Bringing the Midland up to speed

The Sheffield to St Pancras project is tasked with increasing the linespeed on the Midland Main Line, ultimately reducing travel times for a headline sub two-hour journey. Before designs could be produced, an accurate and coordinated topographical survey was needed. A detailed scope of works was developed by Network Rail’s Track Engineer Kevin Mellor and Clients Survey Manager (CSM) Greg Roy, under the leadership of Designated Project Engineer Carl Baker, and was written to ensure conformity with the Network Rail Survey & Mapping Guidance note. The spec had to take into consideration the new topographical survey standards written by the Senior Survey Engineer at Network Rail, Chris Preston.

The primary grid was coordinated over a single intensive day’s work and is one of the largest of its kind ever undertaken in England during a 24-hour period. Other grids have encompassed larger geographical areas, such as those for the West and East Coast main lines, but these were produced over several months or years.

Topographical surveys

To deliver the survey and designs, Network Rail split the route into three design sections and commissioned two consultants – Mott MacDonald and Atkins Rail. Surveys were carried out using a combination of in-house and experienced sub-contract companies such as Severn Partnership. Three-dimensional survey data was required for all the route’s salient features with particular focus on the tracks themselves which had to be detailed to an accuracy of +/-5mm. All other features were required to an accuracy of +/-10mm.

To enable the surveys to be tied together, it was necessary to install and coordinate a network of control stations throughout the route. This took the form of 16 primary control stations in pairs off-track, a minimum of 500m apart and located approximately every 10km between London and Chesterfield. Care had to be taken to ensure there was a good view of the sky for Global Navigation Satellite Systems (GNSS), taking local geography and future vegetation growth into consideration. Primary control had to be installed and coordinated to an accuracy of 1:75,000 – around 15mm per km – in line with Network Rail’s topographical survey standards. In addition to the GNSS observations, a total station distance check, along with an elevation check, was needed to confirm relative accuracies.

The primary grid was supplemented by secondary control markers positioned approximately every 2km throughout the route. Secondary control was also installed in pairs, to an accuracy of 1:50,000 – approx 20mm per km – throughout the project area.

This practice ensures that there is always a reserve marker and allows the surveyor to establish a baseline for traversing purposes.

Tertiary control was installed trackside every 200m, to a similar accuracy as the secondary control. Again, this was in the form of concreted-in monumentation or spigots on mast faces where these exist on the electrified section south of Bedford. The tertiary control markers are the stations from which the topographical survey is carried out. Their maximum spacing is 200m as the maximum survey distance permitted is 100m in either direction.

Coordinating the survey grid

GNSS was used to determine the accurate position of the primary and secondary control grid. GNSS is a more up-to-date version of GPS and incorporates the Russian constellation of satellites ultimately providing a more accurate elevation.

The primary grid was coordinated during a very long day in August and incorporated surveyors simultaneously from both Atkins and Severn Partnership. Clear lines of communication were therefore imperative. Generally a minimum of six geodetic-grade dual-frequency receivers were used, involving a combination of Trimble and Leica equipment. GDOP and PDOP accuracies – a measure of GNSS precision – were taken at regular intervals, with receiver heights measured twice and recorded both manually and digitally. Witness diagrams were also produced to enable control to be found in the future.

The GNSS receivers on the secondary control markers were leapfrogged through their network, giving a minimum observation time on each station of 30 minutes depending on satellite coverage. However this time was increased where necessary.

Office processing

When accurate coordinates for the primary control grid had been obtained, the surveyor could work from ‘the whole to the part’ and hold the primary stations fixed in the secondary network adjustment.

GNSS timelines were edited and ‘cleaned up’ to eliminate any poor data or cycle slips. After the data had been downloaded and baselines edited, the active OS control station data could be imported – this included the precise ephemeris (or precise orbit) data obtained from the satellites’ exact position as tracked by RADAR. This data can take several days to be uploaded to the OS website. When it had all been collected, a least-squares adjustment was carried out on the network, after which the error ellipses were analysed and any discrepancies addressed.

As soon as the QA checks had been completed, the coordinates were processed through the SnakeGrid™ software which holds the scale factor at 1 and negates the effects caused by the curvature of the earth. Final coordinates were produced to the Midland Main Line grid (MML07). A survey report was created in line with the specification and included a list of the primary station coordinates along with a corresponding witness diagram for each new control point. Once the primary grid had been accepted, the next step was to position and install the secondary ground markers, in approximately 200m separated pairs every 1-2km throughout the route.

Double levelling was needed between all primary and secondary control stations. The start and end elevations used were those of the secondary control stations, derived from the GNSS coordination exercise. A check was performed when the levelling run reached the next secondary station in the network. The intermediate secondary levels were then adjusted accordingly where necessary. Each double level loop was checked against itself and the GNSS levels. A good QA check used was to additionally compare the derived elevations against the collected trig-heighted 3D traverse data.

Detail surveys

Following completion of traverse angles to both tertiary control and spigots, detail surveys could begin. There are several techniques available for picking up the required level of detail, however the surveyor generally detailed the track and associated features before moving on to those found in the cess. Where possible, to ensure speed and efficiency, the survey element of the project was undertaken at the same time as the tertiary and spigot coordination.

To supplement the topographical survey, accurate gauging data was required. This information is essential as the project needed to consider route clearances to structures where track lifting or slewing may take place. Several methods of data collection have been adopted on this project, including platform gauging, static laser sweeps and three-dimensional point cloud surveys using laser scanners. Of these, the latter is the most modern technique available to the industry but is currently only employed on upper sector structures such as overbridges.

Quality assurance procedures were followed by the CSM in conjunction with each consultant to ensure that accurate data was produced.

The overall route was split into smaller, more manageable sections or work packages. On completion of each, the data was checked and handed over to the designers. So far the designs have all been produced with little or no difficulty, demonstrating that a good survey and accurate control grid saves time, money and effort for the rest of any project’s lifecycle.

Track design

Track designs have only just started in a very challenging cost and programme environment. The Network Rail team and designers will be pushing design parameters to their limits and using cutting edge high-powered software systems such as 3rd Way for track geometry, Railsys for train performance and Track Ex for rail rolling contact fatigue prediction.

The West Coast project had the luxury of tilting trains for its curves. The Midland Main Line does not have such luxuries and it will be interesting to chart its progress. Watch this space!