1. INTRODUCTION This paper develops and extends the content of the 2017 Harding Memorial lecture (New, 2019) and reflects the development of techniques used on numerous recent projects including HS2 and Thames Tideway. It is inevitable that many construction works will give rise to ground movements which can have adverse impacts on third party assets if these works are not carefully designed, planned and carried out: the presence of sensitive, vulnerable or ageing assets exacerbates this problem. The impact of construction works on buildings throughout the world have been well-documented but there is a lack of literature focusing on the impacts on utility assets.
Through the authors’ longstanding engagement with the Research Working Group (WG2) of the International Tunnelling and Underground Space Association (ITA), it has become apparent that the approaches adopted by both utilities and developers in different countries vary significantly and are influenced by the prevailing social and legal frameworks.
The objective of this paper is to provide a broad guide to the assessment of utility pipelines which are subject to impact from various construction activities. It is intended to provide a structure to impact assessment which includes utility requirements, components of pipeline risk assessment, pipe failure mechanism, masonry, concept of impact strain, and the staged assessment process. The presented methods and strategies are based on the authors’ experience, particularly in London and the Thames Valley. The more risk-based approach puts an emphasis on rationalising the assessment process to reflect the uncertainties about ground and asset conditions, consideration of the consequence of asset failures (see extent of damages shown on Figure 1), and implementation of control and mitigation measures.
The paper cannot cover all types of utility assets and the various conditions encountered throughout the world. However, it is intended to be as informative as possible so that assessors will be able to adapt the presented principles and methods to suit local conditions and construction methods.
2. UTILITY REQUIREMENTS
There are generally three key utility requirements that will have to be considered by developers, designers and contractors throughout the design and construction phases of new development schemes:
? The utility must be in compliance with its statutory and regulatory obligations. In particular, the serviceability to customers should not be threatened in both short and long term. The failure to protect an asset from any construction damage could result in serious health and safety issues, substantial third-party damages, and penalties from regulators.
? Whole-life asset value should be maintained. There should be no damage, loss of capacity or downgrade of an asset to ensure its value is retained to protect stakeholders interests.
? The proposed works should not inhibit or prevent any maintenance or repair works that are required for the asset.
In order to meet these key requirements, developers, designers and contractors are expected to undertake pipeline risk assessments to inform utilities regarding the likely risk to their assets posed by the proposed development, and how the risk might be eliminated, minimised or otherwise mitigated. The assessment cannot be completed without essential information from the utility which includes asset maps (showing the ‘indicative’ alignment and material of the assets); repair and maintenance records; criticality and vulnerability of the assets; as well as the potential consequence to their assets of failure.
The pipeline risk assessment prepared by the developer’s team will be reviewed by the utility engineers and their consultants in order to provide a high-level assurance to senior managers within the utility business.
3. COMPONENTS OF PIPELINE RISK ASSESSMENT
A pipeline risk assessment is often a complex process and some of the required input information may be uncertain or unobtainable. On many occasions, an assessor will have to make certain informed assumptions in order to complete the assessment. Figure 2 shows the six key components of pipeline risk assessment and further details are summarised below:
1. Conditions (properties) of assets
a. Type
i) Water, gas, electric, telecoms and others
b. Location
i) It is important to confirm the line and level (and material) of the asset via survey and exploratory excavations because information provided on utility asset maps may not be accurate or complete. For instance, since the 2011 Transfer of Private Sewer Regulations, sewage undertakers were required to become responsible for many unmapped private sewers.
c. Material
i) Iron (grey or ductile); clayware; brick; concrete (unreinforced or reinforced); thermoplastic (polyethylene (PE) or polyvinyl chloride (PVC)); glass reinforced plastic (GRP)
d. Size
i) Outside and inside diameters; pipe section length
e. Jointing
i) Lead run; Cementitious; Flanged; Welded; Flexible; Age
f. Present conditions
i) Review of repair and maintenance (R&M) history and night line flow records held by the utility will give a better indication of the condition of the pipe network than the occasional trial pit investigation.
ii) Internal survey of sewers allows identification of existing defects which may require urgent attention before commencement of the proposed works.
g. Exclusion zone
i) This is the minimum vertical and horizontal clearances required between the extrados of the asset and the outer face of the proposed works.
ii) This is to allow sufficient space for any repair/ maintenance/upgrade works required for the assets.
h. Ground conditions
i) Geo-environment (e.g. chemical attack on pipe material)
ii) Problematic soil (e.g. alluvial deposits)
iii) Bedding (e.g. poor workmanship)
i. Pre-existing strain (if known)
i) Strain arising from ground movements or other loadings caused by historic works (e.g. cross trenching, tunnelling, basements)
2. Calculations (e.g. ground movements, changes in loading)
a. Tunnels, shafts, basements, boxes and foundations (including any demolition, piling, dewatering works required and any use of heavy construction plant in the vicinity of the assets) and associated vibration issues
b. Geometries and construction sequences/ methods (including both temporary and permanent works)
c. Ground movement and utility damage assessments (e.g. strain, joint rotation and pullout)
3. Consequence of damage
a. Strategic, trunk or distribution main?
