Parametrics in Design and the Project Execution

THOM CHUPARKOFF: I'm Thom Chuparkoff. I'm a project manager with the sports and entertainment firm Populous, based out of Kansas City, Missouri. And Steve Lewis is a structural specialties leader for Walter P Moore, renowned structural engineers here in the United States. Today we're going to talk to you about projects parametric design and project execution. That's going to surround itself around Daily's Place which is a covered practice facility for the Jacksonville Jaguars. And then a 6,000 seat amphitheater that's all-encompassing with the same program.

But first, before we get into the technical components of it in the process, I feel like I really need to set up the project. As you read in the intro to the class, we didn't necessarily use new software. And we didn't necessarily recreate the whole production wheel. But we had an incredibly complex project and an intense schedule. And we just reinvented our workflow. And used our modeling software in new ways and integrated ways.

So that being said, about a year and a half ago, the Jacksonville Jaguars contacted Populous and another sports architecture firm to kind of do a little mini-charette. To envision what an iconic, covered flex field and amphitheater could look like. And how it could systematically attach and be a part of the existing EverBank Field football stadium. Shahid Khan is the owner of the Jacksonville Jaguars. And if anyone knows him or if anyone knows of him-- if you don't, you should look him up. He's obviously a billionaire. But he's very interesting. And he's an industrial engineer by nature.

So the parts and pieces and how the building went together-- from an owner standpoint was very unique because that's what he liked the most. That's what was sexy to him. So when he asked us for an iconic design, in fact, to be honest, we submitted like two or three designs before this. And none of them worked for him. The budget was-- all in number-- was about $60 million. And the projects that we kept showing him that cost $60 million, he didn't like. So I mentioned a mini design competition. So then he says-- well, listen. You guys got to come back. And you've got to show us a building that's iconic.

He lives in Chicago but he's from London. He owns a football club overseas. He takes the Jaguars to London every year. So he's very international. So he wanted it to be prevalent and relative to the national circuit. So he gives us the task to build this building. It's got to be iconic. It's got to connect and have all these unique programmatic pieces. Oh and, by the way-- you have to be done by May of 2017. Or actually April of 2017. So Walter P Moore and Populous started to pre-collaborate before we had any schedules or any contracts or anything established. And we were brainstorming on how that would look. And what kind of schedule that we were confronted with.

So this is a preliminary, just kind of a diagrammatic schedule-- just to kind of define the time period that we're talking about. And why we had to reinvent our workflow and our software usage. So basically, the top of the bar just defines a 6 and 1/2 month design and construction document phase. Within this period of time, along the black diamonds, you see we had different bid packages and different early release packages. So first and foremost, right after schematic design in April, we had a six or seven, eight package that was bolted to which was mill order.

So basically, by the time we're not even done with design development, Walter P Moore and Populous were asked to turn over a tech mill model, a bill of materials model, to our fabricator for fabrication so that they could be ready for building in August. 3 months after our final GNP which is our final construction documents. And then that would give our contractor, which was Hunt-Danis, with construction AECOM. And a vocal partner, Dana's Construction, in Jacksonville.

So knowing that we had this constricted period of time-- this highly complex project-- we needed to find a way to define our process. To find this the people who were doing it. And defining the programs that we're going to use to execute on this. Now we all talk about collaboration. Everyone collaborates. We get it, right? But how do you take a project? And how do you take a distributed workplace? Walter P Moore used offices in Los Angeles, Kansas City, and Tampa. Populous had people in New York City and in Kansas City. We had engineers all over the place. We had our construction manager was in Indianapolis and Jacksonville. We had [? Airup ?] which is out of New York.

So we had to figure out how we could define a workflow and a work plan and create in a more integrated way. Now, Steve likes to refer to this as the London Underground map because it's all over the place. But this is a cool diagram because it shows the 18 different entities with the three different relationships. This project was a design-bid-build project. But the minute we signed a contract, all that went out the window. It was definitely design-assist. It was IPD. It was all those things. So what this shows is Populous is at the top of the page in New York and Kansas City. And then under that we had Walter P Moore. And again, we mentioned all of their offices Los Angeles, New York, Kansas City, and Tampa.

