Chapter 11. The River Tunnels
The New York Rapid Transit Railway Extensions · Engineering News, 1914
The Harlem River Tunnels. The general design of these tunnels is shown in the cross-section, Fig. 76. It was determined by two principal factors: First, the necessity on account of the conditions under which the approaches were located of keeping the tunnels as near the surface as possible; second, the desirability of obtaining a minimum total width to avoid encroachments on valuable private property.
The methods developed in the construction of the tunnels under the Harlem River for the original subway (see Eng. News, Oct. 13, 1904) and at Detroit for the tunnels of the Michigan Central Ry., had shown the practicability of sinking tubes from the surface, and these methods also permitted much closer spacing than would have been possible with shield-driven tunnels, which latter would necessarily or at least most conveniently have had to he circular, with a reasonable space, say 10 ft. or so between each tube. It may be noted, however, that the tunnels which had been previously built by sinking from the surface were for two tracks only, whereas, the new Harlem River tunnel is for four.
Bids were originally called for late in 1910 on two types, H and K. Type H was similar to that of the original Harlem River tubes, and type K similar to the Detroit River tubes. The bid prices were as follows, per lin.ft. of four-track tube: Type K-Lowest $1925, highest $3000, per lin.ft. of four-track tube; Type H-Lowest $2200, highest $2000, per lin.ft. of four-track tube.
Before the contracts were awarded it was decided to change the dimensions; the bids were therefore rejected and the work readvertised, this time calling for bids on three types, H, K and L. Types H and K remained the same except for the changes in dimensions. Type L was a modification of Type H, having the four tubes all together instead of in two pairs.
The bids on the latter two types were all rejected and that of Messrs. Arthur McMullen and Olaf Hoff, the lowest bidders on type K, was accepted; the range of bids at this last letting for the tube-tunnel section having been as follows, per lin.ft. of four-track tube: Type K, $1500 to $1800, 8 bids; Type H, $1650 to $2000, 2 bids; Type L, $1550 to $2000, 6 bids.
The contract price of $1500, equal to $375 per lin.ft. of track, may be compared with the cost of the Detroit River Tunnels, which has been given as $332.29, exclusive of contractors' profits. However, the inside diameter of the Detroit tunnels was 20 ft., as compared with 16 ft. 6 in. for the Harlem River tubes.
The method adopted by the contractors with very few modifications was that developed for the construction of the Detroit River Tunnels, which has been quite fully described in various papers and articles in the technical press. The article in Engineering News, of Feb. 15, 1906, is interesting as showing the development of the process. It consists essentially in the erection of the steel tubes in suitable lengths on shore, bulkheading the ends to get flotation, launching these sections, towing them to the site which has previously been dredged to the required depth, and sinking them in place by filling them with water (see photographs in Fig. 75). The concrete is then deposited around the outside by means of tremies, the sections unwatered and the inner concrete lining placed. This method, of course, obviates the necessity of general work in compressed air, though divers are used to a limited extent.
Fig. 75. Views of the Harlem River tubes during construction. (a) Construction, partial. (b) Construction, completed. (c) Towing to site. (d) and (e) Sinking. (f) Inside.
The accompanying drawings and photographs show the essential details of the structure and the methods of sinking and as the general methods have already been so fully described, it seems only necessary to call attention to such changes and improvements as experience and the particular conditions of the Harlem River work have shown to be desirable. The tunnel was divided into five sections, four of 220 ft. each and one of 200 ft.
Figs. 76-81. Various cross-sections of tunnels. (Click image to enlarge.)
In the Detroit River tubes, the circular stiffening angles, which are spaced about 8 ft. apart, were placed on the inside, as then it was thought necessary to provide temporary interior bracing in the form of the spokes of a wheel. Experience showed, however, that this might be dispensed with, and on the Harlem River tubes the stiffening angles were placed on the outside. This permitted the construction of the braces or struts to the wooden bulkheads or forms at the sides, which materially decreased the necessary thickness of the timber, which latter, in the case of the Detroit tunnels, was 6 in. thick at the bottom and 4 in. thick at the top. For the Harlem River Tunnels, 4-in. plank was used for the lower half and 3-in. for the upper.
The manner of making a tight joint between each of the sections shows an important modification in the direction of simplicity. The old joint with the pilot pin is shown in the drawing, Fig. 77a. This joint was not altogether satisfactory, as it was somewhat difficult to fit and the rubber gaskets are, of course, perishable. The new joint (see Fig. 77b) is a butt joint instead of an overlapping or sleeve joint, and the bolts on the outside are easily placed by divers. The inner plate, which, of course, is riveted in place after the tubes are unwatered, assures practical water-tightness. It will be remembered that the concrete is placed outside this joint before the tubes are unwatered, and tests made with the concrete deposited by the tremies at Detroit showed it to be of very good quality, sufficiently impervious to prevent any leakage of moment. The space between the joint in the shell and the inner plate is to be filled with grout after the latter is riveted in place. There is a pilot pin on each of the two outer tubes, and when both are home, the accuracy of the construction insures a good fit everywhere. The unwatering of the tunnels has shown these joints to be remarkably tight.
