Maintenance budgets get decided by a percentage nobody measures. Rust extent, adhesion and access cost answer the question better.
Every maintenance survey ends with the same argument. Operations wants spot repair because it is cheaper this quarter; the coating engineer suspects a full recoat is inevitable and wants to say so now. Both sides are usually arguing from the size of the invoice rather than from the condition of the steel.
Start with a measured extent, not an impression
The first number that decides the strategy is the proportion of the surface that has broken down, and it is measurable. Rusting is assessed on the ISO 4628-3 Ri scale, from Ri0 for no visible rust to Ri5 for extensive breakdown. Blistering, cracking, flaking and chalking each have their own scale in the same series. A survey that records Ri numbers by zone gives you something to compare against next year, and something to put in front of a budget holder.
Rough thresholds hold up in practice. Below roughly ten percent breakdown, with good adhesion on the sound area, spot repair is defensible. Between ten and twenty-five percent, the decision depends on access and on how fast the breakdown has progressed since the last survey. Above roughly twenty-five to thirty percent, a full recoat almost always wins on whole-life cost, because the labour to prepare and coat thirty percent in isolated patches approaches the labour for the whole area, while the remaining seventy percent still ages underneath.
Then check adhesion, not just appearance
A surface that looks intact can be a shell. Where the existing film has lost adhesion to the substrate, the repair patches will fail at their feathered edges within a season or two, and each repair generates another ring of edge to feather next time. A simple cross-cut or pull-off test on a representative area, plus a tap test around the boundary of sound and failed coating, tells you whether the sound area is genuinely sound.
Film thickness matters as well. Where the original system was applied at half the specified thickness, the sound-looking area has less reserve than its appearance suggests.
The compatibility question comes before the schedule
Overcoating an unknown existing system is the highest-risk part of maintenance work, and it is where most premature failures originate. The relevant unknowns are the generic type of the existing coating, its age and degree of cure, and whether it has been contaminated by chalk, oil, salt or previous silicone contamination.
A test patch answers all of these at once. Prepare and coat about one square metre in the least visible representative location, allow it to cure fully, then check intercoat adhesion by knife or pull-off. A patch that can be lifted with a blade has failed, and the answer is not more surface preparation in the same place but a different coating or a different preparation method.
Preparation is where the money is spent
Spot repair does not mean light preparation. Each repair area must be taken back to a sound substrate, the edges of the surrounding intact coating feathered over a sufficient distance so no step remains, and the feathered zone cleaned of all chalk and contamination. On aged exterior coatings, a light water jet at low pressure followed by a rinse is often the only way to remove the chalk layer without destroying the remaining film.
Where a mist coat is specified for the first pass over a porous or chalked surface, use it. A full wet coat over an absorbent substrate traps air and produces pinholes that look like application defects and are actually preparation defects.
The schedule that survives a decade
- Record Ri grade, adhesion result and dry film thickness by zone at every survey
- Cost the options over the full repaint cycle, not over one financial year
- Price the access, because scaffolding or rope access often dominates the total
- Include the cost of the next survey in both options; spot repair needs more frequent inspection
- Decide the strategy per zone, not per structure; a tank farm rarely has uniform condition
The decision that ages worst is the one taken to fit a budget year. Steel does not reschedule, and the invoice deferred this quarter arrives with interest at the next shutdown.