How a flashback arrestor works: quenching sintered and wet elements, and the valves that back them up

A flashback arrestor stops a flame travelling backwards through a gas line by taking heat out of it faster than the burning gas can replace it. It forces the flame through a mass of very fine passages (sintered metal, packed grains, wire mesh or a column of water broken into bubbles), and it is usually backed by a non-return valve and a cut-off valve that shut off the gas once the flame has been stopped. This page explains the mechanism, the numbers behind it, and how the Indian and international standards define a device that works.

A torch flame sits still only while the gas leaving the nozzle moves faster than the flame can burn back into it. When the local gas velocity in a passage falls below the burning velocity of the mixture, the flame retreats into the torch, the hose and, if nothing stops it, the regulator and the cylinder or pipeline. That is a flashback. The common triggers are well documented: unequal pressures in the oxygen and fuel hoses, a blocked or overheated tip, a cylinder allowed to run nearly empty, a faulty check valve and careless lighting-up or shutting-down (American Welding Society Fact Sheet 28; WorkSafe ACT guidance). The guide to what causes flashback covers them in detail.

Acetylene is the difficult gas for two reasons. Its oxyacetylene flame burns faster, in a thinner reaction zone, than that of any other common fuel gas, so it penetrates smaller gaps. And acetylene can decompose with no oxygen at all: above a limiting pressure a flame can run through pure acetylene in a pipe, and in a long enough pipe it can accelerate into a detonation. An acetylene arrestor must therefore stop a mixture flame coming back from the torch and, on plant and pipeline duty, a decomposition flame or detonation as well.

A flame survives because its reaction zone passes enough heat forward to ignite the gas ahead. Confine that zone between walls set close together and the walls absorb heat faster than the reaction produces it; below a critical size the flame goes out. The critical size is the quenching diameter for a round passage, or the quenching distance between parallel plates. It falls as pressure rises, which is why a device that quenches at low pressure can fail at a higher one.

The figures that matter for acetylene are small:

Two practical conclusions follow. A fine wire gauze does not stop an oxyacetylene flame, because its openings are far larger than 0.12 mm. And although the decomposition flame in pure acetylene is easier to quench at low pressure, the margin closes quickly as pressure rises; at 10 atm a detonation passes a 3 mm tube. The same literature records that a detonation is slowed not as it enters a narrow restriction but as it expands out of it.

The Commission Permanente Internationale de l'Acétylène described the dry arrestor in 1968 in the same terms: a small steel body filled with grains of controlled porosity that stops the flame by cooling it through "a large number of pipes of small section" (C.P.I., Safety in the Production and Use of Acetylene, Paris 1968). The C.P.I. also recorded the pipe-diameter effect underlying every acetylene arrestor: in 0.4 mm capillaries, decomposition propagated only at pressures far above 30 kg/cm², whereas in 100 mm pipe detonation was to be feared above 0.6 kg/cm².

The element stops the flame. It does not stop gas flowing the wrong way, and it does not stop a fire burning on its face if gas keeps arriving. Modern devices therefore stack functions, and IS 11006:2011 Annex B gives each a letter for the body marking:

A dry acetylene device with a flame arrestor and a non-return valve is marked A D F N. The layered idea is not new: the Tri-Sure arrestor described in Engineering in 1962 (vol. 193, p. 425) trapped the flame, closed a diaphragm-operated valve, then locked off the gas flow. A non-return valve on its own is never a substitute for the element; see non-return valve vs flashback arrestor.

Because no practical vessel can be built to contain an acetylene detonation. Measured deflagration pressure rises are about 11 times the initial absolute pressure (11.1 ± 0.6 in sixteen runs, Miller and Penny, IChemE symposium, 1960). A detonation in a long pipe exerts 40–50 times, the cascading described by Sargent (1957) pushes the rise beyond 100–200 times, and Webster and colleagues measured terminal ratios above 320 times in a 12.7 mm tube (IChemE Symposium Series No. 33, 1972). An arrestor at every pipe-to-vessel junction confines the design case to deflagration, which is buildable. IS 11006:2011 cl 4.1 reflects the same thinking: the casing must withstand 10 times its maximum working pressure or 60 bar, whichever is higher, for at least 5 minutes, so a 1.5 bar acetylene device is tested to the 60 bar floor.

IS 11006:2011, first published in 1984, is the Indian specification. Its scope is broad: delivery pipelines, acetylene generators, fuel-gas storage and piping, and welding and cutting systems, wet and dry. It is an indigenous text, not an adoption of ISO, and BIS records no amendment. ISO 5175-1:2017 covers devices incorporating a flame arrestor installed downstream of manifold, cylinder or pipeline outlet regulators and upstream of blowpipes; ISO 5175-2:2017 covers safety devices without one. ISO 5175-1 states that it "does not specify the location of these devices in the gas system" and excludes premixed supplies. A generator-outlet wet arrestor therefore falls under IS 11006, not ISO 5175-1. The flashback arrestor standards page sets out the clauses.

An arrestor that has stopped a flame has done its job once. Soot from the event fouls the element, a temperature cut-off that has operated cannot be reset, and a decomposition can creep slowly through a packed element and re-ignite the gas beyond it minutes later, the subject of the twenty-minute delay. Even a successful quench in a large pipeline sends a pressure pulse downstream.