Partnering with Stantec to Develop AI-Based Total Carbon Analyses

ELLIS HERMAN: Hi, thanks for coming to our presentation-- "Partnering with Stantec to develop AI-based total carbon analyses". I am Ellis Herman. I'm a product manager working on Forma.

CATHY TRUONG: Hi, I'm Cathy Truong. I'm a product designer working on Forma.

JEN COOPER: Hi, I'm Jen Cooper. I am a user researcher working on Forma.

MIKE DEORSEY: And hi, I'm Mike DeOrsey, digital practice manager and an architect at Stantec.

ELLIS HERMAN: So we're going to start off with the safe harbor statement. Most of what you'll see in this presentation will be forward looking and under active development. So please don't make any financial decisions based on this content.

Today, I'm going to start by giving some context on carbon and on Forma. Then Mike will talk about why Stantec wanted to participate in this product discovery work. Jen will summarize our process and our learnings from the partnership. And Cathy will give a demo of the embodied and total carbon analyses that we're developing.

Our learning objectives are first to discover the key blockers in the current carbon-centered workflows, then to learn about the impact of early-phase design decisions on buildings' carbon footprints, to understand the impact of early phase design decisions on product development, and to learn how you can contribute to future sustainability-related product development. So if you see something you like or you don't think is useful or something that hurts the accessibility of one of the tools that we're talking about, please make a note and let us know. The whole point of this presentation is that we want to listen to our users when we're doing product development.

To start with, carbon in early-phase design. You have likely seen these stats before, but some context just so that we're all on the same page. Buildings account for over 40% of total global carbon emissions.

It's about 27% operational carbon, or the carbon resulting from lighting, heating, cooling buildings, and about 15% embodied carbon, the carbon resulting from the production, transportation, and construction of building materials. One important clarification-- in this presentation and throughout AU, you'll hear Autodesk employees refer to total carbon. In this context, that is an Autodesk term that includes operational and embodied carbon but does not include end-of-life processes.

So not only is the building industry the world's largest emitter. But, due to population growth and urbanization, the next 40 years will be the largest wave of building growth in human history. That means that we need to invest heavily in renovation and retrofitting. And we need to be able to build new and low carbon.

Now, a building structure contains up to 50% of its embodied carbon. And its compactness is the most determining factor of its operational energy use. Estimates say that around 65% of a project's total greenhouse gas emissions have already been determined by the end of the conceptual phase of design.

But early-stage decision-makers are usually not carbon experts. And sustainability analyses typically enter the design process only later. When you're determining the amount of reuse, your building's massing and orientation, your building structure and facade, you are severely limiting your range of outcomes if you aren't aware of the carbon impacts those decisions will have.

And our users agree. Yes, you can make some changes and improve your outcomes. But, right now, by the time sustainability analysis enters the design process, the major decisions, primary building form, basic materials, percentage reuse, have already been made.

We're not saying that these decisions should be completely determined by their carbon impact. But considering that this information or something like it is leading about half the talks at AU, I think we can agree that they are a critical part of an informed design process. So, to summarize, today's carbon tools generally target more detailed phases of design.

Once you have a full building modeled, you can run energy usage simulations. You can use its bill of materials to calculate its embodied carbon. But we need tools that guide users before the highest level decisions are made.

Secondly, these tools are usually built for carbon experts. The early-stage decision makers are experts in other fields. We need to create tools that are accessible and understandable to users who are carbon curious but who are not experts yet.

And, lastly, today's carbon tools usually require moving into another software. And you are already using to design your site. This leads to data loss, to time loss, and it means that you have to learn how to use another piece of software. We need these analyses built into the design tools that you are already using.

Now, what is Forma? And what is outcome-based design? Architects and urban planners have to navigate a web of competing constraints, objectives, and stakeholders. And early-stage planners often don't have the information they need when they need it.

Outcome-based design means getting insights into projected outcomes while you're designing. And Forma helps you prioritize it with the automation of previously tedious tasks, like importing terrain and existing building data, by giving you the ability to explore more design options faster with easy to use design tools and, throughout it all, having immediate access to analysis like area metrics and sun. At the end of the day, outcome-based design is about helping designers understand the impact of their decisions when they're making them, through sun, wind, daylight, microclimate, noise.

But maybe, most importantly, we understand that no one is just designing a site for maximum sun. You need to fulfill floor area requirements, to have comfortable outdoor spaces, to create high-performing buildings that still get good evening sunlight. It's about understanding the trade-offs.

And that's the point we want to drive home here. Sustainability is not just low carbon. Sustainability is the successful balancing of all of these criteria-- including usability and accessibility; living qualities, like sun and thermal comfort; and performance. It does still need to be low carbon.

