An important consideration when planning and carrying out tunnelling works in urban areas is the impact that those works may have on third party assets. Substantial ground movements may be caused by the tunnelling itself and by associated works for access shafts and other deep excavations. The problem is particularly acute in mature cities where vulnerable structures overlie old utility apparatus. The impacts of tunnelling on buildings throughout the world have been widely reported but the impact on utility apparatus has not received such detailed and extensive attention in the literature.

The topic of this paper has been further stimulated by interest within the Research Working Group of the International Tunnelling and Underground Space Association (ITA). Discussions within the Working Group over a number of years have indicated a wide range of approaches to the problem in different countries and it is clear that the social and legal framework of a nation influences the approach taken by both Utility Companies (‘Utilities’) and developers.

The principal purpose of this paper is to consider strategies and methods of impact assessment for the protection of utility pipelines and sewers. The content is guided by numerous investigations into pipeline and sewer failures and consideration of the geo-environment in cities which may render conventional analyses intractable or at least unreliable. The guidance presented comes from experience gained and applied during numerous major tunnelling works and forensic investigation of pipe and sewer failures in the Greater London area.

During recent years the greatly improved performance of tunnelling machines in reducing ground movements has in practice shifted the balance of risk analysis from the consideration of ground movement toward pipeline condition and the consequences of failure.

The paper will discuss the present approaches to the problem and suggest more risk based methods which may be considered for future application. The emphasis will be on rationalising the assessment process to reflect the uncertain nature of the pipelines condition, geo-environment and the consequences of damaging impacts and to indicate a strategy for assessment and give guidance to Designers as to how it might be realised in practice.

1 Introduction

An important consideration when planning and carrying out construction works in cities is the impact those works may have on the assets of third parties. For example, works for new underground railways may be particularly damaging if they are not carefully planned and carried out. Substantial ground movements may be caused by the tunnelling itself, the construction of large retaining structures for access and ticket halls, escalators, shafts for access, safety, ventilation, and temporary works.

There are a variety of other tunnelconstruction- related activities that can have adverse impacts on the assets of others. For instance, the transport of abnormally heavy loads and their placement by cranes which may exert very high localised loads beneath outriggers.

Tunnelling impacts on buildings throughout the world have been widely reported. For example, the construction impacts on buildings during works for the London Jubilee Line Extension were extensively monitored and the results of this work have been fully reported (CIRIA 2001).

However, the impacts on utility apparatus such as pipelines have not received such detailed and widespread attention in the literature. The design and construction of Crossrail, which comprises twin 7.2m OD tunnels and numerous major stations beneath the heart of London, has provided an impetus for the further development of impact assessment methodology. Current works which are the subject of extensive assessment processes include the Northern Line Extension, the Thames Tideway ‘Supersewer’ project, High Speed 2 (HS2) and numerous underground station upgrades.

The principal purpose of the paper is to consider and improve strategies and methods of impact assessment for pipelines. The discussion is guided by numerous engineering and forensic investigations into pipeline failures and consideration of the geo-environment of pipelines in cities which may render conventional analyses both unreliable and, in some circumstances, intractable. The complexity of the assessment process is illustrated by figures 1 and 2. Figure 1 (page 21) shows the new tunnels and shafts for the Bank Station Capacity Upgrade in London which is presently under construction. The trunk sewers are shown in blue and comprise various sizes and construction materials the most sensitive of which are made of brick and cast iron and are often over 150 years old. Smaller local sewers are not shown. Further, the figure is a simplified diagram as there are numerous other existing metro tunnels in the immediate vicinity many of which would have created unknown pre-existing states of stress to the sewers and water mains.

Figure 2 (above) shows an example of utility infrastructure beneath the streets of London. The large cast iron pipes are for water distribution, the yellow and green pipe is a major gas main, and the high level ducting for telecoms. Note how the brick utility access chamber has been built partially on top of a water main probably causing the water leakage apparent at the joint.

In many cases the combination of highly complex sources of ground movements and the interaction between the overlying structures and geoenvironment render accurate predictions of damage to be unreliable.

The work has been further stimulated by interest within the Research Working Group (WG2) of the International Tunnelling and Underground Space Association (ITA). Discussions within WG2 over a number of years have indicated a wide range of approaches to the problem in different countries and it is clear that the social and legal framework of a nation influences the approach taken by both Utilities and developers.

A major difference seems to be related to whether or not the utility is privately or publicly owned. In the UK on privatisation the new utility companies became owners of extensive networks of pipelines together with statutory obligations to supply services and strict obligations to comply with the requirements of relevant regulatory authorities. They also became responsible to their shareholders to optimise returns and this led to enhanced motivation in their asset protection arrangements. This has led to developments in the asset protection and assessment processes required of developers.

This report does not attempt to cover all types of pipelines or the widely varying conditions found throughout the world. The emphasis is placed upon the vulnerable pipelines beneath cities and in particular is based on the author’s experience particularly in London. The principles discussed and the analyses given may however be more generally applied with suitable adaption for local conditions and methods of working.

The paper concentrates on the most vulnerable of utility pipelines which are commonly those made of grey cast iron or masonry. Polyethylene (PE) has replaced iron as the material of choice for many gas and water pipe installations and is remarkably resistant to failure caused by ground movements and even direct impact if correctly installed. Ductile iron protected from corrosion is also widely used.

2 Governance

The impact of construction works is usually governed by a framework of responsibilities and obligations enshrined in law. Some of these obligations are spelled out in detailed statute whilst others are only broadly defined (e.g. in tort) and dependant on the Judgements of the Courts.

Stakeholder perspectives may be influenced by very different and possibly conflicting objectives. All will seek to remain within the law but in practice there may be considerable scope for interpretation as to how that might be achieved. Developers will naturally seek to minimise the charges and restrictions to their works whilst Utilities will seek not to be disadvantaged by the disturbances caused by others and often largely out of their control. These objectives may come into conflict and it falls, in the first instance, to engineers from each side to reach equitable agreements.

In the UK, Utilities have a statutory obligation to provide services to their customers. They must therefore seek to ensure that their apparatus is not damaged or its serviceability threatened in the short and long term. The privatised Utilities also have pressures to meet the aspirations of their shareholders and are required to meet rigorous performance targets set by Government Regulators such as the Water Services Regulation Authority (Ofwat).

2.1 Specific Acts of Parliament

Brief consideration of some relevant Acts of Parliament is helpful in indicating the extent of the protection afforded to Utilities by the law. Where major infrastructure works are to be undertaken they may require specific legislation.

Of particular recent interest, is the Crossrail Act 2008 Chapter 18. Crossrail is a new east-west railway linking Reading and Heathrow in the west to Shenfield and Abbey Wood in the east via tunnels beneath Central London.

Schedule 17 of this Act deals with ‘Protective Provisions’ and Part 2 of this schedule specifically with ‘Protection for electricity, gas, water and sewerage undertakers’. The extent and limitations of the protection are rather complex but at section 12(1) the issue of ‘subsidence’ is specifically mentioned.

In paraphrased form the Act broadly requires that if by reason of the construction works, or any subsidence resulting from any of those works, any damage is caused to any apparatus or property of the undertakers (electricity, gas, water or sewerage), or there is interruption in any service provided, the nominated undertaker (in this case Cross London Rail Links) shall bear and pay the cost reasonably incurred by the undertakers in making good such damage or restoring the supply and shall –

(a) make reasonable compensation to the undertakers for loss sustained by them, and

(b) indemnify the undertakers against claims, demands, proceedings, and damages which may be made against them by reason of any such damage or interruption.

