Robust engineering design and a commitment to risk management are crucial steps in preventing deep excavation failure and the consequences of it.

Deep excavations are a key element in the development of urban environments. Excavation of the ground is required for major utilities, basements and new underground transportation tunnels and stations. The ground can be highly variable and excavation can, therefore, introduce considerable hazards to construction projects. The role of the engineer is to identify and understand these hazards and to design and construct a robust solution that will avoid or mitigate the risks to a level 'as low as reasonably practicable' (ALARP). To achieve this, there must be a strong safety and risk management culture ingrained in all of the organisations involved in the delivery of a project.

The consequences of failure, in the form of either collapse or excessive ground movements, can be far reaching and include:

- Injury and loss of life of workers and the public
- Damage or collapse of surrounding structures and utilities
- Project delays or cancellation
- Remedial design and construction
- Litigation
- Insured and uninsured losses
- Negative publicity

The scale of the consequences will vary from project to project, but deep excavation projects are increasingly occurring within congested urban environments - snuggled between historic buildings and sensitive infrastructure - where the consequences of failure can quickly approach an unacceptable level.

Indicative costs of recent deep excavation collapses display a common theme: significant cost despite different causes to which the failure may be attributed (Table 1).

In contrast, there are many examples of successfully completed excavation projects in difficult ground conditions in city centres, demonstrating that deep excavation projects can be delivered safely and successfully when careful consideration and management of the risks associated with the ground are a central part of the project.

Characteristics of the ground

The ground beneath our feet is highly variable. The surface level and thickness of the strata vary as a result of initial deposition and weathering in geological time. The variation can occur over a small area due to localised features; such as buried valleys from historical watercourses or geological faulting, as well as manmade effects, where, for example, there is disturbed ground from backfill around cut and cover structures left from reclamation or construction. Adjacent soil strata can have significantly different engineering properties, which has a considerable impact on their behaviour and, hence, the design of excavations. The presence of groundwater and the ability for groundwater to flow through the different strata will also greatly influence the design.

Compared with concrete, the properties of soil are much less reliable. Normalised strength testing from the soil and concrete indicate concrete is much more consistent as shown in Figure 1. The strength properties of the soil display a large variation in magnitude with a relatively high likelihood of achieving extreme strength magnitudes. The designer should choose engineering soil parameters with due consideration for this variation. Suitable factors of safety, which are typically higher than for concrete and steel, are adopted in accordance with design codes of practice.

Risk management in design

For ground related risks, hazard identification commences with a site-specific desk study that reviews available geological and historical information to identify potential hazards that may be introduced by the ground. Some source examples include historical maps, aerial photography or references to mining, which may have left a void beneath the surface or a localised change in the anticipated stratigraphy at the site due to the backfilling of pits.

Ground investigation is a key part of the design process, providing design information such as the site geology, material design parameters, groundwater regime and contamination. It is also a key stage in the risk management of ground related hazards through the investigation of potential hazards identified in the desk study, such as the presence of harmful gasses like methane and radon.

To determine characteristics such as stratum levels and engineering properties of the ground, an intrusive investigation is undertaken. Depending on the local variability in strata and the coverage of exploratory holes, however, significant localised features such as the buried valley and unidentified stratum shown in Figure 2 may not be identified. These local features represent a hazard to the project when they introduce design conditions that are more onerous than those interpreted from the intrusive investigation.

By its nature, a ground investigation can only ever cover a small proportion of the construction site. In the design of an investigation, the effort to obtain information and the potential value of that information needs to be balanced carefully. A very extensive investigation will reduce, but never eliminate all of the uncertainty in the ground, and will be costly. In contrast, a limited investigation will be less likely to identify potential ground related hazards and, therefore, can significantly increase the risk to construction. Site investigations need careful planning, design and procurement by suitably competent geotechnical engineers to obtain suitable design information and hazard identification with an appropriate level of effort.

When designing for the safety of the excavation and people potentially affected by the construction works, the designer must consider the hazards introduced by the uncertain nature of the ground alongside other hazards such as earthquake, flood, fire and congested working conditions, which are often significant for excavation projects. Several publications provide guidance on potential ground related hazards and risks to excavation projects, including 'Managing Geotechnical Risk' and CIRIA Report C604 'CDM Regulations - Work Sector Guidance for Designers'. Through appropriate design choices, the designer aims to provide a robust design and approach the identified risks with a hierarchy of elimination, reduction and finally control.

Different construction techniques have been developed in response to different ground conditions. By identifying the site specific risks that the particular ground conditions pose, the designer can compare the available options and make appropriate choices of:

- Size and form of retaining wall
- Excavation and construction sequence
- Temporary works (propping)
- Temporary dewatering

Design guidance on these aspects is available in CIRIA Report C580, 'Embedded retaining walls - guidance for economic design' among others.

