Master's lecture in Mechanical Engineering-Rebecca Anne Jones

Master's student: Rebecca Anne Jones

Title: The design of slushflow barriers: CFD simulations

 ___________________________________________

Faculty:  Faculty of Industrial Engineering, Mechanical Engineering and Computer Science

Advisor:  Halldór Pálsson, Professor at the Faculty of Industrial Engineering, Mechanical Engineering and Computer Science

Also in the masters committee: Ásdís Helgadóttir, Assistant Professor at the Faculty of Industrial Engineering, Mechanical Engineering and Computer Science

Examiner: Vigfús Arnar Jósefsson, Mechanical Engineer at Verkís

Abstract

Due to the cost of human lives near avalanche and slushflow prone locations, it is important to understand these flows and their interaction with dam barrier protection in a laboratory set-up. A previous project constructed a small-scale inclined chute that observed different dam configurations (Ágústsdóttir 2019).

In this thesis, the experiment’s observations were used to verify results of 2-D CFD (Computational Fluid Dynamics) calculations, since those are known to have a possible risk of inaccuracy and delusive results. The software, OpenFOAM, was used to construct a mesh, establish boundary conditions, and calculate flow characteristics to replicate the experiment’s high-Reynolds behavior and measurements.

Comparison of experimental and simulated velocity, flow thickness, splash and hydraulic jump heights to find the most accurate match with the experiment. The simulation’s cell size, 0.05 m by 0.025 m, and roughness height of 0.002 m produced the most similar profile with the experiment. For further observations, dams were constructed individually with arrangements similar to the experiment. Parameter results showed that a 95 ◦ simulated dam was most similar to the experiment, and a 34° dam was least similar. With the exception of initial splash differences, the case with two small dams was also similar with the experiment.

This thesis was able to show that 2-D CFD simulations can accurately predict velocity and Froude number parameters, but poorly predict splash and hydraulic jump heights. The simulations can be used to the benefit of fast and low-cost elements at least for velocity and Froude number predictions.