This is very similar to the protection provided in the Channel Tunnel Rail Link Act 1996 (now commonly referred to as HS1) at Schedule 15 Part ii.

More generally Acts may group activities and set out the obligations of the authorities carrying out works that may adversely impact Utility. For instance, the New Roads and Streetworks Act 1991 Chapter 22 sets out the responsibilities of authorities carrying out works for road purposes (Section 83) and for major highway works, major bridge works or major transport works (Section 84): In certain circumstances a Utility may be required to contribute to costs.

A claim may be brought for breach of statutory duty under Section 42 of the London Transport Act 1963. Section 42(1) renders LUL (London Underground Limited) strictly liable to:- Make compensation for any damage caused to apparatus, etc (see 2.4 below).

2.2 Health and safety legislation

The Health and Safety at Work etc Act 1974 is primary legislation covering occupational health and safety in the UK. The Health and Safety Executive (HSE) are responsible for enforcing the Act and this has led to successful prosecutions related to tunnelling works, such as those following the Heathrow Express tunnel collapse in October 1994. The HSE (2006) subsequently published a document on the ‘Risk to Third Parties from Bored Tunnelling in Soft Ground’ which makes clear the difficulties that can, and have been encountered during tunnelling works.

The Construction (Design and Management) Regulations 2007 and 2015 (CDM) are Statutory Instruments introduced as a result of a European Union Directive (92/57/EEC) detailing minimum standards of Health and Safety at construction sites. The Regulations impose legal duties on Clients, CDM coordinators, designers, principal contractors, self employed and workers.

They are comprehensive in scope and effectively mean that a holistic view must be taken in Construction related activities. Whilst dealing specifically with ‘Health and Safety’ issues they must clearly have a bearing on construction impacts on utility apparatus as damage to almost any utility apparatus could lead to serious Health and Safety issues. This may apply not only within the curtilage of the site but to any other location that might be expected to be affected by the works (e.g. pipes beneath a nearby street).

2.3 The Water Industries Act 1991 (Ch. 56, Part vi, Ch. ii)

The Water Industries Act serves as an example of legislation in respect of a particular utility. Section 174 of the Act relates to the protection of Undertakers apparatus in general; it states:- 174 Offences of interference with works etc.

(1) Subject to subsection (2) below, if any person without the consent of the water undertaker—

(a) intentionally or recklessly interferes with any resource main, water main or other pipe vested in any water undertaker or with any structure, installation or apparatus belonging to any water undertaker; or

(b) by any act or omission negligently interferes with any such main or other pipe or with any such structure, installation or apparatus so as to damage it or so as to have an effect on its use or operation, that person shall be guilty of an offence and liable, on summary conviction, to a fine not exceeding level 3 on the standard scale.

The Act provides a general protection to Water Undertakers but is (as always) open to interpretation by the Courts and claims will be judged on the facts in each case.

2.4 Development Consent Orders (DCO)

Under the Localism Act 2011, the Planning Inspectorate (PINS; for England and Wales) became the agency responsible for operating the planning process for nationally significant infrastructure projects (NSIPs).

PINS examines the application and makes a recommendation to the relevant Secretary of State, who will make the decision on whether to grant or refuse development consent. Recently this process has been successfully applied to obtain consent for the construction of the Thames Tideway ‘Supersewer’ and the Silvertown Road Tunnel.

2.5 Legal precedent

The great majority of disputes between Developers (both private and public) and Statutory Undertakers are settled without formal litigation proceeding to a Judgement by the Courts. However, because most settlements contain a ‘confidentiality’ clause the construction community may not benefit from the arguments put forward or details of the events that led to the dispute and its resolution. There is not therefore an extensive body of relevant precedent to assist in dispute resolution although lawyers for both sides will usually cite precedent they believe supports their case.

There is however a seminal case that was pursued to Judgement in the High Court (Technology and Construction) which provides a highly relevant precedent in respect of tunnelling impacts on a Utility. The case was in respect of a claim for losses caused by damage to a large cast iron water main as a result of underground construction works for the Jubilee Line Extension at London Bridge. The Claimant was Thames Water Utilities Limited and the Defendants London Regional Transport and London Underground.

The claim was pleaded as three alternative causes of action, a claim for statutory breach under Section 2 of the London Transport Act (1963), a claim at common law in negligence for failure to use reasonable skill and care in carrying out and completion of the works, and in nuisance for removing and/or undermining ground support from the main as part of the construction of the works.

Judge Wilcox found that, in this case, there should be a single test of causation applicable to the same set of facts, irrespective of whether the case was pleaded in negligence, nuisance or statutory breach.

The crucial test was that Thames Water needed to show, on a balance of probabilities, that the LUL works caused an additional loading on the pipe and that this loading was a material cause of the failure; that is not de minimis (‘the straw that broke the camel’s back’).

The Judgement ( TCC/2004/2021.html) was handed down by His Honour Judge David Wilcox on 18 August 2004 and provides a highly informative document in respect of his deliberations and the application of the law. The Judge found for Thames Water on the issue of liability and causation and the Judgement was not appealed. A description of the legal proceedings and the detailed forensic investigations has been published (New, Knight and Ryan; 2005).

Whilst the outcome of this case was based largely on the specific evidence and analyses submitted by the experts appointed by both sides it is the analyses and the reasoning contained in the Judgement that are particularly helpful in guiding engineers as to the level of disturbance that might constitute ‘material cause’ in respect of ground movements and pipe strains.

3 Pipe failure and causation

3.1 Pipe failure by burst or leakage

Cast iron was originally developed in China during the 5th century BC and early pipes are known to have been made during the 17th century. The majority of grey cast iron (flake graphite) water and gas pipe supply networks were designed and installed beneath our streets during the latter half of the 19th century and the first part of the 20th. This chapter mainly considers the failure of cast iron pipes although the mechanisms discussed may be common in pipes made from other brittle materials such as unreinforced concrete and vitrified clay.

During 1976 and 1977 there were a series of serious gas explosions in the UK and their causes and recommendations were reported by Dr Philip King. His report led to extensive replacement of pipes in the gas supply network. More recently the increasing rates of water main failures gave impetus to efforts such as the Victorian Mains Replacement (VMR) programme in London for water supplies.

It must be remembered that when these networks were originally installed beneath our streets the most demanding loads likely to be applied were from horses and carts not the 44-tonne HGVs that they are routinely subjected to today.

The early pipes also remained relatively undisturbed by ground movements caused by installation and repair at pipe crossings and adjacent trenching. Such works are now ubiquitous in our cities and indeed it is rare to find a pipe that has not been subject to such activities. Also these old pipes are continuously subject to corrosion and their condition is naturally deteriorating with the passage of time.

During the 1950’s ductile iron (spheroidal graphite) replaced grey iron in many applications. In part this was due to its better tensile properties which enabled thinner wall pipes to be produced to meet required mechanical properties. Early ductile iron pipes were not routinely well protected from corrosion and being thinner walled than earlier grey iron tended to suffer from corrosion issues in certain types of soils (Water Research Centre 1986). This deficiency has since been rectified by applying various coatings and other protection to the pipes.

Pipes do not just fail at random. There is always a reason and given a sufficiently careful investigation the cause or causes of a failure can usually be identified albeit with varying degrees of certainty. When a failure occurs it is rare for a full and immediate investigation of all the circumstances to take place. There are a number of reasons for this which include:-

a) When a water main bursts it is common that the ground around and above the main is highly disturbed and often washed away. This means that hard inclusions creating localised loads or other issues related to the pipes surroundings are not easily traceable. It is rarely possible to identify the pipe bedding circumstances at a major failure location.

b) Utilities have a responsibility to restore supplies as quickly as possible and this can mean that the failed pipe remains are not carefully preserved. Indeed, it is common for larger pipes to be broken up in the ground to speed removal by the repair gangs.

c) A detailed investigation may add to the costs of the failure without the obvious incentive of financial return.

