San Francisco’s Central Subway: Part II – Underground Stations
UMS, like most subway stations, is excavated using what is known as the cut-and-cover method. The work process involves digging a trench, constructing a station in the open trench and then covering over the station structure in the trench. Sounds simple, however the “rub” comes when the trench is open, and activities on the ground surface are disrupted. Indeed, a review of the historical record cites few examples of large scale cut-and-cover projects, typical of underground transport structures, from the first recorded large scale work in 2180 BC until 1844 AD.
The first recorded large scale cut-and-cover transportation tunnel was the legendary crossing under the Euphrates in Babylon, commissioned by Queen Semiramis. It remains “Legendary” because, so far, it exists only in the writings of Diodorus the Sicilian in the period between 60 and 30 BC. He describes, in a fair amount of detail, a royal carriageway approximately 660 feet long that was 15 feet wide and 12 feet high to the arch spring-line. The tunnel structure was made of sun baked mud-brick coated with bitumen. My idea of the cut-and-cover work described by Diodoros is shown in the drawings below:
To build this royal tunnel (plus a bridge for the rest of Babylon’s population) the Euphrates River was diverted. The cut would have been a simple one, likely without temporary ground support, since Mesopotamia had limited timber and stone resources available. Builders would have relied on a familiar excavation method that removed soil to its angle of repose. This would have resulted in an amount of disruption to surface activity that was not easy to manage. Given this, it’s understandable why millennia passed until similar proposals were made. In the interim, cut-and-cover was routinely used for smaller scale excavations (aqueducts, drainage, storage space, etc.)
The next “big” tunnel proposal was at the beginning of the 19th Century, a large scale, Semiramis inspired, cut-and-cover tunnel proposed for the Thames Tunnel by Samuel Ware Esq. This proposal was in competition with a bored tunnel option, using Brunel’s shield. Evidently political clout in London was not sufficient to divert the Thames so the tunneling shield, as we discussed in Part I, got its start. As you also may recall from that post, Brunel’s project was not commercially successful and did nothing to sway the prevailing opinion that it was best to keep rail lines at or very near the ground surface.
Conventional thinking did not go unchallenged. Steam powered rail technology advanced rapidly, increasing the pressure on major cities for trains to share the streets with others. However, reliable brake technology was late in coming, so there was a real need to place the train in its own space, at a different level than the street grade. For this reason, in 1844, the Atlantic Avenue Tunnel was built in Brooklyn New York using the cut-and-cover method, arguably becoming the first rail subway line in the world. Twenty years later Londoners were riding in the first subway system, which was grade separated but made close to the ground surface and constructed entirely by cut-and-cover. The surface level disruption associated with the construction of both of these projects was not fondly remembered for many years thereafter. With the success of the Greathead Shield in 1870s (see Part I) London’s engineers, planners and builders embraced the deep “tube” subway. The level of confidence in “the Shield” extended to station construction as well, where a larger specialized shield was used to bore the station space. Although this technique reduced ground-level access to circular shafts, it worked well enough because elevators and circular stairways existed (escalators came later). While less disruptive, the trade-off was that the station shield was complicated to install and use, making it expensive. In the end this method was not readily transferable to other cities without the favorable conditions associated with soil of London Clay. So for many cities a subway with shallower stations was the best option. In turn, these cities worked to minimize disruption of the open-cut by improving the cut-and-cover method.
Although cut-and-cover excavations still looked similar to the Roman cofferdam, the piles were not arranged as palisades, but as lines of “soldiers” and the space in between them was spanned by timber or alternatively the stake-shaped pile was replaced with interlocking corrugated iron or steel sheets. Since these walls were flexible, much bracing was needed, which hindered the builder’s movement as the cuts became larger and deeper. Moreover, excavation became more complex in difficult ground conditions where water was present. For many in the field an “ideal” solution was a watertight wall, stiff enough to require less bracing, more room to move and robust enough to be made permanent. Various ways to achieve this were developed. One, which we will look into further in this post, is a palisade wall made with large piles. Another, the Slurry Wall we will learn about in Part III.