It’s been a long time coming. First envisioned shortly after World War II, Crossrail is a 73-mile-long (118-kilometer) railway serving the greater London area. Under construction since 2009, the £15 billion ($24 billion) project will provide high-frequency passenger transportation for commuters and suburban residents from Maidenhead and Heathrow in the west to Shenfield and Abbey Wood in the east. Crossrail will pass through 37 stations and enable direct passenger connections to Britain’s Network Rail.

The new rail line includes surface track and 26 miles (42 kilometers) of new tunnels passing beneath central London. Depending on location, the tunnels are created by tunnel boring machines (TBM) or cut-and-cover methods. A critical part of constructing the tunnels is maintaining awareness of how the work affects existing rail lines adjacent to the sites. Using a blend of geospatial technologies, a Crossrail construction contractor has created a monitoring solution to provide timely, accurate information for project stakeholders.

Protecting the Rails

Part of Crossrail construction is the C350 Pudding Mill Lane contract, which will link the new Crossrail tunnels to the Network Rail infrastructure in the London Borough of Newham. The roughly £100 million ($160 million) contract, won by Morgan Sindall plc, includes construction of a tunnel portal and new elevated Docklands Light Railway (DLR) station. The work also includes structures to carry the new subterranean railway into the existing above-ground rail network.

The contract required Morgan Sindall to measure the impact of its cut-and-cover tunnel construction on existing lines, including a 1,600-foot (480-meter) stretch of Network Rail track and two sections of DLR track totaling 1,250 feet (380 meters). To detect any tilt or slew of the rails during nearby construction, Morgan Sindall needed to provide 24-hour monitoring of the existing track. The monitoring approach needed to detect movement efficiently and reliably to enable project teams to avoid the possibility of a derailment that could compromise passenger safety and cause delays to project schedules.

Instead of hiring subcontractors to provide monitoring, Morgan Sindall took the work as an in-house function. The company assigned Chief Land Surveyor Nick Giles to handle surveying and monitoring on the Pudding Mill Lane project. Crossrail had provided clear requirements for a monitoring system; It needed to be robust, reliable and repeatable to provide total confidence for those who were depending on it. Working with Morgan Sindall Monitoring Data Manager Dave Gibbs and Monitoring Surveyor Pawel Owsianka, Giles developed a unique two-pronged approach that would provide the required high level of confidence.

Giles’ solution was to have two systems working in tandem. An optical system using total stations and monitoring software would monitor horizontal displacement. Simultaneously, a Flat Mesh bi-axial tilt sensor system made up of several hundred Senceive high-precision wireless tilt meters attached directly to the track would monitor cant and twist.

This approach provided two key benefits. First, the two systems could provide constant backup checks for each other. Second, the dual approach enabled Morgan Sindall to reduce the number of trackside optical instruments needed to provide complete coverage.

To provide optical monitoring, Giles selected 10 Trimble S8 total stations equipped with Trimble VISION technology. The instruments are controlled by Trimble 4D Control software (T4D) running on a central server. The instruments were installed on durable mounts at regular intervals along the Network Rail and DLR tracks. The Morgan Sindall team attached small prism targets directly to the tracks at 3-meter (9-foot) intervals. The arrangement gave each total station a line of sight to up to 60 of the rail-mounted prisms.

Flexibility, Analysis and Consistency

Giles said that the core of the monitoring system resides in the T4D office software. The software enabled his team to create customized operation and analysis, including remotely managing the 10 total stations, all measurement cycles and communications.

The system uses a pre-programmed hourly cycle to provide 24-hour coverage. Each cycle begins with the total stations taking readings to fixed reference targets to confirm each instrument’s location. The instruments then measure to each of the prisms attached to the rails. The data is collected by the T4D software, which computes and records the results into an SQL database. From there, Giles and Owsjanka can extract data to create visual results such as maps, charts and graphs needed for meetings and client reports. Owsjanka, who looks after data analysis and remote access to the instruments, uses the software to create the specific routines he requires for both analysis of the data and report generation.

