QSRA for Old Oak Common HS2 Station: UK Rail Schedule Risk Analysis

A super-hub station designed to connect HS2, Crossrail, Great Western Main Line, and the Bakerloo Line extension in a single interchange, all while construction proceeds above and around an active rail corridor in the heart of West London. If you are managing the schedule for HS2's Old Oak Common station, a single deterministic completion date is not a plan. It is an act of faith in a programme that has already rebaselined its schedule three times.

Quantitative Schedule Risk Analysis (QSRA) is a statistical method that stress-tests project timelines by modelling the impact of uncertainties and discrete risk events using Monte Carlo simulation. It replaces single-point schedule dates with probability distributions, producing a range of possible completion dates at defined confidence levels. This gives programme directors a defensible basis for setting schedule contingency and reporting realistic milestones to the Department for Transport and HM Treasury.

For a station as technically complex as Old Oak Common, where deep-box excavation, platform construction, systems integration, and live railway interface all converge on a single critical path beneath one of London's most constrained sites, QSRA is the difference between a schedule that informs Treasury gate reviews and one that merely provides political cover. The complete guide to Schedule Risk Analysis explains the full methodological framework; this post applies it to the UK's most complex station build.

Here is how IQRM would approach a full QSRA on HS2's Old Oak Common station, step by step.

Why Old Oak Common Needs a QSRA

Old Oak Common is not a typical railway station. It is a deep-box excavation beneath the Great Western Main Line, one of the busiest rail corridors in the United Kingdom, handling over 700 train movements per day. The station design requires construction of a six-platform underground box approximately 850 metres long, built while the existing railway continues to operate above. The interface between HS2 construction and Network Rail's live operations creates a scheduling constraint that has no parallel on greenfield rail projects.

The station's estimated cost has risen from an initial £1.3 billion to between £5 billion and £6 billion, reflecting the complexity that deterministic planning consistently underestimates. The Department for Transport and HM Treasury have both scrutinised the programme's schedule credibility. A deterministic completion date from Primavera P6, without probabilistic validation, provides no information about the likelihood of delivery. IQRM's experience on UK rail infrastructure shows that deterministic schedules on deep-box station projects typically sit at P10 to P20 on the risk-adjusted S-curve, meaning the planned date has between a 10% and 20% chance of being achieved without contingency.

The programme involves multiple tier-one contractors working in sequence and in parallel across civil engineering, mechanical and electrical fit-out, systems integration, and railway systems. Each contract interface is a schedule risk boundary. Without a QSRA that models these interfaces probabilistically, the programme team cannot identify which interface risks dominate the critical path or quantify the schedule contingency needed to protect the opening date.

Network Rail possessions, the planned windows when the live railway is shut down for construction access, are the single most constrained resource on this project. Possessions are booked 12 to 18 months in advance and are shared across multiple infrastructure projects on the Great Western corridor. A missed possession window can delay dependent activities by months, not days. The QSRA must model possession availability as a discrete risk with defined probability and impact, because the consequence of a single missed possession cascades through the entire programme logic.

Phase 1: Schedule Import and Health Check in Safran Risk

The QSRA begins by importing the Primavera P6 schedule into Safran Risk. On Old Oak Common, the P6 file is structured as a multi-contract integrated master schedule covering enabling works, main civils, mechanical and electrical systems, railway systems, and testing and commissioning. Safran Risk imports the native XER or XCR export, preserving all calendars, activity logic, constraints, and resource assignments. Before any risk modelling begins, the imported schedule must pass a rigorous health check.

Remove hard constraints locking milestone dates. On HS2, political target dates for station opening are embedded as hard Finish No Later Than constraints in P6. These prevent Monte Carlo simulation from shifting completion dates realistically. Replace them with soft constraints and note the political target as a reporting reference, not a logic driver.

Close open logic ends at contract interfaces. The civil-to-systems handover is the most common source of broken logic on multi-contract station programmes. If the civils completion milestone is not logic-linked to the systems integration start, the critical path cannot shift between contracts during simulation, and the model will understate total programme risk.

Replace lags with explicit activities. UK rail schedules frequently use large lags between track installation and signalling testing to account for regulatory approval windows. These lags prevent risk assignment. Replace each substantive lag with a named approval or testing activity that carries its own duration uncertainty range.

Verify the testing and commissioning logic. On deep-box stations, integrated testing, trial running, and safety case submission to the Office of Rail and Road are consistently the most compressed and most at-risk phases. Ensure these activities are fully logic-linked with realistic durations before entering the risk model.

