Cast iron pipes can fail in many modes which in general can be summarized into two categories: loss of strength due to the reduction of wall thickness of the pipes, and loss of toughness due to the stress concentration at the tips of cracks or defects. Even in one category there can be many mechanisms that cause failure. The strength failure can be caused by hoop stress or axial stress in the pipes. A review of recent research literature (Sadiq et al., 2004; Moglia et al., 2008; Yamini, 2009; Clair and Sinha, 2012) suggests that current research on pipe failures focuses more on loss of strength than loss of toughness. As was mentioned in Section 3.3.7(b), the literature review also revealed that in most reliability analyses for buried pipes, multifailure modes are rarely considered although in practice this is the reality. Therefore the aim of this section is to consider multifailure modes in reliability analysis and service life prediction for ductile iron pipe. Both loss of strength and toughness of the pipe are considered. A system reliability method is employed in calculating the probability of pipe failure over time, based on which the service life of the pipe can be estimated. Sensitivity analysis is also carried out to identify those factors that affect the pipe behavior most.

Buried pipes are not only subjected to mechanical actions (loads) but also environmental actions that cause the corrosion of pipes. Corrosion related defects would subsequently cause fracture of cast iron pipes. In the presence of corrosion pit, failure of a pipe can be attributed to two mechanisms: (i) the stresses in the pipe exceed the corresponding strength; or (ii) the stress intensity exceeds fracture toughness of the pipe. Based on these two failure modes, two limit state functions can be established as follows.


Steel pipe is manufactured by the pit, horizontal or centrifugal method. In the vertical pit method, a mold is made by ramming sand around a pattern and drying the mold in an oven. A core is inserted in the mold and molten iron is poured between the core and the mold. In the horizontal method, a machine is used to ram sand around horizontal molds that have core bars running through them. The molten iron is poured into the molds from multiple-lipped ladle designed to draw the iron from the bottom to eliminate the introduction of impurities. In the centrifugal method (Figure 3.4), sand-lined molds are used that are placed horizontally in centrifugal casting machines. While the mold revolves, an exact quantity of molten iron is introduced, which, by action of the speed of rotation, distributes itself on the walls of the mold to produce pipe within a few seconds.


Many cast iron pipes made towards the end of the nineteenth century are still in use; their walls were relatively thick and not always of uniform, ‘Spun’ grey iron pipes were formed by spinning in a mould and produced a denser iron with pipes of more uniform wall thickness; they comprise a large proportion of the distribution mains in many countries. Three classes of such pipes were available: B, C, and D for working pressures of 60, 90, and 120 m respectively; classes B and C were more widespread. Carbon is present in the iron matrix substantially in lamellar or flaky form; therefore, the pipes are brittle and relatively weak in tension and liable to fracture. The manufacture of grey iron pipes has been discontinued in most countries, except for the production of non-pressure drainage pipes.


Since cast iron pipes are deteriorating rapidly and causing so many maintenance problems (Section 4.3.2), the distribution network is currently undergoing an extensive replacement scheme with old, leaking and corroded cast iron pipes being replaced by MDPE and uPVC. These new plastic pipe materials are thought to support fewer bacteria than the old hubless cast iron pipe. Their surface is smoother and therefore the surface area smaller and they are not subject to corrosion or biodeterioration.


In addition, the effectiveness of a disinfectant is greatly influenced by the pipe material. Biofilms grown on copper or PVC pipe surfaces were inactivated by a 1 mg/l dose of free chlorine or monochloramine. However, on iron pipes 3-4 mg/l of chlorine or monochloramine was ineffective in controlling the biofilm (LeChevallier et al., 1990) because, as discussed before, the chlorine will preferentially react with the iron surface (LeChevallier et al., 1993). It appears that the option of changing pipe materials to ones with lower biofilm-forming potentials would reduce the biofilm problem.


Many cast iron pipes made towards the end of the 19th century are still in use; their walls were relatively thick and not always of uniform, ‘Spun’ grey iron pipes were formed by spinning in a mould and produced a denser iron with pipes of more uniform wall thickness; they comprise a large proportion of the distribution mains in many countries. Three classes of such pipes were available in the UK: B, C and D for working pressures of 60, 90 and 120 m, respectively; classes B and C were more widespread. Carbon is present in the iron matrix substantially in lamellar or flaky form; therefore, the pipes are brittle and relatively weak in tension and liable to fracture. The manufacture of grey iron pipes has been discontinued in most countries, except for the production of non-pressure drainage pipes.


Lead joint (a) is accomplished by melting and pouring lead around the spigot in the bell end of the pipe. After the lead has cooled to the temperature of the pipe, the joint is caulked using pneumatic or hand tools until thoroughly compacted with the caulking material and made water tight.