i) 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 (distribution main) but nevertheless safety-critical in importance.
ii) Failure of a pipe within a resilient network (i.e. there are sufficient alternative supply routes) will have less impact on the supply to customers.
b. Failure to meet statutory service obligations?
i) There can be a substantial penalty to the utility if it cannot restore its service to customers within a pre-defined period.
c. Loss of whole-life asset value (e.g. damage, loss of capacity or downgrade)?
d. Damage to third-party assets and health and safety issues (e.g. possible flood zones, explosion)?
e. Financial and reputational impact to the utility’s business? The general public usually considers the utility is at fault even though the cause of pipe failure may be associated with other third-party activities.
4. Control (and preparedness)
a. Mitigation
i) Replacement, diversion or strengthening?
ii) Temporary isolation? This will allow the asset to be inspected and/or tested upon completion of the proposed works before it can be safely reopened for use.
iii) Pre- and post-condition survey (including additional inspection to be carried out at critical phases of the works, where necessary).
iv) Locate and exercise of valves controlling the flows to minimise the extent of flooding/ damage associated with pipe failure.
v) Potential risk to workers (e.g. substantial periods of man entry to install structural lining inside a sewer; collapse of brickwork during removal of temporary support)?
vi) Potential risk/disruption to public (e.g. traffic diversion and intrusive traffic management, traffic delays)?
vii Potential risk to projects (e.g. availability of materials, pipes and plants; speed of construction; unforeseen additional costs and time to implement mitigation measures)?
viii) Feasibility of developing a strategy for doing the minimum to preserve serviceability of the assets to reduce potential serious risks associated with mitigation works?
ix) ‘Do-nothing’ approach? This relies on 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.
b. Monitoring
i) Validation of design assumptions relied upon by the assessments
ii) Type (e.g. ground/pipeline movements, leakage, vibration)
iii) Location of monitoring points
iv) Frequency of data collection
v) Data interpretation method
vi) Trigger levels
vii) Action plan (e.g. termination and/or modification of the construction / mitigation processes; change in monitoring and data review frequency).
c. Network resilience
i) Alternative supply routes within the network?
d. Emergency preparedness/contingency planning
i) Review of impact models (possible flood zone and other consequence).
ii) Emergency preparedness plan (EPP) should include an incident recovery plan that will allow rapid repair of a pipe failure. R&M contractors should be made aware of the works and have emergency spares and equipment immediately available for repairs.
iii) Local operations staff notified of the works.
5. Consent
a. Has the appropriate level of analysis and checking been carried out?
b. Is the risk ALARP (As Low As Reasonably Practicable)?
c. Response from utility: ‘Approval in Principle’; ‘No further comments’, or ‘LONO’ (Letter of No Objection)
6. Conciliation
a. Further negotiations with Developer
b. Court injunction (the last resort).
4. PIPE FAILURE MECHANISMS
Pipes do not fail randomly and there is always an explanation for pipe failure. Generally, the following three groups of causes may be attributed to pipe failures:
1. Third-party intervention
a. road works; ground movements; compaction; vibration; abnormal surface or other loads.
2. Weakened pipe
a. Corrosion; casting defects (e.g. incorrect section, wormholes, porosity, joint failure).
3. Operational overload (which is unlikely given a pipe in serviceable condition)
a. Over pressurisation; surge
New (2019) illustrates typical failure mechanisms for pipes made from brittle materials (e.g. cast iron, vitrified clay, unreinforced concrete) and they can be grouped into the following categories:
? Bearing fracture due to concentration of reaction or load (which may occur in any size of pipe).
? Longitudinal cracking (splitting) due to overload (with larger pipes tending to be more vulnerable).
? Circumferential crack due to flexural bending (which is most common in metal pipes with diameters of 300mm or less; or unreinforced concrete/vitrified clay pipe with aspect ratio (pipe length/pipe diameter) greater than two.
? Burst socket due to excessive joint rotation and/or shearing (which may occur in all pipe sizes).
? Spiral failure due to a combination of longitudinal flexural and transverse loading (which is commonly found in pipes between 300 – 600mm diameter).
Experience has shown that the locations of pipe failures rarely coincide with those pipe sections predicted to have the most likely chance to fail. This is because pipe failure depends on various factors which, when present in an unfavourable combination, are sufficient to cause the failure. Therefore, a probabilistic (rather than a strictly deterministic) approach is more appropriate for pipeline assessment.
5. MASONRY
Many early trunk sewers were constructed in brickwork and took various geometrical forms (e.g. egg, circular, horseshoe, arched crown and invert with vertical sidewall). Analyses of such structures can be analytically more difficult, more uncertain and highly condition dependant. A basic understanding of masonry arch theory (see Heyman (1982 & 1995), Szechy (1970) for further reading) will assist in undertaking analyses for assessment purposes.
