This is part 5 of 6
Engineering analysis and implementation
Engineering investigation proved what was already expected: Corbin’s iron frame was designed as a gravity-only structure — fairly typical for the time — and it clearly relied on the various masonry elements of the façade for lateral stability. The plan (Fig 29) shows the main stiffnesses of the lateral system to be along the length of the extreme north perimeter wall and in and around the main building core at the wider east end.
Though lateral east–west loads would clearly introduce some torsional irregularity, equally clearly there was enough solid masonry (over 152ft (46.3m) length) to resist any overturning. However, consideration of lateral north–south loads gave immediate cause for concern: there was no significant solid element anywhere close to the Broadway end of the structure, as the only wall present, the façade, was highly punctured by window penetrations. The desire to form multiple new penetrations in the north wall (Fig 30), the single strongest element of the building, would also change the load paths within the walls and potentially overstress parts of the historic unreinforced masonry.
To assess both the existing condition and the proposed alterations, Arup built two 3-D ETABS structural models, one for the existing and one for the proposed structure (Figs 31–32). These were accurately detailed from topographical survey data, including the north wall’s curvature by over 1ft (300mm) in the middle (presumably introduced at the time of construction as a response to poorly surveyed lot lines).
This allowed the current stress regime in the unreinforced masonry to be reviewed and then compared to stresses after the proposed removals. It was hoped to keep the change of stress within elements to less than +10% when considering lateral loads and less than +5% when considering gravity-only loads, as this would avoid triggering seismic upgrade. (The building upgrade was designed in accordance with the NY State Existing Building Code 2002, which allowed for these modest increases of stress for an existing building as a pragmatic approach to managing old building stock).
But the news wasn’t good. The results indicated Corbin already to be performing badly under north–south lateral loading, and the proposed changes would make it worse.
As the internal floor plan is relatively small (around 2500ft2 (232m2) per floor) and the wedge-shaped geometry further restricts placement of walls (and the existing façade is both original and decorative on both faces), it was considered extremely inefficient and counter-productive to try and reinforce the building within its own footprint. The team therefore decided that tying it to the Fulton Center to resist north–south loading would be more efficient, and allow most of the new lateral load structure to be outside Corbin’s floorplate where there was significantly less pressure on the real estate.
It was still necessary to review the increases and concentrations of stress introduced by the many new penetrations of the north wall. Analysis showed these could be controlled within acceptable limits without intervention above level 2. However, between level 2 and street level, the thresholds previously defined were exceeded.
The solution was borrowed from a flexible approach used to seismically upgrade masonry buildings on the US west coast. The walls between level 2 and street level were encased in a shotcrete layer 4in (100mm) thick on each face. This layer was heavily reinforced in-plane to provide some ductility, and anchored to the existing masonry wall by resin-dowelling several thousand L-shaped reinforcing bars at a 2ft x 2ft (610mm x 610mm) grid across the surface (Figs 33-35). The 3-D ETABS model was used to review and then rationalise and reduce the overall amount of reinforced surface to meet the code overstress criteria.
Lateral stability frame
Linking Corbin with the Fulton Center pavilion at lower levels allowed much of the lateral shear forces to be transferred to the new structure, which could be designed to resist them adequately without the constraints on floor space in Corbin itself.
However, this only partially solved the problem. As the structures could only be effectively tied at levels 2 and 3 because the Fulton Center had much reduced stiffness due to its own geometrical constraints above this level, a means was needed to convey the lateral loads from roof level (9) to the Corbin street level back to the ties. At street level and below, introducing the escalator wellway void also compromised the effective diaphragm action of the floorplate, and it was necessary to replace this action by a series of lateral framing systems described below (“Escalator wellway”).
The solution adopted for the above-grade transfer of lateral loads was a concrete moment frame, which:
(1) would allow east–west passage of both people and MEP services through the frame
(2) was a flexible form of construction that could be field-adjusted to suit existing conditions and potential variability of wall alignment much better than steel
(3) could be easily formed into moment frames without expensive connections
(4) could interface easily with concrete floor diaphragms without difficult or expensive connections
(5) could wrap around existing structural members, allowing them to be retained in situ; this reduced the need for temporary supports to account for existing member removals, and risk of structural movement if existing members were removed.
The frame was located on plan as close to the west end as possible, while still allowing for a horizontal connection to the Fulton Center pavilion. The concrete frame was added into the 3-D ETABS model and the connection to the Fulton Center was modelled as a series of springs.
The Fulton Center superstructure had been modelled separately in GSA, so it was necessary to iterate lateral loads and corresponding spring stiffnesses between the two models, adjusting framing and geometry in each until the results converged satisfactorily, limiting deflections in Corbin to an acceptable level and at the same time minimising additional steel tonnage in the Fulton Center.
One other advantage of the concrete lateral frame was its own dead weight, which helped resist overturning forces and hence the force transmitted to the pavilion. However the frame also required support from a suitable foundation. This had to be carefully co-ordinated into the design, as below the stability frame a new void had been introduced for the deep escalator wellway, with one side of the frame actually sitting on the wellway retaining wall.
