Digital innovation is enhancing the observational method on HS2

Project Overview

£2.8M
predicted material cost savings at Bromford and Washwood Heath
3000t
CO₂ emission savings predicted
Used on complex civil engineering projects, the observational method often delivers dramatic time, cost and safety benefits – but even these can be greatly enhanced with DAARWIN, a new software solution that employs machine learning.

Huge efficiency gains

Since the 1980s, Mott MacDonald’s engineers have helped pioneer a technique called the observational method (OM) to optimise the delivery of major projects, often ones involving deep excavation in difficult ground conditions. These have included London Docklands Light Railway and Limehouse Link, the Heathrow Express rail tunnel, Heathrow Airport Terminal 5 and the Elizabeth Line Moorgate shaft. They showed that The Observational Method could be used to achieve major cost and time savings. On projects the size of HS2, with many kilometres of retained excavations and earthworks, the potential efficiency gains are large.

It’s why HS2 has awarded innovation funding for a trial of a new software solution called DAARWIN, at the site of a tunnel boring machine (TBM) launch shaft and tunnel at Bromford in the Midlands.

Hock Liong Liew: pushing geotechnical boundaries to save cost and carbon

What exactly is the observational method?

Engineering designers and contractors must anticipate the movement of underground structures and existing infrastructure to safeguard the project, protect workers from harm and avoid damage to existing infrastructure. For excavations, ground movement is held in check using piled diaphragm walls, waling beams and struts. In poor ground, jet grouting, deep cement mixing and tension piles or anchors may also be used to prevent base slabs from heaving.

The temporary works take time to install – and ultimately to remove as well, to make way for permanent works. Installation and removal bring additional safety risks: heavy lifting, manual handling and fixing, often at height. And the temporary works impede construction. Excavation machinery and mucking skips, construction materials and workers must operate in a congested cat’s cradle of steelwork.

Finding ways to reduce the extent of the temporary works can produce radical improvements on cost, programme, safety and carbon. And that’s what OM does: it involves predicting movement of the ground and underground structures and then ensuring effective construction control, through effective collaboration of the design and construction teams. A range of predesigned measures can be implemented if measurements show either that safety must be enhanced or that savings can be achieved.

By measuring how the ground and structure responds, we can make sure that the work is carried out efficiently and safely.
Rob Talby
Project principal and geotechnical director, Mott MacDonald

As project principal and geotechnical director Rob Talby puts it: “We have a design and a prediction of how that design is going to perform. By measuring how the ground and structure responds, we can make sure that the work is carried out efficiently and safely. As engineers we want to verify that the design and construction sequences are going according to plan, and we can see this by looking at the predictions versus the measurements.”

 

Observational method benefits

Accurate prediction of ground and underground structure behaviourProgressive design modificationsTime and cost savingsImproved project safety

  • Accurate prediction of ground and underground structure behaviour
  • Progressive design modifications
  • Time and cost savings
  • Improved project safety

Nimble response

The observational method requires a detailed understanding of the ground being excavated and the structures being built. Different soil types have different characteristics – they can variously crumble, slump, shear, flow and swell. Add water, and the performance of each soil type can change dramatically, from bone dry to saturated. Temporary and permanent works designs are based on codes and standards that account for the least stable ground conditions. Designs are checked, reviewed and approved by others – a process known as design assurance – resulting in conservative, ‘belt and braces’ designs.

But conditions are usually not worst-case, and the ground often behaves better than assumed in the design.

The observational method helps us nip in the bud the rare occasions where there is the potential for things to go wrong.
Rob Talby
Project principal and geotechnical director, Mott MacDonald

OM starts with design models that take into consideration potential variation in ground conditions, groundwater and loadings. Design of the excavation sequence highlights when supports systems must be installed. During excavation, ground movement data from instrumentation and monitoring (I&M) is compared against the models to determine if support is required. The key for OM is to create a truly effective and collaborative OM team, comprising experienced design and contracting staff.

A green, amber and red ‘traffic light’ system is commonly used to communicate safety. “If everything’s green then we’re building something safely. If everything is green then we should assess whether the temporary support system can be modified. It’s about seizing the moment to advance the works,” says Rob.

“At amber we can continue without additional support but are on standby to apply it. Red highlights we need to introduce additional support or change the construction sequence. The observational method helps us nip in the bud the rare occasions where there is the potential for things to go wrong, because we’re being proactive in looking at the measurements and we can implement the necessary mitigation measures.”

How DAARWIN evolves design

Ground conditions at every construction site are different and, as work advances, I&M data can be used to recalibrate the design so that it better matches the reality. This process is known as back-analysis. It enables the work sequence and temporary works to be iteratively redesigned and refined as work advances.

