Module 18 · Lesson 18.2

Durability and erection: the two limit states nobody analyses

One is reached slowly over decades, the other in an afternoon halfway through the programme. Neither appears in the frame analysis, and both can govern.

Why this matters

A structure has to survive two conditions that no static analysis describes.

The first is time. Steel in the open corrodes, and the design life is a number somebody wrote in a brief. Whether the structure reaches it depends on the environment, the coating, and — decisively — on details that no calculation sees.

The second is the afternoon halfway through the programme when half the steel is standing and most of the bracing is not. That configuration is a structure. It carries load. It is frequently the least stable arrangement the steel ever finds itself in, and it does not appear anywhere in the design.

By the end of this lesson you should be able to

Treat corrosion as a design-life problem rather than a paint specification
Say why access decides whether a short coating life is acceptable
Name the details that defeat a coating whatever is specified
Quantify how much less stable a partly erected frame is

What you should already know

Frame stability and α_cr (Modules 4 and 16)
Bracing systems and their relative efficiency (Module 16)

Corrosion is arithmetic before it is chemistry

Three numbers decide it.

The environment sets a rate. Corrosivity categories run from C1 — a heated interior, where unprotected steel loses under a micrometre a year — to C5, marine or aggressive industrial, at well over a hundred. That is a range of more than two hundred to one, and it is set by where the building is, not by anything the designer controls.

The coating buys years. A system's life to first maintenance rises with its thickness and falls with the corrosivity.

The brief states a design life. Fifty or sixty years, typically, for a building.

The design question is whether the second covers the third in the environment set by the first — and if not, what happens instead.

Rule classification: engineering recommendation. The rates and coating lives here are indicative figures used to teach the comparison. A real specification is written against a manufacturer's system data and the relevant protective-coating standard, not against these.

Worked example

A 50-year building in three different places

Given

The same structure, with the same 200 µm coating system
A 50-year design life in every case
Only the site changes: C2 rural, C3 urban, C5 marine

Find

Whether the coating lasts the design life at each site, and what follows if it does not

The details that defeat any specification

A coating fails first where the drawing let it down. Five recurring cases:

Water traps. An upward-facing horizontal surface holds water and dirt against the coating indefinitely. The local corrosivity at that spot is effectively a category worse than the rest of the structure, and it is where the failure starts.

Unsealed crevices. A gap between plates draws water in by capillary action and will not release it. The coating never reached inside, and the corrosion proceeds unseen.

Bimetallic contact. Two different metals with an electrolyte between them form a cell, and the less noble one corrodes preferentially. Stainless fixings into carbon steel is the common case; the remedy is an isolating washer.

Sharp edges. Coating pulls back from a sharp edge by surface tension as it cures, so the film there is a fraction of its nominal thickness. Rounding edges is a fabrication instruction with a durability purpose.

Unsealed hollow sections. An open hollow section breathes moist air, condenses it, and corrodes from the inside where nobody will ever look.

Every one of these is a drawing correction, not a specification correction. Writing a heavier coating system on a detail that traps water does not help.

Try it

Does the coating last as long as the building

The environment sets a corrosion rate, the coating buys a number of years, and the design life is what those have to satisfy. Where they do not meet, the answer is rarely more paint.

Corrosivity category

Design life50 years

Coating thickness200 um

Reachable for maintenance

Upward-facing surfaces that hold water

Hollow sections sealed

Corrosion rate, unprotected: 25 um/year
Loss over the design life: 1.25 mm per face
Coating thickness: 200 um
Life to first maintenance: 25 years
Required design life: 50 years
Lasts the design life: NO
Detailing issues: 0

The coating gives 25 years against a 50-year life, so the structure needs maintenance 2 times during it. That is acceptable ONLY if the member can actually be reached — which is a design decision about access, not a specification decision about paint.

Things worth trying

Start at the defaults — C3 urban, 50-year life, 200 um. The coating gives about 25 years: it is HALF the design life, so the building needs repainting once.
That is an acceptable answer only if somebody can reach the steel. Switch 'reachable' to no and read the verdict — the recommendation changes completely while the numbers do not move at all.
Push the thickness up to 500 um and watch how little it buys. Coating life is roughly linear in thickness, so doubling the paint doubles the years — it does not solve a factor-of-four shortfall.
Now change the environment to C4 and leave everything else. The SAME 200 um system falls from 25 years to about 15. The paint did not change; the site did.
Set the thickness to zero in C5 with a 50-year life. Unprotected steel loses 6 mm per exposed face — on a 10 mm plate exposed both sides that is the whole plate.
Turn on the water traps. The number of years does not change and the verdict does — because a detail that holds water defeats the specification locally, and the structure corrodes there rather than where the specification was written for.
Open the hollow section ends. Same effect: an unsealed hollow section takes in moist air, condenses it, and corrodes from the inside where nobody will ever look.
The lesson is the order of the decisions. Environment first, then access, then detail, then paint. Reversing that order produces a specification that cannot work.