There are however certain circumstances when detailed investigations may be carried out:-

a) The Water Industries Act 1991 Section 209 imposes strict liability on water companies if a pipe failure causes loss or damage to Third Parties. However, the Act does not affect the entitlement which a water undertaker may have to recover contribution from others. Therefore, the failure of a water main which leads to extensive flooding and consequential losses may be investigated with such investigations promoted by Insurers with a view to recovery from others if warranted.

b) Because loss of supply to customers is an important performance indicator used by regulators and penalties may result from prolonged outages. This may not be the case where the failure can be shown to have been caused by a third party and beyond the control of the Utility.

c) On occasion the Utility may wish to have a detailed investigation(s) to inform renewal or renovation strategies. This will usually be where the condition of the pipe is suspected as being poor.

The condition (hence need for renewal/renovation) of the undertaker’s assets will be an important factor in negotiations with Regulators when securing future pricing.

It is self-evident that a pipe will fail when the load(s) imposed upon it exceeds its strength. In assessing the probability of failure it is therefore necessary to assess the strength of the pipe and what loads may be imposed upon it. In practice these questions are both very complex and in many circumstances the answers are unknowable for network assessments related to extensive underground works.

Experience has shown that pipes rarely fail precisely where calculations indicate failure was most likely. This is because failure will depend on numerous circumstances which when present in an unfavourable combination are sufficient to cause the failure: It may therefore be considered as a probabilistic rather than a strictly deterministic problem.

Consider cast iron pipeline failures across a water distribution network. They can happen at almost any time or place and the pipe will be subject to loading from various sources. These will normally include the hoop tension load due to the pressure within and ground and traffic loads. In order to be considered fit for purpose the pipe should be able to withstand these operational loads without failure: But failures do occur across the trunk and distribution pipe networks. Generally, these failures may be attributed to three groups of causes:-

1 Third Party intervention. Perhaps the most common of these would be direct impact during road excavations. Others are ground movements, vibration, abnormal surface or other loads.

2 The pipe was weaker than expected. This may be due to corrosion, or casting defects (incorrect section, wormholes, porosity etc.).

3 Operational overload by over pressurisation or surge. This is likely to be rare given a pipe in serviceable condition.

3.2 Types of failure and Causation

This section illustrates the fracture patterns associated with the most common types of failure and the causes usually associated with such failures. Broadly these fall into the following mechanisms:-

a) Localised failures due to point impacts, point loadings and corrosion (see figures 3, 4 and 5). Occasionally failures due to corrosion may occur over larger sections of the pipe frequently at invert level.

b) Longitudinal deformations resulting in flexural bending (Figure 6)

c) Tensile stresses in the transverse plane. Crushing due to excess loading or splitting due to internal pressure usually with corrosion.

d) Socket bursting resulting from joint rotation and/or shearing.

e) Spiral failures believed to be caused by a combination of longitudinal flexural and transverse loadings

Localised failures may occur in any size of pipe. Flexural failures are most common in pipes of 300mm or less and larger pipes tend to be more vulnerable to crushing failures. Spiral failures are most commonly found in pipes between 300mm and 600mm in diameter. Socket bursting by joint rotation may occur in all pipe sizes.

It is accepted that ground movements and changes in ground loading can, and do, give rise to potentially damaging strains in pipelines. These additional loads may go unnoticed or result in catastrophic failure and it follows that the probability of failure in the vicinity of construction works causing ground movement is increased. The requirement is to identify particular areas or locations where the probability of failure is significantly increased by the presence of construction activities and to set these into the context of risk assessment.

3.3 Masonry

Many early trunk sewers were constructed in brickwork and took various geometrical forms (e.g. circular, egg, horseshoe, arched crown and invert with vertical sidewall). Analyses of such structures can be complex and the reader is referred to the classical texts of Heyman (1982 and 1995) and Szechy (1970) for further reading. There are now various software packages that implement arch theory to assess the mechanisms of likely failure but a basic understanding of arch theory is likely to be helpful in carrying out analyses for assessment purposes.

Masonry arch analyses are often confined to assessing the load bearing capacity of the arch and therefore the transverse stability is the primary consideration. However, sewer arches are semi-continuous linear structures that are potentially vulnerable to the longitudinal ground movement curvatures which may be caused by tunnelling beneath in a similar way to pipes.

For assessment purposes the masonry is considered to have negligible tensile strength therefore, when calculating longitudinal flexural strain, the neutral axis should be taken at either the crown or invert extrados depending on whether the ground movements produce hogging or sagging. Tensile strains are therefore developed at crown or invert and the magnitude of the strain considered ‘negligible’ is discussed below. Hogging strains produce tension in the crown which can lead to loss of mortar, then bricks and subsequent collapse.

Figure 15, page 31 shows twin sewer barrels under construction believed to be during the 1920’s. Note that because of their 9 feet span these comprise three brick rings whereas most smaller sewers would be of two rings. Note also the brick infilling of the spandrels which greatly strengthens the structure. It is interesting that this form of construction had little changed since Bazalgette’s works in the 1860’s.

Arch failures are usually associated with some form of structural overload producing tension in the masonry and this is exacerbated where the loads are asymmetrical. Masonry tends to be very strong in compression but is very weak in tension. Critically it is not likely to be the strength of the brick but that of the mortar that determines failure.

Arch design seeks to resolve the loads upon the arch into compressive stresses and thereby eliminate tension. Figure 16, page 31, shows a relatively complex analysis of a five-barrel sewer and illustrates various forms of arch distress. The key issue is to retain the thrust line within the centre one third of the arch thickness where possible as simple statics determines that in this circumstance there will be no tension around the arch. For stability the thrust line should also cross the neutral axis of the arch at least twice. The thrust line should remain within the masonry at all times. Analyses are a matter for statics and where tension is not allowed the limiting factor will be the limiting compressive strength of the masonry. A common mistake is to use proprietary numerical models (FE or similar) and to assume linear elasticity in both compression or tension although some packages may allow a negligible tensile strength to be an input condition.

In practice it is likely to be a loss of mortar or bricks within the crown of the arch which results in failure. Loss of masonry integrity can quickly result in collapse as bricks fall from the crown or elsewhere between the springings and the arch is no longer functioning. Where significant tunnelling-induced movements are expected it essential to inspect the masonry structure and fully repair any defects particularly towards the crown before the works take place.

For assessment purposes some guidance is required as to what level of strain in masonry might be considered as ‘negligible’ and a value of 500 microstrain (0.05%) has become widely adopted. This value comes from work carried out at the UK Building Research Establishment (e.g. Burland and Wroth 1975) and elsewhere and largely relates to tests carried out on vertical walls and the onset of visible cracking at so-called ‘e crit’. It is important to note that the stability of these walls did not depend on the tensile capacity of the masonry but brick arch stability does.

For masonry walls the onset of cracking does not necessarily represent a limit of serviceability but for brick arch sewers it should not be assumed that this is the case. Burland and Wroth point out that ‘There is a great deal of evidence to suggest that tensile ‘failure’, ie loss of tensile strength, occurs at much smaller values of tensile strain than those causing visible damage’.

Masonry sewers are ubiquitous beneath the roads of our cities and are subject to high repetitions of large loadings from heavy vehicular traffic and vibration. A rolling axle gives rise to rotation of principal stresses and such loadings might be expected to produce a fatigue type racking motion within the masonry and failures of these sewers is not uncommon.

