Geodetic monitoring of complex shaped infrastructures using Ground-Based InSAR

In the context of climate change, alternatives to fossil energies need to be used as much as possible to produce electricity. Hydroelectric power generation through the utilisation of dams stands out as an exemplar of highly effective methodologies in this endeavour. Various monitoring sensors can be installed with different characteristics w.r.t. spatial resolution, temporal resolution and accuracy to assess their safe usage. Among the array of techniques available, it is noteworthy that ground-based synthetic aperture radar (GB-SAR) has not yet been widely adopted for this purpose. Despite its remarkable equilibrium between the aforementioned attributes, its sensitivity to atmospheric disruptions, specific acquisition geometry, and the requisite for phase unwrapping collectively contribute to constraining its usage. Several processing strategies are developed in this thesis to capitalise on all the opportunities of GB-SAR systems, such as continuous, flexible and autonomous observation combined with high resolutions and accuracy.

The first challenge that needs to be solved is to accurately localise and estimate the azimuth of the GB-SAR to improve the geocoding of the image in the subsequent step. ... mehrA ray tracing algorithm and tomographic techniques are used to recover these external parameters of the sensors. The introduction of corner reflectors for validation purposes confirms a significant error reduction. However, for the subsequent geocoding, challenges persist in scenarios involving vertical structures due to foreshortening and layover, which notably compromise the geocoding quality of the observed points. These issues arise when multiple points at varying elevations are encapsulated within a singular resolution cell, posing difficulties in pinpointing the precise location of the scattering point responsible for signal return. To surmount these hurdles, a Bayesian approach grounded in intensity models is formulated, offering a tool to enhance the accuracy of the geocoding process. The validation is assessed on a dam in the black forest in Germany, characterised by a very specific structure.

The second part of this thesis is focused on the feasibility of using GB-SAR systems for long-term geodetic monitoring of large structures. A first assessment is made by testing large temporal baselines between acquisitions for epoch-wise monitoring. Due to large displacements, the phase unwrapping can not recover all the information. An improvement is made by adapting the geometry of the signal processing with the principal component analysis. The main case study consists of several campaigns from different stations at Enguri Dam in Georgia. The consistency of the estimated displacement map is assessed by comparing it to a numerical model calibrated on the plumblines data. It exhibits a strong agreement between the two results and comforts the usage of GB-SAR for epoch-wise monitoring, as it can measure several thousand points on the dam. It also exhibits the possibility of detecting local anomalies in the numerical model. Finally, the instrument has been installed for continuous monitoring for over two years at Enguri Dam. An adequate flowchart is developed to eliminate the drift happening with classical interferometric algorithms to achieve the accuracy required for geodetic monitoring. The analysis of the obtained time series confirms a very plausible result with classical parametric models of dam deformations. Moreover, the results of this processing strategy are also confronted with the numerical model and demonstrate a high consistency. The final comforting result is the comparison of the GB-SAR time series with the output from four GNSS stations installed on the dam crest.

The developed algorithms and methods increase the capabilities of the GB-SAR for dam monitoring in different configurations. It can be a valuable and precious supplement to other classical sensors for long-term geodetic observation purposes as well as short-term monitoring in cases of particular dam operations.