THE SHIELD SYSTEM.

Historical Introduction.

—The invention of the shield system of tunneling through soft ground is generally accredited to Sir Isambard Brunel, a Frenchman born in 1769, who emigrated to the United States in 1793, where he remained six years, and then went to England, in which country his epoch-making invention in tunneling was developed and successfully employed in building the first Thames tunnel, and where he died in 1849, a few years after the completion of this great work. Sir Isambard is said to have obtained the idea of employing a shield to tunnel soft ground from observing the work of ship-worms. He noticed that this little animal had a head provided with a boring apparatus with which it dug its way into the wood, and that its body threw off a secretion which lined the hole behind it and rendered it impervious to water. To duplicate this operation by mechanical means on a large enough scale to make it applicable to the construction of tunnels was the plan which occurred to the engineer; and how closely he followed his animate model may be seen by examining the drawings of his first shield, for which he secured a patent in 1818. Briefly described, this device consisted of an iron cylinder having at its front end an auger-like cutter, whose revolution was intended to shove away the material ahead and thus advance the cylinder. As the cylinder advanced the perimeter of the hole behind was to be lined with a spiral sheet-iron plating, which was to be strengthened with an interior lining of masonry. It will be seen that the mechanical resemblance of this device to the ship-worm, on which it is alleged to have been modeled, was remarkably close.

In the same patent in which Sir Isambard secured protection for his mechanical ship-worm he claimed equal rights of invention for another shield, which is of far greater importance in being the prototype of the shield actually employed by him in constructing the first Thames tunnel. This alternative invention, if it may be so termed, consisted of a group of separate cells which could be advanced one or more at a time or all together. The sides of these cells were to be provided with friction rollers to enable them to slide easily upon each other; and it was also specified that the preferable motive power for advancing the cells was hydraulic jacks. To summarize briefly, therefore, the two inventions of Brunel comprehended the protecting cylinder or shield, the closure of the face of the excavation, the cellular division, the hydraulic-jack propelling power, and cylindrical iron lining, which are the essential characteristics of the modern shield system of tunneling. The next step required was the actual proof of the practicability of Brunel’s inventions, and this soon came.

Those who have read the history of the first Thames tunnel will recall the early unsuccessful attempts at construction which had discouraged English engineers. Five years after Brunel’s patent was secured a company was formed to undertake the task again, the plan being to use the shield system, under the personal direction of its inventor as chief engineer. For this work Brunel selected the cellular shield mentioned as an alternative construction in his original patent. He also chose to make this shield rectangular in form. This choice is commonly accounted for by the fact that the strata to be penetrated by the tunnel were practically horizontal, and that it was assumed by the engineer that a rectangular shield would for some reason best resist the pressures which would be developed. Whatever the reason may have been for the choice, the fact remains that a rectangular shield was adopted. The tunnel as designed consisted of two parallel horseshoe tunnels, 13 ft. 9 ins. wide and 16 ft. 4 ins. high and 1200 ft. long, separated from each other by a wall 4 ft. thick, pierced by 64 arched openings of 4 ft. span, the whole being surrounded with massive brickwork built to a rectangular section measuring over all 38 ft. wide and 22 ft. high.

The first shield designed by Brunel for the work proved inadequate to resist the pressures, and it was replaced by another somewhat larger shield of substantially the same design, but of improved construction. This last shield was 22 ft. 3 ins. high and 37 ft. 6 ins. wide. It was divided vertically into twelve separate cast-iron frames placed close side by side, and each frame was divided horizontally into three cells capable of separate movement, but connected by a peculiar articulated construction, which is indicated in a general way by [Fig. 124]. To close or cover the face of the excavation, poling-boards held in place by numerous small screw-jacks were employed. Each cell or each frame could be advanced independently of the others, the power for this operation being obtained by means of screw-jacks abutting against the completed masonry lining. Briefly described, the mode of procedure was to remove the poling-boards in front of the top cell of one frame, and excavate the material ahead for about 6 ins. This being done, the top cell was advanced 6 ins. by means of the screw-jacks, and the poling-boards were replaced. The middle cell of the frame was then advanced 6 ins. by repeating the same process, and finally the operation was duplicated for the bottom cell. With the advance of the bottom cell one frame had been pushed ahead 6 ins., and by a succession of such operations the other eleven frames were advanced a distance of 6 ins., one after the other, until the whole shield occupied a position 6 ins. in advance of that at which work was begun. The next step was to fill the 6-in. space behind the shield with a ring of brickwork.

Fig. 124.—Longitudinal Section of Brunel’s Shield, First Thames Tunnel.

The illustration, [Fig. 124], is the section parallel to the vertical plane of the tunnel through the center of one of the frames, and it shows quite clearly the complicated details of the shield construction. Two features which are to be particularly noted are the suspended staging and centering for constructing the roof arch, and the top plate of the shield extending back and overlapping the roof masonry so as to close completely the roof of the excavation and prevent its falling. Notwithstanding its complicated construction and unwieldy weight of 120 tons, this shield worked successfully, and during several months the construction proceeded at the rate of 2 ft. every 24 hours. There were two irruptions of water and mud from the river during the work, but the apertures were effectually stopped by heaving bags of clay into the holes in the river bed, and covering them over with tarpaulin, with a layer of gravel over all. The tunnel was completed in 1843, at a cost of about $5600 per lineal yard, and 20 years from the time work was first commenced, including all delays.