But back-analysis is complex and time consuming – unless you can improve on traditional techniques.

DAARWIN is the first and only ground engineering software to use machine learning algorithms to link design models with I&M data, accelerating back-analysis and allowing engineers to see whether modifications to the design or construction sequence are needed.

Instead of a large team taking weeks, we’re able to do the same thing in much shorter durations.
Imran Farooq
Ground engineering lead on HS2

Imran Farouq: using adaptive design to achieve whole-life efficiency and resilience

Imran Farooq is the ground engineering lead for the HS2 design JV that comprises Mott MacDonald and Systra. “Historically OM has been managed by a team doing lots of calculations,” he says. “But DAARWIN uses machine learning to process data. Instead of a large team taking weeks, we’re able to do the same thing in much shorter durations.”

DAARWIN is cloud based, enabling very large volumes of data to be managed and easily accessed.

By radically shortening the back-analysis and progressive modification cycles, DAARWIN enables work sequences to progress faster and – when back-analysis shows it is safe to do so – with less temporary support.

DAARWIN features

  • High performance computing in the cloud
  • Genetic algorithms to perform automatic back-analysis
  • Compatible with finite element models
  • Automatic monitoring data acquisition
  • Construction progress tracking
  • Data repository (finite element models, monitoring data, construction progress, images, comments)

Trial and implementation

We initially used DAARWIN to retrospectively analyse inclinometer data gathered during construction of the HS2 TBM launch shaft and tunnel at Bromford Tunnel East Portal (BTEP). Results of our virtual trial showed there could have been programme savings of at least one month for BTEP excavation works, had DAARWIN been used in practice.

£2.8M predicted material cost savings at Bromford and Washwood Heath

After the trial and with HS2 approval, OM was implemented at the eastern end of Bromford Tunnel East Portal employing DAARWIN. This led to the omission of the remaining steel props, which helped BBV to save at least two weeks on its construction programme.

The success of the trial and OM implementation at the eastern end of Bromford Tunnel East Portal helped us to win the Ground Engineering Editor’s Award in 2023.

DAARWIN is now being used live on the Washwood Heath Retained Cut (WHRC), which will connect Bromford Tunnel West Portal (BTWP) to the surface. WHRC is a 1.4km-long retained cut, 26m deep at the eastern end of BTWP.

 

The DAARWIN procedure

  • Upload project information into the platform numerical models including construction progress, monitoring data, images and historical information
  • Analyse multiple design options together with different ground parameter scenarios (from pessimistic, most probable to optimistic) to determine the optimal design option and the most influential geotechnical parameters
  • Compare your design with the monitoring data and assess its performance
  • Calibrate your numerical models to better predict ground behaviour
  • Modify the design

Understanding mudstone

The BTEP and WHRC run through a formation called Mercia mudstone. To date not many large excavations have been carried out and back-analysed in this geology, so designers are reliant on ground investigations. These are likely to lead to a significant underestimation of its strength and stiffness, especially at depth. And that offers a major OM opportunity.

Hock Liong Liew, technical principal and expert in foundations and geotechnics, is leading the OM implementation on the BTWP and WHRC. He says: “Logging mudstones from borehole samples can be challenging. The samples are likely to be disturbed during the drilling process.” That could lead to a more onerous temporary works design. But the mass behaviour of this material is likely to be stronger than assumed.

“As work advances, we’ll gain increasingly knowledge of the in-situ mass characteristic of Mercia mudstone through feedback from field observations and back-analysis using DAARWIN.” OM combined with DAARWIN holds the potential for rapid iterative efficiency improvements.

3000t CO₂ emission savings predicted

Gone: 4km of props and walers

Imran outlines the opportunities: “We’ve initially designed in temporary props and walers at the excavations to meet code requirements for factors of safety. But by using the OM and DAARWIN we expect to be able to demonstrate that, as construction progresses, we’re within the safety margins and can redesign to reduce the number of props.”

Laid end-to-end, the number of props and walers potentially eliminated could add up 4km in length. Without props restricting movement, the project will be able to use bigger excavators and trucks, resulting in a faster and safer dig.

Bigger vehicles mean fewer vehicle movements, which will reduce CO2 emissions by an estimated 3000t. We calculate that construction could be accelerated by four to six months.

“These are the ideals that you’d want to have on any project,” says Imran.

Digital twin

The connection between the engineering design and real I&M data means that DAARWIN effectively provides a digital twin of each project. It can be used to check performance and manage risk; graphic outputs aid communication across the project team and wider stakeholders. The detailed back-analysis provides improved understanding of ground conditions for future projects.