Masonry sewers were generally built to sustain gravity flows without surcharge. As our cities have grown the increasing volumes of sewage passing into the network can produce surcharge pressures. This is highly undesirable as it can result in unacceptable hoop tension in the masonry and result in failure. For assessment purposes therefore it is commonly assumed that the ‘design’ pressure within a sewer is taken as the hydrostatic pressure to the levels of the controlling nearby manholes or other sources of pressure relief.

An additional check on masonry stability is applied by consideration of the confinement/pressure ratio (CPR). That is the ratio of the compressive stress in the masonry exerted by assuming the ground load at the sewer axis to the hydrostatic pressure within. For general guidance it is considered that CPR’s in excess of 1.33 are likely to be acceptable given a sewer in good condition and absent other significant destabilising loadings. A CPR check is also likely to be required for materials vulnerable to surcharge pressures such as unreinforced concrete.

This is similar to the acceptance criteria applied to unbolted water transmission tunnels where CPR’s in excess of 1.33 under surge pressure and 1.5 under operating pressure have been found to be acceptable.

A common requirement in cities is to develop underground space wherever possible. This can often result in major excavations (for basements, stations, ticket halls etc.), which produce potentially damaging unload of confining hoop stresses in both sewers and water tunnels. These may not be acceptable to water Utilities and developers should always consider this at the earliest stage when planning their infrastructure and buildings.

4 ‘Impact strain’ from ground movement

4.1 Definition and constraints.

For the purposes of assessment, the ‘impact strain’ is defined as the strain arising from the ground movements or other loadings caused by the construction works. It does not include strains arising from any other cause i.e. pre-existing ground movement, ground load, operating pressures or normal vehicular loading.

It is not the design strain as classically described by others (see bibliography in Young and O’Reilly 1987) but is specifically that part of the apparatus strain attributable to the works under consideration. This ‘impact strain’ may be that predicted during the assessment process or back-calculated from field measurements carried out during the works.

There are many difficulties in the calculation of impact strain and the results should usually be regarded as indicative only as key factors impacting the pipe performance are likely to be largely unknown. The strain calculated should be compared with the criteria advised by the relevant Utility and will inform but not necessarily dominate the overall assessment of risk.

In an urban situation critical difficulties may include:-

i) Pipe details (e.g. material, condition/defects, size, line/level)

ii) Ground conditions (e.g. geo-environment, bedding, other loads)

iii) Pre-existing strain (e.g. due to historic cross-trenching, deep basements, tunnelling) iv) Complex/intractable ground movement calculations

v) Soil/structure interaction (as iv)

4.2 Key assumptions

For stages 1 and 2 of the assessment process (see Section 5) it is conservatively assumed that:-

i) In longitudinal flexure (bending) the pipe moves with the ground

ii) Either: The joints are stiff compared with the pipe so limiting strain is in longitudinal bending of a continuous linear tube or the joints are fully flexible compared to the pipe so the limiting condition is in joint rotation. Analysis should be carried out for both conditions.

iii) For the calculation of joint rotation (absent site specific information) metal pipes are 12ft (3.66m) long.

iv) Pipe strain is calculated at the pipe extrados and for metal pipes the neutral axis is at the pipe geometric centre. For materials of low tensile strength (e.g. masonry, unreinforced concrete, vitrified clay), the neutral axis is at the extrados.

v) ‘Green field’ ground movement is assumed.

4.3 Calculation of ground Movements

4.3.1 Tunnels

The great majority of utility pipelines are relatively close to the surface, say at a depth to axis of less than 2 metres. For the calculation of green field tunnelling induced movements (displacement, slope, curvature and strain), the required derivatives of Gaussian equations (O’Reilly and New (1982) reprinted in 2015), New and O’Reilly (1991), Leca and New (2007)) are straightforward to apply and have stood the test of time in their application. There is no body of data known to the author that would gainsay their general application for scoping the initial assessment of near surface pipelines.

For movements in stiff cohesive materials at depth the ‘ribbon sink’ assumption made by New and Bowers (1994) may be usefully applied particularly in the vicinity of the tunnel under construction. The equations provided also allow threedimensional movements from complex tunnel geometries to be calculated.

Given the uncertainties in the input parameters given in section 4.1 above it is the author’s opinion that for preliminary route wide assessment purposes more complex models are likely to be unnecessary.

The advent of slurry shield and earth pressure balance tunnel boring machines (EPBMs) has greatly reduced the ground movements due to tunnelling. For the ‘volume loss’ (Vl) value applied to most recent contracts in London the assumption of a 1% loss in London Clay is deemed as ‘moderately conservative’ for such EPBM’s. Similarly, the trough width parameter (K) is usually taken as about 0.5.

The predictions made using the Gaussian models referred to above are of themselves neither conservative or un-conservative and it is a matter for assessors to determine and justify the parameter values used in their assessment. These parameters may vary considerably particularly where onerous ground conditions might be expected/encountered.

4.3.2 Shafts, deep basements, and boxes

Shafts are commonly associated with tunnel construction and can give rise to significant ground movements. New (2017) has proposed a generic equation developed from the original predictive equations of New and Bowers (1994) and an extensive database of measurements taken in London.

The proposed generic equation for settlement Sd at a distance d from the shaft wall is:-

where Sd is the settlement at a distance d from the shaft extrados. The settlement at the shaft wall is taken asH where  is an empirical constant dependent on ground conditions, type of wall and construction method, and H is the shaft depth.

The variable n is a simple multiple of the shaft depth to a distance d from the shaft extrados where settlement tends to zero (For the original New and Bowers equation this requires n=1). The values for n and  are to be determined (by reference to appropriate empirical data) by assessors and will be dependent on ground conditions, construction method, and shaft size. For deep basements and boxes it is suggested the CIRIA C760 is consulted. Total ground movements are simply obtained by superposition as appropriate.

4.4 Longitudinal (bending) strain

Assessment of longitudinal bending (flexural) strain () is calculated from the classical beam equations and is a function of the pipe extrados diameter (D) and the longitudinal radius of curvature (R).

Bending strain at pipe extrados

 = D/2R for a pipe where the neutral axis is at the centre, or

= D/R for a pipe where the neutral axis is at the extrados

The longitudinal curvature (R) is calculated as shown in figure 17, page 32. The red line represents the resultant flexural displacement of the ground (and pipeline) and i the interval between data points. For most practical applications the value for i is taken as 1m but where rapid curvature variations occur this may need to be reduced.

4.5 Joint rotation

Section 3.2 d, and figures 10, 11 and 12, illustrate impacts of joint rotation in bursting open sockets. Although this failure mechanism can occur in pipes of any size it is an important cause of failure in larger diameter trunk mains. For mains in excess of about 12in (300mm) nominal size the strength of the pipe in bending is very large and pure flexural failure is rare. This high bending stiffness can however result in the very high bursting forces generated as the spigot rotates within the socket.

In preparation for the trial described in section 2.5 above the Claimant carried out a substantial investigation to determine the circumferential strain induced in the socket by joint rotation. Two pipe lengths together with their jointing held intact were carefully excavated from the pipeline close to where the socket burst and had occurred.

The pipes were placed in a testing frame in the structures laboratory at the Transport Research Laboratory (TRL). The pipe was extensively instrumented with strain gauges and the pipe rotation measured by an array of LVDT’s (linear voltage differential transformers). Figure 18, page 32, shows the test rig during preparation at TRL. A hydraulic ram mounted on the loading frame applied an increasing pressure onto the wooden saddle via a load cell which caused relative rotation of the pipes.