Transverse stability is usually the primary concern for masonry arch. Some forms of structural overloading or asymmetrical loading often result in the development of tension in the masonry. This can result in a loss of mortar or bricks within the crown of the arch (see Figure 3) or elsewhere between the springings, which can subsequently result in collapse as the arch is no longer functioning because it becomes kinematically unstable.
Arch design seeks to resolve the loads upon the arch into compressive stresses and thereby eliminate tension. Figure 4 illustrates the stress distributions across a masonry arch with the thrust line acting at various locations. Compressive stress is developed across the whole arch thickness when the thrust line is retained within the ‘middle-third’ of the arch thickness. However, tension is developed when the thrust line is outside the ‘middle-third’ limit. This results in the formation of cracks because masonry is assumed to have negligible tensile strength for assessment purposes.
CIRIA C671 (2009) provides useful guidance for a thrust line analysis of a circular masonry structure and further details are provided in Appendix A. The aim for the thrust line analysis is to confirm if the line of thrust can be developed within an arch ring and equilibrate the given loading. Figure 5 shows an example of a thrust line analysis of a 2m internal diameter and 0.6m thick circular brick sewer located at a depth of 8m under three different horizontal stress to vertical stress ratios (KT).
The location of the thrust line is one of the determining factors for masonry sewer transverse stability:
i) Developed within the intrados-third of the masonry arch thickness (see Figure 5a) – tension will be developed at the extrados of the sewer. However, the stability of the sewer could be maintained if there is sufficient external ground pressure to hold the arch in place.
ii) Developed within the middle-third of the masonry arch thickness (see Figure 5c) – tension will not be developed within the arch and the stability of the sewer is maintained.
iii) Developed within the extrados-third of the masonry arch thickness (see Figure 5e) – tension will be developed at the intrados of the sewer and this could result in the loss of mortar and/or bricks around the crown of the sewer.
The ‘ULS envelopes’ shown on Figures 5b, 5d and 5f indicate the minimum depths from the intrados and extrados of the sewer at which the thrust line can be located without exceeding the ULS masonry compressive strength (assuming it to be 2.5MPa in the analysis). For stability, the thrust line should remain within the ‘ULS envelopes’ at all times and also cross the middle axis of the arch at least twice.
The development of tension in the masonry can also be caused by ground movements associated with various construction activities. Given sewer arches are considered to be semi-continuous linear structures, they are potentially vulnerable to the longitudinal ground movement curvatures in a similar way to pipes (see Section 6.3). Further details with regard to transverse strain analysis for masonry structures subject to external ground movements are presented in Section 6.8.
Published test results on tensile and compressive strengths of masonry (see Backes (1985)) indicate that maximum tensile strength can be less than 1% of compressive strength whereas tensile modulus can be significantly less than the compressive modulus. However, there is a common mistake by assessors to adopt numerical models (finite element or similar) which assume masonry as linear elastic material with ‘unrestricted’ tensile strength and modulus of elasticity equal in both tension and compression. It is important to note that a numerical model must not assume tensile properties that do not exist or cannot be justified.
Masonry sewers were generally designed for sustaining gravity flows without surcharge. However, the continuous expansion of population in cities has led to a significant increase in volumes of sewage passing into the network which can generate surcharge pressures. This can result in unacceptable hoop tension in the masonry which can lead to sewer collapse (see Figure 6).
Stability of masonry sewers relies on sufficient confinement during any surcharge events. The consideration of the Confinement/Pressure Ratio (CPR) gives an indication of the vulnerability of masonry sewers when subject to surcharge pressure. CPR is defined as the ratio of soil overburden pressure at the sewer axis level (σvo) to the water pressure within the sewer (P). For assessment purposes, the surcharge pressure within a sewer is commonly taken as the hydrostatic pressure to the levels of the controlling nearby manholes or other sources of pressure relief.
If a masonry sewer is in good condition and other significant destabilising loadings are absent, a CPR in excess of 1.33 is generally considered to be acceptable. This criterion is also applicable to sewers constructed using materials vulnerable to surcharge pressures such as unreinforced concrete (see Figure 7).
CPR check is also a useful tool for assessing potential vulnerability of unbolted water transmission tunnels. CPRs in excess of 1.33 under surge pressure and 1.5 under operating pressure have been found to be acceptable.
Reduction in soil overburden pressure is usually caused by excavations for various structures (e.g. stations, basements) and can produce potentially damaging unload of confining compressive hoop loads in both sewers and water tunnels. Therefore, this issue should be considered at the earliest planning stage for infrastructure and buildings in order to avoid a design which is unacceptable to the utilities.
Many masonry sewers are commonly subject to frequent cycles of large loadings from heavy traffic and vibration. A fatigue-type racking motion within the masonry, caused by rotation of principal stresses and such loadings from a rolling axle, can result in sewer failure.
