Using Horizontal Directional Drilling (HDD) and/or microtunnelling technologies to create long-distance pipelines (up to 10km in one continuous drive) to house multiple high voltage cables, in preference to trenching, in areas of outstanding natural beauty, through urban environments or to connect offshore wind farms to the onshore power distribution network has been considered in a recently completed feasibility study at the Universities of Birmingham and Newcastle in the UK. This paper summarises briefly the outcomes from the project in terms of the electrical and materials challenges (cable support structure, forced cooling, electrical system requirements) and describes in detail the potential solutions to address the associated trenchless technology challenges of creating the ‘cable pipe conduit’. It specifically addresses the use of analytical and finite element modelling respectively to determine long-distance frictional resistance forces for HDD pipelines and the feasibility of extruded microtunnel lining systems for permanent conduit construction respectively, and outlines proposals for new research.
The consumption of power, not least electrical, within the UK and worldwide continues to grow at an alarming rate. Increasingly this power is being imported. If increased power demands from major centres of population and industry are to be met and security of supply guaranteed dramatic steps for rethinking how power is to be supplied are required. The Energy Review undertaken by the UK Government in 2006 was tasked with developing solutions for power generation within the UK that will ensure a secure source of power and contribute towards the Country’s reduction in the emission of greenhouse gasses . Whilst local generation, such as roof-mounted wind generators, solar panels or small-scale power stations, will help to address the problem in part, it is clear that large-scale electrical power generation will be needed to satisfy the majority of this demand. The next generation of major power stations, of any form (clean coal, nuclear, wind, tidal, etc.), will inevitably be built away from the major urban areas, and indeed many of the best sites earmarked are on the coast or offshore, often in very remote locations. Long transmission lines are thus inevitable, and are an indispensable component of a sustainable electricity supply in the 21st century.
It is estimated that the existing high-voltage transmission network within the UK comprises 25,000km of cable (see Figure 1). This network is ageing and will require replacement and upgrading within the near future if the network is to continue to act as the backbone of power transmission within the UK. However, overhead power transmission is not without its critics and is proving unsuitable in various situations. A recently proposed upgrade to the network with the construction between Beauly to Denny of a new high-voltage line (marked as a dotted line on Figure 1) is steeped in controversy as the line has attracted objections from a large section of the community including local councils  and community action groups. Opposition to the new line varies from health concerns to a desire to protect and preserve the areas of outstanding natural beauty within Scotland. The development of potential offshore wind farm sites has gathered pace since 2000 , yet for this source of power generation to become a mainstay within the UK the connection of the sites to the shore must be secure. Overhead lines are clearly inappropriate in these instances and laying the cables along the seabed runs the risk of the power generating facilities being disconnected from the National Grid due to the action of the sea or man. Burying the cables, particularly in the surge zone or where the waters are heavily fished, may reduce the risk of failure. Overhead power lines within urban environments are also facing stiff opposition from local communities and planning authorities with alternative routings being sought. For example, the majority of the power cables for the Olympic development within London are being routed underground in tunnels .
Conventional tunnelling is expensive and, whilst it is desirable to dispense with pylon-mounted overhead cables for buried electricity distribution within sites of natural beauty and urban environments, economic burial methods must be sought. Trenchless technologies have developed to such a point where they may provide a suitable burial solution at reduced cost when compared to conventional tunnelling and cause (far) less disruption to surface process than trenching. Using Horizontal Directional Drilling (HDD) and/or microtunnelling technologies to create long-distance pipelines (up to 10km in one continuous drive) to house multiple high voltage cables, in preference to trenching, in areas of outstanding natural beauty, through urban environments or to connect offshore wind farms to the onshore power distribution network has been considered in a recently completed feasibility study at the Universities of Birmingham and Newcastle, in the UK. This paper summarises briefly the outcomes from the project in terms of the electrical and materials challenges (cable support structure, forced cooling, electrical system requirements) and describes in detail the potential solutions to address the associated trenchless technology challenges of creating the ‘cablepipe conduit’. It specifically addresses the use of analytical and finite element modelling respectively to determine long distance frictional resistance forces for HDD pipelines and the feasibility of extruded microtunnel lining systems for permanent conduit construction respectively, and outlines proposals for new research.
