Setting the Next Interval on a Line With Only One Prior Survey
An 8 in. Sch 40 A106-B line (nominal wall 0.322 in.) carries diesel from a process unit to a tank farm. Minimum required thickness calculated to B31.3, including the corrosion allowance, is 0.220 in. The line has been in service since 1998, but the file contains only one prior survey, taken five years ago, when CML-3 (a 3 in. window at the 6 o'clock position on the straight run, 2 ft downstream of a reducer) read 0.298 in. At the current survey, CML-3 reads 0.248 in. on the same instrument with the same technique. CML-7, located at a pipe support saddle where condensate drips, reads 0.215 in. There are no injection points and no internal lining. The line is currently classified Class 2.
1. Compute the corrosion rate and remaining life at CML-3 and state the latest date you may schedule the next thickness survey. Show the arithmetic and state which rate you used and why.
Read the worked answer
Corrosion rate. (0.298 - 0.248) in. / 5 yr = 0.0100 in./yr. This is a short-term rate only; no long-term (installation-to-now) rate exists because there is no baseline thickness for this CML.
Remaining life. (0.248 - 0.220) / 0.0100 = 2.8 years.
Interval. The next survey is scheduled at not more than half the remaining life: 1.4 years, about 16 months. The class maximum for a Class 2 line is longer than this, so the half-life figure governs. Do not round up to the class maximum, and do not blend CML-3 with better readings elsewhere on the line to soften the answer.
Judgment points.
- Two readings five years apart is a weak basis for a rate. You schedule on it anyway, but you label it provisional and take a third reading early enough to confirm or correct it before the interval is used again.
- The 16-month date is when you need the next data point, not when the pipe is projected to reach 0.220 in. Scheduling to the retirement date itself consumes the entire safety margin.
- If the process changes (rate, temperature, water carryover), the calculation is invalid and must be redone.
2. CML-7 reads 0.215 in. against a t_min of 0.220 in. What are your actions, what options does the owner-user have, and who has to approve what?
Read the worked answer
The reading is below t_min, so the line is not qualified for its present rating and the corrosion allowance is effectively consumed at that location.
- Verify before acting. Re-scan the location, check whether 0.215 in. is a pit or broad loss, and inspect the support contact area. Loss at a water-drip saddle is almost certainly external corrosion, a different mechanism from the internal loss driving CML-3, which means the rate at this point is its own number, not 0.0100 in./yr.
- Treat it as a repair/rerate decision, not an inspector judgment call. The inspector cannot simply accept it and keep watching, and cannot approve continued operation at the existing rating by note.
- Options. (a) Replace the affected span or install a full encirclement sleeve. Weld build-up is not the normal method for restoring pressure-containing thickness on a corroded line. (b) Rerate: an engineer calculates the MAWP at the measured 0.215 in. to B31.3 (with the support loads considered), and if acceptable the operating envelope is reduced, a written rerating record is issued, and the inspection plan is revised. Lowering the pressure without the calculation and the record is not a rerating.
- Look wider. The same water path probably affects the other saddles on this rack. Add TMLs at every support on the line and inspect the contact areas, and recompute the rate at this location once a second reading exists.
3. The unit superintendent argues that because only one prior survey exists, the 0.0100 in./yr figure is probably pessimistic and the interval could be stretched. How do you respond, and what do you change in the inspection plan?
Read the worked answer
The argument cuts both ways and does not support stretching the interval.
- One prior survey is not a history. The five-year window could understate a recent acceleration (an upset, a change in product, water carryover) just as easily as it overstates a slow long-term loss. You cannot know which without more data, and the conservative reading of an unknown rate is the measured one.
- Coverage is the real weakness. A single CML on a circuit is not a survey. Add TMLs covering the corrosion circuit: the reducer downstream, elbows, low points, support contact points, and the straight run upstream for comparison. Each with a documented grid, so the next survey is comparable.
- Confirm, then relax. Take a confirmatory reading in roughly 6-12 months. If it supports a lower rate, the interval can be extended, but never beyond the class maximum, and the extension and its basis must be documented and approved.
- Do not trade data for schedule. Stretching the interval to the class maximum now would leave the line with a 2.8-year margin and no intermediate check, which is exactly the exposure the half-life rule exists to prevent.