Assessing the areas impacted by overland spills of high pour point oils originating from well blowouts, tank breaks and pipeline leaks is highly dependent on the changes in the fluid physical properties as the oil cools down. Since the oil outflow temperatures are often higher than ambient temperature, as oil approaches the ambient temperature, the fluid viscosity and yield stress increase thus restricting the flow and/or stopping it altogether. In this work we discuss the application of a novel approach to overland spills of high pour point oils based on a viscous flow two-dimensional model: OilFlow2D. The model formulation considers that density, viscosity, and yield stress can vary in space and time and involves solving the continuity and momentum equations for free-surface non-Newtonian fluids and the energy equation to determine the oil temperature over time. The numerical solution applies a finite volume method on flexible triangular meshes and is parallelized for Graphics Processing Unit (GPU) devices, offering significantly reduced computational times compared with sequential code. We present a illustrative application to a tank break of a high pour point heated oil that starts flowing in an environment with lower temperature, and subject to wind speed, and solar radiation that vary continuously during the simulation time. The oil is characterized by its specific heat and experimental data representing how the oil density, viscosity, and yield stress depend on temperature. The results show that for a high heat transfer rate, the oil starts cooling as soon as it flows out of the tank and its properties vary as it spills over an irregular topography. As the oil flows away from the tank, it cools down until it eventually comes to a complete stop. We include two scenarios to compare the effect of the wind velocity on the heat transfer of the oil with the environment. The results show that the spill occurring in high-speed wind conditions (20 m/s) have a much shorter runout distance than a spill occurring in relatively low wind speed conditions (2 m/s). We also present maps of oil thickness/depth, oil viscosity and yield stress that show how these properties vary between the two scenarios. The proposed approach, that can also account for the mitigation effect of flow barriers, enables realistic spill simulation of high pour point viscous oils considering complex topography and constitutes a novel and practical approach to oil spill risk evaluation, supporting emergency action plans and contingency plans.