And then the map just kind of shows how much of these collaborations and how much of these communication strands are advisory versus contractual versus prior collaboration. And in other words, when we were asked to take over the ownership of the bill-- the materials model, the fabrication model-- obviously the first thing we had to do is change everything. Open up a dialogue with Banker Steel, who is a fabricator. And Walter P Moore and Populous had to have that conversation. We weren't obligated to. But we knew we had to do it. So you know, this essentially blurs the line.

So the other thing, Steve and I were having drinks last night. We were talking about-- we don't endorse this process. We were put in a position where we had this fun project, this very unique project that was important to us. And we had a period of time to develop it and execute it. Now if it ever happens again, I think that all our bosses will be like, don't agree to that. Because it was just so intense.

So once we defined the period of time, we had our schedule. We know when our drawings were going. We know what big packages we're going to target. But what we did was-- actually Walter P Moore authored this schedule on the left. You don't need to read all the red and blue stuff up top. It's very linear in nature. We were doing things before. Other things were done, taking tremendous risks. Just making sure we're talking to people and be like-- all right. Well, look. If we do that, you're know this has to happen next. And just making sure people knew they were accountable. I mean, we really kind of set a table. And we said look, these six things have to happen. And if one of your things aren't happening, you need to let everyone know as fast as you can.

But what this defines is that it took our schedule and our design time. Now I'm not talking just schematic design design development. I'm talking fabrication. I'm talking that Walter P Moore developed in the Tekla model within 19 weeks. And halfway through that 19 weeks, these guys are building a steel model. Taking the Rhino model, put it intact, doing the Tekla model, providing shops, having dialogue. Flying out to Lynchburg, Virginia. Meeting with Banker Steel. And really trying to figure out the best way for us to take our model and our models and converse and convert it to Banker Steel for fabrication.

Fabrication took 19 weeks. And then erection took 18 weeks. I'm talking an erection of 160,000 square feet of area, which encompasses a 94,000 covered indoor football field. And a 6,000 seat amphitheater connected to the stadium. So the workflow. Everyone's done all this stuff but we did it. And we did it all in a condensed period of time. And we reinvented the way that we were working. So this diagram shows the different parts of the five little wheels of action. Populous and the design team-- including Walter P Moore because that was part of the prior collaboration on the onset-- we come in. We design. We started designing. We defined what was a priority, right?

We knew that it wasn't about the floor plan right away. We knew what the floor plan basically was going to be. We knew what the program was going to be. But we had to start down from the top. We had to start with the iconic nature of the roof. And how the roof was going to be, well, iconic. And that's all, I know, totally subjective. But in this term, Shahid Khan had his version of iconic. And we knew that we had to do something different that he's never seen before. So we took 3,000 tons of steel. And then we hung the roof from it. And then we built a fabric box around it. And then encased all that with concrete.

So now we knew that we were hanging a roof from the structure. So from inside the space, you're looking at the waterproof environment with all of the steel penetrating. Obviously that's a task in its hand. So we started the model out in Rhino and Grasshopper to define it. Define the process and what the shape and the form was. That Grasshopper model was concurrently shared with Walter P Moore on a regular basis. That process went through optimization. Walter P Moore could help us rationalize cost. Obviously I mentioned cost is a big deal. We were sitting at a $60 million budget. We were well over that just based on the fact that it was a competition. And then we knew we had to tighten the reins.

So we knew we had to work on the planar area of the building. We knew we had to work on the surface area because it's undulating. We knew that we had to work on the tonnage. And the curved members versus segmented members. We knew we had to work on our drainage and wind upload. And also, this all happened within the optimization window, using our Rhino model back and forth with the Grasshopper scripting so they could track all the nodes, which we'll get into in great detail in a minute. All that went back through Populous. And well, at this point, it might even have had some specialty. A banker might even have been in the conversation to an extent.