Phase 2: Risk Identification for UK Deep-Box Rail Construction

Risk identification for Old Oak Common requires capturing two distinct categories: estimation uncertainties on every activity, and discrete risk events with defined probability and impact. The project context introduces risk themes specific to deep-box construction beneath a live railway in a dense urban setting.

Ground conditions beneath the Great Western Main Line are the defining geotechnical risk. The deep-box excavation passes through London Clay, Lambeth Group deposits, and potentially the Thanet Sand formation. Ground investigation boreholes define expected conditions, but the QSRA must model the gap between expectation and reality. Unexpected ground water ingress, encountering obstructions from Victorian-era infrastructure, or encountering softer clay lenses than anticipated can each add weeks to individual excavation activities.

Network Rail possession availability is the highest-impact discrete risk. Each possession window is a fixed calendar event. If HS2 construction activities overrun a possession window, the live railway must be handed back regardless of construction progress. The next available possession may be 6 to 12 weeks later. Model each critical possession as a discrete risk event: probability of overrun (based on historical UK possession performance data, typically 15% to 25% for complex possessions) multiplied by the delay to the next available window.

Labour productivity in UK construction is subject to seasonal variation, skills shortages, and regulatory constraints. Unlike Gulf projects where extreme heat is the dominant productivity factor, UK rail projects face winter weather restrictions (concrete pours below 5 degrees Celsius require protection measures), shorter daylight hours from November to February reducing productive shift time, and periodic national skills shortages in specialist trades such as signalling engineers and overhead line equipment installers.

Regulatory approval sequencing creates schedule risk that is unique to UK rail. The Office of Rail and Road must approve the safety case before passenger services can commence. Transport for London interfaces add a parallel approval stream for interchange elements. Each regulatory submission has a defined review period, but the probability of first-time approval is historically low on novel station configurations. Model resubmission cycles as discrete risks with 30% to 40% probability and 8 to 16 week impact per cycle.

Phase 3: Quantifying Uncertainty with the Risk Data Engine

Every duration uncertainty range in the QSRA must be traceable to a data source. IQRM's Risk Data Engine (RDE) methodology replaces expert guesswork with a structured data hierarchy that determines the appropriate probability distribution for each activity group based on the quality and quantity of available data.

For Old Oak Common, the data sources include: historical outturn data from Crossrail's deep-box stations (Paddington, Liverpool Street, Whitechapel), which provide direct comparators for excavation rates, concrete pour cycles, and fit-out durations in London Clay; HS2's own enabling works performance data from the Phase 1 main civils contracts already under construction between London and Birmingham; and industry benchmarks from Network Rail's portfolio of station renewal projects, adjusted for the complexity uplift of a deep-box configuration versus surface-level works.

The RDE selects distributions based on data sufficiency. Where historical outturn data exists for 30 or more comparable activities, a fitted distribution (lognormal or beta) is appropriate. Where data is limited to 5 to 29 data points, a PERT distribution calibrated to the observed range is used. Where fewer than 5 data points exist, a three-point estimate (optimistic, most likely, pessimistic) with a triangular or PERT distribution is the fallback, but these ranges must be explicitly challenged for anchoring bias.

Phase 4: Risk Mapping and Correlation

Risk mapping assigns each identified risk and uncertainty range to the specific activities it affects in the Safran Risk model. On Old Oak Common, mapping must reflect the multi-contract structure: ground condition risks map to civils activities, possession risks map to activities requiring railway access, and regulatory risks map to testing and commissioning milestones.

Correlation is essential for realistic modelling. Without correlation, Monte Carlo simulation assumes that if one excavation activity runs late, adjacent excavation activities are equally likely to run early. This is physically impossible when the same ground conditions, the same workforce, and the same equipment affect all excavation activities on the same alignment. Safran Risk supports correlation coefficients between 0 and 1. For Old Oak Common, apply positive correlation (0.5 to 0.7) between all deep-box excavation activities sharing the same geological stratum, between all activities dependent on the same specialist subcontractor, and between all possession-dependent activities on the same rail corridor.

Ignoring correlation is the single most common error in QSRA models. A model without correlation produces an S-curve that is too narrow, underestimates tail risk, and gives false confidence that the P80 date is achievable. On a deep-box station programme with shared geological, logistical, and regulatory drivers, correlation typically widens the P10-P90 range by 30% to 50% compared to an uncorrelated model.

