Corrosion rate and remaining life

A corrosion rate is a slope through thickness readings at the same condition monitoring location. API practice asks for two of them. The long-term rate runs from the original or baseline thickness to the current reading and describes the average behaviour of the equipment over its life. The short-term rate runs from the previous inspection to the current one and describes what is happening now. They answer different questions: a long-term rate smooths out process upsets, a short-term rate catches a change in service, a new corrosive stream or a failed inhibitor programme.

Where the two disagree, the inspector decides which governs, and the codes require that decision to be recorded with its justification. The default position is to take the higher of the two, because it is the conservative one, but a short-term rate inflated by measurement scatter on a single grid point is not evidence of accelerated corrosion — repeat the readings before acting on it.

Remaining life is then (t_actual − t_required) / corrosion rate. The quality of that number depends almost entirely on t_required, which is not the nominal wall and not the wall minus the corrosion allowance: it is the thickness the pressure design actually needs, from ASME VIII UG-27 or B31.3 as appropriate, plus any structural minimum.

Thickness measurement repeatability sets the floor on what can be trusted. A pair of readings each good to ±0.1 mm, taken five years apart on a line losing 0.05 mm/yr, cannot resolve the rate at all. Where the loss between inspections is of the same order as the measurement uncertainty, extend the interval or move to a grid and treat the statistics, not the single reading, as the measurement.

long-term rate  = (t_initial − t_actual) / years since initial
short-term rate = (t_previous − t_actual) / years since previous
remaining life  = (t_actual − t_required) / governing rate
1 mm/yr = 39.3701 mils per year