Civil 3D + Generative Design = The Corridor Optimizer!

MICHAEL PAVLOVEC: Hello, everyone. Welcome to today's industry talk, Civil 3D plus generative design equals the corridor optimizer. We have an exciting presentation for you. It's class number CES602588. Please jot that number down. You'll be able to go back and search for the recordings following this live performance and use it as a reference point to go back and watch some of the exciting points of the presentation. So jot that down. And let's kick off today's presentation.

So my name is Michael Pavlovec. I've been in the transportation industry for just over 30 years, the last seven years being the most exciting here at GHD as a project manager working on some large transportation projects. You can connect with me on LinkedIn, pull me for a coffee during the conference. But I just want you to take away some important pieces.

Dynamo and Python play a large part of my daily workflows. And I'm always looking to automate things by-- take advantage of Civil 3D's API functionality. I'm leveraging a computational design mindset, generative design tools. I've really looked to enhance and adapt technology into some of our current workflows. With that, I'd like to introduce my co-speaker, Brenden.

BRENDEN PICTON: Thanks, Mike. Hello, everyone. My name is Brenden Picton. I'm really pleased to be here with you today. Just a bit about myself. I've been in the industry for a bit over 20 years, but within the property and building space. I've been with GHD for over 18 years and had the privilege of working in some of the most beautiful parts of the world.

I did start my career as a structural drafter in AutoCAD. But fast forward 18 years, I now find myself managing the automation and emerging technologies team within GHD. Although, as a team, we do manage a lot of the automation tools that get developed within GHD, we have a strategic objective in developing our people and bringing them along the automation journey that we're all part of today. So let's get started and begin with running through our learning objectives.

Our presentation today is divided into four parts. The first part is how to implement generative design on your project. And I'm going to speak on some of the high-level points on this topic. The second section is going to explain how optimization can be your new go-to tool in your projects toolbox. Thirdly, we're going to talk about sustainability and how you can generate some valuable reports to help your client meet their net zero targets.

And lastly, understand how Civil 3D and generative design will change your project outcomes and measurable differences. So there's some exciting stuff there. Let's get started. So before Mike gets into the nitty gritty on the implementation at the technology level, I just want to briefly touch on some of the more foundational pieces associated with implementing generative design.

Not too long ago, I stumbled across this statement that really resonated with me with some of the work we've been doing in the GD space. By its nature, generative design is inherently changing how we approach our design problems. We're able to explore the design space in a way we've never been able to before. So it would be fair to say that GD is changing how the work gets done. If this is true, our focus needs to be on building capability, not just training.

This is how I like to think GHD is approaching the implementation of generative design. It's not just about training our staff. It's about creating a support structure and community where our employees can explore, learn, fail, and grow. It requires time and investment. By building capability, we're able to transform how we do the work that then brings better value to our clients. And hopefully, this becomes apparent as this presentation unfolds.

For GHD, our first exposure to generative design was courteous of the Autodesk Living back in 2018 on a project GHD was working on in the Middle East. You may have seen this before. We had a booth here in 2019. And so it's been around for a while. But this approach to solving a design problem really opened our eyes to the possibility of generative design.

It was around this time that GD was starting to ramp up. And there was a lot of energy in the Revit and the building space. But in 2021, we really wanted to start exploring what was possible in the Civil 3D and linear infrastructure space. Leveraging our EBA partnership with Autodesk, GHD and Autodesk undertook some virtual training sessions specifically focusing on generative design using Civil 3D.

It included some foundational training on core concepts, but also unpacked some real life example problems that designers might face and how they might overcome them. We converted this training into what's known as our Civil 3D Foundations Training course and, ironically, has laid the foundations for where we are today.

To support our users on their GD journey, we've created our internal [INAUDIBLE] engage community, where people can share and chat about the cool stuff they're working on, and our intranet site that caters to all levels of understanding from technical staff who want to get into the nitty gritty to our senior professionals and business development managers who want to understand how we can differentiate and offer a different value to our clients. It's important to understand that building capability extends to all levels from awareness to understanding to delivery.

I just wanted to finish up with this slide. I know we're talking about generative design today. But with the pace of technology change that's occurring at the moment, it's quite staggering. Things were a lot simpler when all we had to worry about was layer naming and hatch boundaries being enclosed. Now we have to be concerned about whether ChatGPT knows what I did prior to September 2021.

But with all the hype and the buzzwords floating around, it's important to be able to understand how these technologies will benefit the projects we deliver. Within GHD, we try and understand, which one of these levels or approaches best resolves the problem we're faced with? Not every problem can be solved by machine learning and AI. Not every problem can be solved by trying to automate it. And not every problem will have a generative design and optimization aspect to it.

