HANGING

OUT in
Salt Lake City
BY PATRICK M. HASSETT, SE, AND JORIEN BAZA, PE

JANUARY 2017
 An elaborate erection and jacking scheme
involving an innovative roof-top truss helps pull together the
Utah capitals newest Class A high-rise.

FROM THE OUTSIDEand from F D C B A

1 2
E
three directions111 Main might look Roof
3
4
5
like any other high-rise. 6 7

But from a steel design and construction Penthouse

perspective, the project was on a whole new
Level 24
level when it came to erection challenges. For
Level 23
one thing, the 25-story, 502,000-sq.-ft Class Jacking Rods at
Level 22 Roof Truss
A office tower, located in the heart of down-
Perimeter Nodes,
town Salt Lake City, extends 46 ft over an Level 21
TYP
adjacent structure. All 18 perimeter columns Level 20

are hung from roof trusses framed with jum- Level 19

bo W14 shapes (with heavy node weldments) Level 18 Level 19
that rest on six structural spherical bearings Level 17 Level 18
at the top of the concrete core, a new scenar- Level 16 Level 17
io for the erection engineer and steel erector. Level 15 Level 16

Level 14 Level 15
Tightening Up Level 14
Level 13
Temporary columns were the obvious
Level 12 Level 13
solution for the north, east and west sides,
Level 11 Level 12
but the south side required a different solu-
Level 10 Level 11
tionone that would allow steel erection
to start at level 5, which would then serve Level 9 Level 10
Temp Bracing shown
as the support base for the floors above. Level 8 Level 9
Red, TYP
The design team at architect and structural Level 7 Level 8

engineer Skidmore, Owings and Merrill, LLP Level 6 Level 7

(SOM) envisioned a saddle cable system that Level 5 Level 6
would employ cables through the core and act 3
4 Level 5
5
to balance the placement of construction loads Level 4
Temp Columns at
while eliminating any eccentricity on the core Cable Jacking
Level 3 North, East and West
Frames
that would affect the stress, strain and creep of Sides, TYP
Level 2
the reinforced concrete core walls. The chal-
lenge for the steel erection team was to come Transfer Level
up with a practical solution of framing and Level 1
cable connections that allowed the tensioning
of the cables to account for sag and stretch, as
the cable loads increased while the building The temporary truss scheme.
was being erected. Working with erector SME,
erection engineer Hassett Engineering devel-
oped a steel-framed concept that minimized Pat Hassett (pat@
the cable length, thereby minimizing cable sag hassettengineering
and stretch. .com) is president and
The temporary framingreferred to as Jorien Baza (jorien@
jacking trusseswas designed to jack up hassettengineering
the building columns while simultaneously .com) is a project
engineer, both with
Hassett Engineering, Inc.
The south side of the building cantilevers
46 ft over an adjacent structure.
All images courtesy of Hassett Engineering
unless otherwise noted.

 Modern STEEL CONSTRUCTION

 SOM

The construction sequence of the saddle cable system.

tensioning the cables. Column lines 3, 4 and 5 would be framed and floor beams would rotate about the core walls. Concrete
with jacking trusses and the saddle cables supporting the south pour strips were implemented adjacent to the core, and spe-
side, and anchored with similar framing on the north side to col- cial pinned beam connections were developed to allow rotation
umns that extended down to the foundation. without adding extra stresses into the system.
After extensive review of available cable types, 4-in.- diame- As is typical for a building of this type, the permanent columns
ter ASTM A586-04 were chosen. The cables were pre-stretched are designed largest at the top where the tension is the highest,
to provide an effective modulus of elasticity of 23,000 ksi and and smaller at the bottom. During construction, however, the
rated with a breaking strength of 997 tons (which translated columns were supported from below and saw large compressive
to an allowable capacity for construction loading of 997 kips). forces most notably in the lower columns. Through collabora-
There were two cables per grid line for a total of six, and they tion with SOM, the columns were checked and resized as neces-
were installed through the 30-in.-thick concrete core walls on sary to handle the temporary construction loads.
a circular radius in a lubricated steel conduit. Column lines 2
and 6 were coupled to lines 3, 4 and 5 through temporary brac- Tier by Tier
ing on the south face of the building between levels 5 and 9, Erection of the composite steel construction followed
and were consequently supported by the jacking trusses as well. a typical sequence of two stories at a time, tier by tier, and
Lines 1 and 7 had a different problem: There was no core to concrete slabs were poured via normal sequencing four floors
anchor through, so a braced frame was designed to carry the (give or take) behind the erection. The column lift (via jack-
hanging and the associated lateral component. ing) was estimated to require 38 in. at every tier, accounting
Transferring load from the temporary shoring system to for deflection due to two slabs and two floors of steel. The
the roof hat truss once the structure was complete was another jacking frame was designed to deflect by rotating about a pin
challenge. The initial concept was to pull up the perimeter col- connection at the core. The two 500-ton jacks at each column
umns using jacks at the hat truss nodes. Through collaborative were designed to push down on the jacking frame while push-
meetings between SOM, SME and Hassett Engineering, this ing up on brackets welded to the flanges of the columns. This
scheme was changed to a lowering of the columns at the base. would provide the elevation adjustment for the cantilevered
The solution allowed gravity do the work as opposed to pull- south side throughout the project. Loads were measured on
ing up, fighting against gravity and the stiffness of the building the jacks and compared to the theoretical loads, which were
itself. With this scheme, the perimeter of the building needed calculated using an ETABS model.
to be erected to a higher point than the final theoretical eleva- However, after a few tiers were erected, the jack readings
tion. Since the core does not move during load transfer, the slab were showing higher loads than expected. Unable to capture

