San Antonio, Texas, has sprouted up along its waterways, some more attractive than others, and for many years, the San Pedro Creek has had about as much appeal as you would expect from a drainage ditch.
But in 2014, a few years ahead of the city’s 300th anniversary, local officials moved forward with a plan to upgrade the channel from a utilitarian flood control system to a cultural centerpiece for the city. The project would restore water and aquatic habitats to a creek that had been practically dry for 20 years, and preserve some historic walls while introducing an undulating and art-centric design with public spaces and art installations.
In essence, “almost every complicated aspect of a linear project is being rolled into an existing drainage ditch,” said Eric Cylwik, virtual construction engineer at Tucson, Arizona-based Sundt Construction.
Not to mention, no 5 feet of the 1.2-mile linear park are alike. “There was no formulaic way we could approach this and say, ‘this is the way we create the channel.’ Every 25 feet, it was understanding what was happening with the design and existing [conditions] and then applying the correct solution to that,” he said.
Twists and turns
With all these variables, it became clear that the traditional, siloed modus operandi wasn’t going to cut it. Omaha, Nebraska-based HDR Engineering, which handled general structural, hydraulics and other design work for the project, revisited the design-bid-build project delivery method altogether after stakeholders called for a “significant” revision to the already over-budget design, said Jeff Mitchell, senior structural engineer at HDR Engineering and project manager for the design team.
The pressure was on to deliver a significant portion of the project by the tricentennial celebration on May 5, 2018, he explained in a recent presentation, and “we realized at that point we couldn’t do that and deliver it in the same way that we had been proposing to.”
HDR got the owner on board with the construction manager at risk (CMAR) framework by pointing to the early coordination that would take place between design and construction, as well as the potential for faster speed to market, better cost certainty and early release packages. Bexar County and the San Antonio River Authority awarded Sundt and joint venture partner Davila Construction a $175 million construction manager at risk contract in 2016, “as we were going from the first 70% to the second 70%” of design coordination, Mitchell joked. The team is eyeing an April 2021 substantial completion date.
Work on the channel was akin to “going in and renovating somebody’s artery while they’re alive,” said Cylwik, and a number of risks were posed just by virtue of the channel’s existing conditions and function as a flood control system. The CMAR method, he said, was better able to allocate this risk in August of 2017, for example, when four inches of rain in four hours submerged some parts of the jobsite in 11 feet of water. Not only were contingency dollars linked to the potential for that type of event, but the risk had been thoroughly addressed by design and construction during the preconstruction phase. A solution to over-excavate and place drain rock at the end of the site had crews back to work in two days instead of a week, Sundt says.
Other challenges the team had to account for were the tight right of way on either side of the project, with historic buildings hugging the channel in some locations, and the haphazard manner that utilities had been added to the channel as the city was built up around it, said Mitchell.
“They ran out of space in the ground and so they just slapped them onto the walls,” he said.
Presenting a far greater risk, though, were surrounding utilities that were out of sight. The team couldn’t risk striking a conduit line during excavation or accidentally damaging a historic wall, for example, so uncertainties had to be addressed thoroughly and early on. To do so, Sundt and HDR leaned heavily on 3D design model coordination from the start of their CMAR reboot.
Breaking down silos to construct new walls
Although it was only contracted for 2D linework, HDR decided that 3D modeling was its best bet to rein in extensive risk on an aggressive project schedule. The firm modeled the elements that it determined would be the most difficult to coordinate — including retaining walls to show the depths and extents of footings, and storm drains — and also knew that the incoming Sundt team would have an appetite for this information, Mitchell explained. In most cases, HDR’s modeling was done to a 300 level of development.
HDR provided Sundt’s VDC team with its AutoCAD MEP, Civil 3D and Infraworks models. But before the contractor began to pull quantities from that data, it had thorough conversations with HDR about the level of detail used for each model element. “That allowed us to stay away from assumptions that weren’t correct about what was in the model,” Cylwik explained.
With any given section of block wall, for example, his team needed to know the number of blocks needed, the volume and length of the footing, the area of geogrid needed to construct it, the volume of gravel backfill, and more importantly, how much they could rely on the existing models to extract these quantities from them.
“There’s a pretty complicated picture that we have to build before we can even start to assign costs to it,” Cylwik said.
Assigning cost to these quantities is simply “outside the realm of design,” Mitchell said, noting that Sundt was not repurposing the design model but rather layering on additional construction-specific data in a parallel modeling effort.
Sundt used Land XML alignments and profiles, Civil 3D corridors and solids, surface data provided by subcontractor Pape Dawson Engineers and CAD linework as inputs to a parametric model that was “more pertinent to constructors,” said Cylwik, to include details like a trench around a utility. The contractor also used Revit to model bridge and gatehouse structures, which it then pulled into Civil 3D and supplemented with retaining wall information. Compared to a traditional 2D quantity takeoff, the 3D environment is “much more reactive,” he said. “If something changes, the model changes with it.”
Sundt pulled survey data and quantity information; cut cross sections every 25 feet, so that sequencing could account for many variations; then added Agtek data to plan for mass excavation. The design model appears to be a truer representation of what the finished product will look like, but omits details like these because they’re “not a pretty part of the project,” said Cylwik, nor are they relevant to what design contributors do.
Nevertheless, the construction model was made possible through design’s decision to work in a 3D environment. “That really allowed us to be able to get in and keep pace with design iterations and explore design options to ultimately bring this into the right budget,” he said. Modeling work that would have taken about 200 hours per iteration took 40 hours instead, thanks to HDR’s willingness to break down a silo and share its design intent data.
Interoperability from design to the field
Interoperability between programs was a key ingredient to the coordination, Cylwik noted, and this extended to model use during the construction phase, as well. Operations employees viewed the models in Navisworks or skimmed them from an iPad using BIM 360 Glue, Chad Yount, senior project manager at Sundt, told Construction Dive in an interview.
The use of 3D modeling also saved his team many of the hassles that can come up on a more paper-based project, he added. Most clashes had been mitigated before reaching the field, and tasks Yount once thought impossible during an initial walkthrough of the site had been mapped out and sequenced.
Furthermore, the model became a powerful visual tool to help workers — who may have joined the project at a mid-way point or only for a few months — to understand their tasks in the context of the full project plans, Cylwik told Construction Dive. One superintendent, he said, would review the model with his crew at the beginning and end of each day to give workers a “clear picture of what they’re working towards and what the expectations are for what the product will look like when it’s done.”
More than a few milestones speak to the success of this model-based and collaborative approach so far, the team says. After several twists and turns, the first phase of the project successfully debuted at the city’s tricentennial.
Another win, the team said, is that the dynamic between stakeholders has been, and continues to be, one of partnership and respect. “We don’t typically fire off an RFI and wait for a response,” Yount said, for example. Instead his team is in the habit of proactive discussions and personal interactions whenever possible, he noted, because “as soon as you can get to know somebody on a personal level … it’s harder to think that they have bad intentions.”
The team returned roughly $3 million to the owner in unused allowance money after the first segment was completed. They were able to circumvent some risks altogether, according to Mitchell, a feat that siloed stakeholders would have a more difficult time achieving.
Finally, the area has since seen ample economic returns in the form of new building construction, school districts and more, according to Yount.
“It’s thrilling as a constructor to be able to say we helped breathe life into this area,” said Cylwik, “in a way that’s responsible with taxpayer money, that really allowed us to be aggressive on a construction schedule and [to] keep a great relationship with the client and design team.”