Firewater main pinholes: MIC forensics and programme response
A terminal firewater ring main is 12-inch carbon steel, 30 years old, kept full and pressurised at 7 bar with treated municipal water. It has flowed only twice in 18 months. Pinhole leaks have appeared at two dead legs and at a 6 o'clock low point of a straight run. Deposits are black and smell of sulfide; a pit beneath one deposit is 80% through wall with a bright metal rim. Water analysis gives 0.1 mg/L free chlorine, 22 mg/L sulfate, 4 mg/L nitrate, 18 °C. A corrosion coupon in the flowing header shows 0.05 mm/y uniform. The maintenance manager proposes continuous biocide dosing and no other change.
1. Set out the forensic sequence and explain how the evidence distinguishes microbiologically influenced corrosion from oxygen corrosion and under-deposit attack without bacteria.
Read the worked answer
Sequence
- Cut out the failed section rather than patching it. Photograph and record the pit morphology before cleaning.
- Sample deposit into anaerobic containers on site; do not let it dry out.
- Analyse deposit chemistry: acid-volatile sulfide, total iron, sulfate, chloride, organic carbon, and carbonate. FeS at high levels with a sulfur smell is the strong marker.
- Microbiological analysis: serial dilution culture for sulfate-reducing and acid-producing bacteria, ATP, and ideally 16S sequencing, since culture recovers a small fraction of the community.
- Assess the rest of the system by ultrasonic B-scan at 6 o'clock and at dead legs before deciding scope.
How the evidence separates the mechanisms
Oxygen corrosion in a full, low-flow main gives general thinning with red-brown oxide. Here the attack is pinhole, located exactly where water stagnates, and under black sulfide-bearing deposit with a bright metal rim - the pit edge that sulfate-reducing bacteria produce as their local chemistry differs from the bulk. The 0.05 mm/y coupon in a flowing header is consistent: general corrosion is low; the damage is under-deposit and localised.
A sterile under-deposit cell would still give black oxide and pitting, so the deposit chemistry and bacteriology are what confirm the biological contribution. Both mechanisms may be present; the location is the giveaway.
2. The system has dozens of dead legs and low points and the budget only covers targeted work. Propose a risk-ranked programme with defined acceptance criteria.
Read the worked answer
Rank by stagnation and consequence. Score every dead leg and low point on length-to-diameter ratio, residence time, accessibility, drainability and consequence of failure (firewater availability at that node). Dead legs of more than about three diameters' length with no flow are the priority; a firewater main that is used twice a year is effectively a stagnant system throughout.
Investigate a sample, not everything. Ultrasonic B-scan at 6 o'clock on maybe 15–20 selected locations across the ring plus each critical dead leg. Sample deposit from two or three of the worst. The aim is to characterise the population, not to certify every metre.
Acceptance criteria
- Free chlorine maintained at the design residual at the furthest point, verified by field test - not assumed from dosing.
- Pit depth trending below a defined fraction of wall at each monitored location; any location reaching that limit gets removed and replaced.
Physical fixes before chemistry
- Remove or shorten dead legs; where they cannot go, add drains and flush them on a defined frequency.
- Exercise the ring on a schedule that is actually executed, below the 7 bar static condition, so the whole circumference is wetted.
- Replace the worst sections; a biocide cannot restore 80% wall loss.
Record the basis so the next round of monitoring has a baseline; without it the programme starts over every year.
3. Assess the maintenance manager's 'continuous biocide, no other change' proposal and state what you would tell the terminal management.
Read the worked answer
It will not work, and it will look like it is working.
- In a stagnant main the biocide diffuses slowly into the very deposits that shelter the bacteria. The biofilm and the corrosion product exert a chlorine demand that consumes the dose before it reaches the metal.
- Continuous dosing at low level is the wrong mode for a sessile community in a deposit. It selects for tolerance and gives a reassuring residual reading in the bulk water while the pits keep growing.
- The 80% through-wall pit is already a mechanical integrity issue. No chemical programme addresses it.
- 0.1 mg/L free chlorine is below what is needed to maintain a residual at the far end of a 30-year-old ring under any mode.
Better programme: a shock treatment with a dispersant to lift the deposit, followed by flushing to waste, then removal or regular flushing of dead legs, then a maintained residual with verification at the extremities. Check the discharge permit before planning the flush - the volume and quality may need approval.
What management needs to hear: firewater is a safety system whose failure is discovered only in an emergency. The risk is not the leak; it is the undetected pit in a leg that is never flow-tested. Give them the ranked list of locations, the cost of the fix, and the date by which it should be done - then get the decision recorded.