So again, taking our Rhino model. So we use Rhino pretty much to package everything out, to share it. Admittedly, we did not use BIM 360 or C4R. It's not because we didn't want to. We were just going so fast that we literally had a script that we kept sharing back and forth. And a model that was just actively being built all the time. So the Rhino was the encompassing package that was shared. And going through the analysis with Steve-- I'll get into some of the things of how we use analysis. Using Excel, especially, is very powerful in this project because Excel can share information. And we talked about good data.

So as long as you have good data, you can do a lot of things with good data. And then back to converging with a Tekla model that Walter P Moore used. And then we took our Rhino model. Converted Tekla for Rhino so that we could illustrate it and show it within the construction documents for permitting. And then that was sandwiched into the production model. This is just a good diagram. It's very illustrative of the parts and pieces. So starting here from the left, I mentioned we had the roof hung PTFE roof, a Sheerfill product if you're familiar.

And then we used our Grasshopper to really organize the process. And to me, it was real important that it was organized. So that we had multiple people in multiple different places using the same Grasshopper script and Grasshopper platform. You know, we always talk about BIM plans, right? This is a thing that should also be a part of that. When you're writing a script or when you're organizing a model, we use this so that each section in the blue-- you chase it up. And then you know exactly where you're inputting your geolocation inputs. And how you're going to modify the primary trusses versus the secondary trusses. In this case, we had primary trusses, bridge and trusses, and then these feed cones.

So we kept a very clear outline within our internal communication. Our BIM plan, if you will, that kind of tried to communicate how you control and draw all this stuff. So the design. So let me just spend real quick and just tell you this design. So the iconic design is performance driven aesthetics. This thing is for the Jacksonville Jaguars and the city of Jacksonville. It was a 50/50 venture. So obviously, Jacksonville-- it's a coastal town of northern Florida. It's bordering St. John's River. Lot of estuaries It's an old shipyards town so it's very industrial.

It's got three of these really beautiful bridges. They're all different colors. They're all illuminated. It's part of their vernacular. So you have this combination of estuaries, these bridges, and the landforms-- because the sea dunes-- and it's just a beautiful place. So Shahid likes that. so we used that as our driver. And then we also knew from a programmatic standpoint, it had to influence sport. It had to invoke their brand. It had to be a part of who they were as a growing team. Finally they're good. So that helps.

We like to blame it on the cover flex field too. And it had to entertain. So now Shahid Khan's trying to blend two programs. Nowhere in the National Football League is there an amphitheater and an indoor training center that systematically unite, which is a one of a kind. Therefore there wasn't much precedent we could take from. We obviously we did tours of amphitheaters and we went all over the country looking up-- flew a half of the country. The eastern side of the country. And we really tried to find a way to bring in entertainment.

It obviously had to interact. He wanted people to move through the building. He wanted it to be iconic. He wanted to be a place-- even if there's no show-- he wanted people to come and hang out. Again, we talked about sculpture. So these are all the drivers that we showed to kind of get him going. We talked about performance from a building, from how it was supposed to operate. But how is building all these parts and pieces we just talked about?

We had to figure out-- so the existing football stadium is connected by the amphitheater. The amphitheater is a three level amphitheater. This section-- when we did a sightline study and we were trying to define how close everything goes-- we studied a lot of different little boutique theaters. Because 6,000 seats is not that big. We're talking Broadway, Radio City Music Hall type of facility. So what we did is we took the Radio City Music Hall kind of [? party, ?] if you will. The section. And we really kind of organized around that. So the sightlines in this place are incredible.

And then that had to abut up to an indoor practice facility. And then of course, there's public and private security reasons that we'll get into. This slide shows kind of the section. And then here shows us the existing stadium, the amphitheater, and the covered flex field with the roof. One thing to note is that EverBank Field has the largest scoreboards. Actually, did Atlanta just take that claim? So let's just pretend that they might-- they're definitely one or two. But they had the largest. Their scoreboards are like 120 feet long on both sides. The scoreboard is this guy right here.