In summary the results of the research indicated that at a joint rotation of about 0.1 degrees the tensile hoop strain was about 100 microstrain. This finding informed future decisions on allowable joint rotation that are consistent with the strain criteria for cast iron given in Section 5.4. An interesting finding was that even when the loading was removed the tensile strains in the socket remained locked in and barely reduced during the course of the trials.

Assessment of the joint rotation () is usually carried out by tracking two pipes along the calculated flexural displacement profile with the joint and the left and right end of the two pipes touching the flexural displacement profile until the point of maximum rotation is found. For a simple Gaussian tunnel settlement profile this will occur when a joint is in the centre of the trough directly above the tunnel axis. The joint rotation will depend on the pipe length which, in the absence of specific information is often assumed to be 12 feet (3.66m)

4.6 Transverse (crushing, bursting or shearing) Strain

To approximate the maximum transverse tensile strain in a pipe subject to uniformly distributed vertical and horizontal loads the following analysis may be applied.

Consider that the pipe is subject to a load uniformly distributed both vertically and horizontally. The vertical stress  being defined as that at pipe axis level and the horizontal defined by the ratio of horizontal to vertical stress K0 (see figure 20). The maximum bending moment M in the pipe wall (at the quarter points) is given by Young and O’Reilly 1987 (equation 14 and figure 9(c)) as

Where W is the total vertical load per unit pipe length (assumed to be on the crown and invert sections of the pipe).


By substitution for W

For a more general and rigorous solution see Roark (1989) Table 17 Case 10 and apply superposition.

From the beam bending equation M/I = 2 E / t where E =  and I = t3/12 per unit length

Where E is the Young’s modulus (typically 80 GPa for cast iron pipe), t = pipe wall thickness

Therefore, by substitution the strain at the quarter points

The above analyses may be applied to ground loads or other uniformly distributed loads. For other loading conditions such a crane outriggers or heavy plant the ground stress at the pipe axis may be approximated by Boussinesq based analyses.

 4.7 Axial (tensile) strain Tunnelling-induced ground movements also produce tensile horizontal ground strains which, when added to bending strains, can often be very significant when making assessments. Even when taking into account ‘reduction factors’ as indicated by Attewell, Yeates and Selby (1986) these ground strains may lead to assessments of a conservative nature.

It must be noted that Attewell et al (1986) effectively acknowledge that the transfer of axial strain into a pipe is highly questionable because they do not allow a similar transfer of compressive strain. Their view is that ‘Fixity conditions are crucial in determining whether tensile or compressive forces are induced…’ The Poulos and Davis calculations relied upon by Attewell et al assume full bonding between pile and soil but for relatively shallow pipes this is highly uncertain (or even unlikely) in practice.

Following review of field experience and the theoretical methods the author has concluded that in the context of Stage 1 and 2 assessments that horizontal ground strains may be regarded as negligible for many types of pipe at relatively shallow depths. The reasons for this conclusion are given below:-

i) The author is unaware of any body of field research or other case history evidence that indicate that significant tensile horizontal ground strains are transmitted to pipes subject to tunnelling induced ground movements.

ii) The basis for the theoretical model suggested by Attewell et al is that developed by others (e.g. Poulos and Davis: 1980) for vertical piles at depth.

The problem with this is that the great majority of pipes are relatively shallow and therefore not subject to the substantial axisymmetric ground pressures encountered at pile depths.

iii) The tensile strain transmitted to the pipe is treated as an elastic soil/structure interaction. However, the movement of the soil as calculated from a Gaussian model is not elastic.

iv) Frictional force transfer at the soil/pipe interface and the effective shear strength of the soil lessen until, when close to the surface, no significant axial tension can be transmitted to the pipe.

v) Many pipes have lead run joints and if significant tensile axial strains are present the joints will tend to ‘slip’ or the lead deform plastically thereby abating axial tension. Other pipes with rubber or similar joints will also relieve any axial tensions.

It is therefore the author’s conclusion that the additional complexity of attempting to include axial tensile strain is likely to be conservative and may be misleading given the uncertainties in the rest of the ground/pipe movement analyses as described above. However, it may be that, in some circumstances, assessors may still wish to include axial strain and that will remain a factor when considering risk particularly to pipes of critical importance.

It must also be noted that axial ground strain may induce relative movement between pipelengths resulting in joint pullout (see Attewell, Yeates and Selby:1986).

5 Assessment for planning and design

5.1 Developer and utility perspectives

Developers employ Designers and Contractors who naturally seek to fulfil the objectives of their Clients in the safest, most expeditious and cost effective manner. The presence of utilities may present direct obstructions to proposed works, require assessment studies, and possibly mitigation or diversionary works. Such activities will give rise to additional project costs which occasionally may be substantial.

Utility requirements may often be simplified and defined under two headings:-

1) The Utility must be in compliance with its Statutory and Regulatory obligations: In particular, there should be no loss of serviceability to customers or damaging impacts to third party stakeholders.

2) There should be no loss of whole life asset value: That is the value of the Utility apparatus should be preserved to protect the interests of their stakeholders.

In order to achieve these broad objectives, Utilities need to know the likely risk to their apparatus posed by the proposed development and how that risk might be removed, minimised or otherwise mitigated. The obligation to do this will normally fall upon the Developer but assistance will be required from the Utility in identifying, locating and specifying the material condition of the asset at risk.

The Utilities will also inform the decision making process by stating the potential consequences of damage to their apparatus. Utility engineers and their consultants will review Developers proposals to provide high level assurance to senior managers within the business.

5.2 Preliminary considerations

A very important part of the planning process for tunnels beneath cities is the choice of line and level for the preferred route and the setting of limits of deviation where applicable. An early part of this process will involve an obstruction survey to identify physical obstructions such as other tunnels, pipelines, sewers, deep basements, piles and natural obstructions related to the geology.

In its broadest sense an ‘obstruction’ may be anything that might affect the new tunnel or its potential impact on third parties. Early review of ‘obstructions’ can lead to changes in line and/or level to remove or at least reduce the impact of the works. At this stage requirements for diversionary and protective works for utility apparatus will be considered. Diversionary works for gas and water supply can usually be accommodated but can be costly. Large sewers however are particularly difficult to deal with because their levels are likely to be inflexible so as to retain gravity operation.

5.3 Components of pipeline risk assessment

The assessment of risk to pipelines is often a complex matter comprising numerous factors many of which may be uncertain or un-knowable. Some of the information required will be essential to any assessment but assessors will often have to make informed assumptions in regard to many details.

A Asset Properties and Conditions
i) Where is it (alignment and level)?
ii) What is it (gas, water, electric, telecoms, etc)?
iii) Material (e.g. iron (grey or ductile), PE, clayware, brick, concrete, GRP) and age.
iv) Size (outside and inside diameters, pipe section length) and jointing (lead run, cementitious, flanged, welded, or flexible).
v) What is its present condition (repair and maintenance history)? vi) What, if anything, is known of the current state of strain?
vii) What is the geo-environment?

B The proposed works and resulting ground movement
Tunnel, shaft, deep box, geometries and construction method. (Final design and temporary works). Calculation of predicted ground movements and derivatives.

C Consequences of damage Strategic, trunk or distribution main?
Failure to meet statutory service requirements?
Loss of whole life asset value?
Damage to third party assets?
Financial and reputational impact to the Utility’s business?