Fig. 125.—First Shield Invented by Barlow.

The next tunnel to be built by the shield system was the tunnel under London Tower constructed by Barlow and Greathead and begun in 1869. In 1863 Mr. Peter W. Barlow secured a patent in England for a system of tunnel construction comprising the use of a circular shield and a cylindrical cast-iron lining. The shield, as shown by [Fig. 125], was simply an iron or steel plate cylinder. The cylinder plates were thinned down in front to form a cutting edge, and they extended far enough back at the rear to enable the advance ring of the cast-iron lining to be set up within the cylinder. In simplicity of form this shield was much superior to Brunel’s; but it seems very doubtful, since it had no diametrical bracing of any sort, whether it would ever have withstood the combined pressure of the screw-jacks and of the surrounding earth in actual operation without serious distortion, and, probably, total collapse. It should also be noted that Barlow’s shield made no provision for protecting the face of the excavation, although the inventor did state that if the soil made it necessary such a protection could be used. The patent provided for the injection of liquid cement behind the cast-iron lining to fill the annular space left by the advancing tail-plates of the shield. Although Barlow made vigorous efforts to get his shield used, it was not until 1868 that an opportunity presented itself. In the meantime the inventor had been studying how to improve his original device, and in 1868 he secured additional patents covering these improvements. Briefly described, they consisted in partly closing the shield with a diaphragm as shown by [Fig. 126]. The uninclosed portion of the shield is here shown at the center, but the patent specified that it might also be located below the center in the bottom part of the shield. The idea of the construction was that in case of an irruption of water the upper portion of the shield could be kept open by air pressure, and work prosecuted in this open space until the shield had been driven ahead sufficiently to close the aperture, when the normal condition of affairs would be resumed. This was obviously an improvement of real merit. The partial diaphragm also served to stiffen the shield somewhat against collapse, but the thin plate cutting-edges and most of the other structural weaknesses were left unaltered. To summarize briefly the improvements due to Barlow’s work, we have: the construction of the shield in a single piece; the use of compressed air and a partial diaphragm for keeping the upper part of the shield open in case of irruptions of water; and the injection of liquid cement to fill the voids behind the lining.

Longitudinal Section.

Cross Section.

Fig. 126.—Second Shield Invented by Barlow.

Turning now to the London Tower tunnel work, it may first be noted that Barlow found some difficulty in finding a contractor who was willing to undertake the job, so little confidence had engineers generally in his shield system. One man, however, Mr. J. H. Greathead, perceived that Barlow’s device presented merit, although its design and construction were defective, and he finally undertook the work and carried it to a brilliant success. The tunnel was 1350 ft. long and 7 ft. in diameter, and penetrated compact clay. Work was begun on the first shore shaft on Feb. 12, 1869, and the tunnel was completed the following Christmas, or in something short of eleven months, at a cost of £14,500.

The shield used was Barlow’s idea put into practical shape by Greathead. It consisted of an iron cylinder, or, more properly, a frustum of a cone whose circumferential sides were very slightly inclined to the axis, the idea being that the friction would be less if the front end of the shield were slightly larger than the rear end. The shell of the cone was made of 1⁄2-in. plates. The thinned plate cutting-edge of Barlow’s shield was replaced by Greathead with a circular ring of cast iron. Greathead also altered the construction of the diaphragm by arranging the angle stiffeners so that they ran horizontally and vertically, and by fastening the diaphragm plates to an interior cast-iron ring connected to the shell plates. This was a decided structural improvement, but it was accompanied with another modification which was quite as decided a retrogression from Barlow’s design. Greathead made the diaphragm opening rectangular and to extend very nearly from the top to the bottom of the shield, thus abandoning the element of safety provided by Barlow in case of an irruption of water. Fortunately the material penetrated by the shield for the Tower tunnel was so compact that no trouble was had from water; but the dangerous character of the construction was some years afterwards disastrously proven in driving the Yarra River tunnel at Melbourne, Australia. To drive his shield Greathead employed six 21⁄2-in. screw-jacks capable of developing a total force of 60 tons. The tails of the jack bore against the completed lining, which consisted of cast-iron rings 18 ins. wide and 7⁄8 in. thick, each ring being made up of a crown piece and three segments. The different segments and rings were provided with double (exterior and interior) flanges, by means of which they were bolted together. The soil behind the lining was filled with liquid cement injected through small holes by means of a hand pump.

Fig. 127.—Shield Suggested by Greathead for the Proposed North and South Woolwich Subway.

Fig. 128.—Beach’s Shield Used on Broadway Pneumatic Railway Tunnel.