And they have these really accentuated steel trusses to mimic the bridges. So our building-- when I said we flipped it, we put the roof on the underside of the structure. Really wanted to complement that and do it in a good way. So again, we have the estuary. How the building flows, how the program flows. We had the expose bridge with the roof membrane. And then the performance defamation. So from a drainage strategy, a passive look for cooling. Building's not air conditioned so we had to create a breezeway effects of the winds. And how the water drains off the building.

Obviously, you can't just put drain scuppers and downspouts around this whole thing because it's the fabric roof with fabric walls. So we had to be very careful on how we shed the water onto the ground. So this shows some parts and pieces that we really started to analyze. So then Walter P Moore came in. And when we started building the model, and we started doing analysis, and we started to really push the limits on our software, we knew what was important to us. And this is just an image of an elevation of the front facade of Daily's place, showing the scoreboards in the background.

One thing also that's important is that the south Side and the north side. You see it. Well I guess I could just-- sometimes I forget I'm short. So the big opening in this area. These are large aircraft hangar doors. They open on both sides basically. So when they're in covered flex field, it can be an outdoor environment. Because they don't want the football team practice in air condition because they don't play an air conditioning.

But they have the doors for security reasons. Here's a rendering that we did of the amphitheater. So the covered flex field is over here. And the football stadium's over there. And it's sandwiched in with an amphitheater. So you have the lower bowl. You have two upper rings of seating and the stage happens here. You see the inside of the roof. We put the roof again on the underside, which Steve still thinks we're crazy And then you see these light ribs, which is secondary trussing that we use just for paying stuff off of essentially. And then a proscenium for the stage.

And then the inside of the covered flex field. So this is a scenario where, from a design standpoint, we really wanted to be complementary of the industrial nature of Jacksonville. But on the inside, it was about sport, right? You go to a museum and the museums are simple. The museum's about the artwork. It's not about the architecture on the inside. This is the same thing conceptually. So we really want it to be clean. We want it to be a big fabric box that glows.

So you see that there's some program. We have goalpost. And we have big ass fans. Those are really what they're called. And then the rest of the stuff is just fabric. And I have some pictures of this in a minute at the end of the presentation. But this is what we sold him. And this is what he really liked. So now we have to figure out how we make a building like that watertight. How we make the building drain properly. How we do it in 12 months, which is still crazy.

Here's a rendering of the building from ISS. You see one of the bridges. The other bridge I showed you was a green one. There's a red one, there's a blue one. The stadium is very industrial. It's a lot of steel trusses. So the framework of the amphitheater-- this is the cover flex for the amphitheater, is back in here. And then the vercade wall fabric which we, by the way-- first building in the United States to use that fabric. So part of our 12 month period, we then had to go in and get code approval for a brand new product out of Germany, which was interesting.

We had some optimization strategies. We knew that we had to reduce the building width, the surface area of the roof as much as possible. We had to find ways to constrict the bowl so that all the cast was in place-- the concrete was as efficient and economized as much as possible. We had to find out ways to phase. Normally we don't get into means and methods. But in this project, that was a big part of our thing. We had to figure out what needed to be drawn, what needed to be modeled, and how we're going to build it. So that even when Steve will talk about [? the Shorn ?] interaction, that was a big part of our design-- big part of his design-- based on how they pick everything.

So surface undulation. How it's all reinforced and all exposure. So we set up a bunch of stuff that gave us priority. So that when we did analysis, we knew what we were looking at. This just shows from the base scheme. We originally started at 192,000 square feet of top surface. And we ended up down to about 157. Just through this analysis, we were able to do it without losing any time. I mean, I would be so bold to say that we would not have been able to do this without the way that the softwares and the analysis that we used in that period of time.