6. CONCEPT OF ‘IMPACT STRAIN’
For assessment purposes, the parameter ‘strain’ rather than ‘stress’ is adopted as the damage criteria. This is because it is displacements both in the ground and of the pipeline which are actually measured and used to monitor and control the works. The conversion to stress can be done but this introduces an unnecessary uncertainty in the choice of appropriate moduli.
It is self-evident that a pipe will fail when the movement/load imposed upon it exceeds its strain limit. In order to assess the impacts on the pipe as a result of the proposed construction works, it is important to understand the existing state of strain within the pipe which is unknowable in many circumstances. This leads to the introduction of the concept of ‘impact strain’ for assessment purposes.
‘Impact strain’ is defined as the strain arising from ground movements or other loadings caused by the proposed construction works. It does not include any strains arising from any other causes (including preexisting ground movement, ground load, operating pressures or normal vehicular loading). This ‘impact strain’ may be that predicted during the assessment process or back-calculated from field measurements undertaken during the works.
The calculated impact strain should be compared with the assessment criteria advised by the relevant utility. This will provide an indication of the risk of damage to the pipe associated with the proposed construction works.
6.1 KEY ASSESSMENT ASSUMPTIONS
For Stages 1 and 2 of the assessment process (see Section 7) it is conservatively assumed that:
1. ‘Green field’ ground movement is adopted.
2. Longitudinal flexural (bending) of the pipe follows this ‘green field’ ground movement (i.e. stiffness of the pipe is ignored).
3. The joints are stiff so that the pipe is assumed to be a continuous linear structure. This allows the calculation of maximum longitudinal flexural (bending) stain.
4. The joints are fully flexible so that the limiting condition is in joint rotation.
5. In the absence of site specific information, a 3.66m (12ft) long pipe is used for the calculation of joint rotation for UK metal pipes. It is also useful to review dimensions provided in the Standards and/or manufacturer specifications for other pipe materials.
6. Flexural strain is calculated at the pipe extrados.
a. Metal/plastic pipes – the neutral axis is at the geometric centre (i.e. lever arm equal to external pipe radius).
b. Materials of low/negligible tensile strength (e.g. masonry, vitrified clay, unreinforced concrete) – the neutral axis is at the pipe extrados (i.e. lever arm equal to external pipe diameter).
7. Total ground movements are simply obtained by superposition, as appropriate.
6.2 SOURCES OF GROUND DISTURBANCES 6.2.1 Tunnels
Gaussian models (e.g. O’Reilly & New (1982) reprinted in 2015; New & O’Reilly (1991); Leca & New (2007)) have been commonly used to calculate green field tunnelling induced movements and associated slope, curvature and strain because they are straightforward to apply and have stood the test of time in their application. The authors are not aware of any body of data that would contradict their general application for initial assessment of near surface pipelines with a depth to axis of less than say, 3m.
The ‘ribbon sink’ model by New & Bowers (1994) is useful for calculating ground movement in stiff cohesive materials at depth (especially for estimating movements in the vicinity of the tunnel under construction). Three dimensional movements from complex tunnel geometries can be calculated using this model.
Volume loss (Vl) and trough width parameter (K) are the two key input parameters for the calculation of tunnelling-induced ground movements based on the Gaussian models. Recent experience in London has confirmed that a volume loss of 1% for tunnelling in stiff cohesive materials (e.g. London Clay) has usually not been exceeded by using Earth Pressure Balance Machines (EPBMs) and this is generally considered as a ‘moderately conservative’ assumption for assessment purposes: this relates to a single tunnel driven in ground undisturbed by other tunnels. Historically however, it is clear from the literature that for twin tunnels the driving of a second tunnel parallel to the first yields larger volume losses.
This is confirmed by recent measurements reported by Wan et al. (2017) during the construction of twin tunnels (7.1m bore diameter, axis depth of 34.5m and 16.3m separation) below Hyde Park in London for Crossrail (now the Elizabeth Line). The surface volume loss during the excavations for the first tunnel being about 0.8% but for the second tunnel increased to about 1.4%. This increase is attributed to a softening of the clay caused by the first tunnel excavations. This would indicate that the impact assessment for the second of twin tunnels a volume loss of greater than 1% should be considered by designers.
It is important to note that the predicted ground movements based on the Gaussian models cited above are neither conservative nor un-conservative. It is the assessors’ responsibility to establish and justify the parameter values adopted in their analyses. Given the uncertainties in the input parameters, more complex models are unlikely to be beneficial/necessary to Stages 1 and 2 of the assessment processes (see Section 7).
6.2.2 Shafts, basements/boxes, dewatering
Tunnel construction often require construction of shafts which are used for tunnel boring machine (TBM) launching, access, and maintenance purposes and the shaft construction itself can lead to significant ground movements. Based on the original predictive equations proposed by New & Bowers (1994) and an extensive database of measurements taken in London, New (2017) has proposed a generic equation for predicting settlement at a distance from the shaft wall:
Sd = αH (1- d/nH)2
where Sd is the settlement at a distance d from the shaft wall.
n is a simple multiple of the shaft depth H to a distance d from the shaft wall where settlement becomes zero. (Note: n = 1 for the original New & Bowers (1994) equation).
