind of curious why there was more buy-in from New York City in terms of the ability Well that's a good question I think there's a really important discussion going on and in Boston we have the same thing what the shadows of these buildings are on parks right so I didn't talk about that I talked about the daylight in that spaces and the immediate surroundings right now you see in the New York Times in fact there's a lot of debate on that and in New York in Boston we have the exact same south of the common argument right now well I think actually the current bylaws are not that effective because they are more looking at when a shadow falls on a piece of parkland I think we can be way more nuanced and with tools such as the UTCI you can actually do that and I think you have to really look at it under perspective so if you have a slim tower then actually shadows move quicker than we think and the less monolithic a building is the less disruptive is a shadow but that's a very very contagious issue right now I think it's partly so contagious and this is why we did this first study here because the evaluation metrics are not out there we really stuck with this all simple shading studies and it's not that simple and I think you see a lot more of that going forward I think we will some of our next studies will just ask people in various parks how they feel about certain situations so that we can mimic it but also the UTCI I'm a big fan of this metric since we showed that it mimics this so well because it looks at more it's not just shadow it's wind it's the mean radiant temperature of everything around us that all influences ultimately how you feel in space yeah perfectly you mean for lighting or for day lighting for no I mean this is of course a very strong ongoing or long ongoing debate that so all standards in the world use a light meter and judge legally speaking the quality or the legitimacy of a space how much light falls on a desk that has nothing to do with how we experience the space so taking luminance based metrics metrics that take view into account would be a lot better but it's also a lot harder we have tried that now for decades to not do it so I think one should look at both spectra one should just do we have any light within the space and is that enough because if I'm under this levels I have some issues but then we need to go on and there's some really interesting work right now there's visual interest metrics that come out of Switzerland right now so there are a lot of new the visual comfort metrics new ways to look at the space on top of just do I have enough lights it will be very hard to get over that I think but again if you don't have the levels then it's probably too dark right and then of course it's good to say is it an appropriate level so if it's a church you obviously don't need 300 lux that's where you have to maybe start yeah so I mean so much what you do is it's covered uses for evidence but how do you see your word word in general oh I love full of thumbs you know that right I didn't talk no I think rules of thumb are absolutely key if we did a survey years ago asking you know eminent architects how they design and the number one responses always oh experience from previous work now it's of course too bad if it's your first building because you don't have that experience so I think these tools are a good way to get you started and that's my favorite quote from from I don't trust any simulation unless I know the results before I run the simulations and I go by a similar metric so if it's not in the right ballpark there's so many ways to get it wrong I probably needed so a solid understanding of what typical numbers are and what you expect is very key so in the case of light that means if you really want to become a light expert and you want to have some quantitative component in your work then taking the Titan emigrange images whenever you are in a space that you like or don't like and get a sense these are good levels that's super useful and that becomes a rule of thumb at one point and of course we have in that handbook a whole series of rules of thumb that get you going and the rules are thumb are important because ultimately the computer doesn't design you design you have to start with something and the computer helps you tweak certain things so it asks more advanced questions Hi Partially I kind of wrote off to that You mentioned that there were some barriers to get over in terms of when the simulation process is built in a kind of molecular setting and those sometimes didn't match up with an ideal setting I'm curious since I say 25 to 30 years from now we have an ideal scenario in which all students in architecture are getting the information what does that look like do you think students and architects at all scales whether they are residential or small-scale are they getting into computers I was wondering what do you ideal do you think that process might look like well it's not 50 years it's in three or four years no I think in many schools of architecture modeling cat three dimension is there and these tools become faster and faster the graphical user interfaces become better I think you will see people in real times moving shaping the building and you see I always think about the ideal cases like when you do your taxes you really want to know do I make money or not so you need this little dashboard and as you design along you get a virtual slap on the finger if you do something where you don't meet code anymore but I think that's where we going to end up and I think you see that more and more you know from the get go if it's going to work and these tools help you to develop this intuition and closer and closer I think it's important to have some basic understandings of how these tools work but we do all of this in the first semester as part of the NAAB class on environmental technologies so I think you can teach this tools along with the basics and I think it's faster than 50 years and these are used in practice even though it's always interesting for us in practice then to our Diva days we have a lot of consultants coming from the larger architecture firms where kind of the sustainability consultants within the large architecture firm do this thank you for coming