Attention may be called to the method of tying the tubes and the partition walls together, as shown in detail. Fig. 76. Reinforcement of 1-in. square rods is placed in the inner concrete lining. Longitudinal rods are spaced 12 in. apart at the sides. It is probable that this might be omitted and still leave the tunnels entirely safe, but is an added precaution thought advisable in view of the comparative novelty of the method.
Before launching a section, the two outer tubes were tightly bulkheaded at both ends, but the inner tubes only about 4 ft. up, as shown in the sketch, Fig. 78, that is, just high enough to provide flotation while being towed to the site. The outer ends of the end sections are, of course, tightly bulkheaded on all four tubes with bulkheads to stand total pressure for the depth, so they will hold when the tubes are unwatered. Photograph 75 (D) shows the southerly end of the first section (A) with all the bulkheads in. At Detroit the tubes were actually launched, by allowing them to slide down the ways as a ship is launched; on the Harlem River, however, it was thought best to build the structure on an open platform over the water, so that flat-decked lighters could be floated underneath to lift them off. The lighters used were water boats which could be filled by the opening of valves provided for the purpose and emptied by pumps. They were floated under at low water, raised by the tide to lift the tubes off the platform, and then when the tubes were moved over deep water were scuttled, leaving the tubes floating. On account of the narrowness of the Harlem River, there would have been some difficulties attending launching, but, in any event, the method used was thought to be better, and proved to be practical and very satisfactory.
The method of depositing the concrete around the outside of the tubes, and tests showing the good quality of concrete so deposited, were fully described in Engineering News, Mar.17, 1910; the method and plant used on the Harlem River was almost exactly the same with one important improvement in the control of the tremies. The tremies are so arranged that they can be raised or lowered to accelerate or retard the flow of the concrete. At the head of each tremie and attached to it, is a platform on which stands the man who controls it. Individual hoists were provided for each tremie, controlled by a continuous rope passing by the platform, so that at any position of this latter the rope could be reached by the man, and the raising or lowering of each separate tremie made almost instantly and as required.
At the Detroit River a steel grillage embedded in concrete was placed in the bottom of the trench at the joints between each section, but at the Harlem River timber bents were driven. There were 4 to 6 bents at each joint; they were framed on shore and driven by two piledrivers, moored facing each other with long followers to reach to the necessary depth. On the first section the bents were driven an inch or two low so that the tubes might be blocked up; it was found, however, that such good control was possible that they were afterwards driven almost exactly to grade.
The method of sinking the tubes is very simple; 12-in. valves are opened in the bottom of the bulkheads in the two outside tubes, allowing these latter to fill gradually with water; the two inner tubes are entirely open. The rate of sinking after the tubes are half full is controlled by air valves at the top of the main tube, if necessary. There is apparently no difficulty in keeping them level, but to aid in this two cross bulkheads are provided, reaching half-way down from the top, providing three sections from which, after the tubes are half full, the air escapes and, consequently, the amount of water entering can be controlled by opening or closing the air valves. The tendency of either end or corner to get out of level was, therefore, easily controlled. As the tubes become completely filled, the flotation is carried by the four cylinders on top, which are in turn gradually partially filled and the excess weight, which is not great, is taken by derrick boats moored on either side during the sinking. The method of control of position is shown in the diagram, Fig. 79.
Some interesting statistics are as follows:
|Weight of steel per lin.ft. of structure||5600 lb.|
|Amount of exterior concrete per lin.ft. of structure||30.0
|Amount of interior concrete per lin.ft. of structure||11.6
|Maximum depth M.H.W. to subgrade||57.2 ft.|
|Buoyancy of four cylinders (on top)||722 tons|
|Excess buoyancy four cylinders||76 tons|
The weight of the structure equipped for sinking, with masts, bulkheads, sheeting, buoyancy cylinders, etc., complete is 646 tons, requiring 19 tons of water in each to overcome buoyancy. One hour is required to fill the structure with water.
Cross passages are provided between the tubes at approximately every 50 ft., in the outer partitions, and two openings in the whole length of the tubes through the center partition. There is a sump in each tube at the lowest point, universal-joint cast-iron pipe being used for discharge. Access shafts, one for each tube, are provided near the ends of the end sections, by which access can be obtained to the interior of the tubes after the outside concreting is completed. The ends of the last sections are fitted with slots (two angles) to take the sheeting of the coffer-dams which are built to connect them with the land sections built in open cut. The connecting coffer-dam is of a single row of steel sheet piling; clay being dumped on the outside, if necessary, to make it tight.
Roof Shields. Just north of the Harlem River the westerly branch of the subway passes under the main line of the N. Y. C. & H. R. R. R., which at this point carries all the traffic from its own lines as well as from those of the N. Y., N. H. & H. R. R., to and from the Grand Central Terminal. The railway has five tracks and is carried on a fill between high masonry retaining-walls. The base of rail of the subway line is to be between 40 and 50 ft. below that of the railway above.