Phase 5: Monte Carlo Simulation and S-Curve Output

Safran Risk runs 10,000 iterations of the schedule, sampling from every assigned distribution and firing discrete risk events according to their defined probabilities in each iteration. The output is a cumulative probability distribution (S-curve) showing the full range of possible completion dates for Old Oak Common station opening.

On a project of this complexity, the S-curve typically reveals a substantial gap between the deterministic completion date and the risk-adjusted forecast. Based on IQRM's experience with UK deep-box rail stations, the expected pattern shows the deterministic date sitting at P10 to P20 on the S-curve, the P50 (median) outcome falling 12 to 18 months beyond the deterministic date, and the P80 outcome falling 18 to 30 months beyond.

The gap between P50 and P80 represents the schedule contingency that HM Treasury expects to see justified through probabilistic analysis. On HS2, the Infrastructure and Projects Authority reviews schedule confidence levels as part of its gateway assurance process. A QSRA that demonstrates a credible P80 date, supported by traceable data and properly correlated risks, provides the evidence base that deterministic planning alone cannot deliver.

Phase 6: Sensitivity Analysis and Risk Drivers

The tornado chart ranks every risk and uncertainty by its contribution to total schedule variance. On a UK deep-box station project like Old Oak Common, the tornado chart typically reveals a pattern that challenges assumptions about where the real schedule risk lies.

The top five schedule risk drivers on Old Oak Common are likely to include: Network Rail possession availability and overrun risk (contributing 20 to 25 percent of total variance), ground condition uncertainty in the deep-box excavation (15 to 20 percent), systems integration and testing duration uncertainty (12 to 18 percent), regulatory approval cycles from the Office of Rail and Road (8 to 12 percent), and specialist labour availability for signalling and overhead line equipment installation (6 to 10 percent).

The criticality index complements the tornado chart by showing how frequently each activity appears on the critical path across all 10,000 iterations. An activity with a 90% criticality index was on the critical path in 9,000 of the 10,000 iterations. On Old Oak Common, the testing and commissioning phase is expected to show criticality indices above 85%, confirming that regardless of how civils construction performs, the final programme phase dominates the path to opening.

The sensitivity analysis gives the programme director a clear investment priority. If possession overrun risk contributes 22% of total variance, and each missed possession costs an average of 10 weeks in delay, then investing in possession planning, rehearsal, and contingency access agreements with Network Rail delivers more schedule certainty per pound spent than any other single intervention.

Phase 7: Pre-Mitigation vs Post-Mitigation Comparison

The QSRA model is run twice: once with the current risk profile (pre-mitigation), and once with planned risk responses applied (post-mitigation). The comparison quantifies the value of each mitigation strategy in terms of schedule days recovered at each confidence level.

For Old Oak Common, effective mitigation strategies typically include: negotiating additional possession windows with Network Rail as contingency access slots, pre-installing ground treatment measures in advance of the main excavation sequence to reduce ground condition uncertainty, establishing parallel regulatory submission tracks so that Office of Rail and Road review does not sit on the critical path as a single serial activity, and pre-qualifying multiple specialist subcontractors for signalling and overhead line equipment to reduce dependency on a single supplier's workforce availability.

The pre/post comparison provides HM Treasury and the Infrastructure and Projects Authority with the cost-benefit evidence they require. If investing £15 million in additional possession contingency windows reduces the P80 completion date by 6 months, and each month of delay to Old Oak Common opening costs the programme an estimated £40 million in extended preliminaries, the investment case is unambiguous. The QSRA transforms risk mitigation from a cost line into a quantified investment decision.

QSRA Reference Table: Old Oak Common vs Standard UK Rail Projects

Best Practices for QSRA on UK Deep-Box Rail Stations

Model possessions as discrete calendar events, not productivity factors. A missed possession is a binary event with a fixed delay to the next available window. Reducing a productivity factor by 5% does not capture this reality. Use Safran Risk's calendar risk feature to model possession windows as constrained access periods with defined overrun probabilities.

Use Crossrail outturn data to calibrate duration ranges. Crossrail's deep-box stations in London Clay are the closest available comparators for Old Oak Common. Outturn data from Crossrail excavation, concrete, and fit-out activities provides an empirical basis for calibrating PERT and lognormal distributions rather than relying on subjective three-point estimates.