α is an empirical constant and αH is the settlement at the shaft wall.
The n and α values are dependent on shaft diameter, ground conditions and construction method. The assessors will have to establish and justify their parameter choices based on appropriate empirical data (e.g. case histories).
CIRIA C760 (2017) provides good guidance on ground movements associated with construction of basements and boxes.
Dewatering (either route wide, local, short-term or long-term) is commonly required to facilitate excavation for shafts, basements and boxes, and this can give rise to significant ground movements. Local dewatering (especially in the presence of alluvial deposits) can result in damaging localised differential ground movements. Designers for the dewatering system should work with the assessors to minimise the risk of damage to the pipelines and other utility assets.
6.2.3 Foundation & heavy/abnormal loading
The application of foundation loadings will result in ground movements and potential overloading of underlying utilities. The construction and use of piled foundations in close proximity to a pipeline may result in damage which can be caused by:
? Excessive pile loading onto the pipeline (e.g. shaft load shedding, end-bearing pressure)
? Uncontrollable ground loss during pile boring within difficult ground conditions (e.g. water bearing gravels)
? Excessive vibration during pile construction (e.g. driven piles, encountering physical obstructions)
The use of temporary casings, sleeving, non-vibratory/ non-percussive piling methods (e.g. silent piling) and redesign of the foundation layout, backed-up by the calculation of associated impacts, are the common ways to mitigate the risks mentioned above.
Construction works normally involve the use of sizeable plant (e.g. cranes, piling rigs, load loaders, self-propelled modular transporters (SPMTs)) which can impose damaging loadings onto pipelines. The ALARP risk approach is to ensure these heavy/abnormal loads are positioned outside the zone of influence of the pipeline, which is an area commonly defined by drawing 45 degree lines upward and away from the pipeline. This can also be achieved by using bridging structures to carry the load away from the pipeline.
If it is not possible to position the loads outside the zone of influence, both longitudinal and transverse analyses for the pipeline will be required. Simple analytical solutions (e.g. Boussinesq approach) are generally adopted for the assessment. It is unlikely to be beneficial to undertake soil-structure interaction analysis given the uncertainties regarding the ground and pipe conditions.
6.2.4 Demolition
Analysis should be undertaken to assess the potential ground movements and associated impacts on the pipeline due to demolition works. Careful control of demolition operation is important to minimise the level of vibration and impact loads transmitted to the pipelines.
6.2.5 Vibration
Construction works may generate vibrations (dynamic ground strains) that can be damaging to local structures, including utility apparatus. However, there are now non-vibratory methods available that can build almost all structures without significant vibration impact. In order to mitigate this risk, it is common for major asset owners (e.g. Transport for London (TfL), Thames Water Utilities Limited (TWUL)) to require an exclusion zone around their assets and this is typically 15m from a vibratory source which is usually piling or vibratory compaction.
Even at greater ranges, some piling works can present difficulties but modern ‘city rigs’ and pre-augured vibratory methods can often meet the requirements of third-party asset owners in restricting PPVs (peak particle velocities) to acceptable levels. Straightforward predictive guidance is given in British Standard BS5228- 2:2009 for piling and tunnelling works and the Highways Agency (Transport Research Laboratory TRL) has published detailed predictive methods for a range of vibratory sources caused by mechanised construction works (Hiller & Crabb, 2000).
Threshold values relating damage to transient and continuous vibration levels for various structural types are given by British Standards BS7385-2:1993. An important factor often overlooked is the possibility that vibration can increase the density of, and cause settlement, in some soils, which can put structures at risk (see BS5228-2:2009 Section B.3: Note 3).
6.3 Longitudinal flexural (bending) strain
Longitudinal flexural (bending) strain (εb) along a pipeline is calculated from simple beam theory and is given by:
εb = y / R
where R is the longitudinal radius of curvature and y is the lever arm to the neutral axis of a pipe.
The calculation of longitudinal radius of curvature (R) is given by:
R = i2 / [2dn – (dn-1 + dn+1)]
where dn is the flexural displacement predicted at the nth data point along the assessment line and i is the interval between the data points (see Figure 8). An interval of 1m is generally adopted for assessment purposes. However, a smaller interval may be required to capture any abrupt changes in curvature along the pipeline.
The lever arm y is a function of the pipe extrados diameter (D) and is dependent on pipe material. Figure 9 illustrates the definition of lever arm for pipe material either capable or incapable of sustaining tension under bending. Examples of circumferential cracks caused by longitudinal flexure (bending) are shown in Figure 10.
6.4 Axial Strain
The calculation of axial strain (εa) along a pipeline is given by:
εa = [(S + ΔS) / S]
where S is the initial distance between P1 and P2, and ΔS is the change in the axial distance between P1 and P2. Axial compression and tension are represented by negative and positive ΔS values respectively (see Figure 11).