At level 7 a step back in the frame was needed, as the architecture called for reinstating the historic floorplan which had a corridor running parallel to the north wall. As this is almost the top of the frame it was easily accommodated (Figs 36–37).
As part of the lateral system upgrade it was also necessary to strengthen the floor diaphragms at each level, which typically comprised wood flooring on timber battens on cinder fill over the terracotta Guastavino arches. As previously noted, upgrading the floor diaphragms had to be achieved without any substantial increase in floor weight if allowable live loadings were to be maintained.
A system in which most of the fill was replaced with lighter cellular concrete (a low-strength stiff material filled with micro-bubbles) allowed key elements at the wrought iron beam surrounds, together with the final wearing surface, to be replaced with heavier lightweight concrete, which also had the strength needed to act as a diaphragm (Figs 38–39). Cellular concrete is relatively uncommon in the US, but had been used successfully in the past on UK heritage projects by members of the Arup team.
The structural modifications for the new deep escalators (Fig 40) were probably the most challenging aspect of the Corbin renovation. The escalators have an overall rise of about 40ft (12.2m), and terminate in a pit nearly 20ft (6.1m) below the existing foundation level. The excavations were almost entirely within Manhattan’s notorious “Bull’s Liver” soil, a vibration-sensitive stratified silt and fine sand that is prone to consolidation, causing settlement under construction vibrations, and rapidly loses strength when disturbed or wetted.
Firstly, the whole of the west end below this level had to be underpinned. To counteract potential issues with the liquefiable soils, the entire perimeter of the underpinning zone had to be stabilised by a system of contiguous jet grouting, which itself caused some minor soil settlement that was reflected in the superstructure. Corbin was instrumented and regularly monitored by a series of real-time strain gauges in combination with a conventional system of readings from static targets strategically positioned on the structure4.
Movements were reviewed daily throughout the underpinning to ensure that the building did not develop any unacceptable tilt or masonry overstress. The team was able to observe daily expansion, contraction, and “tilting” of the building caused by cyclical weather patterns, which were far greater than would have been imagined (up to 0.2in (5mm) vertically and 0.5in (13mm) horizontally). As an additional safeguard visual structural surveys continued in parallel with the monitoring.
Once the grouting was in place, a series of traditional underpinning excavations in maximum 3ft (900mm) wide sections were hand-dug (Fig 41), and the full perimeter of external wall and internal spread footings for the west half of Corbin were underpinned with mass concrete to a level below that of the proposed escalator footings. These works alone took almost a year, and as before, monitoring supervised by Arup helped ensure that they were carried out without approaching an unsafe condition in the field.
The next step required excavation of the soils between the underpinned footings within Corbin’s footprint, and construction of a profiled concrete wellway slab up through the building at a steep angle of around 30˚. This required removal of three levels of internal floor diaphragm: at the street, basement, and sub-basement levels. All these floors carried substantial lateral loading from soil pressure on the south (John Street) masonry wall.
Historically these forces had been balanced by equal and opposite forces from the basements of buildings to the north, but these were removed during construction of the Fulton Center foundations. This created a much deeper three-storey “bathtub”, with contiguous piled retaining walls that only aligned floor levels with Corbin at street and sub-basement levels. It was thus necessary to design a system within Corbin to transfer the lateral loads from the south retaining wall into the new Fulton Center floor diaphragms and consolidate three levels of loading into two levels of support (Fig 42).
The street level support was relatively straightforward, as here the two buildings matched and it was only necessary to create a new steel and concrete floor to span the width of the new opening in the floor and to act as a horizontal beam (Figs 43-44).
For the basement and sub-basement levels a massive ring beam was needed within the escalator wellway to match as closely as possible the Fulton Center diaphragm level. This ring steel was located slightly below the existing foundation level in Corbin. To get the loads from the south basement wall into this ring beam a new concrete retaining wall was formed within the Corbin sub-basement as a collector element (Fig 43). The design for the wall was a delicate balance, as resistance to overturning and soil bearing below the wall base needed to be controlled, but the wall geometry was tightly constrained by the existing heavy masonry superstructure and available space at the building’s very narrow end.
Inevitably, unforeseen conditions arose during construction. It transpired that the existing masonry sub-basement wall had a series of projecting piers, presumably incorporated as stiffeners, which greatly reduced available width for the new concrete retaining wall behind. The geometrical changes were sufficient to make the original design unworkable, as the lever arm for the new wall would now be too small. The solution was to underpin the existing basement retaining wall with concrete needle beams projecting from the base of the new retaining structure so as to mobilise the existing wall’s weight and effectively counteract the negative effects of the shorter base.
A benefit of the escalator wellway beyond its basic function was that the new shaft would allow the public to see a vertical section through the building. The masonry walls and columns are supported off a series of
inverted masonry arches designed to spread the superstructure loads back into the soils more evenly (Fig 44); these are relatively uncommon, and wonderfully aesthetic at the same time. Instead of hiding the structure, the architectural design incorporated the inverted arch foundations as the central theme of the space and mirrored the existing arches in the new masonry liner wall that needed to be formed to the south side.
