The Behavior of Leaning Towers

How to rectify a 58 story sinking condominium sitting on nothing but on a ten foot thick concrete micro-piled mono-slab, is as dubious a notion as correcting the Leaning Tower of Pisa. Of course, I am talking about San Francisco’s notorious Millenium Tower -, the tower has sank some eighteen inches, and tilted fourteen northwest - now in the midst of a maze of litigation, ‘with no fewer than 146 lawyers involved.

I don’t mention the 950 micropiles (above) because - well, they’re not anchored in bedrock, they are literally stuck in the in mud . This is known as a ‘friction-pile,’ as opposed to an end-bearing pile. Such piles were - in Millennium Tower’s case - expected to sink up to 6”. But Millennium Tower sank 18” nonetheless, and settled unevenly - causing the tower to lean. The building is also supported by outrigger piles to the 22nd floor..

Of equal concern are reports that the soil would be subject to liquefaction in the event of an earthquake, where fissures or veins in the substrate separate and allow in water from San Francisco Bay. One only need consult historical documents to learn that the sand substrate was underwater until it was impounded in the late 19th century. Or simply consult the Temblor data for the property, which has the highest risk of liquefaction susceptibility. Dewatering during excavation was an ongoing concern.

The property, located on the corner of Mission and Fremont Streets, is situated atop the Yerba Buena Cove shoreline, as well as adjacent to an bold stream bed. An 1853 map (below) shows that the intersection of Mission and Fremont Streets, was barely above seal level and not as infilled, or impounded.

By 1857, the area was impounded, below.

ARUP conducted several geological studies of the substrate over a period of time to determine vulnerabilities were not imminent risks. The studies were forwarded to city officials, who contend there is no danger of structural failure.

Soil stratigraphy reported by geotechnical Firm Arup in 2012 for TJPA

Before blaming the design engineers, it helps to know that subsequent to completion, adjacent projects took place, and that some will claim that substrate decompression ensued as a consequence - they dug a tunnel next the foundation, removed a few piles - in the soil buffer zone between the two properties.

In any case, there are a few post-occupancy fixes up for consideration, the most plausible so far would involve driving piles diagonal into adjacent property bedrock, which is far more promising than the initial plan - needling 300 micropiles, 300 feet-deep straight down into bedrock (see below).

The micropile to adjacent bedrock approach replaced other more costly ones, such as stabilizing the lower side of the building and letting the other side sink to the same level,  a highly speculative idea that would have cost upwards of the original project GMP - USD$350M.

More recently, San Francisco Chronicle reports “In an application to be filed with the city’s Department of Building Inspection, the Millennium Tower Association laid out plans for a “perimeter pile upgrade,” 52 steel and concrete piles along Mission and Fremont Streets, that would transfer a portion of the building’s load from its existing foundation system to bedrock about 250 feet below.

Some of the lawsuits against the developer allege that the developer knew the tower was sinking during construction, as not previously disclosed. A 2017 ARUP report stated that ”The 58-story high rise portion of the Millennium Tower complex has settled about 17 inches vertically since construction commenced in 2006, with about 11 of the 17 inches occurring since completion of the tower superstructure in 2008,’ also noted the tower was sinking about 1”/year.

Natural settlement is expected for any tower. Structural engineers typically allow for even settlements, but never for uneven displacement, and not to the extent the tower has displaced - which is in excess of 3X over acceptable tolerances. If the tower displaced as it was going up, the fact was either known and suppressed, or unnoticed as a consequence of not being monitored.

The Ramifications of Structure Displacement

If a builder or developer is fortunate, he can detect displacement and settlement as it happens with a vigorous monitoring program. Such a monitoring program would highlight horizon as well as vertical displacements, allowing time to implement mitigation strategies. The nature of the mitigation strategies will be project specific. Even in the absence of a surveyor’s reports, it’s simple enough to use tell-tales - marks and shims - and better than no program at all. To be sure Millennium Tower shows structural stress cracks in its foundations.