So what carbon tools do we have in Forma right now? We've developed an AI-based operational energy analysis that gives results in under a second, allowing users to iteratively experiment with building forms and materials. We think that this not only informs the user but builds intuition. By requiring only massing models and simple user inputs, we give users the relative guidance they need from the absolute earliest stages of design.

This analysis is a machine learning model trained on thousands of basic building forms and Revit insight energy simulations. By extracting features like compactness, orientation, and location from the building you've modeled in Forma, along with a few user inputs, like window-to-wall ratio and insulation, we can almost instantly predict its operational energy use. We think that this type of analysis is ideal for early-stage planning.

While it can't capture the nuances of your building, like later-stage models will, it gives you accurate, relative feedback on the high-level features of your designs while you design them. Our solar energy analysis combines our sun analysis with local weather data to help you understand the potential your site has for energy production. The point that we want to drive home here is that, while some of our analyses, like our operational energy analysis, are technically innovative, simply by incorporating existing tools into the design application you're already using, we can make them an accessible part of the earliest stages of design.

Both of these are starting points for a larger project-long focus on carbon. At the end of the day, our goal for Forma is to start the conversation and prompt a focus on sustainability from the very beginning of the design process. And the rest of this presentation, you'll see the embodied and total carbon tools we're developing so our users can understand the overall carbon impact of their early-stage design decisions. And, in the rest of AU, you'll see the carbon tools other Autodesk teams have created so that users like Stantec can stay informed throughout the entire design process. Now, I'll pass it off to Mike to talk about why Stantec was interested in joining this effort.

MIKE DEORSEY: Thanks, Ellis. First, I'd like to mention a little more about Stantec for those who may not be familiar with us. Stantec is a global design consulting firm of about 28,000 experts operating in 400 locations.

I'm part of Stantec's buildings group. Additionally, we have four other business units, energy and resources, environmental services, infrastructure, and water. Our buildings team is made up of architects, buildings engineers, and other experts that help us to design an inspired-built environment.

Stantec has a strong connection to the communities in which we live and work. And that is the main driver of why we are interested in partnering with Autodesk. We want to empower our designers to be able to create designs that have a sustainable impact on the communities in which they set.

Our buildings group has a carbon impact team that is focused directly on how we can do a better job of designing with carbon in mind. We have many clients that are interested in minimizing their carbon footprint. And we know there are a number of regulatory requirements in this area, both already in existence, as well as coming soon.

And Stantec wants to be well-positioned to be able to advise and guide our clients in both how to meet their own goals, as well as what will be required of them. And the most fundamental level, we want to do what is right, which is a core value at Stantec. For us, being involved in product development is a natural extension of what I've just talked about. We want to ensure that we have the tools that work for our designers and our clients.

In this engagement, we had representative architects, structural engineers, and carbon experts working with Autodesk to ensure a tool that would work as a basis for collaboration between all of our disciplines to help inform project teams. At the end of the day, we want to empower our designers to be making decisions that are best for our clients and for the planet. And now I'll pass it over to Jen to talk in more detail about how our teams work together on Forma.

JEN COOPER: Thank you, Mike. And let me just say that having access to this group of Stantec participants was really beneficial to us at Autodesk because we love to conduct early iterative research. This partnership gave us a head start in creating an effective user experience.

So what we ended up doing with this cohort of Stantec users was to deep dive into their needs and mental models around the unknown areas of our analyses. We were particularly interested in getting feedback on embodied carbon and total carbon workflows. And I'm going to share one of the early findings from this research, along with a set of design principles that we're using as our guiding light, moving into the development phase.

Here are two examples of wire frame prototypes we tested. And, in this case, we were grappling with how to approach material inputs for embodied carbon, which is the part of the screen that is shown in the yellow box. In order to run an embodied carbon analysis, a user must enter building system materials so that the engine can calculate the amount of embodied carbon that the building is going to produce.

We weren't quite sure whether we should pre-fill the material list, as shown on the left, with defaults so that users could run an analysis in one click using those defaults. And the beauty of doing early concept testing is that we can show alternatives and test our assumptions. So, on the right, you'll see a second option, which has the building system inputs in an empty state requiring [INAUDIBLE] all of their materials before running the analysis.

So, through testing, we learned that this manual approach has more benefits to the user. It's especially helpful for novice users who are less familiar with carbon because it gives them a chance to manipulate the material list and see what happens to the embodied carbon as a result. So, ultimately, we move forward with a design, like the one on the right, after hearing the various perspectives of our users.

And, aside from those tactical examples like the one I just gave, we also learned about some core principles to center ourselves around. These are more long-term goals that emerged from listening to our users on a more holistic scale. One is to design for learning.