However, building awareness, knowledge, understanding, and ultimately capability behind each of these levels will help drive your own internal transformation and change how the work gets done. I'll now pass over to Mike, who will go a little bit deeper into the technology aspect and then the more interesting side of our presentation.

MICHAEL PAVLOVEC: Thank you, Brenden. Continuing with the implementation of generative design, I just wanted to highlight some of the important aspects of this. Now we don't have time to go through this in great detail. But thankfully, this morning, Jowenn Lua and Namit Ranjan of Autodesk put on a fantastic class titled Getting Started with Generative Design in Civil 3D: A Beginner's Guide.

It's class number CES600726. Jot that down and watch a presentation to see how you can set up generative design inside of Civil 3D Now, today's talk is going to focus on the 2023 version of both Civil 3D and Revit, bringing that powerful functionality of generative design into Civil 3D. So let's get started.

So why generative design? A fundamental question that we started asking ourselves. A couple of years ago, we just finished construction of a major freeway project here in Ontario, Canada with an overall construction value of $44 million. It was a 32 kilometer long freeway project that included four complex interchanges and two truck layby stations.

During the course of design, actually, late into the design process, we started to see some challenges that resulted in some additional costs being added to the project as we looked to solve some of the complex challenges that we were encountering. Some of these resulting costs were for some of the padding costs being counter to fix some cross-wall, some of the existing infrastructure elements we needed to replace, like guide rail and some of our curb systems. It extended the construction schedule and even led to some traffic staging issues.

We wanted to be able to look into the future to understand some of these project issues at a much earlier stage in the project. Enter generative design. So with Civil 3D, I think those familiar with its corridor building powers understands how your corridors generate some fantastic information for your projects. Now, corridor is a coarser alignment-based that includes both your horizontal alignment and your vertical profile alignment. It also includes the ability to set specific regions or limits within your corridor and, of course, your corridor beginning and end point, your project limits.

It's also based on your existing surface or the original ground surface to start. Civil 3D then begins your corridor. And that's really the basis for the corridor optimizer, that initial existing conditions corridor. Now, as with every corridor, we, of course, load in our sample line set or the interval at which we're going to be extracting some client-required cross-sections. And we're going to build our assemblies. And for those not familiar with an assembly, it's simply the make up of our model, asphalt layers or granular courses, and things like curb and gutter sidewalks that builds our model. Civil 3D is going to go through and build one corridor, just one.

So with the corridor optimizer, we really wanted to take the power of that corridor build and add in generative design's iterative approach to building not just one corridor, but hundreds, even thousands of corridors. And that led us to ask the question, well, which corridors do we need to build? Well, here in Canada, we have thousands of kilometers of rural highways that require rehabilitation. We have new infrastructure that's being built. Each has a different requirement for a corridor makeup.

We started to look at how pavement design and some of the different strategies for pavement design could allow us to build multiple corridors to look at how some of the different strategies would affect the existing corridor elements. So we started to program in and create options that would look at a milling and paving approach, a recycled asphalt approach, processing new asphalt, even things like concrete pavements so that we could actually build corridors based on all the industry-acceptable pavement strategies available to us.

We then looked at varying things like grade raises and no grade raises, the different lift thicknesses and make up. The results were then compared to the existing corridor to really understand impacts to existing infrastructure, all much, much earlier in the process. So starting at the beginning, it really started with the completion of our payment investigation and material testing program. In other words, before we started looking at how many corridors we wanted to build, what types of corridors, we really needed to understand the makeup of our existing condition's corridor.

We input things like the existing asphalt thickness, existing base, or fine granular configuration, or subbase or coarse granulars, so that the corridor optimizer can understand what that initial corridor makeup is all about. This then allows us to calculate the existing corridor strength. And it helps really set some of our Civil 3D corridor regions so we can start to think about linear optimizations as well, as some of [AUDIO OUT] pavement strategy make up in the existing corridor changes throughout the length of the highway.

This then takes us into some of the project requirements or the suitable alternatives we can use to start building these corridors. We can look at straight alternatives, like an asphalt overlay and how many lifts of asphalt we may need. We can look at a pre-milling or paving overlay strategy that includes variables, such as pre-milling depths, required lift thicknesses. We can look at processing and paving, recycling and paving, even concrete pavement mixes. So we're really looking to build all of the corridors, not just one. But which ones then to consider?

So I've chose to talk about granular-based equivalency. Now, there's a number of different empirical methods we can use to calculate an existing embankment strength or an existing roadway strength. This is one of the easier ones. And it's easy to switch in the optimizer, excuse me, in the optimizer should you wish to use a different regional or client-based calculation.