So then we really got into our lady virgin, our capabilities. So how do we take-- we have a Revit model. We have a Rhino model. We have all the Grasshopper script and the support in it. And from that point, the collaboration between Hunt-Danis construction, Walter P Moore, Populous, and Banker Steel. We really had to dig into the parts and pieces. So Steve's going to talk to you a little bit about some of those parts and pieces that were interesting.

STEVE LEWIS: Can you guys hear me? So yeah, just to follow on from Thom. I think there were three kind of unique things about our workflow. I'll just kind of describe the parametric design. One thing we had to do very early was to use interoperability. I think everybody understands that. I think the way we approached it was to obviously use Excel as a container. And really come up with a neutral file format that we could share between Excel, Populous, and the other design partners.

The other unique thing I kind of thought was good about this project was we all talk about digital workflow. We all talk about Grasshopper. But how do you share that digital workflow? So we had this concept. Well, we could describe it as a recipe. And by sharing recipes, that means somebody can take some somebody else's digital workflow. And kind of repeat it. And you can see, for instance here, this kind of describes the inputs, the outputs and all the files associated with that particular workflow. And then all of that information is kind of stored in a recipe.

The other unique thing. I think we kind of embraced interoperability. And we tried to do that in a neutral file format. So the idea of really coming up with a CSV based file format that was able to be shared between our different programs. So as the engineer, I'm mainly concerned about the primary structure. So that's an analysis done in a program called SAP. Very similar to Robot. In terms of fabrication, Tekla's a standard tool with a high level of accuracy. We can share that information with a steel detailers. For documentation, using Revit. And then for the design driver, we were using Rhino.

So how did our process works? So the first thing, similar to what Tom mentioned, was we started off with the design surface. So this is given to us by the architect. And what we do then is, from that driver, we then instantiate what we think is a suitable framing scheme. So this is a sketch of my boss's idea of how to frame it. And then we start exploring these different treatment options. Another key thing, we kind of did something crazy on this project. And it's that we put fabric on the bottom.

And because of that, you have to be very careful about the drainage. So we went through various iterations with Populous to massage that geometry. And one of the things that digital workflow told us here was, OK. Can we adjust the geometry to ensure we get drainage? So that was, I would say, a good parametric exercise which really informed the design. An, it terms of the engineering, this is a hierarchy. So at the bottom we have what we call the V columns. So these v columns provide both vertical and lateral resistance.

Around those columns is the facade wall which is actually tensile skin. And it's supported by back a steel structure about 60 feet. And that's used to transfer wind loads. On top of that, we have the secondary steel. And then below that we have the ETFE. And then on the top we have the primary structure. So for each of these structural pieces, we had an associated tool that we used.

So obviously for fabric secondary, we used a tool called GSA. SAP was generally used for the primary structure. And then we use Revit as the medium to pull all of these pieces together and coordinate between ourselves and Populous. And again, just to really explain, how did parametrics work in a project? So this is just a typical exercise you would have to do. So taking the inputs, which is a design driver surface from Populous, taking our database information about secondary steel size we need. We then set up a recipe of push the information into Grasshopper and then write around scripts, user objects, custom components to pass that and push it straight into analysis model.

So the task we're trying to do here is take Populous's designed surface, rationalize it to a series of measures, instantiate this like secondary steel, and also introduce the fabric seaming which is also important-- when you analyze fabric, to make sure the seams are in the right direction. Another thing we utilized parametric workflow for was trying to get a high fidelity for wind load. So it's two ways you can do this. You can go by codes. You can open a book. You can rationalize this crazy geometry and try and work out what the wind loads are, which generally is going to be conservative.

Well the other approach is you get a wind tunnel consultant involved. And you try and map these wind pressures to your analysis model. So we had a workflow here where we could take an Excel datasheet from the wind consultants. And then we mapped it to the design surface. And through our workflow we pushed it through to our analysis model. The key thing here as well, like Tom said, is we already bought in to a way to steel. So by having more fidelity in our analysis, we can make sure we're keeping that steel tonnage constant through the design. So you know this is a high level of fidelity but I think in this case it was worth it.

