Numerical and Experimental Investigations of Horizontal Turbulent Particle–Liquid Pipe Flow

Abstract

A Eulerian–Eulerian computational fluid dynamics approach is used in conjunction with appropriate auxiliary models for turbulence and solid dynamic properties to study the complex turbulent flow of particle–liquid suspensions in a horizontal pipe. Numerical simulations of the detailed flow field are fully and successfully validated using a unique experimental technique of positron emission particle tracking. The study includes nearly neutrally buoyant as well as dense particles, ranging from small to large at low to high concentrations, conveyed by a Newtonian liquid. Results are analyzed in terms of radial particle and liquid velocity profiles as well as particle distribution in the pipe. The approach provides predictions with a high degree of accuracy. Particle behavior can be classified into three categories depending on their size and particle–liquid density ratio. An analysis of the forces governing the two-phase flow is used to interpret the phenomena observed.

Numerical study of steady pipe flows and head loss coefficients

“…Having created the pipe geometry, a structured mesh consisting of a finite number of hexahedral cells was generated. This mesh is considered to have better accuracy than an unstructured mesh [9,10]. In the case of pipe with the bend, the fluid velocity on the inlet, uinl, was set to 0.12 m/s.…”

Experimentally trained hybrid machine learning algorithm for predicting turbulent particle-laden flows in pipes

“…Full experimental details and protocols can be found in our previous papers. [35][36][37][38] The experimental particle-liquid flow conditions under which the data bank utilized for training the ML algorithm as well as assessing its performance was generated are listed in Table I.…”

“…The tracer instantaneous velocity is first estimated by a regression analysis using a number of locations and time obtained from the raw PEPT trajectory data. 35 As shown in Fig. 1, the mean velocity inside a given cell ( u m ) is calculated by averaging all instantaneous velocities (u i ) in the cell.…”

“…First, accurate Lagrangian phase trajectories were obtained from the raw experimental PEPT data (t, x, y, z) using a regression analysis. 35 Combining these Lagrangian PEPT trajectory data (x, y, z, u x , u y , u z ) with the associated experimental conditions (C s,m , q r , d p ,), a new data array was generated. Since the flow in the pipe section visualized by PEPT was steady and fully developed, the axial x position was irrelevant.…”