Britain Rides Bikes It Doesn’t Build. So Where Does the Risk Live?

Ask anyone to name the most dangerous place to cycle in Britain and they will picture somewhere loud and obvious. A snarl of buses and lorries in a big city centre. The sort of road you would never put a child on.

We went to the data instead. Not out of curiosity about cycling, but because of a failure we watch repeat in our own work. The hazard that hurts people is rarely the one everyone is watching. It is the one sitting in a number nobody thought to question.

Our analysis of the Department for Transport’s 2024 casualty figures, broken down by local authority, put the City of London at the top. The square mile, not the sprawl. Of 184 road casualties there that year, 88 were cyclists. Nearly half of everyone hurt on those roads was on a bike.

Outside London, the number that stopped us was Reading. It is a busy commuter town with heavy traffic, not a place that looks safe for cyclists. But Reading sits 11th in the country and first outside London, with cyclists making up 28.88% of all road casualties.

Proper cycle lanes, segregated routes and engineered junctions push that share down. Where they do not exist, cyclists absorb more of the risk even on roads that look no different from any other. The data does not tell you a place is dangerous. It tells you whether the controls in place are protecting the people they are supposed to protect.

That is a process safety argument. We just made it with road data.

We went looking for the risk behind Britain’s cycling boom in the places bikes are built, and found almost nothing to look at.

Around 1.6 million bikes are sold here every year, and only about 160,000 of them are built in Britain, by a small and stubborn set of makers in places like west London and the West Midlands. The rest arrive in containers. We import roughly three times the value of bikes we export, much of it from China and the rest of Asia. Britain is one of Europe’s biggest bike markets and one of its smallest makers.

The hazard of building the bike went offshore with the work.

But the bike is changing. The fastest-growing segment is the e-bike, a machine built around an energy-dense battery rather than a mechanical drivetrain. Unlike the conventional frames that went offshore with manufacturing, the cells that power them are now being built here. The hazard that comes with them, combustible dust and thermal runaway, is not something a bicycle workshop has ever had to manage.

The e-bike is not really a bicycle. It is a bicycle wrapped around a battery, a small energy-storage system you park in your hallway. The fires that gut a flat in minutes are not a cycling problem. They are a battery problem that happens to be riding a bike.

And batteries, unlike bikes, we have decided to build here.

The map of where that happens is filling in fast. A cell plant of more than 15 gigawatt-hours is now running in Sunderland, next to the car plant. A four-billion-pound facility is rising at Bridgwater in Somerset. Another has planning permission in Coventry. These are some of the most energy-dense manufacturing sites this country has built in a generation, and the hazards inside them are the two we spend our working lives on.

The first is dust. Making a cell means handling fine powders and flammable solvents at industrial scale, and powder handling like that is a recognised combustible-dust scenario, one that has to be assessed rather than assumed.

Building lighter vehicles adds another. Machining aluminium and magnesium throws off metal dust, and metal dust is among the most violent explosion hazards there is. None of this is exotic. Powder-coating a bike frame, a trade that never left, is a combustible-dust process too.

The second is heat. Battery cells, lithium-ion being the most prevalent but not the only chemistry in production, hold a remarkable amount of energy in a small space, and under the wrong conditions release all of it at once, a failure called thermal runaway. Once it starts, it is almost impossible to stop.

At the scale of a hallway, that is the e-bike fire on the evening news, now the fastest-growing fire risk some UK fire services track. South Wales alone logged 136 lithium-battery fires in five years, most of them in the last two.

At the scale of a factory, it is something else. Two years ago, on 24 June 2024, battery cells began to explode at a lithium plant in Hwaseong, South Korea. The fire tore through a floor holding around 35,000 cells in moments and killed 23 people. Investigators later described a plant that had been rushing production and had gone five years without a safety inspection.

The hazard was never one cell. It was one cell multiplied by everything stacked around it, and that multiplication is close to invisible at sample scale and total at production scale. This is what a responsible cell maker characterises and designs out before a battery ships, using thermal tests like DSC and ARC, the methods that show when and how fast a material lets go of its energy.

Again, none of that is new. Dust has exploded and reactions have run away for as long as we have made things. What is new is the scale, and the speed it is being built at.

The misrepresentation of DSEAR regulations is one of the most consistent gaps we see in practice. DSEAR is treated as a one-time obligation, completed at the point of setup and filed. A hazardous area classification done when a facility handled one set of materials does not automatically extend to cover what that facility handles now. Across Britain, warehouses, workshops, and distribution sites are being retrofitted to receive, store, and service e-bikes in volume, none of them originally designed with lithium batteries in mind. The hazardous-atmosphere profile has changed. In most cases, the hazardous area classification on file has not.

A Hazardous Area Classification, the assessment under the Dangerous Substances and Explosive Atmospheres Regulations that decides where a flammable or explosive atmosphere can form, produces a drawing of your site marked up in zones. Zone 20, 21 and 22 for dust. Gas and vapour have their own zone classifications. This is the hotspot map of the plant.

Your zoning drawing can pass every audit and still be wrong. A hazardous-area classification is only as good as the material data underneath it. If the dust figures behind it, the Kst, the Minimum Ignition Energy, the Minimum Ignition Temperature, were borrowed from a handbook rather than measured from the powder in your silo, then your map is only a guess.

Two failures are possible. In the first, you assume your powder is non-combustible and it is not, so there is no zone on the drawing where one should exist. In the second, the severity is higher than the handbook assumed, so the explosion protection you have specified falls short of what the actual material demands. In either case the drawing looks complete and is wrong. The zone boundaries and the protection they demand are only correct if the data behind them is real.

The fix is not complicated, but it must be real. Test the actual material you handle, in an accredited lab. Accredited data, the kind with ISO 17025 behind it, is a different category of evidence from a figure that was close enough.

That difference shows up where it counts. An HSE inspector follows it first. An insurer increasingly requires it at renewal. Without it, cover becomes harder to obtain and more expensive to hold. And if the worst ever happens, it is the line between a decision you can defend and one you cannot. There is no such thing as a close-enough powder when the one in your silo is the one that ignites.

This is the work we do. We test the materials in your process, in an ISO 17025 accredited laboratory, and replace the assumption with data that will hold up under scrutiny. Whether it is a Kst and minimum ignition energy figure that needs to inform a zoning drawing, or DSC and ARC data that tells you how a cell behaves under thermal stress, the number on its own does not tell you what to do next. We understand both the testing and the process safety decisions that follow from it, and we close the gap between a result and an action. We do not hand over a number and leave. We work through what it means for your risk picture, what it changes in your assessment, and what you need to show an inspector or insurer to demonstrate the decision was made on real evidence.

• When was your DSEAR zoning last revisited against a process or material that has changed since the drawing was made?
• If a new line, powder or battery process arrived this year, is it on the map yet?
• Could the number behind your most important safety decision survive an inspector asking where it came from?

If the honest answer to any of those is “I am not sure,” that is usually a capacity problem, not a competence one. The map is large and the team is small. Knowing which hotspot to check first is where the work begins.