THOM CHUPARKOFF: Yeah. And I would like to also add-- a secondary benefit of this was our owner was trying to insure a building that's never been done before. There's obviously the insurance company asking all kinds of questions. Your roof is 160,000 square feet of fabric. What happens when a panel collapses? What happens? How can you assure me? You know, Hurricane Irma came through and I'm proud to say we had not one tear or loosened fabric. So this also allowed us to be very conscientious about the realities of building construction, and on a coastal climate. So we got a lot of value out of this. And this, the wind engineer was BMT out of London. So predominantly, our New York office is coordinating with them to do that.

STEVE LEWIS: Yes, just to talk a little bit further to what Thom said. So the other thing we have to think about is, we have the fabric on the structure. And we have to make sure that the strength of that fabric can withstand some of the wind pressure which is similar to the hurricane season, which is pretty serious. And as an engineer, we try and present these results graphically. So we can have a conversation with the client and tell them, OK. We have some issues here in that the capacity of the fabric may actually be not as great as the demand.

So what I mean by that is we always compared demand versus capacity. And provided capacity is more than demand, you're OK. So this is a graphical illustration to make sure the fabric is working for the applied wind loads. And again, we also have to sometimes consider extreme load events. So what would happen-- which we're pretty confident will never happen-- what would happen for fabric panel fail? So the way you approach that is just to identify some key zones. And really remove the fabric from the analysis and just see what it did to the fabric stresses.

So this isn't an actual-- what I would call-- a service lowcase. This is us saying, OK. If a panel failed, are we comfortable that the steel support in the other areas and the fabric don't get compromised? Because if you take a panel out, the forces have to go somewhere else. And then as Thom said, we're moving 1,000 miles per hour in this project. And we have to look at ways to save tonnage. We have to look at ways to minimize risk.

And this is an example here where we had to really see, well-- in our original design, we had this transverse secondary steel members. And the potential issue with that is there's more opportunity for problems with waterproofing. So one study we were able to do very quickly in parametrically was had the analysis model set up. And then we could just explore options of taking the steel out. Is it compromising the fabric strength?

THOM CHUPARKOFF: I mean, there's other variables too. It's like Banker Steel-- so like February we said-- hey, we have this building. There's 3,000 tons of steel coming your way. And we need it in six months. And those people-- most of the times, people already have a backlog. A, they're not available. B, they're not interested. And C, they just don't have the material.

So because we're all undulating in steel, there was a period of time where it was like, look. We can't get all that curved steel. Or we can't roll all that steel. Can we find ways to incorporate segmented steel without compromising the design? So that was another part of this analysis that happened concurrently with us just developing and modeling our project.

STEVE LEWIS: So, yeah. Another key problem here is we're moving very fast. And we're having to almost run two parallel models together. So the way we approach that is to consider the modeling of the secondary steel and the fabric and to call GSA. The primary steel was done in SAP. And these two analysis were happening in parallel with one another. And at some point, we have to pull them together and make sure that the forces imposed from the secondary steel on the primary is acceptable. It's not overstressed. And in terms the primary structure, we have what we call the longitudinal trusses. So I think the span, typically, of these was about 300 feet.

And the load from the longitudinal trusses are then resolved into what we call these mega trusses. In terms of the fabric stress, that is carried and resolved into the transverse trusses, which then you carry back to longitudinal. And back to the mega trusses And really, what we're doing here is we have already agreed on a steel tonnage, which we've bought off on. And we're trying to maintain that tonnage as we explore in all these different design options. And this kind of gives you an idea of some of the optimization required. Every single color represents a different steel section. So there's very little regularity in the steel shapes and weights.

THOM CHUPARKOFF: The only perpendicular steel is like the column and in the foundation plate. Everything else of that was an obtuse angle.