JANUARY 2017
 The cable jacking frames and building core.

the behavior in the ETABS model, the team determined

that some load transfer (by plate action of the slabs) was
occurring between jacking lines 3, 4 and 5 and braced
frame lines 1 and 7. Connections for the jacking trusses
and the bracing for lines 1 and 7 were reinforced to accom-
modate the additional loading, and the jacking magnitude
at each adjustment was minimized to keep the floor eleva-
tions within acceptable tolerances while ensuring that the
jacking frames would not be overstressed.
As with any high-rise project, the team had to pay close
attention to the differential shortening or compres-
sion strains of the perimeter columns relative to the core
structure. But in the erection scheme for this project, the
columns were in compression, then finally in tension, and
the floors lost elevation due to the deflections of the roof
truss. Anticipating this in the design phase, the team per-
formed analysis to determine what, if any, column length
changes would be needed. Column lengths per tier would
be required to be 18 in. shorter than theoretical. However,
SMEs field experience told us that there would be ap-
proximately 18 in. of weld shrinkage at the column splices.
Therefore, the columns were detailed to their theoretical
lengths, thus eliminating complexity in the modeling for
the column shop drawings.
Since the columns would compress during shored con-
struction, then stretch and deflect (due to hat truss defor-
mation) after hanging, a specific, unique set-high eleva-
tion was required at each column of each floor that would
vary at any given stage of construction. The erection plan
also required the bottom level of each column line erected
to an initial set-high elevation, varying between 3 in. and
4.5 in., to account for deformation of the temporary steel
during construction, final column stretching and final hat
truss deflection.
Similarly, the perimeter nodes of the roof hat truss were
erected to a tip-up elevation between 1.5 in. to 2 in. to
account for the final truss deflection alone. To complicate
matters, the schedule required the curtain wall glass panels
to be installed during steel erection, prior to jacking down
the building and transferring the building weight from the
shoring to the roof truss. The glass panels were designed
to allow greater than normal tolerances due to movement
after installation. As such, general contractor Okland
Construction performed extensive survey monitoring on
a weekly basis so that fine adjustments to the preset eleva-
tions of each column could be made in order to ensure the
integrity of the panels. The target maximum differential
vertical movement between adjacent columns was +/- in.

A jacking frame and column bracket.

 Modern STEEL CONSTRUCTION

 A 1
C 3

D 4
Erect Downriggers shown in
Green, on Grids C and D, 1
Plumb up and begin welding 5
ASAP

F
6
Penthouse
7

Level 24

Level 23

Ensure Dywidag Rods are

fully installed in nodes and
nuts are tightened
@ 1,7 / C,D.

The main axial members of the hat truss (in green).

Making Adjusments Relying on the carefully developed survey data provided by

Erection proceeded quickly up to level 24, with the south side Okand Construction, and correlations made with ETABS erec-
column elevation adjustments made close to the theoretical estimate tion sequence modeling, adjustments at the roof were done in
of 38 in. per tier. As the stretch of the cable continued, it would slip four phases:
through the core, occasionally breaking friction with a noticeable 1. The four outliers on the north were jacked in.; these
bang. As the roof truss began erection with the heavy members and column grids then could provide some pretension for the
nodes, the concept was to start at the core, landing the nodes on the supporting hat truss, acting as an abutment
structural spherical bearings and work out towards the four perim- 2. The four outliers on the south were jacked 1 in.
eter column lines. This procedure allowed welding to start at the 3. The columns on line A at 3, 4 and 5 were jacked in. to
core and work outward, thus allowing relatively symmetrical and provide further pretension for the next jacking.
unrestrained weld shrinkage. In addition, the main axial members of 4. The columns on line F at 3, 4 and 5 were jacked 1 in.,
the trusses could soon be able to support their own weight, carrying in. and 1 in., respectively. By making these small final
load from the perimeter toward the core. adjustments at the roof level, the roof trusses were ef-
Jacking rods were installed on the bottom chord of the roof fectively preloaded and the erection team was able to set
truss as a backup plan to make further elevation adjustments, the geometry closer to the theoretical elevations prior to
which would involve pulling up the columns and pulling down jacking down the building
the trusses. This backup scheme proved to be very useful in ad- Once the columns were welded to the nodes at the main truss
justing the floor elevations, as the floor elevations at the south lines (3, 4, 5, C and D) and the rods were engaged at the out-
side and outlier columns were lower relative to the other grid lier columns, the building was ready to be lowered into position.
lines. These outlier columns were columns outside of the These outlier columns were expected to deflect more than the
main truss lines and were supported by perimeter trusses off main truss line columns. Therefore, they were left un-welded to
the main trusses at the roof. Because the measured loads during the nodes so that they could be jacked later for a final floor eleva-
jacking indicated that more jacking would stress the saddle ca- tion adjustment at those grids.
bles beyond their intended design loads, the decision was made Once the roof welding and inspections were complete, the
to leave those three columns relatively low. building was ready to be jacked down and the load transferred to

