Packed-tower flashback arrestors for large acetylene pipelines: Raschig rings gas velocity, packed height and the upstream bursting disc

A packed-tower flashback arrestor is the answer to a problem small arrestors cannot solve: a long, wide acetylene pipeline at low pressure, where a sintered element would cost too much pressure and a decomposition can still travel kilometres. The tower is a vessel full of Raschig rings, sometimes irrigated with water, that offers a large cooling surface to the flame while leaving wide passages for the gas. Its design rules come from tests made after the Hüls pipeline explosion of 1954, and they still set the numbers: ring size, gas velocity, packed height and an upstream vent.

Because acetylene in a wide pipe is far less stable than in a narrow one, and the pressure available to push gas through an arrestor is small. On 2 January 1954 at the Chemische Werke Hüls in Germany, chlorine was admitted into an acetylene line by a sequence of mishaps. The decomposition it started travelled through more than 2 km of 15, 30 and 60 cm pipe at only 1.03 atm absolute, below 0 °C, wrecked several large items of plant and injured a number of people (Dechema Information Bulletin, 7 March 1954; Chemical and Engineering News, 1954).

The pipe-diameter data explain why. The C.P.I. recorded that in 100 mm pipe a detonation is to be feared above 0.6 kg/cm², against more than 7 kg/cm² needed for full propagation in 10 mm pipe (C.P.I., Safety in the Production and Use of Acetylene, Paris 1968). Kogarko and colleagues proposed limiting transport to 1.4 atm absolute in 20 cm pipe and 1.05 atm absolute in pipe up to 38 cm (Khimicheskaya Promyshlennost, 1962). Long lines of 15–60 cm carrying acetylene at 0.03–1.5 atm gauge existed at a number of locations, and they needed an arrestor with very low pressure drop. Some used towers containing porous metal plates; the design that was tested and widely adopted was the Raschig-ring tower. Containment is not an option either: Carver, Smith and Webster measured terminal pressure ratios above 320, a lower bound, in a 12.7 mm tube (IChemE Symposium Series No. 33, 1972). The pipe diameter and detonation guide covers the underlying data.

Schmidt and Haberl tested a ring-packed tower against real detonations (Technische Überwachung, 1955, 423; Chemie-Ingenieur-Technik, 1956, 677):

Water makes a large difference. Irrigated rings quench almost at once because the water film and spray add a heat sink to the metal or ceramic surface, the same reason a wet hydraulic seal works. Dry rings still work but need far more depth.

The rules drawn from that work, as compiled by S. A. Miller (1965), are short:

The ring rule is stricter than the test. The tested tower used 2.5 cm rings in a 25 cm shell, a ratio of one-tenth; under the rule, a 25 cm tower would take rings below about 1.7 cm.

The 60 cm/s rule sets the tower diameter for a given flow. Taking it as the velocity through the empty cross-section at line pressure, the arithmetic is:

Flow here is actual volume at line pressure. At 0.5 atm gauge, 1 m³ of line gas is about 1.5 m³ at atmospheric pressure, so a 25 cm tower handles roughly 160 m³/h measured at atmospheric conditions. These figures are arithmetic on the published rules, not tested ratings.

Because an empty space inside the housing is a combustion chamber. If a flame from the pipeline can enter a void above or below the packing before it meets the rings, it can burn there as a deflagration, and a deflagration in a closed vessel raises pressure about elevenfold: 11.1 ± 0.6 times the initial absolute pressure in sixteen runs by Miller and Penny (IChemE symposium, 1960). With rings packed right to the nozzles, quenching starts at once and the housing never sees that load. Check any tower drawing for unpacked plenums, distributor spaces and settling gaps, and allow for rings settling in service.

A tower that stops the flame still passes a pressure wave. In Schmidt and Haberl's tests the gas in the long feed pipe had been heated to about 2,800 °C and raised to 72 atm before reaching the tower, and it did not instantly return to ambient conditions. More than 10 atm passed into the pipe beyond, although no explosion happened there. A bursting disc set at 10 atm, mounted in a tee near where the explosion entered the tower with its outlet pointing upwards, gave the gas an escape and limited the transmitted pulse to about 3.5 atm gauge.

The disc is a partner to the arrestor, not a substitute. Miller and Penny measured peaks of 50–150 atm near a disc set to burst at 32 atm, and a disc cannot outrun a detonation front. Design the tower and its upstream vent together, and rate the downstream pipe for the residual pulse. See bursting discs vs flashback arrestors.

Rarely in welding distribution. IS 15190 (Part 1):2002 covers acetylene pipelines up to 155 kPa gauge and DN 100, recommends restricting pipe to DN 50, and excludes pipelines in chemical plants where acetylene is a process feed. Within that scope ordinary wet and dry arrestors serve, and cl 4.12.1.5 allows several units in parallel where one cannot carry the flow. Towers belong on large-bore plant and process lines, gasholder connections and long transfer mains outside that envelope. IS 11006:2011 names delivery pipelines and acetylene generators in its scope, but its flashback type test, five flashbacks with 32–35 % acetylene in oxygen, was written for compact devices; a tower should also be proved against pipeline decomposition, as Schmidt and Haberl did.