Watermain as-built conflict with a gas main crossing
A contractor is laying a 2.4 km DN300 watermain in a road reserve. The design specifies a minimum cover of 0.9 m and a minimum separation of 0.40 m from an existing 100 mm gas main that crosses the route at chainage 1+240. Setout used a total station from two control marks on a traverse that closed 1:12,000. An as-built survey, done on a later day by GNSS RTK, reports the watermain crown at 0.75 m depth at 1+240 with a 0.09 m lateral offset from design; everywhere else the as-built agrees with design within 0.02 m. The contractor's QA states 'within plus or minus 50 mm horizontal and 25 mm vertical tolerance, therefore compliant'. The gas main's recorded depth at the crossing is 1.15 m.
1. Is the contractor's compliance argument sound? Work through the geometry and decide whether the crossing complies.
Read the worked answer
No. The tolerance and the minimum cover are different tests, and passing one does not satisfy the other.
Cover. The as-built crown is at 0.75 m against a 0.9 m minimum. That is a 150 mm deficiency - six times the plus or minus 25 mm vertical tolerance. The tolerance describes how accurately the pipe was meant to be set out around the design; it cannot be used to justify a result that breaches the minimum cover. Cover fails.
Separation. Using nominal DN300 (0.30 m outside diameter):
- Watermain crown 0.75 m depth, so centre 0.75 + 0.15 = 0.90 m, invert 1.05 m.
- Gas main recorded at 1.15 m depth. Even taking that at face value as a crown depth, the clear vertical gap to the watermain invert is 1.15 - 1.05 = 0.10 m, against a 0.40 m requirement.
So the vertical separation fails by roughly 0.30 m before the lateral offset is even considered. The 0.09 m lateral offset is itself inside the plus or minus 50 mm tolerance, but a crossing separation is a three-dimensional clearance: the lateral movement changes the crossing geometry and reduces the effective separation further.
Critical caveat: a recorded depth for the gas main is not a surveyed depth. Before any decision, the gas main must be potholed and surveyed at the crossing. A 0.10 m apparent clearance is also inside the danger zone for future maintenance excavation, so the asset owners will treat this as a hit risk, not a paperwork issue.
2. The discrepancy appears at one chainage only and does not build progressively along the line. What does that tell you about the source of error, and what records would you call for?
Read the worked answer
The error signature is diagnostic.
- Progressive error (growing with distance from control) points to the control network, a scale factor, or a datum/transformation problem. At 1:12,000 over 2.4 km the traverse misclose is only about 0.2 m total, and it would not be localised.
- Localised error at a single chainage points to something specific to that location or that day: a misread staff or tape, a wrong invert recorded on the as-laid sheet, a displaced or disturbed peg, a laser/level misclosure within one section, a hard spot under the pipe, or a processing/attribute error in the as-built data.
Do not overlook the as-built method itself. It was RTK, on a different day, at a location with a 0.09 m lateral disagreement elsewhere. RTK under street trees and buildings is exactly where multipath and poor geometry produce 50-100 mm scatter, so part of the 'discrepancy' may be measurement noise in the check rather than a construction error.
Records to call for: setout records and as-laid sheets for the section around 1+240; level books and the control coordinates and adjustment report; the design invert and grade for that chainage; the pipe supplier's actual outside diameter (DN300 nominal is not the manufactured OD); the potholed and surveyed position and depth of the gas main; and the RTK as-built raw observations showing fix status, satellite count, PDOP and residuals at that point.
Then re-observe the crossing with the total station before concluding anything.
3. Assume the cover and separation deficiencies are confirmed. What is the practical outcome and how do you document it?
Read the worked answer
The pipe cannot be accepted as built in that location. Options, in order of preference:
- Re-excavate and relay the affected length to restore 0.9 m cover and 0.40 m separation - the only option that restores the design intent.
- Relay to a lower grade locally with a redesign of the vertical alignment, if the surrounding cover permits.
- Protection - a reinforced concrete slab over the watermain and/or the gas main, with the asset owners' written approval. This manages the risk of third-party excavation but does not restore separation, so it is a concession, not a cure.
Documentation: an as-built deviation report to the water network owner and the gas network owner, showing the surveyed positions, the pothole results, the agreed remedy and the approvals. The as-built plan and asset data must record the actual depths and offsets, not the design values. If a concession is accepted, it belongs on the property/asset file so the next person who excavates there knows.
Note also the process failure: the crossing should have been potholed and surveyed before the watermain was laid past it, and positive separation confirmed then.