Results

The project was submitted for and included in the Case Study Library of the OECD's Observatory of Public Sector Innovation.

The hope was that the project would demonstrate that INSAR can be used to assess and monitor deformations in bridges. Some of the results are positive for the application of INSAR, while others make it more difficult to deploy this technique immediately.

Key results that would enable the use of INSAR: 

• Although EGMS (European Ground Motion Service) now offers a free, highly useful service that performs an annual INSAR analysis across Europe (https://egms.land.copernicus.eu), the study shows that high-resolution data is needed to interpret the deformation of bridges. This is because, on the one hand, the horizontal positioning of the measuring points is much better with high-resolution images and, on the other hand, because bridges are relatively small structures, where a higher number of points provide a better picture of their behavior.
• The analysis showed that for some bridges (e.g., the Vilvoorde viaduct), monitoring with INSAR should be possible. The viaduct is not currently being monitored due to difficult accessibility, and the INSAR analysis yielded good results here (for one of the bidders). Larger structures in particular allow for long-term monitoring with INSAR.
• Because so many measurement points are available over time (1 measurement/14 days with InSAR instead of 1 measurement/3 years in the best case with levelling), the behaviour of a bridge can be better understood. The influence of the seasons on the behavior of the bridge, as well as its long-term behavior in the event of any settlement, can also be detected more quickly. Currently, there are sometimes displacements due to thermal changes, which make it difficult to know whether a bridge is actually deforming or simply deforming due to temperature variations.
• An annual INSAR analysis would provide significantly more information for interpreting the deformations than the current water level measurements.
• A fairly large area (with a large number of bridges) is covered by a single analysis.
• It is possible to draw up a report for each bridge, making the measurements quickly and clearly available to those who need to interpret them. It should be possible to link this to the management application used internally by the government. The shapefiles of the bridges from the management application were already used in the implementation of this project.
• Several known locations with larger, known deformations also clearly emerge from the analysis. The results here correspond to expectations based on other measurement techniques.
• There are certainly major safety advantages for those carrying out the measurements.

Key findings that still hinder the use of INSAR:

• The POC did reveal a difference between the results of the various tenderers. This means that the analysis depends on the software used and the settings chosen (reference point, cut-off values for a reliable measuring point, atmospheric correction, elevation model correction, etc.). The chosen measurement period also has an impact on the results.
• There was a difference between the analysis during the POC and the analysis of the same area during the implementation of the project. In addition to the impact of the chosen reference point, it appeared that some manual interpretation/analysis had been carried out during the POC, which is not easily scalable.
• Not all bridges have the same number of measuring points available. The dimensions, but also the material from which the bridge is made, the orientation of the bridges and any neighboring buildings/trees that may cause a shadow effect, influence the number of measuring points available on a bridge.
• The exact position of the measuring points in the horizontal plane depends on certain choices that are made (including the elevation model used) and is therefore also dependent on the analysis.

Key findings that complicate the decision to purchase high-resolution images on a large scale:

• Due to the short period for which the high-resolution images were available for analysis (just under two years), mainly/only larger displacements/deformations can be detected. The deformations of bridges are often relatively small/slow, making it difficult to estimate whether the deformations of the bridges are being captured.
• Due to the limited measurement period and the long interval between the levelling measurements, it is difficult to compare the available results of the levelling measurements directly with the results of the INSAR analysis. 

Opinion of end users

The engineers who have to use the technology in their daily work are somewhat sceptical. These engineers are used to working with highly accurate measurements and very well-known measurement points. They have many questions about a technique whose workings they do not fully understand (the satellite flies at a great distance from the Earth, never flies in exactly the same position, and the results are said to be accurate to the millimetre per year). A major stumbling block appears to be the position of the measuring points in the horizontal plane. Both the fact that the location of the measuring points cannot be chosen and the fact that it is difficult for them to know which point/zone causes the reflection is difficult for the engineers.

Further roll-out / next steps

Firstly, a decision must be made as to whether we want to purchase high-resolution satellite images within Flanders/Belgium or whether we believe that the service offered by EGMS is sufficient to answer the questions posed in advance (can this data be used to optimise the use of topographical measurements, can the data be used to detect abnormal behaviour at an early stage/more quickly)? In this case, we could opt to purchase satellite images only in densely populated areas/areas with a lot of infrastructure works/areas where we know that the subsoil is susceptible to subsidence and/or instability, etc. We are thinking in particular of larger cities (outskirts of large cities). A “mixed” scenario seems ideal, whereby images are purchased for a few specific locations and EGMS is used for the other areas.

If the decision is made to purchase high-resolution images, a framework contract for analysing the measurement data can be drawn up within two years. However, it will be a challenge to achieve the desired quality of analysis.

It is essential that, in the short term, consideration is also given to how and where the EGMS analyses can be stored and visualised within Flanders/Belgium (the annual analysis always includes the analysis of images from the past five years, which makes it useful to also store and visualise historical analyses).