2. INSTALLING AND OPPERATING THE CABLES WITHIN THE PIPE
The installation and operation of the cables within the small diameter pipes (600mm to 2000mm) over long distances was addressed at Newcastle University and this investigation is briefly summarised herein. Installing a cable or groups of cables over long distances in pipes that are too small for man-entry raises a number of limiting factors that principally relate to the joining of the individual lengths of cable on-site and the need to drag the fabricated cables without over-stressing them. Unless the cables are installed offshore, or are only intended for low capacity networks, then the cable installed into the tunnel is likely to contain a number of joints. These joints tend to be larger in diameter than the cable itself and can be considered the weak points within the system. To install these effectively without over-stressing the joints, and thus potentially limiting the performance of the network, would require a lubricated sled system that transmitted the load, thus greatly reducing the load in the cable, yet was flexible enough to allow the cable to form the spiral shape commonly associated with large cables that have previously been installed on reels, when the load on the cable is removed.
Once the cables have been installed and are operational there is an issue of how to transmit power. AC is commonly used, however in confined spaces, where AC conductors are in close proximity, the capacitive current becomes prohibitively large and limits the rating. An alternative to AC is high voltage DC (HVDC), which is a more efficient method of power transmission than AC over long distances. The main obstacle to wider use of HVDC transmission is the high cost of HVDC switching, but recent developments in power electronics are promising and costs are expected to fall within the next five years.
3. THE USE OF HDD AND MICROTUNNELLING OVER LONG DISTANCES
Prior to the installation and operation of the cables the pipe must be installed and there are a number of factors that limit the application of either HDD or microtunnelling over the lengths considered within the feasibility study (2.5km up to 10.0km). These factors have been outlined in previous No Dig publications [6, 7], are summarised in Table 1, and will form the basis for the remainder of the paper.
Table 1: Factors Limiting HDD and Microtunnelling Drive Lengths 
| ||Horizontal Directional Drilling ||Microtunnelling |
|Cutting the bore || |
- Achievable drive lengths are directly linked to the power of the drilling rig
- Boulders can deflect the cutting head
- Cutting tools can struggle to cut through geological interfaces (e.g. soil to rock)
- Changes in geology can cause problems with guidance of the cutting tool (the tool ‘wriggles’)
- Cutting tools may wear over time, necessitating the removal of the drillstring from the bore for repairs
- Position of the cutting tool is controlled from the drilling rig at the entry point
- Cutting tools wear over time, which may cause the drive to stop
- Worn tools must be replaced automatically
- Changing geology can affect the performance of the cutting tools
- Boulders can prove difficult to cut with a TBM, especially when in a clay matrix (e.g. boulder clay)
- In weathered rock the characteristics of the rock mass, rather than the material properties of the rock, contols behaviour
- Maintaining line and level of the TBMs can be difficult in soft ground
|Stabilising and flushing the bore || |
- Granular materials can collapse onto the drill string, increasing the pull-in force required
- Large volumes of drilling fluid are required to stabilise granular ground conditions
- Changes in geology can result in loss of drilling fluid, resulting in collapse of the bore
- Excessive pressure within the drilling mud can result in hydraulic fracture of the soil
- The diameter of the tunnel prevents the use of more than one set of carts to extract the cuttings, reducing the rate of advance
- Slurry systems require intermediate pumps that may obstruct access to the TBM (i.e. delivery of cutting tools to the TBM)
- If pipejacking is used, large volumes of drilling fluid are required to lubricate the pipestring through granular conditions
|Lining the bore || |
- ‘Wriggles’ within the bore increase the frictional forces acting on the pipe
- Fabrication of the pipe during pull-in can result in increased frictional forces on the pipe due to the 'stop-start' motion of the installation
- Increased frictional forces result in reduced drive lengths
- Plastic pipes may be damaged if the bore collapses during installation, potentailly reducing their engineering life
- Pipe jacking is sensitive to misalignments within the bore and is a relatively slow technique, limiting the rate of advance
- Collapse of the bore onto the pipestring will increase the force required to install the pipeline, resulting in the need for interjacks, and may slow the rate of installation
- Increase in the force required to jack the pipestring into place and/or missalignments may resulting in spalling
- Segmental or extruded lining must be installed by the TBM without supervision from personnel•
|Logistics || |
- Large volumes of drilling fluid are required to stabilise an HDD installation (up to 3 times that of the bore)
- The drilling fluid must be stored, prepared and filtered of cuttings on site
- Possible problems with disposal of the drilling fluid if the ground is contaminated
- Curing rates of an extruded lining may have an impact upon the drive rate
- Segments must be transported to the TBM for erection. This activity must not conflict with other processes
- Unless the umbilicals are attached to the pipe barrel then they must be dragged behind the TBM, or the TBM must automatically install them
Three of the limiting factors identified within the desk study were subsequently investigated with the use of numerical modelling. The first issue was the interaction between the drilling mud and the surrounding ground, particularly when driving though mixed conditions or at interfaces between varying geological formations, and was explored in simple terms using discrete element modelling (DEM). The second related to the pull-in forces during HDD installations over long distance and was investigated using Pipe-Force 2005 , see Section 4. The final issue was the ability to create an extruded lining behind the TBM that would not deform excessively and would not have a detrimental effect upon the drive rate of the TBM due to curing time requirements. This was addressed by creating a model in Plaxis, a commercially available Finite Element Modelling (FEM) package.