STEVE LEWIS: So in terms the power of parallel analysis, one thing we found useful for the parametric design workflows we had is to not use it just as a means of building geometry. But just use it as a means of QC in our model. So simple thing we're doing here. We're just making sure that the applied load applied to our GSA model, which is a secondary fabric model, ties up with that load. You then put it on a parallel model, which is in SAP. So quick QC check within your workflow. Maybe you have a point cloud-- a vector associated with the point cloud, which is your force. You can very quickly sum them up gives you a total load

And then what we had to do then is in a secondary steel model, we have a whole point cloud of vectors representing a point of application of force. So we have to map that through analysis model. So workflow is very good, though we weren't actually building a model. But we're using it to kind of transfer information. And make sure the fidelity was pretty high. So what that means is we're just making sure we're tracking the ID of the vector in one analysis of it to another.

So that's almost like Revit. You have unique IDs. And that allows the user or the engineer in the primary structural model to make sure, OK. You've got a one to one relationship. I'm happy that our forces have been translated correctly. This kind of gives you an idea. This is just a one load case. So for each load case, there's probably 4,000 point reactions. Each one is different. We would go into this fidelity because we've already bought off on a steel tonnage. So we want to make sure that we're taking as much opportunity off parametric and digital workflow as possible to make sure the fidelity of the data is 100% accurate. The other thing to think about is it's not always just about optimizing. You've got to think about structural performance.

So the other key driver we had to always meet was to make sure the structure was within what we call a frequency. That's an idea of how much a structure vibrates. So a low frequency means it's really soft. High frequency, stiff. So as we go into these design explorations with Populous, we're always producing kind of a summary sheet which just tells us, OK. Yes, we've optimized the steel. But it is the structure still stiff enough? And then again, setting up a design matrix where we make some different design options for Populous. For instance, can we remove cables for bracing and introduce rods?

The simple issue there is if you use a rod, you can't get the same amount of tension. And the frequency of your model may actually go down. So I would say, working parametrically and with an architect and engineer, you can sometimes have to set up more performance-based criteria that aren't immediately obvious when you start the design. And then what this shows again is this idea of always tracking the model. Always tracking the weight. Always tracking the frequency.

And when we started the project, I think we agreed to a tonnage of about 3,729 clips. And just for these five different iterations, you can see we managed to reduce that but still keep the stiffness

THOM CHUPARKOFF: Yeah. This is another application where Hunt-Danis construction, our CM-- they were leveraging our track in our parametrics. Also because there is two prices they were getting from a steel fabricator. There's one price for the more complicated stuff. And then there's a different price for the easy stuff that they can get off the line.

So when we're going through this, we worked in tandem with our CM to make sure that the numbers that they thought they were tracking, that the Banker Steel was saying that they're ordering-- that we could always keep those two things simpatico, if you will.

STEVE LEWIS: Then the other piece we got involved in is not just the engineering. But we have to make sure that all these pieces can come together. We have to make sure we can fabricate it. And the complexity of the structure means it's a high level of detailing required. That is, how does the fabric look connected to secondary? Do we have the waterproofing?

THOM CHUPARKOFF: So this is an area where we blurred the line. Remember I told you it was a design-bid-build process. But we just completely threw out the window as soon as we got going. And Walter P Moore, Populous actually took on the onus of the shop drawings for the fabric. So instead of us doing a structural steel layout, giving it to the contractor, a contractor gives it to the fabricator they do shop drawings, they give them back to us, and we look at it-- we did the shop drawings for them. So they basically moved money from the construction budget.

We wrote an ASR, an architectural service request, to do that service on the front end. So it was Walter P Moore and Populous, we're doing that from the onset. And then furthermore, when we did the fabrics, we then we hired a spec. There's a company called StructurFlex that did all the fabric roofing. And additionally, instead of them hiring somebody else to do the detailing in shops for them, they hired Walter P Moore to do that. So that they were one entity to do everything. So there was no shop drawing process. Their drawings from the onset were the fabrication drawings.