So here, based on our inputs for the existing corridor, the optimizer understands what the existing strength is. We then input our proposed strength based on traffic loading and life cycle expectancy of our strategy. And the corridor optimizer then understands the strength it needs to develop to meet our design criteria conditions. That takes us into our optimization phase.

With the optimization, we're really looking at taking our corridor strategies and really fine-tuning them to meet our project goals. Now in order to do this, we first need to select the corridor elements we wish to vary or the elements we wish to start changing slightly to see how those changes affect our overall corridor and its performance.

But with each variable we select to vary, it's important to set two important aspects, being the step value and a constraint. So if we're talking about asphalt as a variable, the first thing we want to do is set a step value to evaluate. That will allow us to look at evaluating, say, a 40 millimeter lift of asphalt versus a 50 or a 60 as opposed to 41, 42, 43.

It's important that when you select a variable, you select an appropriate step value in consideration. Along with that, the constraints are equally important. For asphalt, again, as an example, we can then look at setting a minimum lift thickness of 40 millimeters and a maximum lift thickness of 70. That will help constrain our study.

You can also set variables for maximum milling depths and ranges. It's really based on your imagination as to which elements of the corridor you're looking to vary. And again, all of this is done to exceed your roadway strength requirements. So the corridor optimizer is not going to select alternatives that don't meet that minimum strength requirement.

So when we look at the optimization, how do we actually optimize the strategy? How do we evaluate which corridor alternatives we've just built affect the existing environment? Well, that's where, again, we come back to the power of Civil 3D corridors. Now with each corridor, the modelers in the audience are going to understand this, it creates a surface for each of the different elements within our corridor. And it's also going to generate quantities.

Now, starting first with the surface and, again, for each one of the hundreds or thousands of different corridors we generate, we're going to take that design surface and we're going to interrogate it against the existing environment so we can understand how it impacts our existing corridor and affects things like curbs and gutter, guide rails, even overhead structures and overhead signs. So with each of those surfaces from each of the alternatives, we're able to see how it affects the existing corridor, in a sense, looking into the future to answer some of those questions.

Now, I mentioned quantities and surfaces. But to me, the most exciting part is the quantities. Now, with every corridor, it generates quantities for all the different elements in your assembly, again, the asphalt, granulars, curbs, sidewalks. The minute we add a unit price to each of our pay items, we get a total cost. And that allows us to generate project-specific costing for each of the hundreds or thousands of alternatives we're generating.

And these are straight inputs. So we can use very specific project costing, whether that's from a contractor in a design build element or based on a regional requirement and project site. And very similar and straightforward from a programming standpoint is how we can implement an input production rates and use the same quantity calculation with a production rate to determine the construction schedule.

So we can input things like how many tons of asphalt a contractor can place in a day, how many linear meters of curb can be constructed in a day or a week, even things like pavement markings, and signings, and expected production. We can program in the sequencing or order of operations to ensure that we, of course, do the excavation first, regular pay-- placement then paving. All allows us to calculate a powerful construction schedule for each one of these alternatives. That takes us into sustainability. Brenden.

BRENDEN PICTON: Thanks, Mike. On a personal level, I think it's great that sustainability is starting to become or starting to be at the forefront of everyone's mind. But again, obviously, there's mixed opinions on this topic across all levels of society. But in some parts of the world at government levels, we are starting to see some policy action.

So if you're from Australia, you will understand what the New South Wales government is. If you're not from Australia, New South Wales is the state where the Sydney Opera House is located. But recently, they've just implemented a policy that all buildings, both residential and non-residential, now must include the assessment of embodied carbon as part of the design and approval process. Not really across all the details, but coming from the property and building space, this policy has had some architects absolutely scrambling.

But we started to think about this in the linear infrastructure space and wondered, how long would it be until the same policy is introduced into the development of our roads and highways? And how would we go about calculating these values? And what information would our clients want to know? And then, obviously, how we could implement it into the corridor optimizer.

MICHAEL PAVLOVEC: Thank you, Brenden. So when we think about sustainability, and our corridors that we're building, and of course, through our corridor optimizer, the hundreds or even thousands of corridors we're building, again, we come back to the quantities. And we started to look at how we could input a carbon footprint for each material that we're using in construction. We're inputting values for kilograms of CO2 per every cubic meter of material. That allows us to create a carbon footprint for each one of the alternatives that we're generating through this optimization process.

Now, this allows for a very quick comparison of the alternative during, say, a preliminary design phase, or the auctioneering of different workflows, or even during the design build environment. Or during the bid phase, you want to know, which options, perhaps, are more sustainable than others? This is going to help us drive to better sustainable solutions and, of course, allows us to make data-driven decisions.

