Part 2 Underwater Cutting Explosions Oxygen’s Role

Examining the risk of explosion given specific conditions.

underwater cutting explosions

In the first article of this series, we have seen how and how much H2 is produced during underwater cutting. But is this gas by itself susceptible to explode? The answer is no. How, then, does an explosion occur?

We are generally taught that to trigger an explosion, we must bring together the three elements that are included in the fire triangle. These three elements are an oxidizer, a fuel and a source of ignition. While it is true that the mere presence of these three elements is enough to start a fire, it will not be enough to automatically start a gas explosion. To happen, it will also be necessary that the fuel / oxidizer mixture is confined in any enclosure in the form of a gas mixture in suspension and that the concentration thresholds in oxygen and fuel are conducive to explode.

If these six conditions aren’t met simultaneously, then the risk of explosion then becomes zero. Thus, to explode, our H2 must be mixed within a certain range with an oxidizer, which for us will always be the oxygen coming from the torch or eventually our breathing air in case of welding.

This range is defined on the one hand by the lower explosive limit (LEL) and on the other hand by the upper explosive limit (UEL). Below this LEL, there is too little combustible gas in the mixture to trigger ignition, while above this UEL there will be too little oxygen in the gas mixture to sustain the explosion.

However, not all flammable gases have the same explosive range and for the same gas it varies depending on the oxidizer (oxygen or air). For what concerns an H2 / O2 mixture, the explosion domain is between 4 and 95%.

Does this mean that there will be an explosion as soon as 4% hydrogen is reached? As nothing is found in our manuals, we have carried out in a series of tests on small vessels containing from 4 to 133 ml (0,24 to 8,11 in3) of H2/O2 mixture at ratios ranging from 5 to 100% and thanks to it we were able to trace an explosion pressure profile.

a figure hexagon of explosion
Fig 1: Hexagon of explosion.
a figure explosivity range


Fig 2: Explosivity range.
a figure tank for the h2o2 mixtures explosion tests


Fig 3: Tank for the H2/O2 mixtures explosion tests.
a graph h2o2 explosion profile curve


Graph 1: H2/O2 explosion profile curve.
a figure residual gas recovery system and test method


Fig 4: Residual gas recovery system & test method.

By looking at this profile, we can confidently say that the answer to the question posed above is no. From 4 % hydrogen, there is effectively a reaction with oxygen in the form of a “flash over” but without any sign of overpressure. A slight rise in pressure starts only from 8% hydrogen, and then accelerates from 12%.

We also note that the more the percentage of fuel gas increases, the more the explosion pressure will increase until it reaches a maximum when the ideal mixture (stoichiometric mixture in which all the oxygen as well as all the fuel is burned) is reached. Then the explosion pressure begins to drop until the upper explosive limit is reached and prevents any explosion. Now that we know that at least 8% of H2 is needed to generate a beginning of an explosion, let’s see if that percentage is reached during underwater cutting. Very little concrete information on this subject is currently available, but for the few that are, the results are diametrically opposed. On the one hand, there are those mentioned by two well-known trade associations that claim without further precision that this percentage of hydrogen is very high and on the other hand, the results of a few rare residual gas analyzes which on the contrary indicate extremely low percentages always below the LEL. So, which one to believe? The best was therefore to find the results by ourselves.

We have therefore realized another series of tests (15) in various cutting conditions into which the residual gases were recovered.

Some (9) of the gas samples were submitted to a flame to verify if they could generate an explosion while the gases from the other samples (6) were analyzed. If we except the explosion of a bag filled with a mixture of welding gases and air, no other bag exploded.

As for what concerned the gas analysis, none of the samples had a fuel percentage above 3%, which was therefore more in accordance with the results found on the internet.

a graph oxygen flow nl without depth correction in function of the rod type


Graph 2: Oxygen flow (NL without depth correction) in function of the rod type.
a figure explosion test


Fig 5: Explosion test.

How can such a low fuel percentage be explained?

Whether the cutting is performed using an oxy-arc or ultra-thermic electrode or by using a gas cutting torch, these two types of method always use oxygen.

This will allow, on the one hand, carrying out the combustion of the metal and on the other hand, thanks to the action of the pressure jet, to eliminate and drive these oxides out of the kerf. To carry out this flush, a fairly high cutting pressure is generally used (4 to 7 bars), which as can be seen on the graph generates a more or less significant consumption of oxygen.

