Numerical Modelling of the Interaction Between River Ice and Control Structures

«ABSTRACT: Northern hydro-systems are greatly influenced by ice processes and especially river ice breakup events that occur in early spring. Those events are characterized by a large ice floe discharge in rivers that can sometimes damage rivers’ infrastructure or lead to ice jams and related floods. Some regions are regularly affected by severe ice jam events that threaten residential areas and damage ecosystems. Thus, such events pose security issues and engineering challenges when designing river infrastructures. Many cities have installed ice control structures (ICSs) often made of piers to deal with river ice breakup-related issues. These structures aim to mitigate ice jams by holding back ice floes upstream from specific locations. River ice breakup events are characterized by many discrete ice floes evolving in a highly dynamic river flow with a fragmented and deformed free surface. Understanding and predicting how ice floes interact with these control structures in this context is essential to optimize the design of ICSs. To this end, numerical models which are more flexible, more transferable, and less expensive than experimental studies have been used as powerful tools to study those kinds of phenomena. Nevertheless, the complexity of such multiphysics phenomena that involves dynamic flow features and the interaction of many solids challenges the numerical methods in terms of accuracy and computational cost. Classical Eulerian mesh-based methods that are often used for fluid resolution in many mechanical problems have shown their limitation in dealing with such highly dynamic cases where free surface detection is challenging. On the other hand, Lagrangian particle-based methods such as Moving Particles Semi Implicit (MPS) or Smoothed Particles Hydrodynamics (SPH) have shown their efficiency in dealing with violent flows thanks to their mesh-free and Lagrangian nature. Coupled with the Discrete element method (DEM), which has proved its efficiency in dealing with solid motion and contacts, it provides a fully mesh-free Lagrangian model suitable for the simulation of such phenomenon. The simulation of large-scale events with fully Lagrangian methods remains an open research problem as it challenges the stability of those methods and remains computationally expensive. To address this knowledge gap, we developed and used a threedimensional fully Lagrangian numerical model to study ice-control structures interaction in a river breakup context. The model was developed using the open-source code DualSPHysics coupled with the open-source code Project Chrono. DualSPHysics deals with the fluid phase using the SPH method while Project Chrono uses the DEM for the solid motion resolution. The parallel implementation of the solvers used allows the simulation of large and detailed cases in reasonable computational time.»