A bridge expands and contracts by an amount you can calculate from its length, its material and the temperature, and the bearings let it. Bearing 3 moves four millimetres more than the arithmetic allows, and only when the air is below zero.
I am a bridge inspection engineer for a county council and I have a hundred and forty structures.
The arithmetic first, because it is the whole argument.
Steel expands by about twelve millionths of its length for every degree. Concrete is about ten. So a deck of sixty metres, over a forty degree swing from a January night to an August afternoon, moves about thirty millimetres, and that movement has to go somewhere.
It goes into the bearings at each end and into the expansion joints. A bearing is a sliding or rolling assembly that lets the deck move horizontally while carrying it vertically, and you can measure where it is sitting with a ruler, or better with a dial gauge on a bracket, which reads to a hundredth of a millimetre.
This bridge is a four span steel composite structure carrying a B road over a river. Eight bearings, two per pier line, numbered one to eight from the north abutment.
We have had dial gauges and temperature loggers on all eight since 2016, because the deck was strengthened and the monitoring was a condition of the design.
Seven of the eight do what the arithmetic says, within a few tenths.
Bearing 3 matches them down to about plus two degrees.
Below zero, it goes four millimetres further than it should.
Not a drift. It is there on a cold night and it is back in line by the following afternoon, and the gauge traces it going out and coming back.
I have two hundred and eleven nights of it.
What four millimetres means on a bearing of that type.
It is not dangerous and I am not going to pretend otherwise. That bearing has thirty millimetres of travel each way and it is using thirty four on a cold night in one direction, which is within its range.
It is impossible rather than unsafe. The deck is one deck. The other bearings on that pier line, two metres away, carrying the same steel, move the amount the steel moved.
So one corner of a rigid deck went four millimetres further than the rest of it, and came back.
What has been checked, and there is a lot of it because a council will pay for monitoring and will not pay for a new bearing.
The gauge. Swapped, recalibrated, and mounted on an independent bracket off the pier rather than off the deck, which is the correct way and which rules out the bracket moving.
The logger. Two temperature sensors, one in the air and one drilled into the steel, because the steel temperature is what drives the movement and on a clear night the steel goes colder than the air.
That one mattered, and it is the only honest explanation anybody produced. If that corner of the deck were getting colder than the rest, it would contract more. So I put four sensors on the deck steel, one per corner.
On a frosty night all four read within nought point three of a degree of each other.
Nought point three of a degree on sixty metres of steel is two tenths of a millimetre. Not four.
What is under it.
Pier 2 is in the flood plain on the north bank, on piles, and the pier that carries bearings 3 and 4 is the one nearest the old mill leat.
I have had the pier surveyed for settlement with precise levelling, twice, once in frost. It does not settle, and a settling pier would take both bearings with it anyway.
The river is nine metres away. The scour survey is clean.
I have had a drone over it, a diver under it, and an access platform on it for a week.
What happens in the paperwork.
It is a recorded anomaly on the structure's record, with the data attached, reviewed at every principal inspection, which is every six years.
The 2022 principal inspection report, written by a consultant who is better at this than I am, says the movement at bearing 3 exceeds the theoretical value under low temperature conditions, that the cause has not been established, that the bearing remains within its design travel, and recommends continued monitoring.
That is a correct professional judgement and I signed the covering letter.
What I do, and the one thing I stopped doing.
I read the gauges in person twice a year, which is more than the monitoring requires, because the logger gives you a number and a gauge gives you a number and a face.
The cold readings I used to take at night, because that is when the movement is there, and you cannot see a four millimetre difference at two in the afternoon.
I have taken that reading eleven times in the dark, on my own, with a head torch, standing on the maintenance gantry under the deck at pier 2.
The last time was February 2023, and the reading was four point one millimetres out, and the gantry at pier 2 is an open steel walkway about two feet above the water with a handrail on one side.
What I recorded in my notebook, and what I have never put on a form, is that the deck over my head ticked and moved while I was writing, which steel does all the time when it is cooling and which I have heard a thousand times, and then the dial on the gauge in front of me went round another half millimetre in about a second, which is not cooling, because nothing cools in a second, and I was the only warm thing on that pier.