As part of creating the customized system, Giles defined alerts that would be issued when the software detected changes or motions that exceeded specific levels. The alerts, sent by email or text messaging, provide the ability for rapid reaction to a change in the field. One concern was the potential to send alerts to stakeholders who did not have the expertise required to interpret them. This could result in delays to works with action being taken when none was necessary. Owsjanka created customized alarm triggers to provide specific information including tolerances for displacement at any point with pre-programmed alert levels at 8, 15 and 20 millimeters (0.3, 0.6, 0.8 inches). This allowed him to observe trends rather than ad hoc measurements. Other sets of alarms alert him to factors such as instrument maintenance, damage to an instrument or prism or if the line of sight to a prism has been obstructed.

By using the software’s filtering process to manage who receives the alarms, Owsjanka allows the Morgan Sindall monitoring team to establish a new level of security. If an alarm is triggered, an email alert is sent to just three selected members of the Morgan Sindall monitoring team. One of the three team members can then immediately log onto the system, even from home, and analyze the data before deciding if action is necessary and if additional parties need to be alerted.

Typically, an initial alarm is triggered if measurements show prism movement of 8 millimeters (0.3 inches). This information is compared to the findings from the Senceive tilt sensor system and the prism is then “watched” over the next three monitoring cycles, a period of three hours, to detect any changes.

Giles described a recent example of the system’s performance. An alarm had triggered at 8 millimeters, alerting the three selected members of the Morgan Sindall team. They then closely monitored the affected prism through its subsequent cycles. No further movement was detected over the next three months. Network Rail was made aware of the 8-millimeter alarm from the outset through the regular update meetings. But because no action was necessary, Morgan Sindall’s filtering process prevented distraction of other parties. Subsequently, when a second alarm was triggered at 15 millimeters (.06 inches), the monitoring team immediately alerted Network Rail and all stakeholders were informed with full confidence that the alarm was genuine. The track was realigned in less than 12 hours.

Turning Data into Information

Morgan Sindall produces a daily monitoring report for Crossrail. The report is a part of the Shift Review Group made up of Crossrail, DLR and Network Rail that meets every morning. The textual report details any recorded movement (including a graph if an alarm has been triggered). A report generated by T4D software is also used at a weekly review panel meeting. Additionally, all the data is uploaded, in Crossrail’s preferred format, to the underground construction information management system (UCIMS) developed to monitor construction data for Crossrail. Giles notes that using customized reports has enabled Morgan Sindall to observe and analyze trends quickly and efficiently and present their findings in formats preferred by project stakeholders.

Owsjanka says that flexibility is key to smooth operations of the optical system. He can set up algorithms to provide customized analysis, which is especially useful for detecting and examining trends. For example, if settlement on a prism is detected, he can immediately set up a graph to compare its behavior with the five other prisms in closest proximity to provide a wider picture. As the project progresses toward its target end date, Owsjanka can provide increasingly sophisticated information and analyses.

Giles and his stakeholders are happy with the system performance. “Managing the monitoring on this project in-house has brought us innumerable benefits,” he says, “not least of which are satisfied clients and stakeholders. Should a significant movement in the track occur, there is no ambiguity. Consequently, in the case of a genuine alert, trains can be stopped immediately.” Network Rail and DLR have reported their confidence in the monitoring as well. Morgan Sindall has delivered on the requirements for a robust, reliable and repeatable system.

Postscript

Construction on the Pudding Mill Lane project reached completion in May 2017 and Morgan Sindall decommissioned its monitoring system. Giles says that the system performed well for the entire project and that the solution has become a benchmark for automated monitoring of rail track.

Morgan Sindall plans to relocate the monitoring equipment to install a new monitoring system on a railway improvement project for Network Rail on England’s east coast with the potential to upgrade to high-speed rail service.