It is evident from the investigation that although DEM could be an important tool when predicting the flow of liquid through a porous medium (replicating drilling mud entering a soil), it requires considerable further development before it can realise this potential. In the time available during the feasibility study undertaking such development was not possible and the attempt to model the interactions between the drilling mud and surrounding formation with DEM was postponed. The modelling with Pipe-Force 2005 and FEM was more productive and these are described in the following sections.
4. PREDICTING THE PULL-IN FORCES WITH LONG DISTANCE HDD INSTALLATIONS
Pipe-Force 2005  was developed to predict the pull-in forces in HDD applications. Whilst this model is unlikely to predict localised increases in pull-in forces due to problematic ground conditions or stoppages in the installation of the product pipe (as experienced in the field), it is able to predict baseline pull-in forces associated with HDD installations. The model describes a HDD installation as a series of straight sections (as opposed to a series of parabolas). Nodes are introduced where the lines intercept one another (reflecting a change in drilling angle) and Pipe-Force 2005 analyses the tensile force at each of the nodes, see Figure 3. An idealised HDD installation would require only four nodes and three straight sections; the nodes would describe the entry and exit points as well as the two transitional points, where the angle of the bore changes from an acute entry or exit angle (generally considered to be between 6º – 20º from horizontal) to the horizontal section of the drive. Additional nodes may be inserted within the borepath to reflect a more complex borepath, deviations from line and level or collapsed sections of the bore.
The theory that governs the calculation of these forces (based upon the theory of flexible beams) is described by Lasheen and Polak . In summary, the model calculates the tensile pull-in force at a node by summing the tensile forces created due to frictional forces on the ground surface (as the pipe is pulled into the bore), the frictional forces acting on the pipe within the bore (a function of the unit weight of the pipe, pipe buoyancy, overcut ratio and the nature of the surrounding formation), the forces (due to bending moments) arising as the pipe navigates curves within the borepath (a function of the pipe diameter, wall thickness, the Young’s modulus of the pipe material, overcut ratio and the nature of the surrounding formation) and the forces (fluidic drag) induced as the pipe is pulled through the supporting drilling mud (a function of the viscosity of the drilling mud and the speed of installation).
Modelling was conducted in three phases; a parametric study (not reported herein), the inclusion of deviations and collapsed sections in the bore path and the variation of the bore length. The parameters used in the study are listed in Table 2.
|Parameter ||Introduction of Deviations ||Long Distance Drives |
|Drive Length |
Entry and Exit Angles (º)
External Pipe Diameter (m)
Standard Dimension Ratio (SDR)
Friction Coefficient Between the Bore and Pipe
Friction Coefficient Between the Ground and Pipe
Young's Modulus (MPa)
Yield Stress of the Pipe Material (MPa)
Failure Stress of the Pipe Material (MPa)
Viscosity of the Drilling Fluid (Ns/m2)
Rate of Installation (m/s)
Specific Weight of the Pipe (kN/m3)
Specific Weight of the Drilling Mud (kN/m3)
0.6 to 1.8
|up to 10000 |
0.6 to 1.8
Table 2: Definition of the Parameters used in the Pipe-Force 2005 Study.