We know that carbon is complex to understand. We also know that it's complex to interpret and regurgitate. So we want a tool that is inclusive of all users, regardless of experience with sustainability. We learned that there are opportunities to use this tool to teach our users about the trade-offs involved when making decisions that account for carbon efficiency. The example I just gave about the inputs in the material list illustrates the designing for learning principle because we're helping users build that intuition through seeing the effects of their material selection.

Another one of our design principles is to design for collaboration. Even though architects are our primary end user, structural engineers and sustainability experts are also providing feedback and consulting along the way, even to the point of potentially using the tool themselves to help contribute ideas and alternatives. Ultimately, we heard that we must treat carbon analysis as a team sport to bring the most value to our users.

And the third big takeaway from our research was to design for the client. We realized that client presentations are an important step in this process and that, in order to properly convey comparisons and trade-offs to the client, the architects need to build a narrative around carbon efficiency. Our tool aims to support users with client-ready outputs.

We recognize that, for this tool to demonstrate its value, it really needs to be engaged with at the user level as well as the client level. So, as I mentioned, these design principles are our long-term goals. However, we aimed to incorporate them into what we've built so far. And, at this point, I'll pass it over to Cathy, who's going to share our product demo of carbon analysis in Forma.

CATHY TRUONG: Thanks, Jen, for wrapping us up to the product demo here where I'll be showing the analysis designs within Forma for operational carbon, embodied carbon, and total carbon. During this demonstration, I'll be showing the workflow and tying back to the key takeaways Jen just shared earlier that directly influence the user experience. So this is a product demonstration.

So the numbers here are for example only. So this might not reflect a real-life situation. So please keep this in mind.

And, lastly, we are open to hearing your feedback about the design. So feel free to discuss with one of my Autodesk colleagues here or even sign up to be a part of future research to share your thoughts. I'll share how to sign up to be part of the research community at the end of this product demo.

To begin this is a video showing Forma where I've already set up a site, outlined here in red, and modeled several volume studies. On the right panel, I can find analyses represented by graphic icons at the top, and I can see that I have the operational carbon analysis open. The analysis takes results of the operational energy analysis that Ellis demonstrated earlier to calculate an approximate carbon value based on the emissions from the electricity grid in the region and the end-use energy to operate the building.

The rapid analysis works very quickly to give me results. When the results are computed, the buildings are colored with their approximate operational carbon emission per unit area. The value of each building can be determined with the colors legend at the bottom of the screen.

Because this analysis is based on the results from the operational energy analysis, we allow the user to open up a small version of this analysis to refer to the information and settings for operational energy. We think nudges like this helps the user understand the dependencies between the two analyses relating back to key takeaway number one-- designed for learning. As you can see, we're starting to build our total carbon picture and operational energy and operational carbon are a large part of this. Let's now take a look at its counterpart-- embodied carbon.

Now we switch analysis on the right-hand panel to begin looking at embodied carbon. Like our operational energy analysis, the embodied carbon analysis will be a machine learning model. That means that it uses extracted features of your buildings, the number of stories, floor area, location, as well as the user inputs we will talk about here, to quickly predict embodied carbon.

We think providing a quick analysis like this makes embodied carbon accessible to all Forma users regardless of time, carbon expertise, and level of detail in the project. This way, users can start thinking about the impact of their material choices right from the earliest stages of design. So let's take a look at the inputs then.

We see on the right panel here that we're required to input function, envelope, and structure systems before we can start running the analysis. Following the takeaway from our research, we found it was a learning opportunity for users to make decisions themselves and see how their choices affect embodied carbon. We wanted the tool to empower architects to make these important design choices and acknowledge there are other trade offs for these choices, like design expression or cost, that should be considered other than just the embodied carbon impact.

So, before we dive into these inputs, I want to show what's under the additional settings dropdown. We heard from many of our research participants it was important to capture other ways of reducing carbon, including adaptive reuse of existing building elements and carbon storage. We won't go into in this in this presentation but hope users will experiment.

So let's take a look at functions. Setting the primary program or the use of the building will have an effect on embodied carbon. Many buildings are mixed use. And not all buildings on a site need to have the same program. But, for this demo, we'll show that all the buildings are in multi-residential complex.

Now if we look at the envelope settings, you can see that there are two missing fields. We worked with users to figure out the right balance between providing detail and flexibility, meanwhile not overwhelming the user with too many inputs. We landed on cladding and structure as required fields which allows users to mix and match cladding and structural backing materials.

We set reasonable defaults for the window-to-wall ratio with double-glazed windows with an easy way to override these values. As Jen mentioned, we interviewed sustainability experts as well. We understood they bring their expertise on projects to consult on material choices with respect to sustainability.

We wanted a tool that was accessible but also flexible for experts so that there are default calculations or assumptions users can override to get more accurate results. This brings us to key takeaway number two-- design for collaboration. Architects can start consulting with subject matter experts early in the design process to refine these values with experimental or specific material choices.