Cement joint (b) is started at the bottom with the cement mixture, and the mixture then caulked. Pipe with cement joints must not be filled with water until after 12 h has elapsed.

Roll-on joint (c) requires a round rubber gasket that is slipped over the spigot before it is pushed in the bell. Braided jute is tamped behind the gasket, after which the remaining space is filled with a bituminous compound.


Push-on gasket joint (d) is made by seating a circular rubber gasket inside the contour of the socket bell. The slightly tapered pipe end permits the gasket to fit over the internal bead in the socket. A special lever action tool, manually operated, then allows the bell and spigot past the gasket, which is thereby compressed as it makes contact with the bottom of the socket.


Mechanical joint and pipe joint should be thoroughly cleaned to remove oil, grit, and excess coating and then painted with a soap solution. Cast iron gland is then slipped on the spigot end with the lip extension toward the socket (or bell) end. The rubber gasket, also painted with the soap solution, is placed on the spigot end but with its thick end toward the gland. The entire section of the pipe is pushed forward to seat the spigot into the bell; the cast iron gland is moved into position for bolting.

The Putney gas explosion was a real wake-up call, and accelerated the replacement of old gray ductile iron pipe fittings by polymers such as medium-density polyethylene (MDPE), high-density polyethylene (HDPE), and unplasticized polyvinylchloride (UPVC). HDPE has a tensile strength of ≈20–37 MN m?2 (which is more than adequate for typical internal pressures). Most importantly, though, it has a Young’s modulus which is ≈150–300 times less than cast iron. This means that HDPE pipes can deflect under misalignments of the kind experienced in the Putney explosion without reaching the fracture stress. Even better, over a long time the polymer also creeps, which further dissipates the stresses caused by misalignment. Polymers are also very resistant to corrosion, so should last indefinitely in the ground.

But how are lengths of polymer pipe joined together? The following clip shows how:


http://www.youtube.com/watch?v=83PTUoFBq9s&feature=related


The steps in the process are shown in Figure 27.11. First, the ends of the pipe to be joined are machined flat and parallel using a double-sided rotating disk planer. Then the ends are heated with an electric hotplate. Finally, the hot faces are pushed together using a hydraulic ram. The softened thermoplastics fuse together, making a high-strength leak-proof joint. This is a quick, reproducible method, which requires little skill on the part of the operator—in marked contrast to the lead-filled spigot-and-socket joints of the old cast iron system. Figures 27.12 and 27.13 show an alternative joining method, where one end of the pipe has an enlarged bore into which the mating pipe can be inserted. This overlapping joint can then be fixed and sealed with polymer adhesive. It would be hard to envisage any replacement materials so well adapted to this challenging environment than thermoplastics.

The earliest oil pipelines in the United States, laid in the 1860s, were typically constructed of 2-in cast-iron pipe threaded and screwed together in short segments. Oil was propelled through the pipeline using steam-driven, single cylinder pumps, or by gravity feed. These early pipelines, seldom more than 15 mi in length, were prone to bursting, thread stripping at the pipe joints, and frequent pump breakdowns mainly due to the percussive strain on the lines caused by each stroke of the pump which “resembled the report of a rifled gun.” Development of the four-cylinder Worthington pump revolutionized the transportation of petroleum by pipeline with its constant flow and uniform pressure (The Engineering and Building Record, 1890; Scientific American, 1892; Herrick, 1949; Williamson and Daum, 1959).


By the 1870s, a 2000-mi network of small-diameter gathering lines connected the oil-producing areas with regional refineries and storage points on the railroads and rivers where the oil could be shipped to refineries via railcars or ships and barges. Typical crude oil trunk lines were constructed of 18-ft sections of lap-welded wrought steel pipe fittings 5 or 6 in in diameter joined with tapered, threaded joints manufactured specifically for pipeline service. The pipe was generally buried 2 or 3 ft below the ground surface. Worthington-type pumps were used as the motive power for the lines, and the pumps were powered by steam generated by coal-fired boilers. Pump stations were spaced as needed to maintain the flow of oil over the terrain crossed by the lines. At the pump stations, oil was withdrawn from the lines and passed through riveted steel receiving tanks some of which were 90 ft in diameter and 30 ft high holding about 35,000 barrels (The Engineering and Building Record, 1890; Scientific American, 1892; Herrick, 1949). Diesel-powered pumps began to replace steam power around 1913–1914 (Williamson et al., 1963).

