Bridges breathe. Daily and seasonal temperature swings expand and contract the deck, and the bearings and expansion joints must accommodate every millimetre of that movement. Get the thermal actions wrong and you either oversize the bearings - wasteful - or, far worse, under-provide for the movement and lock the deck, forcing it to crack or distort. EN 1991-1-5 sets out how to quantify the effect; the interesting question is how conservative its ranges really are.
The two components of thermal action
EC1-1-5 splits the deck's thermal behaviour into two parts, because they do different things:
- The uniform temperature component is the overall warming or cooling of the whole deck relative to its construction temperature. It drives the expansion and contraction - the total length change the bearings and joints have to absorb.
- The temperature difference component is the gradient through the depth of the deck - top hotter than bottom on a sunny day, or the reverse at night. A restrained gradient induces curvature and internal stress, not just movement.
The two are combined in specified proportions, because they rarely peak together.
Why the gradient is the dangerous one
The uniform component is intuitive and mostly about providing enough movement capacity. The difference component is subtler and easier to get wrong:
Get the gradient wrong and you under-size the bearings - or worse, lock the deck and crack it. A restrained temperature difference has nowhere to go but into stress.
A deck that is free to curve simply bows slightly; a deck that is restrained against that curvature builds up internal stresses that the static load cases never showed. This is why the gradient deserves as much attention as the headline expansion length.
How the code ranges compare to reality
Instrumenting a deck over two summers gives a useful reality check against the EC1 profile:
- The measured uniform component stayed within the EC1 envelope across the monitoring period - the code's range for overall temperature comfortably bracketed what the deck actually experienced.
- The peak gradient briefly exceeded the code profile on clear, still afternoons - the specific conditions (strong sun, little wind to mix the air) that drive the sharpest top-to-bottom difference.
The lesson is not that the code is unsafe, but that it is a careful average, not an absolute ceiling. The design profile is calibrated to be reliable across the population of bridges and conditions, and brief, local excedances of the gradient under extreme-but-rare conditions are part of how a probabilistic code is meant to work.
What to carry into design
- Provide generously for the uniform movement - it is cheap insurance, and getting the bearing and joint travel right avoids a locked deck.
- Respect the gradient, especially where the deck is restrained - integral and continuous decks turn a temperature difference into real stress.
- Read the code profile as a calibrated design value, not a measured maximum; the comfort comes from its reliability across many cases, not from never being exceeded in any single afternoon.
The EC1 thermal-action provisions sit in the standards reference. The related load-combinations article covers how thermal actions combine with the other variable actions on the structure.