Deviations were introduced to the horizontal section of a previously developed idealised borepath with the use of nodes placed in a quasi-sinusoidal pattern (see Figure 4). The magnitude of the deviations and the spacing between deviations were varied during the investigation. The magnitude of deviations represented in Figure 5 describe the total change in height as the quasi sinusoidal pattern was applied, for example 600mm deviations refer to a ±300mm pattern from line and level.
The magnitude of the deviations was varied, using the same spacing as that for the baseline borepath in Figure 5, and the distance between deviations was reduced. It can be seen that varying the spacing between deviations has had the greatest impact on the pull-in force (for a relatively small deviation from line and level); the relative increase in pull-in force at 20m, 15m, 10m and 5m spacings is 1.23, 1.44, 2.16 and 3.96, respectively, when compared to the baseline pull-in force.
Collapsed sections were introduced with the addition of nodes within a pre-defined borepath and the coefficient of friction was increased between pairs of the nodes. The length of the simulated collapse borepath and the coefficient of fiction within the collapsed were varied, see Figure 6. The length of the collapsed section would appear to be of less significance to the change in pull-in forces when compared to the change in friction coefficient within the bore. Indeed, the change in friction force, even when over a relatively short section of bore in relation to the total length of the drive, has sufficient impact on the pull-in force that the installation could fail in some circumstances.
Four drive lengths (2km, 5km, 7km, 10km) were investigated using four pipe diameters (600mm, 1000mm, 1400mm and 1800mm), which were selected in light of the outcome of the heat dissipation modelling for given power ratings, pipe lengths and cooling methods (not described herein), see Figure 7. It would appear theoretically possible to install a 600mm diameter pipe over the 10km drive when considering the maximum pull-in capacity of current technology (approximately 6000kN). However, this would see the drilling rigs operating at its limit and as the borepath modelled within Figure 7 is idealised, with no deviations or collapsed sections, any changes resulting in an increase in pull-in forces would make it impossible to install the pipe over 10km using conventional drilling practices. Thus accurate control of the cutting head and maintaining the stability of the bore become paramount if such an undertaking is to be achieved. Increasing the pipe diameter to 1000mm results in a reduction in maximum drive length to less than 5km and a 1400mm diameter pipe would be limited to approximately 2km drives.
5. DEVELOPMENT OF A MODEL TO SIMULATE EXTRUDED LININGS BEHIND THE TBM
A numerical model to predict the behaviour of a concrete lining extruded behind a microtunnelling TBM as it drives forward was developed in ‘Plaxis’, a commercially available finite element package suited to 3D analyses. The analyses are performed in Plaxis using plastic calculations of the advanced ultimate level type. The TBM is assumed to be 8.6m long (with formwork) and 2m in diameter, the axis of the tunnel is assumed to lie 10m below ground surface level and 7m below the phreatic surface. The model is 20m wide (x-direction), 20m deep (y-direction) and extends over a 40m length (z-direction), see Figure 8.
Figure 9 illustrates how the movement of the TBM is simulated within Plaxis. The movement of the TBM is considered in discrete steps; Stage 1 - the TBM is present, Stage 2 - the TBM is moved forward by its entire length, from Slice 1 to Slice 2 and a contraction factor is applied to the rear of the TBM (Plane A) and the rear of the Liner (Front Plane). In addition a reduced interface for adhesion and friction is applied to the soil surrounding the TBM, The initial properties of the liner (liner B) reflect a concrete still in the early stages of curing. In Stage 3 the process is repeated, with the TBM occupying Slice 3, the strength of the lining (liner A) in Slice 1 is increased to reflect changes due to curing of the concrete lining, and an additional section of the weaker lining (liner B) is introduced in Slice 2. The procedure may then be repeated for subsequent stages. As the number of analytical steps increases, additional sections of liner are incorporated within the model and the strength of the existing sections of liner are increased until they reach that of the fully cured material. The outcomes of modelling in Plaxis include normal, torsion and shear forces, bending moments (not reported here) and displacements (see Figures 10 and 11).