Some level of ground movement is an inevitable result of excavation projects. Where the predicted ground movements indicate an undesirable or unacceptable level of damage to surrounding property, protective measures are adopted to reduce the predicted level of damage to an acceptable level. These protective measures should follow a hierarchy of:

- At-source measures, such as the method of construction or excavation sequence
- Ground treatment measures to reduce or modify the induced ground movements
- Structural measures to increase the capacity of the structure to resist the ground movements

For large urban underground infrastructure projects, such as the Jubilee Line Extension in London, the costs of such protective measures can be significant and must be allowed for in budgeting for the project.

From design to construction

A strong link between design and construction is crucial for a successful risk management process. Regardless of the procurement method or project organisation, safety critical assumptions and risk related measures identified during design need to be communicated to the construction team, as they will be responsible for protecting and carrying through the measures. Initially, this is achieved through the documentation required by the Construction (Design and Management) Regulations, such as the risk register for the project. Effective communication between the designer and contractor is necessary throughout the excavation process to respond to changes identified during construction and, where appropriate, to verify that the key design assumptions are in keeping with the ground conditions revealed during construction.

Monitoring the movement of the ground and adjacent buildings during excavation is fundamental to validate the many design assumptions relating to ground and groundwater conditions and likely mechanisms of failure. Unexpected or more onerous movements identified during excavation will require a review of the design assumptions, as they may be indicative of more onerous loading conditions or additional possible failure mechanisms. Where appropriate, it may be necessary to introduce contingency measures, such as additional propping, other alterations to the construction sequence or additional corrective measures, for example, grouting for the adjacent buildings.

Risk management culture

In response to the insurance industry's concern over the performance of tunnel construction projects, British Tunnelling Society and the Association of British Insurers in 2003 produced the 'Joint Code of Practice for the Risk Management of Tunnel Works in the United Kingdom. This document demands an equitable sharing of risk between the parties involved in tunnelling projects and sets standards for all parties, including the client. To ensure adoption of the risk management approach required by the code, adherence to the code is a prerequisite for obtaining contractor's all risk insurance and third party liability insurance for tunnelling contracts in excess of £1 million or where there is a significant risk to third parties. The International Tunnelling Insurance Group has now prepared an international code of practice version which complements the UK version.

Loss prevention risk audits are an important activity within the risk management process on major projects. They are usually undertaken by independent and experienced construction practitioners on behalf of the international insurance markets and aim at ensuring compliance of the project (management, design, construction and co-ordination of third parties) with appropriate codes and standards. It is a benchmarking exercise, carried out at intervals throughout the project, and represents a positive step towards loss prevention.

Safety and risk management are fundamental to the minimisation of the likelihood of deep excavation failure. For safety and risk management approaches to work and not just be a paperwork exercise, there must be a cultural mindset within the industry towards safety and the control of risks. This culture must be present at all levels of construction - from the client down to individual operatives for a site. The client's demonstrable commitment to this is fundamental, as without it, the safety and risk management of a project will be compromised.

Sara Anderson is a civil engineer and an associate in Arup.



Clayton, C (2001) Managing Geotechnical Risk, ICE/DETR
Health and Safety Commission (1994) Managing construction for health and safety, Construction (Design and Management) Regulations 1994. Approved Code of Practice, HSE Books, Suffolk, England.
British Tunnelling Society /Association of British Insurers (2003) The Joint Code of Practice for Risk Management of Tunnel Works in the UK
International Tunnelling Insurance Group (2006), A Code of Practice for Risk Management of Tunnel Works
Nicholson et. al. (1999) The Observational Method in Ground Engineering: Principles and Applications, CIRIA Report 185
Gaba et. al. (2003) Embedded Retaining Walls - Guidance for Economic Design, CIRIA Report C580.
Arup & Gilbertson, A. (2004) CDM Regulations - Work Sector Guidance for Designers CIRIA C604
Grose, B. (2006) Tunnelling, risk management and collapses - Where are we going?, International Conference on Deep Excavations, 28-30 June 2006, Singapore.

Case history

Nicoll Highway Collapse, Circle Line, Singapore

In 2004, a deep excavation for the underground Circle Line in Singapore collapsed, killing four construction workers and causing severe damage to the adjacent Nicoll Highway. It was fortunate that, at the time of collapse, the normally busy highway was temporarily empty in that area and no members of the public were injured. The delays due to redesign and the inquiry process have been significant.

The committee of inquiry into the collapse of the excavation identified flaws in the design of the excavation, but also criticised the safety culture and failures in dealings with risk on the project. Nicoll Highway is not alone in this criticism, as many investigations into the collapse of deep excavations find a failing in the project's commitment to safety and a lack of an identifiable safety culture.