APCOM & ISCM 11-14 th December, 2013, Singapore 1 Simulation of Bubbly Flow in a Vertical Pipe Using Discrete Phase Model *H.Y. Li 1 , J. Lou 1 , Z. Shang 1 and H. Tang 2 1 Institute of High Performance Computing (IHPC), Agency for Science, Technology and Research (A*STAR), 1 Fusionopolis Way, #16-16 Connexis, Singapore 138632 2 School of Mechanical & Aerospace Engineering, Nanyang Technology University, 50 Nanyang Avenue, Singapore 639798 *Corresponding author: lih@ihpc.a-star.edu.sg Abstract Bubbly flow is widely encountered in many engineering applications, such as those in chemical and nuclear systems, bubble column reactors and oil transportation pipes. Therefore, understanding of bubbly flow in a bubble-liquid flow system is extremely important. In this paper, bubbly flow involved with thousands of bubbles in a vertical pipe is numerically simulated. The motions of the bubbles are tracked using a Discrete Phase Model (DPM) and bubble-bubble interactions are simulated through the model of discrete element method (DEM). The effects of bubble diameter on the bubble flow trajectories are studied. Comparisons are made on the flow field with and without considering bubble-bubble collision. Keywords: Bubbly flow, Discrete phase model, Bubble trajectory, DEM collision. 1. Introduction Bubbly flow is widely encountered in many engineering applications, such as oil and gas pipes, chemical and nuclear systems (Oolman and Blanch, 1986; Chen et al., 1994) and bubble column reactors (Jakobsen, 2001). In these systems, millions of bubbles are dispersed into a continuous phase which is the carrier fluid. The movements of these bubbles have significant effects on the flow fields as well as the pressure drops in the systems. Therefore, understanding the dynamics of the bubbles is essentially important to know bubbly flow. Experimental investigation of bubbly flow has been performed extensively (Liu and Bankoff, 1993a, 1993b; Gnotke et al., 2003; Daeseong et al., 2010). For experimental study, it generally requires large length scale test rig and high resolution measuring instruments to provide convincing data. These would lead to an extremely high cost. Meanwhile, it is rather difficult to capture the physical phenomenon occurred for each individual bubble in the experiments. In view of this, theoretical studies, in particular numerical simulations, play an essential complementary role in understanding the bubble dynamics in bubbly flow. Bubbly flow generally involves two phases which are the carrier fluid and the bubbles. The carrier fluid is usually treated as the continuous phase in the numerical simulation. Bubbles can be treated either as a continuous phase or a discrete phase based on the methods bubbles are handled. These methods include Eulerian-Eulerian (EE) two fluid method, Lagrangian-Eulerian (LE) method and Direct Numerical Simulation (DNS) (Hirt and Nichols, 1981; Unverdi and Tryggvason, 1992; Shan, 1997; Osher and Sethian, 1988; Quan and Schmidt, 2007). EE (Drew, 1983; Enwald, 1996) two fluid model assumes bubble as another continuous phase which the average size and average velocity are chosen to represent the information for all the ranges of bubbles. Although EE model can be applied in the large scale system with both spatial and time, it is not able to represent two streams of bubbles with different velocities at the same location. The interactions among bubbles are usually not considered either. This results in the unrealistic simulations of the physical phenomena observed in the bubbly flow. Unlike EE model, LE model and DNS treat the bubbles as a discrete phase. DNS can reveal the useful detailed insights of bubble behavior and bubble interactions. Generally, it can only be applied in a system where a small number of bubbles are considered. For bubbly flow which involves thousands of bubbles, LE could be the most appropriate choice. In LE model, bubbles are represented in a Lagrangian reference frame while the carrier phase is represented in an Eulerian frame. Under such a treatment, the movement for each individual bubble in bubbly flow could be traced. The interactions among bubbles such as bubble