LNG Processing Train
Commissioning interface between EPC and operations slipped by 22 weeks. Root cause: commissioning resource plan never validated against the EPC handover schedule.
- Value
- $1850M
- Duration
- 60 months
- Outcome
- Late & over
- Commissioning resource gap visible at month 36
- EPC forecast and operations forecast diverged
- Validate the commissioning resource plan against EPC handover dates
- Two parallel forecasts mean one of them is fiction
- Reserve float for commissioning, not just construction
Project context
A 4.5 million-tonne-per-annum LNG processing train added to an existing brownfield LNG facility on the north-west Australian coast. Scope included a new gas treatment unit, a new cryogenic heat exchanger train, a new condensate stabilisation unit, two new LNG storage tanks and a tie-in to the existing jetty and loading arms. Contract was a lump-sum EPC with a Tier-1 international contractor, with separate contracts for the operator's commissioning team and the long-lead equipment (the cold box, the main compressors and the storage tanks). Total installed cost AUD 2.65 billion (approximately USD 1.85 billion). Programme at award was 60 months from notice to proceed to first LNG production. The single biggest source of risk on a brownfield LNG addition is not the EPC scope itself, which is well-understood, but the interface between construction completion and operations commissioning.
Two forecasts that should have been one
From month 18 onwards the project was running two parallel schedules: the EPC contractor's schedule, which forecast mechanical completion at month 48, and the operator's commissioning schedule, which assumed full handover at month 46. The two-month gap was visible in both documents but neither party owned the reconciliation. The EPC schedule was driven by the engineering, procurement and construction sequence and treated commissioning as a downstream activity. The operator's schedule was driven by the introduction-of-hazards plan and the safety case re-approval and treated mechanical completion as an upstream assumption. Nobody was explicitly asked the question 'which of these two dates is right' until month 36, by which time both had drifted further apart, with the EPC at month 52 and the operator at month 48. The lesson is that whenever a project carries two parallel forecasts of the same milestone, at least one of them is fiction. The remediation is governance: a single integrated handover schedule, owned jointly, reviewed monthly.
The commissioning resource gap visible at month 36
By month 36 the operator's commissioning team was 38 people. The integrated handover plan, when it was finally produced in month 38, required 96 people from month 44 onwards to meet the original mechanical-completion-to-first-LNG window of six months. The 58-person gap could not be closed inside 12 months in the regional labour market — LNG commissioning engineers with brownfield experience are a thin pool. The operator started recruiting in month 38 and reached 72 people by month 48, which was insufficient. The result was a 22-week slip in the commissioning sequence, which translated directly to a 22-week slip in first LNG. The lesson is that the commissioning resource plan must be validated against the EPC handover plan at every quarterly review from month 12 onwards, not from month 36 onwards. A 12-month lead time on commissioning recruitment is the minimum for this type of asset.
Cost overrun and the cascade
The 22-week slip in first LNG cost the project approximately AUD 340 million in standing costs, demobilisation and remobilisation of construction trades, extended preservation of installed equipment, and lost margin in the first cargo window. The EPC contractor's lump-sum contract did not pass these costs to the contractor because the slip was on the operator's side of the interface, not the contractor's. The operator's contract with the EPC contractor included a contractual mechanical completion date that was met within five days, which meant there was no contractual lever to recover the cost. The lesson is that on EPC-plus-commissioning structures, the commercial protection on the construction side is often water-tight while the commercial protection on the commissioning side is paper-thin. The project economics live or die on the commissioning interface, and the contracts rarely reflect that reality.
Float reserved for commissioning
The original schedule reserved 14 weeks of float between mechanical completion and first LNG. On a brownfield LNG project of this size, that is approximately one-third of what is needed. The lesson learned, written up by the operator's project assurance team after first LNG, was that the float reserve for commissioning should have been at least 24 weeks, and that the reserve should have been protected as a separate management reserve owned by the operator rather than absorbed into the EPC contractor's contingency. On the next train at the same facility, the float reserve was set at 26 weeks and the commissioning team was mobilised 18 months ahead of mechanical completion. That train hit first LNG within three weeks of the planned date.
What this project says about EPC-plus-commissioning structures
On large process facilities, the most likely failure mode is the interface between the EPC contractor and the operator's commissioning team. The contracts are typically well-drafted on the construction side and weak on the commissioning side. The schedules are typically rigorous on the construction side and aspirational on the commissioning side. The result is consistently that mechanical completion is achieved close to the contractual date and first product is achieved months later. The fix is structural: a single integrated handover schedule owned jointly, commissioning resourcing planned 18 months ahead, and explicit float reserved for the commissioning sequence as a protected management reserve. None of these are technically difficult. They are governance choices that have to be made in the first 12 months of the project and protected through every cost pressure that follows.