STEVE LEWIS: So the way we approach that is that we have the idea of this kind of common file format that we're distributed between different tools and between different partners. And you know, very early on, we kind of set up a way of IDing a lot of these elements and making sure we're able to track that ID as we go through different programs.

I mean this is kind of useful because, you know, if you're looking at stiffening in a structure, maybe you want to stiffen two members. You want to make sure you can track which member you've upsized. And you want to make sure that goes through to the fabrication model.

So here we had a workflow where we would push the information from the Excel database into GSA. And that information flowed through to the design model in Rhino, which kind of is indicating what these IDs are. And really, that also helps with documentation.

I mean, as Thom said, there's no-- how do you say? No real grid lines in this project. Everything is organic. There's no-- I don't think there's any symmetry in it. There's just nothing.

And the only way we can really approach that is to kind of document it in terms of identifying unique IDs or node points and, through an indexing strategy, be able to coordinate those geometrical data points between the fabricator, ourselves, and also the fabric contractor.

And really, this gives you a sense of the fidelity of some of the fabrication and digital modern we had to get into. So the image to the left is really a Tekla model. This is an image of the mast being fabricated in the shop. And then this is how it all came together in the field. And you can see you've got a pretty good resolution between this and the digital.

So this is all done by Walter P. Moore running in parallel. And again, this idea of indexing, the importance here is, for this connection, it has to be engineered. So that means we have to be able to transfer our connection forces from our analysis model. And obviously, if you know what the IDs or the index are of those models, you can easily extract the forces. You know, you're coordinated with the detailing.

And again, you know, we have our kind of basic database of IDs, section shapes. We kind of bite off between ourselves what information we're going to carry between the different platforms. And then we're able to very quickly push the information say from a SAP model into a Tekla model, obviously utilizing the API, writing your own custom object. Or it probably wouldn't be a custom object. It would be your own custom component.

And then that information was pushed directly to a Tekla model. So every element was IDed. And really that's the bare bones model. So a Tekla model isn't just about the steel. You have to build all the detail in on top of that.

And this kind of shows I would say the complexity of it. So you can see we're starting to instantiate some connection details. So each connection detail needs to be coordinated. It needs to be engineered. And it needs to be represented in 3D.

The problem is, if you remember, you've got fabric on this model. So there's a lot of coordination, a lot of busyness going on. Another few screenshots.

So as Thom said, we actually had to take some ownership of what we're calling the tertiary detailing. So every tab that the fabric was attached to, that was also brought and put into the Tekla model. We used Rhino and a Grasshopper script there to instantiate the tabs. And then I think that was just exported as a IFC or [? IGES ?] into Tekla.

And then you can really start to see the complexity and the coordination needed. So we have a very basic, primary structure. So we try to save costs by regularizing the primary structure size, which is segmented.

And then we put the kind of expense in rolling the secondary steel. So the secondary steel is where we make sure we have enough curvature. We want to make sure there's drainage. And then all of this has to be coordinated.

So I think you'll have some shots to show us.

THOM CHUPARKOFF: Yeah, yeah, yeah.

STEVE LEWIS: And very quickly, it's all very well being able to say, OK, this structure is actually erected. We're comfortable all the forces and deflections are adequate. But you have to think about how do you construct it.

And you know, you can see just from this rendered image that there's a huge degree of complexity here. I mean, how do you construct something like that? And having this ability to kind of label, ID, and group elements through a common kind of platform, it allows us to explore different erection schemes. So much like in Revit where you can filter and create worksets, maybe by attributes, if we know what our index are for some of these elements, we can start exploring different erection schemes.

And the key thing here is, when you explore the erection scheme, you have to analyze it. What would happen if one your elements is over-stressed at this point? Well, you've got a problem. Right?

So sometimes, your erection sequencing can control your member sizing. And the way we approach this is to set up what we call a stage. And then we do a staged analysis. So you literally think about how you want to construct it, and you go through all these different stages, one after the other, to make sure that your stresses are adequate.

And I think the image on the right kind of shows the erection sequencing going on for the installation of secondary steel.