D Control and preparedness
Mitigation (replacement, diversion, strengthening)
Emergency preparedness/ contingency planning Network resilience

E Assurance leading to consent
Has the appropriate level of analysis and checking been carried out?
Is the Risk ALARP (As Low As Reasonably Practicable)?
Response from Utility:- (‘Approval in Principle’, ‘no further comments’, or ‘LONO’ (letter of no objection))

F Objection
Further negotiations with Developer Court injunction (the last resort) 5.4 Assessment guidance and Criteria

It is recommended that all Developers contact the relevant Utilities as early as possible in their planning and design processes. Early consultation will often avoid abortive work and the consequential delays and unexpected costs to projects. Utilities may advise Developers and give guidance on acceptable procedures and criteria that may be applied to indicate impact on their apparatus. For example, Thames Water offer ‘Guidance on piling, heavy loads, excavations, tunnelling and dewatering’ on their website at https://

Tables 1 and 2 are taken from that document and provide a starting point for the Developers impact assessment process. The criteria are for guidance only and represent changes below which the risk of significant damage may be considered negligible for apparatus in average to good condition (they are not pipe design criteria). The impact criteria actually applied may vary according to circumstances and are a matter for the Developers’ designers.

Useful guidance on the assessment of strain in cast iron pipelines is given by Attewell, Yeates and Selby (1986) and Bracegirdle et al (1996).

6 The staged assessment process

This section develops the current assessment process to a more risk based approach which may better reflect the balance between the likelihood and consequences of an adverse event and the ability to recover during and after the works.

Tunnel and shaft construction techniques have improved greatly over recent years and this has effectively changed the risk balance between likelihood (based on lower predicted ground movements) and the consequences which have remained essentially the same. It also seeks to minimise risk by improving preparedness so as to react promptly and proportionately to an event to contain and mitigate any losses.

Tunnelling impact assessments for buildings are commonly carried out in three stages: These stages are described by Mair, Taylor and Burland (1996) and are not repeated here. A similar, but not the same, three stage approach has been retained for utility apparatus.

6.1 Stage 1 assessment scope and information Retrieval

Stage 1 is a preliminary assessment that seeks to identify all utility apparatus that may be affected by the proposed works. The usual method to achieve this is to apply a filter based on green field settlement estimates derived from conservative assumptions of likely ground movements (see section 4.3). For instance, apparatus outside estimated settlement contours of less than 1mm might be excluded from further consideration. Commonly when consent is given for tunnelling the acceptable variation in line and level is specified in terms described as ‘limits of deviation’. The determination of the zone of influence should consider the worst case locations within these limits until the final location is fixed.

This has been the approach applied to tunnelling for most structures (buildings, bridges, tunnels and pipelines) along a tunnel route. It has the advantage that third party asset owners can easily relate to ground settlements expressed in millimetres. The difficulty is that settlement, of itself, does not damage structures. For example, in some mining areas very large settlements occur but do not damage overlying properties. Occasionally ‘slope’ or tilt is related to building damage but again it is not necessarily damaging.

For most structural damage assessment, the critical parameter is ground curvature and for buildings this is commonly interpreted by calculation of ‘deflection ratio’ (Burland 2001). However, pipelines should more generally be considered as continuous linear structures (rather than short beams) or at least structures having a large aspect ratio (length/diameter): For assessment purposes the critical ground movement parameter dictating longitudinal bending (flexural) strain and joint rotation is curvature (1/R) not deflection ratio.

For Stage 1 pipeline assessments it is therefore suggested that contours of settlement on plans of the route are supplemented by contours indicating minimum radius of curvature (R) or maximum curvature (1/R) (see figure 21). Curvature is the second derivative of flexural displacement. It varies not only with distances away from the tunnel but it also depends on the angle intersected between the tunnel alignment and the line of assessment. This is simply calculated and plotted routinely using software available to most civil engineering consultants. Note that in Figure 21 the pipe direction is normal to the tunnel axis and therefore the curvature represents a maximum and is calculated from the settlement profile.

Review of these contours may significantly reduce the number of utility assets that require further analyses and highlight those most likely to need mitigation at the earliest stage. This can be very important where major mitigation or diversionary works are required as they can take an extended period to arrange and implement.

The stage 1 process should also begin the gathering of the information needed to carry out the subsequent stages of the risk assessment. There are two main areas that will have greatest influence on the subsequent risk assessment process:-.

6.1.1 Likelihood of damage

The Utility should be requested to provide all the information that it has available regarding the type and location of its apparatus and also its repair and maintenance (R & M) history. Settlement and maximum curvature contours should be mapped onto plans showing the location of the utilities apparatus within the zone of influence of the works. Figure 21 gives a simple example of a section of tunnel with contours of settlement and contours showing the maximum sagging radius (in kilometres) in red and the maximum hogging in blue. The curvatures are for the worst case where the pipeline (green line) runs normal to the tunnel direction.

Table 3 shows strains for various imposed ground movement curvatures (1/R) and highlighted in orange and red are pipes that would exceed the guide ‘negligible’ criteria and which would be considered in more detail at Stage 2.

The calculation of settlement contours supplemented by contours of maximum curvature provides the zone of influence of the proposed works and initial guidance as to apparatus that might be impacted and an indication of severity. So far this is essentially a paper exercise but in some circumstances it may necessitate the opening of trial pits to determine the apparatus location and material, particularly where this is uncertain for strategic assets.

A critical issue which determines the vulnerability of pipelines is its condition and pre-existing state of stress. Many pipelines will have been in the ground for over 100 years and will have been subject to historical loadings and other sources of ground movement producing locked-in strains that are effectively unknowable.

For some major tunnelling projects attempts have been made to determine the condition (extent of corrosion, etc) by in-situ testing of sections of pipe typically 1 metre long. Water companies use such testing to evaluate the state of their pipe networks usually within a particular supply district metered areas (DMA’s). The number of sections tested is, or should be, determined by ensuring that a statistically significant number of sample lengths are investigated. Results may then be used to produce information which, when taken together with leakage data (sometimes referred to as ‘night line flow’ or ‘unaccounted for water’), will inform decisions on pipeline renewal.

However, while of value for statistical network analysis purposes it is the author’s experience that such testing rarely produces a reliable indication of the vulnerability of particular pipelengths within the ‘zone of influence’ of tunnel construction. The testing carried out basically measures the extent of corrosion induced loss of section (usually by graphitisation and/or tuberculation) of the pipe walls but provides no indication of pre-existing stress conditions in the pipe. This is because failures by corrosion (and/or high pre-existing stresses) often occur at random locations within long sections of pipe which are in otherwise very good condition. The chances of locating specific areas of particular vulnerability by sample testing are therefore low.

Water companies also routinely collect and analyse data on ‘night line flow’. Records are taken in the early hours of the morning when flow in the network is likely to be minimal and is used to indicate DMA’s where leakage could be a problem. Strictly ‘night line flow’ is more properly referred to as ‘unaccounted for water’ but nevertheless would be expected, in most cases, to reflect the condition of the pipe network in a particular DMA.

The author therefore suggests that an improved method of assessing apparatus vulnerability is to analyse the repair and maintenance (R & M) history and the night line flow records held by the Utility. This has the advantage that particular areas of pipeline may be considered holistically for susceptibility to failure by burst or leakage and this should better reflect the overall sensitivity to new ground movements and other loadings.

Knowledge of night line flow before and after the works will also be helpful in indicating whether works may have increased leakage which otherwise may not be apparent.

6.1.2 Consequences of damage

A critical part of the assessment process is to obtain early utility advice on the importance of their apparatus within the zone of influence. This information will be very important when evaluating risk in terms of the ability of the utility to perform its statutory service obligations and the consequences of damage to third parties.

