Surveying Crossrail’s tunnels from space

Using satellite radar, a team of engineers was able to monitor how Crossrails tunnelling impacted the buildings above. Andrew Wade reports.


The subterranean world beneath London is some of the most heavily tunnelled of any major city. From Bazalgette’s sewer system to the ever-evolving Tube network, the UK’s capital has for centuries plumbed its depths to provide all manner of utilities for the populace. With recent projects such as Crossrail and the Thames Tideway Tunnel adding yet more complexity underground, the need to assess the impact of tunnelling on surface structures has never been greater.

At the University of Bath, Dr Giorgia Giardina has been employing a new technique that uses satellite imagery to monitor surface building deformation, verifying the method by way of the Crossrail tunnels.

Combined with traditional monitoring techniques, it’s a breakthrough that could provide engineers with a more accurate and comprehensive tool for assessing the impact of subterranean construction. The technology –known as InSAR (Interferometric synthetic aperture radar) –has evolved over the past decade with the launch of X-band satellite constellations such as the Italian Space Agency’s COSMO-SkyMed.

“Civilian X-band satellites are characterised by improved spatial resolution together with a reduced revisit time, varying from a couple of weeks down to a few days,” said Dr Giardina (below), lead researcher on the project and lecturer at Bath’s department of architecture and civil engineering.

“The high-resolution of the COSMO-SkyMed X-band SAR data results in an improvement of 320 per cent and 550 per cent with respect to (the longer-serving) RADARSAT-1 and ENVISAT data C-band satellites. The X-band wavelength also enables unprecedented accuracy of the order of millimetres in the InSAR monitoring of ground and structural deformations in urban areas.

“Second-generation satellites were expected to provide more accurate data and a short-term monitoring of natural and man-induced hazards, like earthquakes, landslides and glacier movements. What was not clear until the actual processing of ground deformation data in urban areas was the full extent of the X-band satellite potential for structural engineering applications and in particular in integration with damage assessment tools. We are filling the gaps between two branches of electrical and structural engineering previously unexplored.”

Anyone working in central London during the Crossrail tunnelling will likely have seen teams of engineers using geodetic prisms and manual levelling points to monitor building displacement. In Soho and other areas particularly vulnerable to damage, total stations were deployed that can monitor up to 100 levelling prisms on surrounding buildings, providing constant observation while the tunnel boring machines carved their way under the city.

However, for the majority of buildings above large-scale tunnelling projects, only a few points can be monitored constantly. Integrating InSAR with ground-based levelling allows virtually every point on every building to be observed, providing much more insight than just using traditional techniques on their own.

“During the Crossrail project, several commercial companies provided satellite-based monitoring, comparing them to ground-based data to prove the high quality of satellite measurements,” Dr Giardina told The Engineer. “In our study, an initial comparison with ground-based data was only used as a validation of the InSAR measurements.

“InSAR data was then integrated to structural assessment procedures to prove the higher spatial resolution of satellite measurements with respect to the ground-based measurements, which were typically available for the buildings located along the Crossrail route, and also the additional insight –enabled by the use of InSAR data –on soil structure interaction mechanisms. In both, ground-based measurements would not have been sufficient to obtain the same results.”

To assess the expected impact of ground movements on existing structures, an estimate of the shape and magnitude of these movements must first be made. This requires information on the source of ground movement, such as whether it was caused by underground construction or a landslide. The features of the source must also be considered. With tunnelling, for example, one needs to know whether cut-and-cover or bored tunnels were used. Lastly, the soil type plays a key role, as anyone who has dug through London’s mix of clay, chalk and alluvium can attest.

The shape and magnitude of the deformations affecting a building are also influenced by the characteristics of the building itself, including its dimensions, weight and stiffness. This soil-structure effect is typically not accounted for in damage prediction, potentially leading to an inaccurate damage assessment. Using InSAR, Dr Giardina and her team were able to accurately determine how buildings and soil were interacting with each other.


In terms of out and out precision, the sub-millimetre accuracy of InSAR compares quite favourably to 3D prism technology, though is not quite in the same league as precise levelling techniques that can measure changes of fractions of millimetres. Current satellite revisit times mean that InSAR cannot provide real-time monitoring, though future constellation expansions could go some way to addressing that.

Nonetheless, the radar technique can provide high-resolution imagery day and night, independent of weather conditions. Another advantage is the fact that it is not susceptible to some of the more bizarre –though not necessarily infrequent –earth-based interference, such as vandalism and fouling from pigeons. Interestingly, the InSAR technique can also be retrospectively applied using older satellite data, allowing structural engineers to evaluate the response of buildings which were not part of an original monitoring plan.

“The high spatial resolution and accuracy of InSAR monitoring makes it already comparable to ground-based techniques, with the advantage of reduced costs, large area mapping and retrospective application,” Dr Giardina explained.

“The current satellite revisit time of a few days only allows near real-time monitoring, while ground-based techniques can provide real-time measurements. Future satellite constellations are expected to fill this gap between InSAR and ground-based monitoring by shortening the revisit time even further, potentially offering a substitute for ground-based techniques.”

Although the use of satellites for structural monitoring is in its early days, the Bath team is hopeful that its research and the success of the Crossrail project could help give the technology wider recognition across the industry.

Longer term, Dr Giardina hopes to develop a type of early warning system that could be deployed over large urban areas, where huge amounts of InSAR data could alert engineers to problems much more rapidly. This would require some level of automation whereby relevant measurements were filtered from the millions of data points collected.

“Thanks to their higher resolution and reduced revisit time, second-generation satellites, like the X-band satellites used in this study, can provide millions of monitoring points for a single city,” said Dr Giardina.“This huge amount of data is a limiting factor to its effective use. An automatic tool able to detect the most significant measurements would reduce the number of points to be analysed from millions to thousands, allowing data analysts to focus only on the most relevant areas.”