Trench collapses still kill people in the UK, and almost none of those deaths are down to a lack of engineering knowledge.
We have understood how to support a trench for decades. The guidance exists, the equipment is available on hire from a yard down the road, and the mechanics are not in dispute. And yet, most years, someone is buried — usually in a trench that was only going to be open for twenty minutes, in ground that stood up fine yesterday.
A cubic metre of saturated clay weighs around a tonne. A wall collapse drops several tonnes onto someone faster than they can react, and the usual cause of death is asphyxia — the soil prevents the chest expanding — not crushing. This is why the subject is worth getting right rather than getting through.
This article covers how to select trench support on a UK site: the real difference between a trench box, shoring and battering, what the standards actually require, and where the decision goes wrong.
People talk about trench support as if it were one thing with a few product variations. It is not. There are three fundamentally different approaches, and confusing them is where problems start.
Battering means cutting the sides back to a stable angle so the trench supports itself. No equipment goes into the excavation because the geometry does the work.
It is often the safest option where there is room, because nobody has to rely on a system performing. But it needs space — a deep trench battered to a safe angle in soft ground consumes a lot of width — and it needs ground that will hold the batter. It also generates far more spoil and backfill, which has a cost.
Shoring supports the trench walls. Hydraulic struts, waling frames or timber close-boarding bear against the faces and actively hold the ground in place.
The important word is support. Shoring is intended to prevent the walls moving, which matters enormously when there is a service, a structure or a road next to the trench that cannot tolerate ground movement.
A trench box protects the people inside it. It is a two-sided shield, spanning between panels on spreaders, that provides a safe zone for operatives if the walls collapse.
Here is the distinction that gets missed on site: a trench box does not necessarily support the trench. The ground outside the box can still move. The box protects whoever is standing in it; it does not guarantee the walls stay where they are. If there is a live service running parallel a metre away, a trench box may keep your operative safe while the ground movement damages the service anyway.
So the first question is never “which box do we have.” It is: am I trying to protect the operative, or hold the ground still? Those have different answers, and reaching for a box out of habit answers a question you may not have asked.
Several documents sit behind trench work, and they do different jobs.
HSG185, Health and Safety in Excavations — Be safe and shore, is the HSE's key guidance. It sets out the well-known principle that excavations are dangerous and must be made safe by support, battering or another suitable method.
CIRIA Report 97, Trenching Practice, is the recognised design guide for temporary support in trenches up to 6 m deep. It takes the guesswork out of selecting and designing a support system, and is written for designers, temporary works coordinators and supervisors alike.
BS EN 13331 covers proprietary trench lining systems — the product standard for the boxes and shoring you hire.
CDM 2015 places the duties. Regulation 22 requires excavations to be checked by a competent person at the start of every shift, after any event likely to have affected stability such as rainfall, frost or vibration, and after any accidental fall of material. Those inspections must be recorded.
There is a widely-quoted rule that trenches deeper than 1.2 m require support or battering. It is a useful prompt, but it is routinely misunderstood.
Depth is not the trigger for danger — ground is. Soil can collapse from a trench well under 1.2 m, particularly if it is loose, waterlogged, recently backfilled or surcharged. The honest position is that support is required whenever the ground could collapse, and 1.2 m is a threshold below which people wrongly assume they are safe.
Treat 1.2 m as the depth at which complacency becomes indefensible, not the depth at which risk begins.
A new drainage connection in an established residential street. The dig is 2.4 m deep, in mixed made ground over firm clay, running along a carriageway with parked cars and a live water main shown on the records roughly a metre from one side.
The instinct is to drop a trench box in and get on with it. Work through it properly and the picture changes.
The box protects the gang, but not the main. A trench box will keep the operatives safe. It will not stop ground movement reaching the water main a metre away. If that main cannot tolerate movement, the requirement is shoring that actively supports the wall, not a shield that protects the people.
The surcharge is significant. Parked cars and passing traffic sit right at the trench edge, and traffic loading close to a trench substantially increases the pressure on the walls. That surcharge is a design input, not a detail — a support system adequate for an unloaded trench may be inadequate with an HGV passing a metre away.