But is that enough? We can look at a footprint and input our project's requirements for granulars, asphalt, concrete, even things like steel and wood. But is that really enough? We had to take a step back and really look at how a material source could affect your overall carbon footprint. In other words, would it change how sustainable a strategy is?

We started to look at the circular economy approach to make sure that we understood how the embodied carbon in total should really be properly evaluated from a raw material acquisition, hauling it to a processing plant producing that material, hauling it to the site, and then, of course, during the construction phase itself.

That really led us to leveraging the power of ArcGIS. Now, inside of Civil 3D, I'm sure most of you are familiar, we have the connector for ArcGIS. We allowed-- we utilize that powerful connection to bring in some of the GIS functionality to allow us to start to take a deeper look at how we could approach sustainability. We created a tool through the corridor optimizer to allow us to select a project's location or, in the case of a linear project, its limits. We could then select the pits and quarries that we're going to be sourcing some of the raw materials from.

We then select processing plant locations for our asphalt plants, our concrete ready mix plants, and any other significant material source. That allows us to start generating some hauling routes from the raw source to a processing plant, again, from the processing plant, your project's location. So we could start to look at a more complete embodied carbon impact for each pay item and for each of our hundreds or thousands of corridor alternatives.

So the corridor optimizer, I wanted to show a couple of screenshots of what that actually looks like for those of you that are not familiar. The standard study screens, in order for you to create your optimization, allows you to pick the different inputs you wish to vary. Now, here we've selected all of our payment strategies. So we can cycle through the different iterations.

I've selected things like our surface course thickness, our binder course thickness, our removal depths. So we could cycle through all those different permutations and combinations of those strategies to see how it affects the existing corridor. We then look at and select which outputs we wish to use. Or in other words, how are we going to go through and actually weight the different alternatives it calculates and generates for us?

So we select things like our strength, minimizing the minimum allowable strength for the alternative we're trying to target. We want to minimize the grade changes and, of course, costing. You can also select the course sustainability metrics or working day schedule, even things like quantity optimization, to look at balancing your cut and fill quantities.

So depending on how many outputs you select, it'll determine the size of your study. Next, you want to select which outputs you want to constrain. Now, here we've put in a range of 250 plus plus or minus to limit the amount of alternatives it's going to generate around our target strength. You don't want to generate too many options that exceed that strength. Again, we're trying to optimize and choose the corridor alternative. We've also selected grade raise and trying to minimize that. Now we've selected a range. Again, this helps control the size of the study.

Jumping to the results-- before we jump into the results, I just wanted to show a quick video of how that looks inside of Dynamo. So inside of Dynamo, we're looking at how to vary some of the refinery inputs, how it cycles through different pavement thicknesses through different asphalt strategies, milling, paving, concrete pavements, a little insight into how it actually looks at varying it once you select your study and start to run that optimization.

As far as the results are concerned, there's some very powerful metrics. And, again, once you select the outputs you wish to analyze, it's going to give you some incredible dashboard metrics you can use in discussions with your clients and stakeholders to really come up with what you consider to be the preferred alternative for your project.

Now, in a preliminary or even a detailed design environment, this is going to allow for a more complete evaluation of your alternatives and, certainly, generate a more accurate lifecycle costing of the alternatives. It's considering, of course, your new corridor design, but also how that corridor affects the existing corridor elements.

Now, in a design build element, you can use this during the pursuit phase. So as you're working with your contractor to develop your bid price, you can run through these corridor alternatives and, again, assessing some of the impacts using contractor-specific unit rates or even production to calculate values to help your contractor partners understand what should be built, how long it's going to take to build. And this, of course, is going to generate fast, risk-free quantities.

Again, this is using your sample lines, your standard end areas, to calculate quantities. This is accurate, risk free quantities for bidding. It's going to generate incredibly power sustainability reporting, again, driving that journey to net zero and, of course, quantity optimization, balancing your cut/fill on a project.

And of course, environmental mitigation. We haven't even really talked about how optimization of your slopes on a corridor can help us stay not only inside our right of way along our corridors, but outside of some of our environmentally sensitive areas. Now, how does this all change our project income-- project outcomes?

When we first started looking at optimization, we leveraged it on a Highway 127 project in rural Ontario with our contractor partner, GIP, Green Infrastructure Partners. Now, this was a design build project from the Ontario Ministry of Transportation. And it was a 23 kilometer long project in rural Ontario. Not a high complexity project. But due to the project's location and challenges, optimization was the method we chose with GIP to really help fine-tune our bid price.