So following up to this point of collaboration are the structural settings. We know architects are not structural engineers, so we only have one required input to specify the primary structural system. However, if you go into the advanced setting to mix different systems, we think structural engineers can actively participate in this design process with the architect, either directly experimenting with choices in Forma or commenting on the results of the analysis.

So, now that I've completed all the required inputs, I'm ready to run the analysis and see the embodied carbon of my proposal. So now we're looking at the embodied carbon analysis results for my proposal. The analysis takes a snapshot in time of the decisions I made, including the materials, settings, and building form.

Similarly to operational carbon, the analysis colors the buildings with the approximate embodied carbon results expressed in tons of carbon dioxide equivalent or CO2e for short. It's a unit of measurement that allows us to compare the emissions of different greenhouse gases based on their global warming potential by using CO2e we can understand and compare the impact of these gases in a standardized way. So we worked with users on this results page to make sure they had results that were easy to grasp from both a carbon nonexpert and expert scenario.

These statistics are a key place for users to learn about embodied carbon. The circular graph quickly illustrates the breakdown of embodied carbon by material and category. And we can quickly see what has the largest impact.

We also provide the breakdown by material. And this shows that concrete is a large contributor to embodied carbon. So let's try to play around with different systems to see how it impacts embodied carbon.

So here I change the structural system from reinforced concrete to mass timber. We can already see that the embodied carbon went down significantly by the New colors of the building. Their colors lie closer to the smaller values of the legend.

Perhaps this isn't the right structural solution for the project. But now I can take these results to a structural engineer to begin having a conversation on structural systems and how we can reduce embodied carbon. We want to encourage users to compare runs like these together to see how different choices affect embodied carbon.

We have a specific Compare tool that is designed to make this comparison of different analysis runs easier. I'll show this later in the presentation. So, now that we have our operational carbon and embodied carbon results, we're ready to take a look at the total carbon analysis.

So, unlike the other analyses, we looked at the total carbon analysis takes account of the carbon emissions over the lifetime of the building and not only the carbon emissions of the opening year. For first-time users of this analysis, we provide this context in the information box. And we also recommend to use this tool in conjunction with the Compare tool in order to maximize the benefits of this analysis.

Under the info box, you'll also see the different aspects of carbon we saw earlier that feed information into total carbon. So let's take a look at the results. The resulting analysis colors the buildings with approximate total carbon emissions from the extraction of materials, construction, and operations phase over 30 years.

What's new here is the chart on the right-hand side demonstrating that the embodied carbon remains more or less the same over time minus some materials for renovation compare-- and repair. What you'll notice is that the operational carbon increases over time as the building gets older and needs more energy to operate. We hope this is a learning opportunity for users to think more about their operational energy choices as it can have a large impact on carbon emissions.

So let's take a look at the Compare tool here, which we believe is a super useful utility when it comes to talking about carbon. So this Compare tool already exists in format to visually split the screen so we can compare multiple analysis together. However, we took a separate look with our Stantec participants to see if we can extract more value from this feature.

We showed users various layouts and tried to understand how much of the 3D scene and the statistics we should show. The answer was to remain flexible and have an ability to expand statistics or scenes depending on the use case. Users shared a variety of uses for this tool that expanded our initial understanding.

At first, we thought the tool would be for internal discussions. But it became more clear that the view was valuable to show to the client. This takes us to the final research takeaway-- design for the client.

Architects spend a lot of time building a narrative and convincing clients to make better choices for building design. Oftentimes, they export screenshots from programs like Forma to create graphics and put them into their own presentation, which takes up a lot of time. What we heard from users is that we can expand this tool to help them build this story.

They shared many ideas about showing the design process, analysis runs that didn't work, design options, and even bolding some numbers in the data that they want to highlight in a presentation with the client. So these findings are not only applicable to a carbon analysis. This could be comparing different analyses of the design together, like sun, wind, and noise conditions. In conclusion, there are many ideas from users from this extensive research process that we will continue to be inspired by moving forward in future Forma features.

So this brings us to the end of our presentation. To summarize, we presented an example of a product development cycle from research to delivery as it applies to carbon analysis in Forma. We'd also like to highlight the amazing partnership and commitment by Stantec to help us validate these designs.

We hope you're excited to test out these carbon analyses in Forma. If you have any feedback from this presentation or after testing the carbon tools, please post on our Autodesk community forum with your comments or suggestions. And if you'd like to participate in future research studies in Forma, here's how you can register.

So make sure your voice is heard by signing up for the Autodesk research community by visiting the website in the QR code. Benefits for signing up include an exclusive look at upcoming projects and the ability to influence the future of [? BIM. ?] Thank you for your attention, and we look forward to any questions or further discussions.