The bending strain (εb) and axial strain (εa) are usually combined together to calculate the total ‘impact strain’ for a pipeline when undertaking assessments. This can sometimes result in very significant impact strain even after the axial strain reduction factor proposed by Attewell et al. (1986) has been applied.
Following review of field experience and various calculation methodologies, New (2019) concludes that axial strain may be regarded as negligible for various types of pipes at relatively shallow depths (say 2m) for the Stage 1 and Stage 2 assessment processes. However, assessors may still wish to include axial strain when assessing pipes of critical importance. It is important to note that the potential axial compressive and tensile strains cannot be relied on to reduce flexural tensile and compressive strains respectively (i.e. the beneficial axial strains are ignored) for assessment purposes.
6.5 Joint rotation
Assessment of joint rotation (?) is typically undertaken by tracking two pipes along the predicted flexural displacement profile with the left and right end of the two pipes and the joint following the flexural displacement profile (see Figure 12) until the point of maximum rotation has been established. The magnitude of joint rotation is dependent on the pipe length L which is usually assumed to be 12ft (3.66m) if as-built records are not available.
Joint rotation is an important cause of failure in larger cast iron pipe (with diameter in excess of about 300mm or 12in) because of its higher bending stiffness. This allows the generation of very large bursting forces as the spigot rotates within the socket. Figure 13 shows an image of a reconstructed cast iron pipe after suffering from socket bursting caused by joint rotation.
6.6 Joint pullout
The axial ground strain may induce relative movement between pipe lengths resulting in joint pullout (see Figure 14) which should be assessed because excessive joint displacement can cause leakage. Assuming the pipeline is rigid and the joint is free to accommodate axial movement, the joint pullout will not be more than the differential ground movement between mid-points of adjacent pipes (see Attewell et al. (1986)).
For an optimally aligned spigot and socket connection, a pullout limit of 3mm can be considered for cast iron pipes as a joint leakage criterion for assessment purposes.
For non-flexible joints (e.g. bolted flanges), the pipelines should be considered as continuous linear structures with a check that the connections are not overstressed.
6.7 Transverse (crushing) strain for pipe
The problem with large diameter pipe subject to load from concentrated sources from above is crushing. This can result in longitudinal cracking (see Figure 15) which is usually developed at the quarter points of the pipe wall. The maximum transverse bending strain (εt) at the quarter points of the pipe wall is given by New (2019) as:
εt = 3 σv r2 (1 – K0) /2 E t 2
where σv is the vertical stress at the pipe axis level. This can be estimated by Boussinesq-based analyses.
K0 is the ratio of horizontal to vertical ground stress.
r is the mean radius of the pipe.
E is the Young’s modulus of the pipe (typically 80GPa for cast iron pipe).
t is the pipe wall thickness.
6.8 Transverse strain for masonry sewer
For assessment purposes, it is conservative to assume that masonry has negligible tensile strength and each part of the outer circumference of the masonry sewer moves with the ground (full bond). The transverse sewer movements are calculated around the periphery of the brickwork at appropriate intervals.
Figure 16 provides the nomenclature used in the calculations of transverse strain of a circular masonry sewer. The vertical and horizontal movements derived from the methods described in Section 6.2 are resolved into movement in the radial (δN) and hoop (δT) (circumferential) directions:
δN = δy cos α + δz sin α
δT = -δy sin α + δz cos α
The sewer is divided into a number of segments at appropriate intervals anticlockwise around the circumference as shown. The axial strain in each segment is calculated from the change in length of each segment indicated by δT and flexure is derived by analysis of a triplet based on δN (see Figure 8).
A similar assessment approach can be applied to non-circular masonry sewers.
The key to masonry sewer stability is to prevent excessive tensile strains in the haunches and crown so as to prevent bricks falling out giving rise to arch instability. Depending on the sewer position relative to the proposed construction works, the sewer may be in either egging (i.e. increase in vertical and decrease in horizontal dimensions of the sewer) or squatting mode (i.e. decrease in vertical and increase in horizontal dimensions of the sewer) and both have to be considered during the analyses of the transverse strains so as to ensure that the worst case is detected.
Example charts showing the predicted distortion, displacement and tensile strain profiles of a circular masonry sewer, which runs normal to the tunnel direction including assumed parameters, are shown in Figure 17. Note that the maximum transverse tensile strains are found to be in the crown/invert section and springings, with the sewer under ‘squatting’ and ‘egging’ modes respectively. The hoop compressive strain in the sewer due to the existing ground load is to be considered and this reduces the maximum transverse tensile strains currently shown on the charts.
6.9 Assessment guidance and criteria
Early consultation between the developers and the relevant utilities (especially during the initial project planning stage) is highly recommended. This will allow the developers to get a better understanding of the acceptable procedures and criteria which can be applied for the impact assessment. This can reduce the risk of abortive work and the associated delays and unexpected costs to projects.