The structure that nobody analysed

Here is a frame that reaches its design intent: four braced bays, α_cr of 12, comfortably non-sway by the criterion of Module 4. Everything checks.

Now consider the afternoon when 80% of the steel is standing and one of the four braced bays has been completed.

The lateral stiffness is a quarter of the finished value. The vertical load is 80% of it — so the critical load factor falls on both counts at once.

Critical load factor at an erection stage

α_{cr, s t a g e} \approx α_{cr, co m pl e t e} \cdot \frac{n _{in s t a l l e d}}{n _{co m pl e t e}} \cdot \frac{1}{f}

What it calculates: The critical load factor of a partly erected frame, as a first estimate

$α_{cr, s t a g e}$: critical load factor at this stage of erection
$α_{cr, co m pl e t e}$: critical load factor of the completed frame
$n_{in s t a l l e d}$: braced bays installed so far, including any temporary bracing
$n_{co m pl e t e}$: braced bays in the completed frame
f: fraction of the frame's steel already standing

This assumes

Lateral stiffness is roughly proportional to the number of braced bays
The bracing installed is distributed similarly to the completed system
Vertical load scales with the steel erected

In plain terms: For 1 of 4 bays with 80% of the steel standing this gives about 3.75, against 12 for the structure everyone checked — the stage is 3.2 times less stable than the finished frame, and it is nowhere in the design.

Rule classification: derived. This is proportionality, not a code rule: α_cr scales with lateral stiffness and inversely with the vertical load that stiffness has to stabilise. It is a first estimate for reasoning about sequence, not a substitute for analysing the stage properly once it has been identified as critical.

Try it

A frame is least stable while it is being built

The completed frame has all its bracing. The half-erected one has whatever has been installed so far — and the steel is already standing. Watch how far below the finished structure the intermediate stages sit.

Permanent braced bays installed1 of 4

Steel erected80 %

Temporary bracing0 equivalent bays

Completed frame: 12.0
This stage: 3.75
Without temporary bracing: 3.75
Fraction of the completed value: 31 %
Worse than the finished structure: yes
Adequate at this stage: NO

With 1 of 4 braced bays installed and 80% of the steel standing, α_cr is about 3.8 against the completed frame's 12.0 — 3.2 times worse. There is no temporary bracing in this assessment. NOT adequate. The erection sequence is a structural design and this stage needs temporary bracing designed for it — the governing case for that bracing is frequently more onerous than anything the completed structure sees.

Things worth trying

Start at the defaults — 1 of 4 bays braced, 80% of the steel standing, no temporary bracing. Alpha-cr is 3.75 against the finished frame's 12: this stage is 3.2 times worse than the structure anybody checked.
Note that the finished frame at alpha-cr = 12 is comfortably non-sway. This intermediate stage is not, and nobody analysed it.
Set the bracing to 2 of 4 and the steel to 100%. Half the stiffness with all the load gives exactly half the alpha-cr — 6. Still short of 10.
Now add 2 equivalent bays of temporary bracing to the default stage. Alpha-cr goes from 3.75 to 11.25 and the stage becomes adequate. The readout keeps reporting what it would have been without.
That is the whole point of the lab: the temporary bracing is not a convenience, it is the load path. Somebody has to design it, and it belongs to the erection sequence.
Walk the 'steel erected' slider from 10% to 100% with the bracing at 1 bay. The curve falls steadily — the load goes up faster than the restraint does.
Look at the shape of the curve. It is not a straight line from zero to the finished value, and there is no stage at which you can interpolate the answer. Each stage is its own structure.
Set the bracing to 4 of 4 and the steel to 100%. You recover exactly the completed frame — which is the only configuration anybody usually checks.

Practice

A frame will finish with 4 braced bays and α_cr = 12. At one stage, 1 bay is braced and 80% of the steel is standing, and the erection engineer proposes temporary bracing equivalent to 2 permanent bays. What is α_cr at that stage with the temporary bracing in place?

Check yourself

The same 200 µm coating system gives 66 years in C2 and under 10 years in C5. What follows for a coastal building with a 50-year design life?

Predict first

A designer specifies a heavy coating system, and the detailer draws a bolted splice with an unsealed gap between the cover plates. Where does the structure corrode first?

Summary

Corrosion is three numbers: an environmental rate, a coating life, and a required design life
The same 200 µm system gives 66 years in C2, 25 in C3, 9.8 in C5
A short coating life is acceptable only if the steel can be reached — which is a building decision
Water traps, crevices, bimetallic contact, sharp edges and open hollow sections defeat any specification
Those are drawing corrections, not specification corrections
A partly erected frame with 1 of 4 braced bays and 80% of the steel up sits at α_cr ≈ 3.75 against 12
Every erection stage is a structure, and its temporary bracing is a load path somebody must design

Progress is kept in this browser only.

This is educational material. It uses simplified examples to teach principles, and must not be relied on for real design or safety-critical decisions. Module overview and checkpoint