Apply correlation across all activities sharing common drivers. Ground condition, possession, and workforce risks affect multiple activities simultaneously. An uncorrelated model on a deep-box station programme will produce an S-curve that is 30% to 50% too narrow, giving false confidence in the P80 forecast.

Update the QSRA quarterly during peak construction. A single QSRA at the start of the project becomes outdated as risks materialise and new risks emerge. Quarterly updates during the civils phase and monthly updates during testing and commissioning capture the evolving risk profile and validate that mitigation strategies are delivering the expected schedule recovery.

QSRA Delivers Treasury-Grade Forecasts for UK Rail

Old Oak Common station is the most complex single structure on the HS2 programme. Its schedule is shaped by forces that deterministic planning cannot model: ground conditions that vary metre by metre, possession windows that constrain access to the busiest rail corridor in the country, regulatory approval processes with uncertain timelines, and contract interfaces where delay in one package cascades through four others.

A QSRA built in Safran Risk, grounded in Crossrail outturn data through the Risk Data Engine, with properly correlated risks and calibrated distributions, gives the HS2 programme team, HM Treasury, and the Infrastructure and Projects Authority what deterministic planning cannot: a defensible, probability-weighted forecast of when the station will open, which risks drive the schedule, and where investment in mitigation delivers the greatest return in schedule certainty.

The complete guide to Schedule Risk Analysis covers the full QSRA methodology. For UK rail-specific applications, the principles demonstrated here for Old Oak Common apply equally to other HS2 stations, Network Rail enhancement programmes, and the emerging pipeline of mass transit projects across UK cities.

Frequently Asked Questions

What is a QSRA for a UK rail station project?

A Quantitative Schedule Risk Analysis (QSRA) for a UK rail station is a Monte Carlo simulation that stress-tests the Primavera P6 schedule by modelling duration uncertainties, discrete risk events such as possession overruns and ground condition surprises, and correlation between related activities. It produces a probability distribution of completion dates, allowing the programme team to report realistic milestones to HM Treasury and the Infrastructure and Projects Authority instead of relying on deterministic dates that carry no probability information.

Why does Old Oak Common HS2 station need a QSRA?

Old Oak Common is a deep-box excavation beneath the busiest rail corridor in the UK, involving multi-contract interfaces, Network Rail possession constraints, complex London Clay geotechnics, and dual regulatory approval from the Office of Rail and Road and Transport for London. Deterministic schedules on projects of this complexity typically sit at P10 to P20, meaning the planned date has only a 10% to 20% chance of being achieved. A QSRA quantifies this gap and identifies the specific risks driving it.

What software is used for QSRA on HS2?

Safran Risk is the industry-standard tool for QSRA on UK rail infrastructure programmes. It imports the native Primavera P6 schedule, supports correlation modelling, calendar risks for possession windows, and produces S-curves, tornado charts, and criticality indices. Safran Risk preserves all P6 logic, calendars, and constraints during import, ensuring the risk model is built on the same schedule the programme team manages daily.

What confidence level should UK rail projects report?

IQRM recommends P80 as the standard reporting confidence level for UK rail infrastructure, aligned with HM Treasury Green Book guidance and Infrastructure and Projects Authority expectations. For external commitments to parliament or lenders, P90 provides a conservative upper bound. The P50 is useful for internal planning but should never be presented as a committed delivery date for a programme of Old Oak Common's complexity.

How does possession risk affect QSRA results on UK rail?

Network Rail possessions are fixed calendar windows booked 12 to 18 months in advance. If construction activities overrun a possession, the live railway must be handed back regardless of progress, and the next available window may be 6 to 12 weeks later. On Old Oak Common, possession risk typically contributes 20% to 25% of total schedule variance, making it the single largest risk driver. Modelling possessions as discrete calendar events with defined overrun probabilities is essential for realistic QSRA output.

How often should the QSRA be updated during construction?

IQRM recommends quarterly QSRA updates during the main civils phase and monthly updates during testing and commissioning. A single QSRA at project sanction becomes outdated as risks materialise and new risks emerge. Regular updates capture the evolving risk profile, validate mitigation effectiveness, and provide the Infrastructure and Projects Authority with current confidence levels for gateway reviews.

Written by Rami Salem, Quantitative Risk Management specialist with 15+ years of experience in oil and gas, EPC/EPCM, and infrastructure projects across the UK and GCC. Approved consultant for Saudi Aramco and ADNOC.