Thames Water Utilities Limited (TWUL), the largest water and wastewater services company in UK, offers ‘Guidance on piling, heavy loads, excavations, tunnelling and dewatering’ on their website (https://www. thameswater.co.uk/developers). The assessment criteria provided in that document are summarised in Tables 1 and 2 and they assist developers in preparation of impact assessment reports. The criteria are for guidance only, and it is based on early work by Attewell et al. (1986) and years of experience by TWUL. It is intended to represent a level of risk of damage which may be reasonably regarded as negligible for a pipe in average to good condition. It is a fact however that any pipe is potentially vulnerable to any increase in strain and joint rotation. The developers’ designers may adopt alternative criteria values, provided they can justify that the risk of damage remains negligible.
Attewell et al. (1986) and Bracegirdle et al. (1996) also offer useful guidance on the assessment of strain in cast iron pipelines.
7. THE STAGED ASSESSMENT PROCESS
A more risk based, three-stage approach for pipeline assessment has been proposed by New (2019). The aim is to demonstrate the risk balance between the likelihood (based on evaluation of ground movement and ‘impact strain’ presented in Section 6) and consequences of an adverse event, and the ability to recover (i.e. emergency preparedness and resilience) during and after the works.
Details for this staged assessment approach will not be repeated here. However, a comprehensive summary of this staged assessment process is presented on Figure 18 and the key considerations are listed below:
Stage 1 (Assessment scope and information retrieval)
a. Identification of all utility assets which may be affected by the proposed works:
i) Calculation of likely green field ground movements based on conservative assumptions.
ii) Assets falling outside the 1mm settlement contour are likely to be excluded from further analysis.
iii) Contours indicating minimum radius of curvature (R) or maximum curvature (1/R) can assist in identifying those assets most likely to need mitigation at this earliest stage.
b. Gathering of information needed for subsequent stages of the risk assessment (see ‘Conditions (properties) of assets’ and ‘Consequence of damage’ in Section 3 for further details).
Stage 2 (More detailed analysis)
a. Longitudinal and transverse analyses (based on moderately conservative assumptions and calculation methodologies as presented in Sections 5 and 6) for those assets not eliminated by the Stage 1 preliminary assessment.
b. Produce charts showing (axial, lateral, vertical and flexural) displacements, curvature, (axial, flexural, combined tensile and compressive) strains, joint rotation and pullout along the particular section of the pipe under consideration.
c. Identify those assets likely to be particularly vulnerable based on R&M history, night line flow history and condition survey.
d. Identify those assets which are of critical importance and with high consequence of failure.
e. Those assets which fail to meet the utility assessment criteria and/or with high criticality/ vulnerability are further considered in the Stage 3 risk analyses.
Stage 3 (Evaluation of risk)
a. Risk/remedy based approach
i) Acceptable level of risk? ALARP concept (i.e. the risk shall be ‘As Low As Reasonably Practicable’)
ii) Importance of collaboration between developer and utility to develop a ‘reasonable and optimised’ solution which is acceptable to both parties
iii) Establish/justify the need for intrusive diversionary or mitigation works (see ‘Control (and Preparedness)’ in Section 3 for further details regarding inherent risks associated with these works and alternative options).
iv) Consideration of consequential losses (in terms of ‘health and safety’, failure to meet statutory requirements, and ‘commercial’).
v) Risk of incorrect analysis or human errors during the formal checking and approval of reports.
b. Need for soil-structure analysis?
i) Use of numerical models can be helpful for certain applications but it is generally considered to be unnecessary for pipeline assessment, given the uncertainties about the ground and pipe conditions and difficulties in establishing an appropriate set of input parameters.
ii) Assessors should justify the necessity of undertaking numerical analysis (finite element or similar) and the results should be corroborated by closed form (analytical) solutions together with case history data where possible. Further guidance on the management of advanced numerical modelling in geotechnical engineering can be found in CIRIA C791 (2020).
Mitigations
a. There are various engineering solutions to mitigate risk to pipelines. However, there will be risks associated with the implementation of the mitigation measures. Further details are summarised under ‘Control (and Preparedness)’ in Section 3.
b. Pre- and post-work inspections of the assets are recommended. This can inform the assessors regarding the current condition of the asset so that an appropriate set of criteria will be adopted in the impact assessment. Also, any damages/defects identified after the completion of the proposed works can then be repaired in order to restore the whole-life asset value.
c. It is important to have an Emergency Preparedness Plan (EPP) in place throughout various phases of the construction works. An incident recovery plan should be included that will allow rapid repair of a pipe failure and have provision of equipment and spares required to restore serviceability to customers promptly while repair works are being undertaken (e.g. overpumping equipment covering a sewer failure scenario).
d. It is important for developers to provide a RAMS (Risk Assessment and Method Statement) package for early review by the utility to ensure the works will be carried out within the constraints defined by the utility and be ALARP.