JANUARY 2017
 Jack North Outliers
A 1
+1/2" @ A2, A6, B1, B7 0.2
2
Jack North Side
B 0.4 +1/4" @ A3, A4, A5
3

C
4

D 5
Jack South Outliers
+1" @ F2, F6, E1, E7

0.3 6
E
7
F
Roof

Penthouse

Level 24
Jack North Outliers
0.2 +1/2" @ A2, A6, B1, B7
Level 23

Level 22

Level 21

Level 20
0.5
Level 19 Jack South Side Jack South Outliers
+1" @ F3, 0.3 +1" @ F2, F6, E1, E7
+3/4" @ F4,
Level 18 +1" @ F5

Level 17

Pre-jacking adjustments to the hat truss.

the two-way roof hat truss system. A hydraulic control system required in. of jacking-down at level 5, a total jack movement of
was placed on level 5 at the south (line F) to control the six jacks 3.5 in. was required. This was due to the jacking trusses moving up-
at the three jacking frames, and another hydraulic control system ward (since they were unloaded) at the same time the building was
was placed at level 1 (ground level) to control the 22 jacks and 11 moving downward. After each 1-in. lowering step, jacking forces
columns. As the jacking-down process proceeded, surveyors were and survey elevations at levels 5, 15 and 24 were reported.
placed at level 5 and at the roof and were in radio contact with Elevations and loads were compared to theoretical predictions,
Oklands superintendent. SME, Hassett Engineering and SOM and the curtain wall deflections were checked to be within toler-
had continuous communication between the ground and level 5 ance before proceeding to the next step. This process was repeated
jacking, giving the go-ahead for each simultaneous jacking-down until the jacks were unloaded and the building was fully supported
of 18-in. increments. The glass curtain walls and MEP systems by the hat truss above. Total jacking-down at the columns on the
were being monitored by their respective subcontractors as well. ground varied between 3 in. and 4 in., and total jacking at level 5
As the load was transferred to the hat truss and the load on the was about 8 in. for a column movement of approximately 1.7 in.
temporary cables was reduced, cable slippage back through the Cable movement through the core was between 3 in. and 3.4 in.
core was observed as the cables occasionally broke friction with Further adjustments to column elevations were inevitable
noticeable bangs, as with initial erection. since the floors would continue to deflect due to the pouring of
the level 25 concrete slab, the removal of the temporary braces
Winding Down and the completion of the faade and remaining dead load. The
Jacking-down was performed in small increments until an ac- worst predicted deflection locations were at the outlier columns,
cumulation of 1 in. at the ground and in. at level 5 was achieved. as they were supported by the more flexible parts of the hat truss.
Column elevations and jacking loads were then recorded, reported After load transfer, about in. of jacking was performed at the
to the team and reviewed; upon approval, jacking-down would roof level on the four outlier columns on lines A and F, north and
recommence. A three-to-four ratio was used since previous fine south, respectively. Removing the temporary braces or releasing
adjustments at the roof jacking rods had preloaded the south side the connections caused a redistribution of forces throughout the
more than the north, resulting in less required jacking-down at the building and resulted in further deflection, predominantly at the
south for full load transfer. Furthermore, in order to achieve the outlier columns on the south, where the braces were retaining

 Modern STEEL CONSTRUCTION

 A tension release at a braced
end-plate connection.

much residual tension load. This was done

by loosening the tension rod nuts in the
end-plate connections of the braces, with a
resulting gap spread of between 116 in. and
3
16 in. between end plates.
Finalizing the roof slab pour only added
to the deflections. In response, and in an-
ticipation of the remaining faade load and
other dead loads, a final adjustment of 0.4 in.
was performed at the two south outliers on
line F in order to bring the floor elevations
closer to theoretical. These outlier columns
were then welded off, essentially securing
the building into its final state and intended
structural system load path. The faade, hav-
ing survived several changes in elevation and
geometry, was given a final adjustment.
The steel structure of 111 Main was
built from the ground up and went from
being supported at the bottom levels
to hanging from the hat truss above
basically 18 in. at a timeproving that
steel was ideal not only as a framing system,
but also in terms of constructability for an
especially intricate erection operation. 
Owner
City Creek Reserve, Inc., Salt Lake City
General Contractor
Okland Construction Company, Salt
Lake City
Architect and Structural Engineer
Skidmore, Owings and Merrill, LLP, San
Francisco
Erection Engineer
Hassett Engineering, Inc.,
Castro Valley, Calif.
Steel Fabricator, Erector and Detailer
SME Steel Contractors, Inc.,
West Jordan, Utah

Jacking adjustments at the roof node.

JANUARY 2017