The wellway is also a great place for the public display of salvaged terracotta from the roof and the old cast iron boiler doors, and these elements were incorporated into the new liner wall to the south side (Fig 45).
Guastavino floor strengthening
As discussed above, the typical floor upgrade used a combination of cellular concrete fill and lightweight concrete slab to control overall floor weight, but part way through construction the client decided to change the proposed use of levels 2 and 3 from offices to retail, with consequences for floor loading. This decision was driven partly by the location at corresponding levels of retail space in the Fulton Center pavilion; this would enable connectivity through both buildings for a larger retailer, thus adding value to the project.
Individual strengthening of the wrought iron beams with continuous steel plates at mid-span or similar would have been costly, visually intrusive, and inefficient, as the weights of the remedial plates would have impacted overall floor loads. A solution of minimal weight but increased strength was needed, and the weldability test results showed that it was also highly desirable to avoid welding to the wrought iron, as the necessary preheat would have proved costly.
The team developed a solution with a proprietary Hilti product, originally aimed at the new-build/metal deck market. By using Hilti shear connectors screw-fixed to the existing beam flanges by self-drilling screws, the team proved a 30% increase in overall beam capacity without changing any other floor diaphragm details; the design was verified using Arup’s in-house Compos software. This was a flexible system that could easily be installed by the contractor without any site-welding. Hilti also made field tests to verify that the anchor capacities reached published values.
Cast iron corbels
Upgrading the existing floor capacity from 75lb/ft2 (365kg/m2) to 125lb/ft2 (610kg/m2) at levels 2 and 3 required the increased forces to be successfully transferred back to the vertical load-bearing structure. The existing columns were square hollow cast iron, with uniform wall thicknesses. Arup’s field investigation had established that the floor beams were typically supported off cast iron corbels (or brackets), cast integrally with these columns; they took two distinct forms, either single T-section corbels in the column face for secondary beams, or double TT-section corbels aligning with the column perimeter walls for the primary beams.
While researching contemporary design methodologies for cast iron and wrought iron sections, Arup had referenced the typical 1890s “Engineer’s Pocketbooks” — a prime source of information on safe design, in the absence of any nationally-published design codes or guidance. One such contemporary guide referred to testing of similar cast iron corbels by the NYC Department of Buildings that had yielded surprisingly low results, potentially invalidating the perceived wisdom of the time (Fig 46a).
Given the critical nature of the connection, this clearly required further investigation and prompted Arup’s structural team to work with its advanced technology group (ATG) to try to replicate the unexpected failure modes and gain further insight into the problem (Fig 46b).
The results, to be published in a forthcoming paper5, demonstrate how the engineers of 1890 had limited understanding of the behaviour of shear versus flexure, not to mention more complex biaxial and triaxial states of stress, and how this may have led to the under-design of similar connections in thousands of buildings throughout the US.
The direct result for Corbin was to adopt a reinforced concrete shear head detail cast within the slabs, transferring a proportion of the loading between the relatively weak single T and the much stronger double TT brackets. This also avoided any awkward upgrade of the bracket detail, which would be made doubly difficult by the lack of weldability of the section, and particularly undesirable as the brackets were to be left exposed for aesthetic reasons.
29. Plan view showing narrowness and plan irregularity, which dictated the need for lateral bracing.
30. The north wall at the beginning of the project.
31–32. 3-D ETABS structural model used to establish levels of stress in masonry.
33. Drilling of north wall to accept epoxy-anchored reinforcement prior to shotcreting lower portion of wall.
34. New in-plane reinforcement in the north wall.
35. North wall with completed strengthening works up to level 2 and new egress connections (two per floor).
36. New concrete lateral frame.
37. Lateral frame solution to reinforce weak end of building.
38. Forming the new concrete frame at level 2: secondary beams were left in place while large girder was encased; original floor construction can be seen in the background.
39. Strengthening floor diaphragms:
a) original Guastavino floor;
b) typical structural upgrade;
c) composite beam upgrade;
d) Hilti X-HVB shear connector.
40. The top of the new escalators.
41. Underpinning in progress.
42. Creating the deep escalator.
43. Construction of new escalator foundations with temporary ring beam steelwork in place.
44. Creating the escalator wellway through and below the original foundations.
45. Down the deep escalator, with salvaged terracotta on the left and the original inverted arch foundation exposed to view.
46. Modern analysis compared with historical evidence:
(a) Images of column testing carried out in 1890 by NYC Department of Buildings engineers;
(b) Arup LS-Dyna analysis of corbel failure.
47. Column head before restoration.
48. Restored cast iron central columns and corbels in typical interior space.
Originally published in the Arup Journal, November 11, 2014. Authors include: Ian Buckley, Craig Covil & Ricardo Pittella.
Coming up next: The Corbin Building, Fulton Center: Rediscovering and Renewing An Architectural Gem – Part 6