It was at first thought that this work might be carried out in open cut, carrying the railroad on timber falsework, but the acute angle of the crossing, depth and character of material would have made this a somewhat hazardous undertaking, and it was finally decided to adopt the method shown in the accompanying drawings.
Timbered drifts have been driven, as shown in Fig. 80, the center one, as will be noted, being considerably higher than the other two on the outsides. The material encountered has been mostly rock, but the work was rendered quite difficult in parts by reason of the fact that the top of the rock was just below the top of the drifts, requiring the support of the earth overhead and blasting of the rock below.
In these drifts the side and center walls are to be built of concrete and then the balance of the excavation is to be taken out under the protection of the double segmental roof shields, detail of which are shown in the drawings, Fig. 81. These shields, as will be seen, are quite unique in design and form. It is intended to work each independently of the other, shoving one at a time, but, of course, not to be the extent of one entirely clearing the other, as they necessarily react on each other to take up the side thrust.
The writer is especially indebted to Mr. Olaf Hoff, of the firm of McMullen & Hoff, the contractors for this work, for the above information, for the plans and details of these shields, and of the Harlem River tubes, he being principally responsible for the design and execution of this portion of the work.
The East River Tunnels. The additional connections between the new lines in Brooklyn and those in Manhattan are to be by means of two pairs of tunnels under the lower end of the East River, as described in Engineering news, Apr.30, 1914. The contracts for all four tunnels were recently awarded to the Flinn-O'Rourke Co. for a total amount of about $12,500,000, and work was actually started about Nov. 1 on the sinking of the shafts.
The tunnels are to be driven by the shield method and thcre are two novel features which are to be tried to which attention mav be called at this time.
In several places the roofs of the tunnels are quite close to the river bed, so that additional cover must be provided. Owing to the swiftness of the tidal current at this point and its scouring action, the problem of retaining a clay blanket in place according to the method heretofore used in East River tunneling seemed to be one of some difficulty. The method proposed, however, will not only probably retain the clay, but by placing the material at this time (November, 1914) it will settle well into position and become nearly impervious, by the time the tunnels are driven. A comparatively narrow, thin blanket of clay is first deposited on a line on each side of the location of the tunnels. Rock from any of the numerous excavations always going on in and around New York is dumped on top of this and a clay blanket varying in thickness from 5 to 15 ft. is then dumped between these piles of rock, and is finally covered with other rock, as shown in the sketch Fig. 82. This blanket will be approximately 125 ft. in width over all. It is believed this clay blanket will stay in position and effect the desired purpose. The contractors have been fortunate in being able to obtain an excellent grade of clay from dredging in progress on the Hudson River, near Edgewater, N. J., which ordinarily would have to be towed to sea for disposal.
Fig. 82. Method of covering river bottom over East River tunnels. (Click image to enlarge.)
The second feature is a method of filling the annular space around the outside of the tunnel left behind the tail of the shield, when the latter is shoved ahead. The outside diameter of shields used for tunneling is usually from 6 to 8 in. greater than the outside diameter of the tunnel, this leaving a space of 3 or 4 in. all around to be filled by the movement of the surrounding material or in some other manner as the shield is forced forward. This movement, while not always of importance, does tend to produce distortion to the cast-iron lining and to settlement of the ground above the tunnel, which latter may be undesirable by reason of possible damage to overhead structures, as in the street approaches to these East River tunnels, and as happened in Joralemon St., Brooklyn, during the construction of the Battery tunnel for the first subway line to Brooklyn. In such material as that under the East River, which is mostly sand and gravel and especially with the tunnels so near the surface, any disturbance of the ground above the tunnels, even in the river with no buildings above, is undesirable. Its prevention will probably also tend to lessen the waste from escaping air.
The shield is to be built with a double skin of 0.5-in. plates separated by a space of 1.25 in. The clearance between the shield and the tunnel is 0.75 in. The two skins are separated by 1.25x3.5-in. separators. Eight rectangular pipes 7/8x1 1/8 in. inside, 3/8 in. thick, project through the back of the shield, and gravel similar to that used for roofing purposes, is blown through these pipes by air pressure to fill the space as the shield is shoved ahead.
Experiments on a small scale have already been made which show fairly conclusively the feasibility and practicability of this method for preventing any movement of the surrounding material into the space left by the shield, but a full-size shield is now nearing completion with which final tests are to be made with the complete apparatus.
If the two improvements above described succeed in any marked degree in overconling the difficulties usually experienced in tunneling through water-bearing loose sand and gravel, with light cover, by reason of blowouts, the generally quite considerable loss of air, consequent heating of the tunnel, and the settlement of the ground above, they must be considered as a distinct advance in the art of subaqueous tunnelihg.
The writer is indebted to John F. O'Rourke and W. Gray, who have developed these methods, for the above information.