Pipe size can be a good indication of its importance to the supply of water in a particular area. However, resilience measures such as alternative supply routes within the network can considerably reduce this. Sometimes large diameter pipes are less important than they might appear because alternative supplies have become available since, for instance, the building of the Ring Main tunnels in London.

Consequences of pipe failure are not necessarily related to pipe size. For instance, a water supply to a hospital or fire mains may be in a relatively small diameter pipe but nevertheless safety critical in importance.

6.2 Stage 2 – More detailed analysis

A second-stage assessment will consider the apparatus not eliminated by the Stage 1 preliminary assessment. For pipelines the object of this phase is usually to estimate the strain, joint rotation and pullout for moderately conservative assumptions in respect of ground movements and assuming that the pipeline moves with the ground (see section 4.3). This will usually include charts showing resultant flexural pipe movement and derived charts of curvature and consequential strain along the particular line of the pipe under consideration (see examples shown on Figure 22, and Figure 23, below).

These results will then be compared with the criteria provided by the Utility. Apparatus that fails the comparison with the utility criteria will be identified and further considered in more detailed Stage 3 risk analyses.

Stage 2 should identify apparatus that is likely to be particularly vulnerable as identified by a poor repair and maintenance or night line flow history as this may suggest that it may be subject to significant corrosion or has been subject to large strain producing events in the past.

Stage 2 must also identify apparatus that is particularly important and the consequences of its failure. These will include health and safety issues, major financial impact due to damage (e.g. flooding or explosion), or inhibit the ability of the Utility to fulfil its statutory or commercial obligations.

6.3 Stage 3 – The evaluation of risk

A more detailed third-stage assessment will be applied to apparatus that fails to meet the acceptable disturbance criteria advised by the Utility. At present this may take the form of a more detailed soil/structure analyses. However, it is the Author’s experience that because of the difficulties and uncertainties inherent in such a process a more risk/remedy based approach might be both more effective and simpler to achieve.

It must be accepted that all tunnelling involves risk. For Utilities, risk is a complex a function of probability (likelihood of occurrence) and consequence both of which may be difficult to assess for the reasons discussed above. The question immediately begged is what risk is acceptable and to whom? Those are difficult questions which may be answered in part by the application of the ‘ALARP’ concept which may be simply stated to be that the risk shall be ‘as low as reasonably practicable’.

The difficulty here is that what one party may regard as ‘reasonable’ the other may not. That will depend on the perspectives of the tunnelling promoter and his advisors and the Utility and in practice these may be quite different (see above). Both groups will want the works to be carried out safely and without incident but the Utility will be focussed on preserving its assets and ability to meet its obligations and the promoter on completing the works on time and to budget. Both are reasonable positions but may nevertheless result in conflict.

For tunnelling and construction work in general risk assessments and method statements (RAMS) are required and are often prepared by the Contractor as one is dependent on the other. That often means that there is likely to be late consideration of the temporary works necessary and in some cases the permanent works if revised. In practice the Contractor best understands the proposed works and the Utility the consequences to their apparatus and third parties. It is therefore a document that will be optimised if fully informed by both parties.

In many cases the risk analyses are something of a hybrid between deterministic and probabilistic in nature. At their worst risk analyses take the form of a ritualised recital often handed down from earlier projects sometimes with the benefit of some project specific concerns pasted in. They are often long on PPE (personal protective equipment) requirements and short on even mention of the consequences of utility damage.

A common approach is to attach a number from 1 to 5 as to the possible hazard and how likely it is to occur, and to multiply this by another number from 1 to 5 reflecting the possible consequences. These numbers are often based on highly uncertain and unresearched ‘engineering judgements’ and their product used to inform all as to the extent of the risk. It must be accepted that the problem is a difficult one but they can sometimes amount to little more than best guesses.

An important purpose of risk analyses for utility apparatus is to establish the need for diversionary or mitigation works. These works of themselves may involve substantial risk to life and/or property and it is important that this is taken into account and a holistic appraisal is achieved. They can be highly disruptive and require extensive traffic management with the consequential inconvenience to the public.

It is not possible to be prescriptive in a detailed sense as to how the overall risks are best assessed. However, guidance based on current best practice, is possible. The process must be collaborative between the teams representing the promoter of the works and the impacted Utility. For instance, the Utilities will require input from their operations teams to assess the impact on the networks of damage, either during works or before, if intrusive mitigation or diversionary works are likely to be required.

All will agree that health and safety is paramount both to the utility, construction workers and third parties.

There will be occasions when there is a requirement to divert apparatus due to physical clash conflicts. For most pipelines this is usually possible but may incur substantial costs and yield risk of its own.

Where damage to utility apparatus is anticipated as a result of ground movements informed by condition (R&M history and leakage indicators) there are various approaches to the problem. These will need to consider the importance of preserving the serviceability and whole life asset value of the apparatus, the network resilience, and the option of doing nothing and repairing during the works or after.

Consequence considerations will involve both H&S (e.g. hospital supplies, fire mains, flood, explosion) and commercial losses (e.g. loss of supply to industrial premises or a wide domestic area). It is the Author’s experience that the only realistic solution to deal with the risk is by close contact between the Utility and the developer to arrive at an agreed solution that is truly ALARP and takes a holistic view of risk.

A major risk to the assessment process is that due to incorrect analyses (either in concept or application) and the failure to pick up errors during the formal checking and approval of reports. Due to competitive pressures it sometimes occurs that reports are prepared by relatively inexperienced or junior staff. It is therefore vital that all ‘checkers’ and ‘approvers’ must ‘value their signature’ and take full responsibility for the content of the reports they sign.

It has become common practice to resort to numerical models relatively early in the stage 3 process. There are applications where this can be helpful but for assessments these are likely to be rare. Where numerical models are used they should be justified as necessary and corroborated by closed form solutions with case history data where possible.

6.4 Mitigation

A detailed discussion of mitigation methods and the risk of such works is beyond the scope of this paper but some general guidance is appropriate. There are numerous engineering solutions to mitigate risk to pipelines including pipe replacement and slip-lining. This will create risks to workers and such solutions will often involve traffic diversion and intrusive traffic management which of themselves will create additional risk to all including the public.

For other structures, such as large brick sewers, mitigation may require permanent structural lining or temporary support by steel ribs and mesh. Either may require substantial periods of man entry and that introduces further risks to the project. Recent experience has been that mitigation works can create serious risks and that where it can be safely applied a strategy of doing the minimum to preserve the serviceability of the assets may be the preferred option.

As a minimum ‘mitigation’ of risk to water supply pipelines and tunnels should include:-

i) Location and exercise of all valves controlling the flows

ii) Review of impact models (this may include review of possible flood zones and other possible consequences)

iii) Monitoring of the ground/pipeline movements

iv) Pre and post works inspection by man entry (where possible) or CCTV or ‘smart pigs’

v) Inspection at critical phases of the works where necessary

vi) Monitoring of leakage

vii) Preparation of a robust EPP (Emergency Preparedness Plan)

viii) Local operations staff notified of the works

The EPP should include an incident recovery plan that will allow rapid repair of a pipe failure. In particular R & M contractors should be made aware of the works and have emergency spares and equipment immediately available for repairs.

6.5 Monitoring, control, and validation of design Assumptions

An essential corollary to the impact assessment process is that the works are carefully monitored during construction and that the monitoring is previously agreed with the Utility. Such an agreement will include the location, type and frequency of data collection, analysis and actions. The results will provide ongoing assurance to the Utility and will be used to inform measures controlling the works. Monitoring is to confirm that the design assumptions relied upon by the assessments are valid or that the disturbances are similar to, or less than, those estimated. If the design assumptions are found to be incorrect then the results of the monitoring will be used to terminate and/or modify the construction/mitigation processes so as to retain the integrity of the apparatus and ensure an uninterrupted supply to customers.