Leaning Tower of Pisa taken circa 1890-1900, Library of Congress

Yet, some builders don’t always survey the edge of slab as the building goes up - especially if there is space between it and the curtain wall to take up any shifting. One reason they don’t survey for plumbness is for lack of reference points - such as a plumb line taken from a nearby structures. There isn’t always  availability and access to these points, e.g., the top and bottom edge of slab of an adjacent structure.

Many of downtown New York City’s structures are built on similar substrates, and in the same way - mono-slab with or without piles. Determining just who has the final accountability say-so in final design can be a slippery slope. Typically, an engineer will overdesign, rather than underdesign. If the local Building Departments accept mono-slab sans piles, developers will be the first to value-engineer them away.

In most cases, displacement mitigation is a crap-shoot that effectively serves as a band-aid approach for conditions that cannot be reconciled, or only in part. The unknowns of future displacement make the timing and expected efficacy of mitigation theoretical and risky. The Leaning Tower of Pisa was stabilized in, 1999, for at least another 200 years. This was accomplished by placing ballast on the north side of the tower, and removing the soil below to allow that side to shift load - not unlike one of the Millenium Tower proposed fixes discussed above.

Millennium Tower’s structural deficiencies also affect fitout: displaced elevator rails could affect the elevator travel (alignment in the shaft). Displaced elevator shaft openings could inhibit the ability to realign crooked elevator rails. Pressure on plumbing and mechanical piping could increase causing rupturing and bursting decades down the road.

Had the Millenium Tower vertical displacement been noted sooner, the developer might have had a better opportunity to rectify and control future displacement. But surveyors often do not always shoot the vertical axes of edge of slab (EOS) as they do deck levels. In the case of  Millenium Tower, the curtain wall brackets could take up any minor vertical displacement.

As the distance away from furthest-point EOS increase, the embeds and brackets may no longer be adequate to support increased cantilever needed to reach the curtain wall panels. Curtain wall corners would not meet as a consequence of the distended building footprint, and require extension of the last panel. A builder then would have to expose the embeds and retrofit a bolted or welded brack, which would have to be designed by a structural engineer.

In such case, a developer may attempt to saddle the builder with the cost and responsibility of engineering redesigning the curtain wall bracket. In order for that to happen, he would have to demonstrate that displacement was caused by the builder. In the time it takes to assign responsibility, design, order, fabricate, and install new brackets, the builder will have increased his schedule duration exponentially.

At any time, legal problems could arise for encroaching on a neighboring property line or setback. They may also arise because the gross square footage of the floor space is reduced: if a building is 18” out of alignment, it shifts that 18” from usable floor space on the short side, and transfers it to the distance between EOS and curtain wall - unusable floor space.

As the footprint of the exterior changes, the relation of the interior lay-out to surveyor axes lines no longer corresponds properly, which will necessitate in builder RFIs and designer sketches showing revised layouts. The designer will expect the trades sketchers to redesign all of their installations, as well, so as to eschew having their own engineer. Stairwells no longer align, balconies shift into the interior, and riser core offsets are not plumb - requiring bothersome concrete chipping to enlarge penetrations.

In addition, many designers like to align partitions with the curtain wall shadow-boxes designed to obscure shear walls - or to cover an interior demising wall perpendicular to the curtain wall. If these shadow boxes shift laterally, partitions will no longer align. Neither will any datum- alignment such as a counter, stone joint, head or sill maintain its relationship with the interior program. The result is a lot of discombobulation where few of the elements align.

Finally, building true, plumb, safe structures has as much to do with timing as it does bad luck. If the problem is noticed before it has become insurmountable, there is still a fighting chance. In order for that to happen, engineers need to be more thorough in their soil investigations - designing to exceed the worst risk scenarios, and builders must be more diligent in monitoring structures to ensure they go up straight, and not crooked.