The made ground is the weak layer. The firm clay lower down is not the problem. The upper made ground — variable, possibly loose, possibly containing old fill — is where a collapse initiates, and it is exactly where the operatives climb in and out.
Plant standing-off matters. The HSE rule of thumb is to keep tracked plant back from the edge by at least the depth of the trench. For a 2.4 m dig, the excavator tracks want to be 2.4 m clear — which, in a narrow street with parked cars, may not be achievable without traffic management that nobody costed.
None of this is exotic. It is the ordinary reality of a routine dig, and it is precisely the kind of job where corners get cut because it looks simple.
Confusing a box with shoring. Using a shield where the requirement is genuine wall support, and discovering the difference when the ground next to the trench moves.
“It stood up yesterday.” Ground that held overnight is not proven safe. Moisture, vibration and time all reduce stability, and a face can fail without warning.
Ignoring surcharge. Spoil heaped at the edge, plant tracking close, or traffic passing all load the wall. Spoil should be kept back from the edge, not piled against it.
Getting in and out over an unsupported top. The support stops short of the surface, and the operative climbs through the one unsupported zone every time. Access arrangements are part of the design.
No water management. A trench that floods unexpectedly is working under conditions the support was never designed for. Dry at the surface tells you nothing about a permeable layer at depth.
Installing or removing support with people in the trench. The sequence of getting a box in and out has its own risks, and it needs planning rather than improvisation.
Lone working. Deep trench entry needs a second person at the surface, not one operative out of sight in a hole.
Treating inspection as a formality. The start-of-shift inspection by a competent person is a legal requirement and a genuine control — not a box to initial.
There is no safe depth below which support can be assumed unnecessary. The widely-quoted 1.2 m figure is a prompt, not a threshold of safety — ground can collapse from a shallower trench if it is loose, wet, backfilled or surcharged. Support is required whenever the ground could collapse.
A trench box (or shield) protects the people working inside it if the walls collapse, but does not necessarily support the walls or prevent ground movement. Shoring actively supports the walls to hold the ground in place. Where a nearby service or structure cannot tolerate movement, shoring is required rather than a box.
Proprietary systems are usually supplied with a manufacturer's design, which must be confirmed as suitable for the actual depth, ground and loading. Larger, multi-sided or unusual excavations should be designed by a competent engineer. CIRIA 97 provides the basis, and the work sits under BS 5975 temporary works procedures.
As a rule of thumb, keep tracked plant and excavated spoil back from the edge by at least the depth of the trench — for a 2 m trench, 2 m clear. Loads near the edge act as a surcharge on the trench walls and significantly increase the risk of collapse.
CIRIA Report 97, Trenching Practice, is the recognised UK guide to the design and use of temporary support for trenches up to 6 m deep. It is written for designers, temporary works coordinators, supervisors, estimators and planners, and covers identification of ground conditions and selection of support.
Under CDM 2015, a competent person must inspect at the start of every shift before work begins, after any event likely to have affected stability such as rain, frost or vibration, and after any accidental fall of material. The inspection must be recorded in writing.
It can, but water changes the loading and the behaviour of the ground, and the standing water conceals the trench floor, services and any collapse. Water management is required, and the support system must be suitable for the saturated condition rather than assumed adequate.
Because the failure is almost never a lack of knowledge. It is a shortcut under programme pressure — a trench that will “only take a few minutes”, a gang that has always done it a certain way, a box used where shoring was needed. The gap is behavioural, not technical.
Trench support is one of the best-understood problems in temporary works and one of the most consistently fatal, and those two facts are connected. Because the knowledge is not the issue, better calculation does not fix it.
Decide first what you are protecting — the operative, the ground, or both — because that selects the approach. Match the system to the real conditions rather than the contents of the yard. Treat surcharge and water as design inputs. Keep plant and spoil back. And inspect properly, because ground that stood up yesterday is not evidence of anything.
The safest trench is rarely the one with the most equipment in it. It is the one with the right equipment for the conditions, installed by people who asked what they were actually trying to prevent.
LMT.Pro runs temporary works calculations and produces the design brief, calculation report, risk assessment and RAMS — ready to issue.
Try LMT.Pro free →