Monitoring, Control, and Validation of Design Assumptions
a. A well-thought-out and carefully planned monitoring system will provide useful information for validation of the design assumptions relied upon by the assessors and provide ongoing assurance to the utility. The monitoring results will also be used to confirm adequacy of the measures controlling the works.
b. The Instrumentation and Monitoring (I&M) plan should include the type and location of monitoring points, frequency of data collection, data interpretation method, trigger levels and action plan.
c. The action plan should include a strategy to cover what to do in an event of a breached trigger level (e.g. termination and/or modification of the construction/mitigation processes; change in monitoring and data review frequency).
d. Continuation of post-work monitoring until cessation of significant ground movements has become normal practice. For some recent major infrastructure projects, the monitoring works have been terminated when the settlement rate is less than 2mm/year.
8. CONCLUSIONS
This paper provides a guide to the assessment of pipelines subject to impact from various construction activities. In summary, a pipeline risk assessment comprises six key components including ‘Conditions’, ‘Calculations’, ‘Consequences’, ‘Control’, ‘Consent’ and ‘Conciliation’, and these form the components of the three-stage assessment processes. The presented methods and strategies are drawn from the authors’ experience of the assessment processes developed through their works in London and the Thames Valley over the last two decades.
This paper is intended to be as general and informative as possible so that the assessors will be equipped with appropriate tools to go through the impact assessment processes and have a positive dialogue with the utility. The ultimate goal is to achieve a ‘reasonable and optimised’ solution which is acceptable to both parties, is truly ALARP and most importantly, takes a holistic view of risk.
9. ACKNOWLEDGEMENTS
Any views expressed in this paper are those of the authors and not necessarily those of Thames Water Utilities Limited or any other utility.
APPENDIX A: THRUST LINE ANALYSIS OF A CIRCULAR MASONRY SEWER
Background
The thrust line analysis aims to confirm if the line of thrust can be developed within an arch ring and equilibrate the given external loading (see Figure A.1) without exceeding the masonry compressive strength. If this is the case, the arch ring will not collapse under this loading in accordance with the lower bound theorem of limit analysis. The following calculation is mainly based on the methodology documented in CIRIA C671 (2009).
Assumptions
The self-weight of the brickwork is ignored for simplicity. Also, a quarter of the masonry sewer is considered in the thrust line analysis because of symmetry (see Figure A.2). The tensile strength of the masonry is considered to be negligible for assessment purposes.
Input parameters
re External radius of the sewer
rc Radius along centerline of the sewer
t Sewer thickness
σm Masonry compressive strength
σH External horizontal pressure at sewer axis level
σV External vertical pressure at sewer axis level
KT Ratio of external horizontal pressure to external vertical pressure (σH/σV)
Calculations
Thrust at crown is:
HH = σH × re
Horizontal and vertical force components at a given section located by the angle (θ) are:
RX(θ) = HH – σH × (1 – cosθ) × re
RY(θ) = σV × re × sinθ
The minimum depth from the intrados/extrados of the sewer at which the thrust line can be located (dmin) without exceeding the masonry compressive strength is:
which is based on the compressive stress distribution shown on Figure A.3
where
This defines the locations of the ULS intrados and extrados envelopes around the arch. Within the bounds of the ULS envelopes, a further check is undertaken to confirm if the thrust line can be developed within the ‘middle-third’ of the arch thickness (i.e. dmin = t/3) so that there will be no tension within the arch.
Bending moment at a given section located by the angle (θ) is given by:
Axial force at a given section located by the angle (θ) is given by:
AA (θ) = RX (θ) × cosθ + RY (θ) × sinθ
Eccentricity of the thrust line (i.e. offset from the centerline of the arch) at a given section located by the angle (θ) is given by:
Worked example
Thrust line analysis of a 2m internal diameter masonry sewer with an axis depth of 8m:
Step 1 Calculation of input parameters
re External radius of the sewer = 1.6m
rc Radius along centerline of the masonry arch = 1.3m
t Sewer thickness = 0.6m
KT Ratio of external horizontal pressure to external vertical pressure = 0.7
σV External vertical pressure at axis level = 20kN/m3 × 8m × 1.4 = 224kN/m2 (ULS, load factor of 1.4 for overburden with soil unit weight of 20kN/m3).
σH External horizontal pressure at axis level = KT σV = 156.8kN/m2 (ULS).
σm Masonry compressive strength = 2.5MPa (ULS).
Step 2 Calculation of ULS envelopes.
Step 3 Calculation of ‘middle-third’ envelopes.
Step 4 Calculation of thrust line based on dmin = t/3.
Step 5 Checking if the thrust line is compatible with the masonry compressive strength.
Step 6 Conclusions.
The analysis indicates that the thrust line:
i Remains within the ‘middle-third’ of the masonry arch thickness and no tension will be developed within the arch.
ii Remains within the ULS envelopes at all times.
iii Crosses the middle of the arch at least twice.
Therefore, the arch ring will not collapse under this loading case in accordance with the lower bound theorem of limit analysis.