The movement of the TBM and the curing of the lining material are time-dependent phenomena and these processes can only be approximated when modelled in Plaxis. The rate of migration of the TBM has been assumed to be 10-15 millimetres per minute (described by Fourie , as a typical rate in poor ground conditions), which equates to 0.9 metres per hour. The length of the TBM, with formwork, has been assumed to be 8.6m, resulting in a time of 9.6 hours for the TBM to move forward over its entire length. This time, and distance travelled, has been adopted as a single step phase in the analyses. Assuming such a timeframe for each of the analytical steps has an impact on how the lining material is modelled within Plaxis, because the lining material extruded at the start of the 8.6m drive will have had a longer period of time to cure than the material extruded at the end of the 8.6m length. Thus a difference in strength exists between each end of the segment created as the TBM moved forward. To model this, the strength values for the lining material along each section are averaged to form one value for analysis. This assumption does not reflect reality; for this to reflect real conditions within the field the time and distances adopted in each step would have to be significantly reduced. However, it does allow for the investigation to determine whether modelling within a FEM package can describe the extruding process.
To test the robustness of the model, a parametric study was undertaken to investigate the performance of the simulated tunnelling method for Stages 1 to 3 when changing: the contraction factor, face pressure, soil types and position of the phreatic surface. Where possible the displacements derived were recorded and compared to the values predicted by empirical solutions. Figure 10 illustrates the effects on the displacements induced with the change of contraction factor applied to the newly formed lining as well as the effects of the TBM face pressure on the resultant displacements (Stage 1 results shown only). In both cases the change in the value of the respective parameters results in very little effect on the settlement profile.
Figure 11 illustrates the effects of changing the level of the phreatic surface (in clay) and soil type for a Stage 1 analysis (N.B. ‘mixed face’ conditions refer to 10m of clay overlying sand). Doubling the depth of the phreatic surface from 3m to 6m has little effect on the maximum displacement yet appears to decrease the overall trough width. Changing the soil type resulted in changes to the displacement profile, e.g. the maximum vertical displacements, at the front plane (Figure 9), were: clay (6.7m), mixed face (6.3mm), sand (5.6mm) and stiff sand (1.0mm).
Using empirical solutions to predict the maximum horizontal and vertical displacements in a clay soil returns values of 0.8mm and 1.3mm respectively. Comparison with Figure 11 indicates that the FE model is overestimating the displacements induced by the creation of the tunnel. It was presumed that this overestimation is due a combination of the constitutive model selected and the lack of refinement of the mesh, i.e. the trends and relative performance of the analysis will be correct, though the absolute values may not be. Since only four months were available for modelling work in the feasibility study, it was not possible to address this phenomenon and therefore refinement will be made during the next phase of research. What is important here is that extruded liners can be modelled in a simplified manner. Moreover the performance of two extruded lining materials; a concrete and a polymer concrete, can be compared – a key criterion for this research (see below).
Concrete is commonly used to create cast-insitu linings within tunnels . However, it slow curing rate and logistical complexities have prevented its adoption in small diameter tunnels. Preliminary findings (not reported here) suggest that the curing rate and initial strength of the concrete are insufficient for exposure (i.e. removal of temporary support) at the drive rate considered (10-15 millimetres per minute). In contrast the curing rate and initial strength of the polymer concrete investigated would appear to be suitable to act as an extruded liner and would appear not to have a detrimental effect upon the drive rate of the TBM. Further research is required to validate these initial findings.
6. PROPOSED CONTINUATION OF RESEARCH PROJECT
The feasibility study into the factors that limit the use of HDD or microtunnelling over long, continuous drives has highlighted a number of areas that require research and development before such applications can become a reality. The research must address the ability to create the tunnel lining and the subsequent methods of installation and operation of the cables. In order to create the conduit a combination of experimental and numerical studies must be carried out to address: drilling mud interaction with the formation, control of the frictional forces within the bore and the guidance of the cutting tool (all necessary to permit HDD to be used); the creation of the lining behind a TBM and the extraction of the cuttings through small diameter tunnels over long distances are the primary issues for microtunnelling. The development of HVDC switching is a parallel future need.
Numerical analysis has been undertaken, as part of a larger study, to investigate the interaction between drilling fluid and surrounding formation (DEM), the pull-in forces associated with HDD installations over long distances (Pipe-Force 2005) and the development of a FEM model to investigate extruded linier.
