Confined space entry on agricultural and processing sites

What counts as one

Broadly, and you should read the definition in the regulation that applies to you rather than this paragraph, a confined space is an enclosed or partially enclosed space that is not designed or intended primarily to be occupied by a person, that is or is likely to be entered by a person, and that carries a risk from a harmful atmosphere, from oxygen being too low or too high, or from engulfment.

Size is not the test. A large tank with a top hatch can be a confined space. A small storeroom with a door and normal ventilation generally is not.

The other thing people miss is that a space can move in and out of the definition depending on what is in it and what is happening. A vessel that is safe empty and open is a different proposition part filled, or mid ferment, or straight after a clean.

Where they are, on the sites I work on

Grain silos and field bins. Fermentation tanks and wine vats. Product tanks, mixing vessels and anything that gets cleaned in place. Effluent pits, sumps and wet wells. Auger and conveyor pits. Road tankers and milk tankers. Chemical and gas storage vessels. Some pump and valve pits.

None of these look dangerous from the outside, and most of them are entered by someone who has done it many times before without a problem. That is the risk profile that worries me most: a task with a long history of nothing going wrong, right up until the day the atmosphere is different.

What causes the harm

Atmosphere is the main one. Oxygen displaced by carbon dioxide during fermentation. Hydrogen sulfide from effluent and organic sludge. Nitrogen used to blanket a tank. Ammonia from a refrigeration leak finding a low space. Residues and vapours left behind by a cleaning chemical, particularly where an acid and a caustic step run close together.

Engulfment is the other. Grain behaves like a liquid when it is flowing and it can take someone down faster than they can react, and bridged or crusted material can collapse under a person's weight without warning.

Then the mechanical hazards that happen to be in the same space: agitators, augers, blades and anything that can start while someone is inside because it was never properly isolated. Heat, too, in a closed vessel in summer.

The control that matters most is not entering

Every hour spent improving an entry procedure is worth less than an hour spent removing the need for the entry.

That can mean cleaning in place rather than by hand, inspection ports and cameras rather than a person, sampling from outside, fittings that can be serviced without going in, or changing a maintenance routine so a task that used to need entry does not.

It is not always possible. It is possible more often than most sites assume, because the entry became routine long before anyone asked whether it was necessary.

If entry is genuinely unavoidable

A written entry permit for that space, that task and that day, issued by someone with the authority to refuse it. A generic permit signed in advance is not a permit.

Atmospheric testing before entry, and monitoring while people are inside, with equipment that is calibrated and in date. Testing once at the hatch tells you about the air at the hatch.

Isolation and lock out of everything that could move, fill, drain, heat or pressurise the space, verified physically rather than assumed from a drawing.

Ventilation appropriate to what you are dealing with, and enough time allowed for it to work.

A standby person outside, in continuous communication, who does not have another job to do at the same time and does not enter.

Signage and physical barriers so nobody wanders in behind you, and access control on the openings.

Respiratory and other protective equipment chosen for the atmosphere the testing found.

And entrants, standby people and rescuers who are trained for the work. That training is something you arrange. My part is making sure the permit system records who holds what, and checks it before the hatch opens rather than afterwards.

Rescue is the part that goes wrong

Regulator guidance across jurisdictions makes the same point over and over: the person who goes in to help is often the next casualty. Someone collapses, a workmate goes in after them, and the atmosphere that took the first person takes the second.

A rescue plan has to be written down, has to be specific to that space, and has to be capable of getting a person out without a second person entering unprotected. That usually means retrieval equipment rigged before entry.

If the plan is "call triple zero", then you need to know how long that takes to your site, what the responding service can do when it arrives, and what happens in the meantime. On a farm an hour from town, that is a very different answer from a plant in an industrial area, and it needs to be worked out before the day.

Where to check the current position

Safe Work Australia publishes a model Code of Practice for confined spaces, and each jurisdiction publishes the version or the equivalent that applies there. Victoria publishes its own compliance code.

The relevant Australian Standard on confined spaces is also worth having if you are designing an entry system rather than following one.

Definitions and duties do change. Check the current instrument for your jurisdiction before you rely on a detail here.