Headspans were necessary but now improved resilience is needed

Generally, the railway electrification schemes that first emerged in Britain, before the introduction of more recent AC designs, relied on fairly heavy equipment. Multitrack supports were generally of the portal type – that is, a heavy ‘goalpost’-style arrangement.

As new designs emerged for the development of the 25kV AC system in the 1950s, once again multitrack overhead line equipment (OLE) situations were met by solid structural arrangements. A common example on the West Coast main line south of Weaver junction was the BICC ‘Welded Road’ portal.  Other portals were of double-channel, structural steel beams.

When British Rail (BR) looked to complete the electrification of the West Coast main line northward from Crewe to Glasgow, the cost of electrification was being seriously challenged by the government. To gain approval for those northbound extensions, BR undertook an extensive review of its existing designs with the intention of reducing both capital costs and the ongoing maintenance that painted portals needed.

What emerged from that review was a new, lightweight, modularised, headspan-based support system. This new design was used on the northern extension of the West Coast main line electrification, opening the route up to electric traction.

Despite its advantages in terms of cost, the headspan does have issues with resilience and reliability. In particular, the lack of mechanical independence between registrations means that, in the event of a problem such as a de-wirement, the impact is significant, affecting multiple tracks and increasing the time to reinstate the equipment.

To help resolve these issues, engineering consultant Arup has successfully completed a ‘Headspan to Portal’ conversion project on the East Coast main line south of Peterborough. Working closely with Network Rail, Arup has provided a portal conversion design that could largely be installed during ‘rules of the route’ access.

Arup’s head of electrification, Jonathan Ridley, invited Rail Engineer to meet him in York and explain the details.

Headspans

Already described as a ‘goalpost’, a portal consists of two vertical masts which support a single horizontal beam that spans the railway. The OLE is supported from this beam, one assembly for each line.

An alternative to the portal is the headspan arrangement. This structure still comprises two vertical masts, but, instead of the beam, two horizontal tensioned wires (the upper and lower cross-span wires) are strung between them to locate the OLE.  A third, top wire is a profiled headspan wire, and this provides support to the overall arrangement.

The headspan does have the advantage of being generally cheaper and easier to install than the equivalent portal. However, the headspan is a load-balanced system where the tensions in the wire runs themselves contribute to the geometric stability.  If one wire run breaks, the design geometry will be lost, since all other wires will be out of balance.  This type of structure is therefore not mechanically independent and a failure on one track can well mean all four tracks are out of service.

Headspans require regular maintenance to check the span wire tensions, and adjustment of the equipment tends to lead to the design and replacement of assemblies.  On high-speed lines, a mechanical wave, created by the passage of a train pantograph, also affects the adjacent wire runs.

In addition, headspans can require larger foundations than a portal, so as to resist a heavy overturning moment caused by the transverse wire tension. 

Headspan wire corrosion issues have also been experienced, as well as some other disadvantages. For instance, a mid-point anchor (MPA), where the OLE wires are fixed in position at their midpoint to keep the contact wire stable, cannot be a single point restraint due to the flexible nature of the system. This results in a distributed MPA, where the catenary is restrained over several structures to distribute the load.

Because of these reliability issues, it is now apparent that headspans are best suited to lower-speed applications or circumstances where low capital cost would be more important than high availability or performance.  In the UK, headspans have been installed in large numbers, but their less-than-reliable performance means they are no longer installed for new designs on main lines.

Improving resilience

With the continued drive to improve resilience within the railway system, and in view of some of the shortcomings and restrictions of the headspan solution, there has been a growing move to seek alternatives or replacements for the wire-based cross-track structure described here.  For some time, where infrastructure projects included major reconstruction, headspans have been replaced with new portal structures. 

An early example was the construction of Luton Airport Parkway station, within the original Bedford to St. Pancras electrification project, where several portals of a new design were installed. Modern designs of portal boom have replaced the welded rod format of the early years.

Recent experience of a significant number of catastrophic mainline failures has led to the consideration of the wholesale replacement of headspans, in order to improve performance and reduce disruption.

During works connected with enabling Crossrail connections to the Great Western main line on the approaches to London Paddington, there arose a need to make numerous alterations to the overhead line configuration due to staged track layout changes.

Arup became involved in the proposals, as designer for the scheme. The ‘Old Oak Common and Paddington Approaches’ (OOCPA) phase of the works involved complex staging, with continuous rearrangement of track layout and the accompanying stage-by-stage rearrangement of the OLE. However, the existing electrification scheme utilised headspan structures, as that was the standard design protocol at the time of construction in the early 1990s.

Analysis carried out by Arup confirmed that multiple sequential rearrangement of OLE on a headspan was not practical, as the balanced-cable arrangement would not allow for the easy rearrangement of individual wire runs, whereas small part steelwork (SPS) on a fixed portal beam could be adjusted relatively easily as the track alignment changed during construction staging. 