Yet of all the volume sent through the electrode (or through the gas torch nozzle), only a small amount of the total volume of oxygen is burned during the oxidation of iron. Without entering into the details of a chemical equation, it is estimated that between 2 liters (122 in3) and 3 liters (183 in3) of O2 will be needed to cut 1 cubic centimeter (0.06 cu in) of steel. Clearly, this means that under normal conditions of use, the percentage of unburned oxygen present in the residual gases is relatively high and, according to published studies, generally exceeds 97% (thus less than 3% hydrogen).

On the other hand, this percentage of oxygen present in the residual gases can decrease if for one or another reason the cutting pressure (and therefore the flow rate) is reduced below the pressures prescribed by the manufacturer. Not only does this decrease the performance of the electrode, but it will therefore indirectly increase the percentage of hydrogen present in the residual gases.

This risk is then particularly present when the diver uses oxy-arc electrodes, because already in normal condition of use and due to its design, this type of electrode consumes as can be seen in graph 2, less oxygen than an ultra-thermic rod, and therefore less unburned oxygen is present in the gases which rise to the surface.

a table of hydrogen in function of the oxygen flow with an exothermic rod
a table of hydrogen in function of the oxygen flow with an oxy arc rod

To fully understand what is described above, we can take inspiration from the following example:

Knowing that it takes +/- 3 l (183 in3) of oxygen to cut 1 cc (0.06 cu3) of steel (iron) and that this same volume of steel (1 cc) will produce by thermolysis about 8 ml (0.49 in3) of hydrogen, let’s check what the % H2 in the residual gas will be when a diver cuts a steel plate 35 cm (13.78″) long × 1 cm (0.39″) thick.

In our example, the diver will make 2 cuts at different pressures using a Ø 9.5 mm exothermic electrode (1.1 cm (0.43″) wide kerf) (table 1), as well as 2 cuts using a oxy-arc electrode Ø 7.9 mm (0.9 cm (0.35″) wide kerf) (table 2).

Finally, for comparison, the % H2 on another 35 cm (13.78″) × 5 cm (1.97″) thick cut (1 cm (0.39 ″) wide kerf) is also calculated (table 3).

As can be seen, the H2 concentration can rapidly increase if the diver is using a low oxygen pressure, but also if he cuts thick steel. Concerning the normal condition of use and in view of the results obtained during our different tests as well than those published on the web, it seems a priori that the risks to get a high concentration of hydrogen into the residual gases is limited. But CAUTION, this does not mean that explosions caused by hydrogen are not possible. Indeed, it would seem that in practice and during a gas cloud explosion in the open air, a freshly contained explosive mixture is sometimes characterized by a non-uniform composition in which various concentrations of the mixture can be met, some of which being explosive.

a table hydrogen in function of the steel thickness with an oxy arc rod

It is therefore not impossible that such a same situation can also be encountered underwater. Now, it must be known that the underwater cutting gas explosions are not always caused by hydrogen. They can also be caused by other fuels (mainly methane). Sometimes also unexplained explosions can occur incidentally when the gases are not confined at all. These are called “steam explosion” and may be due to a violent reaction of the incandescent slag with a cold liquid, which then causes the explosive fragmentation (nucleation) of the slag. Although these are possible with carbon steels, this type of explosion is however encountered more frequently when cutting so-called exotic metals such as zinc, aluminium, certain type of bronze, and even stainless steel, or thick steel pieces.

As you can see from these two articles, cutting underwater can generate the risk of explosion, the consequences of which are sometimes fatal. This is why this technique should only be undertaken after having carried out a correct analysis of the risks that may arise and by making sure in particular that no gas accumulation can occur or is already present near the cutting or welding area.

If this risk exists, it is then imperative to evacuate the gases which are present or which may accumulate, by practicing evacuation vents and by making sure to sufficiently ventilate the concerned area with an inert gas or, failing that, with compressed air to dilute the flammable mixture and keep it below its lower explosive limit.

If despite these measures, the risk analysis shows that there is still a risk of gas accumulation, then DO NOT CUT and choose another work technique.

If you want to know more about these underwater cutting explosions the you may read the following document which can be downloaded for free:

Underwater Cutting Explosions Causes/Effects/Consequences & Prevention

DIVE SAFE and as a well-known burner says VENT! VENT! VENT!