It was not until May 1879 that the Tidewater Pipe Company, Ltd. began operation of the first long-distance crude oil pipeline covering the 100 mi between Coryville and Williamsport, Pennsylvania, to connect with the Reading Railroad. The line was constructed of 6-in wrought-iron pipe laid on the surface of the ground (except when crossing cultivated land) and relied on only two pumping stations, one at Coryville and the other near Coudersport. The expansion of the oil under the hot summer sun caused the line to shift as much as 15–20 ft from its intended position, knocking over telegraph poles and small trees, but no serious breaks occurred. In the spring of 1880, Tidewater buried the entire line (Williamson and Daum, 1959).


The success of the Tidewater pipeline set the pattern for the construction of other long-distance crude oil “trunk” lines which sprang up in the early 1880s connecting the oil regions of Pennsylvania with refining centers in Cleveland, Pittsburg, Buffalo, Philadelphia, Bayonne, and New York City (Williamson and Daum, 1959).


By 1905, the oil fields in the Oil Regions of Appalachia stretching from Wellsville, New York, through western Pennsylvania, West Virginia, eastern Ohio, Kentucky, and Tennessee were becoming depleted. The new oil fields discovered during the early 1900s in Ohio, Indiana, Illinois, southeastern Kansas, northeastern Oklahoma, and eastern Texas were quickly connected by trunk lines to the eastern refining centers as well as the new western refineries in Lima, Ohio; Whiting, Indiana; Sugar Creek, Missouri; and Neodesha, Kansas (Johnson, 1967).


The proximity of the prolific Spindle Top Field to the Gulf coast made the area around Houston, Port Arthur and Beaumont, Texas, and Baton Rouge, Louisiana into a petroleum refining center. Regional pipelines were built to carry crude oil the relatively short distances to the Gulf coast refineries (Johnson, 1967). The oil tanker ships operating from the Gulf coast ports competed for and obtained control of most of the long-distance oil transport to the refineries and markets along the eastern seaboard by the mid-1920s (Williamson et al., 1963; Johnson, 1967).


Until the 1930s, when large-diameter steel pipe was in widespread use, the carrying capacity of oil pipelines was increased by laying an additional line or lines alongside the original pipe within the same right-of-way. This practice was known as “looping.” The carrying capacity of 8-in lines was about 20,000 barrels per day, while 12-in lines handled 60,000 barrels per day. Since the largest refineries operating in that era were designed to handle crude at the rate of approximately 80,000–100,000 barrels per day, the carrying capacity of the pipelines built by a refiner were carefully gauged to support the refinery with little excess capacity to offer to others (Wolbert, 1979; Willson, 1925).


By 1941, just prior to the United States’ entry into World War II, there were about 127,000 mi of oil pipeline in the United States composed of about 63,000 mi of crude oil trunk lines, about 9000 mi of refined product lines, and about 55,000 mi of crude gathering lines (Frey and Ide, 1946). From February through May 1942, 50 oil tankers serving the Atlantic seaboard were sunk by German submarines. The continuing attrition of the tanker fleet by enemy action and the diversion of tankers to serve military operations abroad caused a tremendous increase in the use of pipelines to transport both crude oil and refined products to the east coast which consumed about 40% of the petroleum produced in the United States. In June 1941, before the Pearl Harbor attack, pipelines delivered about 2% of the petroleum needed by the east coast; by April 1945, pipelines carried 40% of this critical supply (Frey and Ide, 1946).

The wartime expansion of the pipeline network added more than 11,000 mi of trunk and gathering lines, repurposed over 3000 mi of existing pipelines in new locations and reversed the direction of flow of more than 3000 mi of other lines (Frey and Ide, 1946). One of the pipelines converted from products delivery and reversed in flow direction to convey crude oil to east coast refineries during the war was the Tuscarora pipeline. After the war, it was reconverted and its direction of flow was again reversed to convey gasoline from the coastal refineries to the interior (Johnson, 1967).


Noteworthy wartime pipelines owned by the federal government were the “Big Inch” crude oil line, the largest pipeline in the world at that time measuring 24 in in diameter for much of its 1254 mi length; and the “Little Big Inch,” the longest refined products pipeline in the world at 1475 mi of 20-in diameter pipeline (Frey and Ide, 1946). Only during World War II did the federal government finance oil pipeline construction (Johnson, 1967).


With the proven success of long, large-diameter crude and refined products pipelines during World War II, the rapid growth in demand for petroleum products in the post-World War II era prompted a great expansion in construction of large pipelines. The number of refined products pipelines increased about 78% from 9000 mi in 1944 to 16,000 mi in 1950. Crude oil trunk lines expanded from about 63,000 mi in 1941 to about 65,000 mi 1950. The postwar increase in the diameter of the crude oil trunk lines, and therefore their carrying capacity, far outweighed the relatively modest increase in mileage (Johnson, 1967) (Table 24.1).