DEM has the potential to predict the interaction between drilling mud and the surrounding ground, and thus act as a powerful tool for those working with HDD. However, it is apparent that it requires considerable further fundamental development before it can realise this potential.
Modelling of the pull-in forces during HDD installations suggests that it is theoretically possible to install a 600mm diameter, or smaller, pipe over 10km, although the results of the modelling would suggest that an increase in the coefficient of friction over a relatively short distance, due to a collapse or deviation from line and level, would increase the pull-in forces jeopardise the viability of the installation.
A 3-D FEM model was developed in Plaxis to simulate the behaviour of extruded linings within a day of being exposed from the formwork as the TBM drives forward. Several parametric studies were undertaken to investigate the potential for modelling this in Plaxis. It has been shown that this is possible and that the comparison of two materials, a concrete and a polymer concrete, produced interesting and encouraging results. For the purpose of future research it is proposed that it would be better to use an epoxy concrete liner which can be designed to have increased curing rates, since it appears the relatively low strength of juvenile concretes may not be sufficient to transmit the loads without suffering excessive deformation. Without some means of rapid strengthening, the use of extruded liners would be likely to reduce the drive rate of the TBM (thus increasing the cost of the construction method). Conversely, with relatively high juvenile strengths and rapid curing rates, polymer concrete is a potentially suitable material for an extruded lining. These are aspects for further investigation.
The authors wish to acknowledge their other academic partners on the HV Cable project from Newcastle University, UK. We also wish gratefully to acknowledge the financial support provided by the Engineering and Physical Science Research Council (EPSRC) in the UK. The industrial partners for the project are Advanced Specialist Moulders, Charador and Plastic Pipes, and their support in developing many of the ideas reported herein is also gratefully acknowledged.
 POST (2007). “POSTNOTE: Electricity in the UK.” Parliamentary Office of Science and Technology, February 2007, POSTNOTE No. 280. Online Report (Last accessed June 2007). http://www.iop.org/activity/policy/Publications/file_21079.pdf
 BBC (2006) “Council Objects to Pylon Proposal.” British Broadcasting Corporation Online news report. (Last accessed June 2007). http://news.bbc.co.uk/1/hi/scotland/tayside_and_central/4974504.stm
 OWEN (2007). “Offshore Wind Energy in the UK.” The Offshore Wind Energy Network, Online (Last accessed June 2007). http://www.owen.eru.rl.ac.uk/
 McKenna, J. “Power lines to hang over Olympic Park until 2009” New Civil Engineer, p10, Issue Date 13th April 2006
 Kotsikos, G (2006). “Personnel Communications.” Dr G Kotsikos, Newcastle University
 Royal A. C. D., Rogers C. D. F. and Chapman D. N. (2006). "Installation of Cables Over Long Distances Using Trenchless Techniques." Proceedings of the 24th International Conference on Trenchless Installation of Utilities No Dig Conference, Brisbane, Australia, 29 Oct – 2 Nov.
 Royal, A. C. D., Rogers, C. D. F. and Chapman, D. N. (2006). “Practical Limitations of Traditional Trenchless Technologies for Long-Distance High Voltage Cable Installation.” Online Proceedings of the No Dig Live Conference, Coventry, September 26-28. (Last accessed October 2006) http://www.nodiglive-pipelines.com/conferencepapers
 Cheng, E. and Polak, M.A. (2005). “Modelling Installation Loads for Pipe In Horizontal Directional Drilling.” Proceedings of International No-Dig 2005, Rotterdam, The Netherlands.
 Lasheen, A. and Polak, M. A. (2001). “Predicting the Behaviour of HDPE Pipes in Horizontal Directional Drilling.” Underground Infrastructure Research: Municipal, Industrial and Environmental Applications. Edited by Knight, M and Thomson, N. A.A Balkema Publishers ISBN 90 2651 820 X
 Fourie. B. (2004). “TunnelTEQ system for Wellington.” Australasian society for trenchless technology ASTT, newsletter, March 2004, 12 pp.
 STAJ (2007). “Extruded Concrete Lining Method.” Shield Tunnelling Association of Japan. . Online Report (Last accessed June 2007). http://www.shield-method.gr.jp/english/ecl/index.html
1 Department of Civil Engineering, University of Birmingham, Birmingham, United Kingdom.
2 Department of Civil Engineering, University of Waterloo, Canada.