Designers considered installing new portal structures, but they also examined the feasibility of utilising the existing steel support structures and landing a new portal beam on them.

Learning from a Network Rail trial project, at Potters Bar on the East Coast main line, Arup produced proposals for a practical method of converting the OOCPA headspans to portals, which was progressed over a small number of strategic OLE structures. As the headspans in the site formed a mid-point anchor, a new mid-point portal was installed for practical reasons, but the conversion of adjacent headspans went ahead as per the Form A design.

Valuable experience was gained from these OOCPA conversion works. For example, one of the masts on a headspan was found to be rotated by around nine degrees – not a problem when supporting flexible span wires but quite inconvenient when supporting a stiff fixed portal beam. In addition, the installation of the portal beam impacts the type of loading on the supporting steel.

Portals for Connington

The Connington area, on the East Coast main line south of Peterborough, experiences relatively high windspeeds and, as such, is prone to dewirements, resulting in significant train delays over recent years. Network Rail identified this area for conversion to portals to help build resilience into the OLE system. Arup was commissioned to prepare a design study to look at replacing several headspans in the area. Again, the provision of installing new structures was rejected and a detailed analysis of the possibility of reusing the existing support masts was undertaken.

Using Arup’s well-developed tools for assessing geotechnical issues and ground conditions, a close study of the foundations was made, to discover whether they would cope with the varied stresses and loads from the new portal geometries.

Economy would best be achieved by the reuse of the existing masts, but these would have been installed in a manner that facilitated the cross-track wires.  Using point cloud surveys, the precise positions of the masts had to be recorded, along with any skew or twist as had been seen at Paddington, which would have an impact on the loading of the finished portal.

Ground engineering studies of the foundations were essential as the original foundations would have been installed to take cross-track stress rather than the new loadings imposed by the beams. This involved a detailed structural analysis to determine the existing foundation loads and compare them with proposed portal loads. Arup’s proprietary software was used both to model the new loads imposed on the foundations and to check the stresses in the Series 1 boom and connection angle to the masts due to the loads of the UK1 OLE, a design first used on the West Coast main line for higher train speeds, and the OLEMI (OLE Master Index) equipment which continued to support the OLE in the conversion.

In summary, the emphasis was on the strength of the concrete, reinforcing bar cover and the general suitability of the foundation for the portal conversion. However, without the cross wires, the bending stress on the vertical structures is reduced.  In all cases, the foundations at Connington were side-bearing concrete – no piles were involved.

Following these initial design considerations, a detailed design for thirty structures was prepared, covering this high-risk area on the East Coast main line. Performance aspirations would suggest that all headspans in a tension length should be changed, but the complexity of replacing items such as mid-point anchors and neutral-section supports drove the decision to convert only the simpler, multitrack headspans within the tension length. Mid-point anchors, booster transformer structures, and switching structures were among some of the structures deemed too complex for this phase of the headspan-to-portal conversion projects.

Installing the Portal booms

Many design and construction meetings were held with the Network Rail’s works delivery team, accompanied by drawing revisions, such as extra dimensions, to suit the construction team’s needs on site. Installation was carried out on site by Works Delivery, acting as principal contractor. Design acceptance was similarly eased by working with the E&P (electrification and plant) and structures route asset managers, and the crane provider was brought in at an early stage.

Staging of the work was very important, particularly on such a heavily used route. The design, therefore, detailed all of the stages of the conversion process, not just the finished result, taking into account the analysis of both mast and boom orientation, the detailed construction methodology and a view of simple versus complex structure types.

First, after the headspan wire was removed, the new boom was landed on the two main steel masts, with the two cross wires retained. After this interim stage, the SPS was modified, in a staged process, to fit in with site availability and line possessions. 

Lifting in the new boom was a complex procedure, each one weighed in excess of a tonne and had to be manipulated into the final position with existing wires in situ, but, once it was in place, the processes became more self-contained. An initial stage-by-stage approach could have led to one road being upgraded at a time, but Jonathan Ridley pointed out that, in practice, all four roads were completed at once.

If an incident occurs, damage is now usually limited to the single track involved and the equipment can be returned to normal operating condition in less time than if the failure occurred on a headspan structure.

With the conversion of 30 headspans completed, Arup can look to the future. The next step is to convert a complete tension length, including all of the complex structures that were left out in this conversion project. Further evaluation of the complex structures will be required as part of a new feasibility study and the performance improvement gain is expected to be considerable.

Whilst the reasons for the original switch to headspans can be readily understood, when their low capital cost contributed to obtaining authorisation for important electrification schemes, it resulted in reliability that was less than that of the original portals. Now the design has gone back full circle, to the cost-effective conversion of those headspans back to portals, delivering the reliability that today’s busy railway needs.