It has become practice that monitoring will be continued until significant ground movements have ceased: For some recent major tunnelling projects this has been interpreted as when settlements have lessened to below 2mm/year.


This paper documents methods and strategies that may be applied to the assessment of Utility apparatus subject to impact from subsurface construction activities: It is based on wide experience of the processes adopted during works in London during the past 20 years. It is not intended to be prescriptive but to inform, where appropriate, engineering assessments made by Consultants on behalf of their Clients. Any guidance suggested is necessarily general as each development will be different and likely to require specialist individual appraisal. The paragraphs given below are presented as a summary of the more significant findings and, where possible, to draw conclusions from them:-

a) Forensic investigation of numerous pipeline failures has provided a good understanding of failure mechanisms and the very presence of various Statutes often assists in the resolution of disputes between Utilities, Developers and other third parties. Unfortunately, much of this information is unpublished because most disputes are settled out of court and a common stipulation of such settlements is that they should remain confidential to the Parties.

b) However, in a seminal judgement the High Court proceedings described in section 2.5 have been most helpful in defining ‘material cause’ of loss in respect of subsurface works and pipeline failure. Albeit indirectly, these proceedings have lent support to damage criteria which have become widely applied.

c) Tunnelling and associated construction techniques have improved considerably in recent years and the likelihood of damaging ground movements has diminished accordingly. Many old cast iron pipelines have been replaced by polyethylene pipes which, when properly installed, are extremely resistant to damage. However. London still relies heavily on Bazalgette’s (Bazalgette 1865) brick sewers and iron works for the disposal of sewage and his paper on the works is still a valuable record of construction details.

d) Nevertheless, that part of the pipeline infrastructure that has not been renewed is steadily deteriorating by corrosion and ongoing disturbance by nearby trenching, tunnelling and deep basement construction. It is therefore important that Regulatory authorities allow Utilities to undertake reasonable renewal programmes and that a holistic view of the positive impact of these works is properly accounted for in their review of strategic business plans.

e) The benefits of renewal are not only the reduction of leakage and reduction of failures but also the longer term mitigation of risk and costs arising from repairs and consequential traffic and pedestrian inconvenience. Renewal can also offer Utilities the opportunity to rationalise and improve the resilience of the networks ther y improving reliability of supply.

f) In recent years the extensive tunnelling works beneath London has stimulated the development and revision of pipeline assessment processes. In particular, the move to an improved review of risk has indicated that a less intrusive mitigation strategy is likely to be of benefit. In particular, the overall assessment of risk must not only consider the impacts on the utility pipeline but also the often significant risks associated with the mitigation works themselves. Unnecessary works will also be a charge on urban conurbations as a whole as well as the tunnel sponsors and Utilities.

g) For Stage 1 assessments developers should consult the Utilities at the earliest opportunity during the formulation and design of the works. Besides the usual plans of settlement contours it will be helpful to include a contour of maximum hogging and sagging curvature as this is a critical parameter in the pipeline assessments (see section 6.1.1). This may provide better advance warning where major diversionary, replacement or strengthening works may be required and/or to reduce the number of individual pipelines requiring detailed analysis.

h) To inform the latter assessment of risk associated with pipe condition the developer should request utility repair and maintenance history records which will provide an overview more relevant than the occasional trial pit investigation. Further request of area night line flow data will indicate areas where leakage (‘unaccounted-for water’) is present and may support Utility discussions if leakage increases during the course of the works.

i) Stage 2 assessments should contain charts showing clearly the predicted resultant flexural ground movements and the assumptions made in their calculation. Charts showing ground curvature and pipeline strains should also be provided together (see Figures 22 and 23) with estimated joint rotations.

j) Section 4.7 discusses the use of axial strain resulting from tunnelling and the Author’s conclusion is that in many circumstances it is likely to be unhelpful and overly conservative particularly for shallow pipelines. However, this will be a matter for individual designers to discuss in preliminary meetings with Utilities. k) Stage 2 should open detailed dialog between the Utility and developer in particular respect of risks to strategic assets. It is important that both sides to this discussion can agree that the protective works are proportionate to predicted disturbances.

l) A major uncertainty in tunnelling and subsurface works in general is the possible encounter with unexpected ground conditions. Even though site investigations are thorough and modern techniques applied recent experience has proven that this can still happen and has given rise to much larger tunnelling induced ground movements than predicted.

m) The concept of the risk being ALARP is helpful but in some circumstances it is possible that the potential consequences of a pipe or tunnel failure are so severe that mitigation measures will be put in place even where the likelihood of a damaging event is relatively low. For instance, where a tunnel is being constructed close to a pressurised water tunnel or gas transmission line the consequences may be so severe, both in terms of loss of life and/or cost, that carefully considered mitigation is put in place. If programmed into utility operations the isolation of strategic assets as the tunnelling works pass nearby is likely to be possible without disproportionate cost. The asset can then be inspected and/or tested and safely reopened for use.

n) Stage 3 assessments will be critical in determining the overall risk and any required mitigation works in respect of each asset where exceedance of utility criteria is predicted or a strategic asset lies within the zone of tunnelling influence as determined during Stages 1 and 2. This process can be challenging because the asset is beneath the ground and therefore difficult to assess in terms of condition and to inspect during and after the works. However, it is at this stage that decisions must be made and these may fall into two categories as follows.

o) The first is where the risk to a strategic asset is such that it can be agreed between the parties that mitigation works are required. If agreement cannot be reached it is likely that the Utility will prevail due to their statutory responsibilities and they are better positioned to judge the consequences of damage. Also at this stage it should be clear which other non-strategic assets are at significant risk and that failure is likely. These assets are also likely to be mitigated.

p) Secondly there may be numerous assets that while failing to meet the criteria suggested by the Utility might not require physical mitigation and an alternative course of action can be agreed. It is likely that this alternative will be decided on the basis of a robust emergency preparedness plan and an undertaking from the developer that, on completion of the works and stabilisation of the ground movements, any damage will be either repaired or the asset replaced.

q) The degree of criteria exceedance acceptable without physical mitigation will be a matter for individual Utilities to decide. It is suggested that, where consequences are anticipated to be small and rapid restoration of services is possible, an increase in acceptance criteria may be considered. For instance, the criteria for no pre-works physical mitigation that might be required for a cast iron pipe may be 200 microstrain rather than 100 microstrain. For masonry sewers in good condition the 500 microstrain criteria might be increased to 1500.

r) For brick sewers it is critical to ensure that the condition before any works are undertaken is good and that disturbance producing tension strains, particularly between the springings, is to be avoided. A strain of 500 microstrain has become accepted as a reasonable working criteria for assessment purposes but damage can occur even where this criterion is met.

s) There will be other factors which will significantly affect such decisions. For instance, where a pipeline has a good R&M and leakage history the criteria might be extended. To the contrary where the pipe has a poor history it may need mitigation in any event.

t) Recent experience in London is that pipeline damage due to tunnelling induced ground movement is rare where satisfactory assessments indicating ‘negligible’ impact are achieved and tunnelling performance targets are met. Damage has occurred when unexpected ground conditions are encountered and when unplanned physical impact (e.g. from piling or impact breakers) occurs.

u) For the reasons explained above it is the opinion of the author that less physical mitigation measures should be considered wherever possible not least because the risks associated with mitigation works often pose significant additional risks in themselves: A holistic assessment is required.