Now, we were obviously successful. And at the conclusion of this project, not too many months ago, we were able to come back and really look at the lessons learned from generative design and how this highway optimization really changed the outcome. We were able to recycle over 25,000 tons of granite-- granular material, reducing cost from our overall bid price.

We looked at designs that minimized grade changes to avoid impacts of some of the existing infrastructure, in other words, maintaining some of the guide rail and changing its mounting height to meet our new grades as opposed to replacing it. And due to this project's challenging cross-wall and super elevation values, we ended up coming up with a processed asphalt alternative to help generate some material we needed to reshape the highway. That allows us to fix the cross-wall and superelevation with materials generated on-site, an incredibly sustainable alternative.

Lastly, we only considered alternatives that would have met or exceeded our strength requirements, which really allowed us to hone in and pick the preferred processing alternative in how many lifts of asphalt, what thickness of asphalt. Again, all during the bid phase. Essentially, looking into the future, giving us fantastic insight of what some of those costs or some of the impacts would be. And we were able to include that within our submitted price.

Lastly, it allowed us to look at how to source our materials, where quantities are being generated, where they would need to be hauled to. We had a number of very significant culvert replacements on this project, which required some temporary detouring. So through the optimization project, we were able to calculate where materials could be used from one culvert replacement, and hauled, and used during the staging of the next replacement. It was our first successful use of the corridor optimizer.

We also leverage it on another project, an 18 kilometer rehabilitation project in rural Ontario, Highway 124, to really help us balance our cut/fill quantities. Now, this project helped us utilize the materials during the staging, again, of some of our culvert replacements and widenings of the highway. That allowed us to save over 25,000 tons of grain material. Again, a significant savings.

The design allowed us to minimize our grade changes. So we were able to tie into things like curb, and gutter, and traffic islands, some of our drainage structures. That avoided costly replacements, curves, and adjustments of our catch basins and manholes. It also allowed us to retain some of the existing safety elements. It allowed us to adjust guide rail and safety barriers as opposed to replace it.

And most importantly, it allows us to maintain some very specific structural overhead clearances. By looking at minimizing our grade raises, we were able to ensure that we weren't raising the grade and violating clearances to overhead signs and overhead structures. Again, we used the processing of asphalt approach to generate the kind of material we needed to correct the cross-wall and superelevation on this project.

We've also used the corridor optimizer during our pursuits. It allows us to leverage generative design to really look at what some of the possibilities will be. So as we're preparing our proposals for various pursuits, we're really able to look at what our project needs are, what our client requirements are, and allows us to zoom in and understand how material movement and cut and fill balance will affect how we approach a project.

Understanding infrastructure impacts, how some of the different corridor alternatives will generate conflicts with existing services, water mains, sanitary sewers, and of course, underground and overhead utilities and some of those relocation needs. It generates powerful corridor quantities which, again, allows us to generate project costing construction schedules and sustainability reporting, as we talked about earlier. It really allows us to immerse our client into all of the possible solutions.

We haven't really talked about a subsurface optimization. But the corridor optimizer, again, it's building a corridor, hundreds of corridors, even thousands. And that includes some of our subsurface elements. We've used it on a project to include some of our existing infrastructure, gas mains, water mains. So we can look at how different alternatives and the overall thickness of those alternatives may affect or even violate some of our working clients to gas mains and water mains.

It also allowed us to look at excess soil requirements. Now, with new legislation, especially here in Ontario, we wanted to make sure we had a full understanding of how much excess soil we would be generating on a project and how we would manage that material. Again, making decisions based on data much, much earlier in the project.

Now, as we're generating hundreds or even thousands of corridors, it allows us to take those corridor models, turn them into solids, and visualize them to help immerse our clients into some of these alternatives and what they actually look like as they impact the existing environment. Some fantastic uses for the corridor optimizer.

Next, it really is more than just a corridor optimization. It helps us, of course, with our corridor optimization process. It includes the services, as we talked about. But also, more importantly, it helps us optimize grading [AUDIO OUT] for ponds, even structures and embankments leading up to some of those overpass sites. So the corridor optimizer has a number of different uses that we can leverage on our projects.

Now, in closing, we've tried to come back to our learning objectives and really separate this into four important parts of our presentation. The first, of course, was to help share how to build capability in Civil 3D using Dynamo and generative design functionality. Secondly, how to optimize your project, how you can balance your quantities, generate cost effective and more efficient designs for your client.

Next and most exciting, perhaps, is sustainability and how you can introduce sustainability into your project metrics to help deliver more sustainable solutions to your client. And lastly, of course, how to implement generative design to add more value to your projects and change the way work gets done. That concludes our presentation. Thank you.