Quantitative Risk Analysis — New-Build Energy Facility

The challenge

SOMA Project Controls was engaged to deliver an independent Quantitative Risk Analysis — incorporating both QSRA and QCRA — on a £300m new-build energy facility, ahead of final investment decision and contract award.

The client required assurance that the contractor’s proposed 24-month construction programme and cost estimate were realistic, and that schedule and cost contingency were appropriately quantified — not based on assumptions or optimism, but on evidence.

The stakes at final investment decision are particular. Once the contract is signed, confidence levels cease to be a discussion topic and become a benchmark against which performance is measured. A P80 that was quietly optimistic at sanction will surface as variance in the first quarterly report, and the conversation thereafter becomes about recovery rather than about the risk profile that was accepted. Getting the QRA right at this stage was not about modelling exercise — it was about the realism of the baseline the client would live with for 24 months.

Our approach

We conducted a detailed review of the contractor’s baseline schedule and cost estimate, validating logic integrity, sequencing, the realism of activity durations, and the treatment of cost uncertainty. This ensured the model inputs were robust before quantitative analysis was applied.

Three-point estimates were developed in collaboration with the project team for both schedule activities and cost line items. Monte Carlo simulation modelled the full range of possible outcomes, establishing P50 and P80 forecast positions for time and cost.

Our QRA methodology follows AACE International recommended practice. Risks were captured in two distinct categories: inherent uncertainty in activity durations and cost line items (modelled as three-point distributions on the base estimate), and discrete risk events (modelled as probability-weighted impacts that only apply if the risk materialises). Keeping the two categories separate is what allows the model to answer two different questions — how uncertain is the base plan, and how much additional exposure comes from things that might go wrong.

Correlation was handled explicitly. Activities that share a resource pool, a weather window, or a common subcontractor do not move independently, and treating them as if they do systematically understates confidence levels. We applied correlation groups where the evidence supported it, and tested the sensitivity of the P80 position to the correlation assumption — a transparency check that ensures the headline number does not depend on a hidden judgement call.

What we found

The analysis revealed that the contractor’s baseline programme and estimate carried material risk that had not been quantified:

• Programme duration: contractor baseline 24 months vs QRA P80 position 26.2 months — a 2.2-month schedule contingency requirement that was not in the submission.
• Project cost: contractor baseline £300m vs QRA P80 position £324m — an additional £20m cost contingency beyond what was included.
• Prolongation exposure: the 2-month schedule extension exposed the client to £3–5m in prolongation costs — extended preliminaries, site management, and temporary facilities — that were not visible in the contractor’s submission.

What the numbers actually meant

A 2.2-month schedule extension and a £24m cost uplift are not forecasts of certain overrun — they are the level at which the client should be 80% confident that actual performance will be better, given the risk profile identified. The difference matters. It frames the client’s decision as one of risk appetite rather than one of contractor selection: either accept the P80 position and size contingency accordingly, or renegotiate the scope and risk allocation to bring exposure down.

We presented the results as a decision support package rather than a QRA report. For each significant risk driver, the report showed which risks were inherent to the scope (and therefore non-negotiable), which were allocated inefficiently in the draft contract (and therefore worth renegotiating), and which were best mitigated by a specific change in sequencing or procurement strategy. The client used this to run a focused contract review before signature — the QRA was not just an input to the confidence level, it was an input to the contract itself.

How this work differed from a standard QRA

Independent QRA at final investment decision sits in a particular place in the project lifecycle. The model has to be good enough to satisfy an investment committee, robust enough to withstand a client challenge, and practical enough to inform contract negotiation within days rather than weeks. That combination is hard to achieve without an integrated cost-and-schedule model, a transparent audit trail, and a clear separation between model input and model output.

We built the model in a form the client could continue to use post-sanction — re-running the analysis against actual performance as the programme advanced, updating the distributions as uncertainty resolved, and maintaining a live view of the confidence position rather than freezing the P80 at contract signature. This matters because a QRA delivered as a one-off document loses value within three months; a QRA delivered as a maintained model remains a governance asset throughout delivery.