The remarkable success of the London Tower tunnel encouraged Barlow to form in 1871 a company to tunnel the Thames between Southwark and the City, and Greathead, in 1876, to project a tunnel under the same waterway known as the North and South Woolwich Subway. Barlow’s concession was abrogated by Parliament in 1873, without any work having been done. Greathead progressed far enough with his enterprise to construct a shield and a large amount of the iron lining when the contractors abandoned the work. From the brief description of his shield given by Greathead to the London Society of Civil Engineers, it contained several important differences from the shield built by him for the London Tower tunnel, as is shown by [Fig. 127]. The changes which deserve particular notice are the great extension of the shield behind the diaphragm, the curved form of the diaphragm, and the use of hydraulic jacks. Greathead had also designed for this work a special crane to be used in erecting the cast-iron segments of the lining.

Fig. 129.—Shield for City and South London Railway.

While these works had been progressing in England, Mr. Beach, an American, received a patent in the United States for a tunnel shield of the construction shown by [Fig. 128], which was first tried practically in constructing a short length of tunnel under Broadway for the nearly forgotten Broadway Pneumatic Underground Railway. This shield, as is indicated by the illustration, consisted of a cylinder of wood with an iron-cutting-edge and an iron tail-ring. Extending transversely across the shield at the front end were a number of horizontal iron plates or shelves with cutting-edges, as shown clearly by the drawing. The shield was moved ahead by means of a number of hydraulic jacks supplied with power by a hand pump attached to the shield. By means of suitable valves all or any lesser number of these jacks could be operated, and by thus regulating the action of the motive power the direction of the shield could be altered at will. Work was abandoned on the Broadway tunnel in 1870. In 1871-2 Beach’s shield was used in building a short circular tunnel 8 ft. in diameter in Cincinnati, and a little later it was introduced into the Cleveland water-works tunnel 8 ft. in diameter. In this latter work, which was through a very treacherous soil, the shield gave a great deal of trouble, and was finally so flattened by the pressures that it was abandoned. The obviously defective features of this shield were its want of vertical bracing and the lack of any means of closing the front in soft soil.

Fig. 130.—Shield for St. Clair River Tunnel.

[Larger illustration]

Longitudinal Section.

Cross Section.

Fig. 131.—Shield for Blackwall Tunnel.

[Larger illustration]

With the foregoing brief review of the early development of the shield system of tunneling, we have arrived at a point where the methods of modern practice can be studied intelligently. In the pages which follow we shall first illustrate fully the construction of a number of shields of typical and special construction, and follow these illustrations with a general discussion of present practice in the various details of shield construction.

Transverse Section.

[Larger illustration]

Longitudinal Section.

[Larger illustration]

Fig. 132.—Elliptical Shield for Clichy Sewer Tunnel, Paris.

Longitudinal Section.

[Larger illustration]

Cross Section.

[Larger illustration]

Fig. 133.—Semi-elliptical Shield for Clichy Sewer Tunnel.

Mr. Raynald Légouez, in his excellent book upon the shield system of tunneling, considers that tunnel shields may be divided into three classes structurally, according to the character of the material which they are designed to penetrate. In the first class he places shields designed to work in a stiff and comparatively stable soil, like the well-known London clay; in the second class are placed those constructed to work in soft clays and silts; and in the third class those intended for soils of an unstable granular nature. This classification will, in a general way, be kept by the writer. As a representative shield of the first class, the one designed for the City and South London Railway is illustrated in [Fig. 129]. The shields for the London Tower tunnel, the Waterloo and City Railway, the Glasgow District Subway, the Siphons of Clichy and Concorde in Paris, and the Glasgow Port tunnel, are of the same general design and construction. To represent shields of the second class, the St. Clair River and Blackwall shields are shown in [Figs. 130] and [131]. The shields for the Mersey River, the Hudson River, and the East River tunnels also belong to this class. To represent shields of the third class, the elliptical and semi-elliptical shields of the Clichy tunnel work in Paris are shown by [Figs. 132] and [133]. The semi-circular shield of the Boston Subway is illustrated by [Fig. 134].

Half Transverse Section A-B.

Half Rear-End Elevation.

[Larger illustration]

Details of Casting Supporting Ends of Jacks.

Details of Castings under Ends of Girders.

Longitudinal Section C-D.

[Larger illustration]

Fig. 134.—Roof Shield for Boston Subway.

Prelini’s Shield.

—In closing this short review mention will be made of a new shield designed and patented by the Author and shown in [Fig. 135]. It is an articulated shield composed of two separated shields whose outer shells overlap each other. The shields are connected together by means of hydraulic jacks attached all around the two diaphragms. Between these diaphragms is a large inclosed space called a safety chamber, where the men can withdraw in case of accidents and where the air can be immediately raised to the required pressure. This is an advantage in case of blow-outs, because the flooding of the tunnel is prevented, while the accident is limited to only a few feet from the front. On account of the shield being advanced half at a time it is always under control and is thus better directed through grade and alignment. Besides, this shield will not rotate around its axis and consequently it can be built of any shape, thus permitting the construction of subaqueous tunnels of any cross-section and even with a wider foundation, which is impossible to-day with the ordinary rotating shields of circular cross-section.

Fig. 135.—Transversal and Longitudinal Section of Prelini’s Shield.

[Larger illustration]