Report

Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Master’s Dissertation Defense Carlos M. Teixeira Supervisors: Prof. José Carlos Lopes Eng. Matthieu Rolland 17th July 2013 Outline Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Introduction Objectives State of the Art Methodology Results and Discussion Conclusions FEUP/IFPEN 2 Introduction Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Catalysts performance evaluation Performed in units at pilot scale The trend is to reduce the size of testing units (economic and safety reasons) Catalyst size remains constant (customer demands) Consequence Reactors with low tube-to-particle diameter ratio (1 < / < 5) FEUP/IFPEN 3 Introduction Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Reactors with low tube-to-particle diameter ratio Pseudo Homogeneous Models may not be valid Local Phenomena are dominant Wall Effect Packing Effect FEUP/IFPEN 4 Introduction Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Example of Packing Effect Problem Description 0.8 Packing of eight cylinders with 0.7 different arrangements Laminar regime Cylinders with constant concentration in their surface Transfer solute to the fluid FEUP/IFPEN 0.5 0.4 0.3 0.2 Particles in contact the inlet flows through the packing Cout/Csurface Fluid with zero concentration at 0.6 0.1 0 Normalized outlet concentration for the different arrangements 5 Outline Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Introduction Objectives State of the Art Methodology Results and Discussion Conclusions FEUP/IFPEN 6 Objectives Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Study the phenomena of single phase fluid flow through fixed-bed reactors at low particle Reynolds number Understand how the packing structure affects the flow FEUP/IFPEN 7 Outline Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Introduction Objectives State of the Art Methodology Results and Discussion Conclusions FEUP/IFPEN 8 State of the Art Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing CFD Simulation of Fixed-Bed Reactors Benchmark Method: Lattice Boltzmann Finite Volume method has been successfully used by many authors In most published works, the ratio of tube-to-particle diameter is low (/ < 10) FEUP/IFPEN 9 State of the Art Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing CFD Simulation of Fixed-Bed Reactors Coupling between Hydrodynamics, Heat Transfer and Chemical Reaction: Less works on the literature Applied in small size problems (dozens of particles) Particle shape: mostly spheres FEUP/IFPEN 10 Outline Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Introduction Objectives State of the Art Methodology Results and Discussion Conclusions FEUP/IFPEN 11 Methodology Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Coupling between DEM and CFD GRAINS3D (Packing Simulation) FEUP/IFPEN PeliGRIFF (Fluid Flow Simulation) 12 Methodology Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Grid Refinement Studies Relative Error in U inlet 1 10-1 0° Re=0.01 0° Re=50 45° Re=0.01 45° Re=50 90° Re=0.01 90° Re=50 10-2 -3 10 10 100 d p/h 1000 Relative error in the inlet velocity as a function of the grid resolution (ε=0.799, l/dp=1) FEUP/IFPEN 13 Outline Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Introduction Objectives State of the Art Methodology Results and Discussion Conclusions FEUP/IFPEN 14 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Structured Packed Beds Unit cell approach (a) (b) A packed bed of simple cubic arrangement of spheres. a) Unit cell b) Alternative FEUP/IFPEN representation of a simple cubic unit cell. 15 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Structured Packed Beds of Spheres Validation Case Comparison between the simulated dimensionless pressure drop and results from Hill et al. (2001) for a dilute array of spheres (ε=0.799) FEUP/IFPEN 16 Direct Numerical Simulation of Fixed-Bed Reactors: Results and Discussion Effect of Random Packing Flow through Structured Packed Beds of Cylinders Effect of cylinder orientation 100 0° Dimensionless Pressure Drop, ϕ 45° 90° 10 1 0.1 1 10 100 1000 Redp Effect of cylinders orientation on dimensionless pressure drop (ε=0.799, l/dp=1) FEUP/IFPEN 17 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Structured Packed Beds of Cylinders Transition from laminar regime to unsteady and chaotic flow Particle Reynolds number as a function of time for 45º orientation (ΔP=10 Pa) FEUP/IFPEN 18 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Case ID FBR1 FBR2 FBR3 Nº of particles 540 200 100 0.451 0.444 0.467 Porosity, ε FEUP/IFPEN 19 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Simulated Packed Beds Grid parameters and computing times on 128 processors ( = 1) FEUP/IFPEN 20 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Pressure Drop Dimensionless pressure drop as a function of porosity. Comparison between simulations and Ergun correlation predictions (Redp=1). FEUP/IFPEN 21 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Spatial Velocity Distribution Three different zones are identified: Recirculation zones in the packing top and bottom and in the wake of the particles (with negative velocities) High velocity zones where the void fraction is small and the velocity increases up to a factor of 15 Low velocity zones near the particles surfaces FEUP/IFPEN 22 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Statistical Velocity Distribution 1.4 Inlet Z2 Z3 Z4 Z5 Z6 Z7 Z8 Z9 Z10 Z11 Z12 Z13 Z14 Z15 Outlet Entire Domain 1.2 P (U z /U inlet) 1 0.8 0.6 0.4 0.2 0 -2 0 2 4 U z/U inlet 6 8 10 Probability density functions of normalized z-velocity in different zones of FEUP/IFPEN the fixed-bed. 23 Direct Numerical Simulation of Fixed-Bed Reactors: Results and Discussion Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Statistical Velocity Distribution (link with porosity) 0.85 Porosity, ε 0.75 0.65 0.55 0.45 0.35 0 Inlet 0.2 0.4 z/L 0.6 0.8 1 Outlet Axial average porosity profile FEUP/IFPEN 24 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Statistical Velocity Distribution (link with porosity) Probability density functions of normalized z-velocity for different porosities ( = 1) FEUP/IFPEN 25 Results and Discussion Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds of Cylinders Statistical Velocity Distribution Probability density functions of normalized x-velocity for different porosities ( = 1) FEUP/IFPEN 26 Outline Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Introduction Objectives State of the Art Methodology Results and Discussion Conclusions FEUP/IFPEN 27 Conclusions Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Structured Packed Beds The methodology was validated with well-established cases from the literature Dependence of Pressure Drop across Packed Beds of cylinders on its orientation was studied Transition from steady laminar flow to time oscillatory and chaotic flow was observed at ≥ 60 FEUP/IFPEN 28 Conclusions Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Flow through Randomly Packed Beds Good agreement between Ergun’s pressure drop predictions and simulation results were found Velocity distributions were analyzed and three different zones were identified Velocity distributions appear to follow the average local porosity: the length to establish the flow is identical to the length to establish the porosity FEUP/IFPEN 29 Direct Numerical Simulation of Fixed-Bed Reactors: Effect of Random Packing Thank you for your attention www.ifpenergiesnouvelles.com FEUP/IFPEN 30