Experience of using FEM for real-time flood early warning systems: Monitoring and modeling Boston levee instability

• Presentation of a real-time, physics-based FE module for dike stability analysis as a functional part of the UrbanFlood early warning system workflow.
• Discussion of functional interaction of the FEM module with other real-time components of the early warning system.
• Validation of the FE module on the test site – the Boston levee (UK) prone to slope slippages
• Analysis of the recorded sensor data.
• Results of the numerical slope stability analyses: comparison of the conventional LEM (limit equilibrium method) technique to the FEM simulation; discussion of the error sources in LEM.

Boston levee has a documented history of slope slippages under tidal fluctuations reaching 6-m range on spring tides. A finite element model of the Boston levee has been developed in the off-line mode; after that, it was integrated into the common information space of the UrbanFlood early warning system and connected to live sensor data stream registering tidal fluctuations of river level and corresponding response of pore pressures and media temperatures. Stability analysis was carried in a real-time mode.Besides finite element method, limit equilibrium analysis of levee stability was used in parallel. Real-time assessment of stability was performed by interpolating between safety factors pre-computed for a number of different tidal phases. FEM results indicate instability and agree with real-life observations, while LEM predicted safe condition of the dike. The mismatching is caused by a simplified procedure of pore pressures calculation which is typically used in LEM. For clayey levees, this simplification can lead to over-estimation of slope stability.The Boston levee case has become one of the pilot sites for validation of the UrbanFlood early warning system (EWS) for flood protection. This validation case has shown that the FEM module for levee stability analysis successfully operates in the real-time workflow of the EWS and correctly predicts levee instability.