...based on following the prevailing 2007 building code rather than the 2001 code, which Vi�oly used, and involved breaking up the building into separate sections. Little was lost in terms of the quantities of chemicals allowed for research.

“Marianne O’Brien’s idea for the occupancy change was the great contribution that won them the job,” says Bade.

The DPR team found other ways to cut costs and simplify construction, including eliminating Vi�oly’s “underbelly of the beast,” an expensive-to-build mechanical level hung below the two-story lab. The solution embedded the mechanical room into the office level. “We could do this because, with design-build, we knew exactly how big the mechanical units were,” says SmithGroup’s O’Brien.

Eliminating the underbelly had a positive ripple effect, she adds. The gross area in the bridging documents was 72,416 sq ft, with 45,986 net-assignable sq ft, which translates to 63.5% efficiency. In DPR’s proposal, the gross area was 67,613 sq ft, with 46,223 sq ft assignable, for 68.4% efficiency. The area ended up at 68,500 sq ft, with 67.4% efficiency. “This is the most space-efficient lab we have ever done,” says O’Brien.

The team found other ways to trim costs. The structural concept had the superstructure—a conventional braced frame—sitting on an external steel substructure. The architecturally expressed substructure consisted of a series of space trusses, sitting on seismic base isolators.

The space trusses, which had few vertical members and many sloped columns, would have required a great deal of falsework during erection, which would have been in the way of other work, slowed the pace and added cost.

So in the proposal stage, Forell/Elsesser, steel contractor Schuff Steel Co., San Diego, and DPR concentrated on developing a more constructible scheme.

The result is still a space truss, but it has more vertical members—one at every grid line. The vertical columns provide a direct load path with diagonal braces acting as lateral-load-resisting elements rather than as sloped columns. “Adding columns appears counterintuitive, but it eliminated shoring and helped the schedule,” says Forell/Elsesser’s Naaseh.

Structurally, the building consists of four connected pods that step up 27 ft with the grade of the hill, from east to west (see drawing, p. 74). The two-story superstructure is an ordinary concentric-braced frame. The substructure is a space truss using ordinary concentric-braced- frame connections.

The lab’s east end cantilevers 30 ft; the west end cantilevers 45 ft. The north side is a 24-ft propped cantilever. On the south side, the structural concrete foundation consists of 22 drilled piers, 2.5 to 5 ft in diameter. On the north side, each isolator sits on one of twenty 5-ft-square pedestals. Pedestals vary in height from 5 to 25 ft, to take into account the different grades, and sit on pile caps over small-diameter drilled piers.

High Seismic Zone

The concept included base isolators, which let the building “float” over the foundations during shaking, because UCSF wanted a facility that could be occupied after a major seismic event. Base isolation, which absorbs seismic loads, also allowed the use of unusual and irregular shapes in the high seismic zone because it allows conventional superstructure framing. “Without base isolation, this project would have been cost-prohibitive because of the high cost of foundations associated with a fixed-base structural system at this site,” says Naaseh.

The project’s isolators, which consist of opposing curved-surface dishes sandwiching an articulating slider, allow the building to move 27 in. in any direction (see drawing, below). The dish’s curvature provides the restoring force that lets the building settle into position. The lower dish is bolted to the foundation; the upper dish is bolted to the steel structure.

With base isolation, “nothing but gravity” keeps the upper dish and the building above it from sliding off the foundation, says Steve Marusich, Forell/Elsesser’s project manager.

But the lab is long and narrow, with little weight on the south side. Analysis during the proposal phase revealed that, under certain shaking conditions, the entire south side would lift up to an inch for more than one second.

The building would behave much like a seesaw, with a fulcrum at the lower line of isolators on the heavier north side, says Marusich. “This is very unusual [behavior] for an isolated building,” he adds.

The engineer realized an uplift restraint device, to absorb tension, was needed. Initially, the idea was to order an isolation device that would double as a restraint. But the system needed only eight restraints, which made the custom items cost-prohibitive. The long lead time also raised schedule concerns.

Instead, Mason Walters, a Forell/Elsesser principal, designed a restraint that Schuff produced. It’s unusual for fabricators to make “Swiss watches,” he says.

In the building, a restraint is paired with every third isolator along the south grid line, between a pier and the steel structure. The restraint had to be honed to resist 100 tons of uplift anywhere while exactly tracking the performance of the isolator without restraining it from performing horizontally. The restraint also had to be something the fabricator could make quickly, says Walters.

Each restraint consists of two track beams perpendicular to each other, with roller assemblies (see drawing, p. 75). The 5.5-ft-long lower beam, bolted to the foundation, has a curved upper flange; the upper beam, bolted to the superstructure above, has a curved lower flange. A movable connector assembly, like a chain, is between track beams.

The whole device articulates in two directions, potentially at once, with its paired isolator, says the engineer. “It acts like a big clamp, but only when the building requires it,” says Walters. “Unless there is uplift, the building does not know the rollers are there,” he explains.

To convey the design to DPR, the school and Schuff, Walters made a one-third-scale wood model, complete with...