what‘s happening inside fixed-bed reactors? berlin –chemical and process engineering wehinger,...

TU Berlin – Chemical and Process Engineering

Wehinger, Jurtz, KraumeSTAR global conferenceBerlin, 2017/3/6-8

Slide 1

What‘s happening inside fixed-bed reactors?

Gregor D. Wehinger, Nico Jurtz, Matthias Kraume

Chemical & Process Engineering

Technische Universität Berlin

Berlin, 06. März 2017

Ysop



Slide 2

Syngas production from marginal gas resources

Crude oil Natural gas

Syngas (H2 + CO)



Slide 3

How to produce syngas?

Steam reforming of methane (SRM)

Partial oxidation of methane (CPOX)

Dry reforming of methane (DRM)

• Catalytic endothermic process:

CH4 + CO2 ⇋ 2H2 + 2CO Δ𝐻 = 247 𝑘𝐽/𝑚𝑜𝑙

• Mainly cheap nickel-based catalysts

• Performed in fixed beds at 700-1000 °C

• Drawbacks: hotspots and coke

formation

➢ What’s happening inside the reactor?



Slide 4

How to design DRM fixed-bed reactors?

• Highly endothermic reaction

• Heat transfer into the reactor small

tube diameter

• High gas velocities and small pressure

drop large particles

• Small tube-to-particle-diameter-ratios

(1 < D/d = N < 16)

D

d

N=1.8 N=4 N=8 N=16



Slide 5

How to model small N fixed beds?

• Inhomogeneous bed structure for small

N fixed beds:

➢ Significant wall effects

➢ Local backflows

• Large axial and radial gradients

• Local interactions between kinetics and

transport phenomena

• Conventional description based on plug

flow and pseudo-homogeneous kinetics

➢ Adequate modeling with full CFD and

detailed reaction models



Slide 6

What do we expect from a CFD model?

Morphology

Local velocities

Heat transfer

Heterogeneouscatalysis

Validate against experiments



Slide 7

Workflow of particle-resolved CFD simulations

DEM simulation to generate random packing

Eppinger et al. (2011) Chem. Eng. J. 166(1), 324-331Wehinger et al. (2015) Chem. Ing. Tech. 87, 734-745



Slide 8


Meshing of calculation domain

Polyhedral cells

Prism layers




Slide 9


CFD simulation




Slide 10


Morphology

Local velocities

Heat transfer





Slide 11

Morphology & local velocities: spheres (N = 10)

Wehinger & Kraume (2017) Chem. Ing. Tech. 89(5), 1-3

Experiments from Giese et al. (1998) AIChE J., 44 (2) 484-490



Slide 12

Morphology & local velocities: cylinders (N = 10)

Experiments from Giese et al. (1998) AIChE J., 44 (2) 484-490Wehinger & Kraume (2017) Chem. Ing. Tech. 89(5), 1-3



Slide 13


Morphology

Local velocities

Heat transfer





Slide 14

Heat transfer in profile reactor

Horn et al. (2010) Rev. Sci. Instrum., 81, 064102Wehinger et al. (2016) AIChE J., 62, 12, 4436-4452

D = 18 mm N = 18

H =

25

mm

Wall temperature: 500-1000 °CVolume flow: 500-2500 mL/min

quarz capillarydC = 0.7 mm

d = 1 mm

a-aluminaspheres



Slide 15

Workflow: particle-resolved CFD

Capillarydc = 0.7 mm

DEM particleinjectiondp = 1 mm

solid

gas



Slide 16

Heat transfer in packed bed of spheres

Velocity [m/s]

Without reaction, wall temperature: 635 °CFeed: ሶ𝑉𝐴𝑟 = 2500 mLN/min

capillary

Inlet

Outlet

Hotwall

Wehinger et al. (2016) AIChE J., 62, 12, 4436-4452



Slide 17


Morphology

Local velocities

Heat transfer





Slide 18

DRM in profile reactor

Horn et al. (2010) Rev. Sci. Instrum., 81, 064102Wehinger et al. (2016) AIChE J., 62, 12, 4436-4452

D = 18 mm N = 18

H =

25

mm

quarz capillarydC = 0.7 mm

d = 1 mm

a-aluminaspheres

Wall temperature: 835°CFeed: xCH4

/xCO2/xAr = 0.32/0.40/0.28

Flow rate: 500 mLN/min

1 mm

Nickel catalyst

CH4 + CO2 ⇋ 2H2 + 2CO Δ𝐻 = 247 𝑘𝐽/𝑚𝑜𝑙

Dry reforming of methane (DRM)



Slide 19

DRM microkinetics for CFD simulations

• Microkinetics describe heterogeneous catalysis:

• Adsorption

• Surface reaction

• Desorption

• DRM microkinetics on Ni from Delgado et al. (2015):

• 52 elementary-like reaction steps

• 14 surface species, including Carboxyl (COOH),

and 6 gas phase species

• Formulation in CHEMKIN style with modified

Arrhenius equation:

Delgado et al. (2015) Catalysts, 5, 871-904

ሶ𝑠𝑖 =

1

𝐾𝑠

𝜈𝑖𝑘𝑘𝑓𝑘 ෑ

𝑗=1

𝑁𝑔+𝑁𝑠

𝑐𝑗𝜈 𝑘𝑓𝑖 = 𝐴𝑖𝑇

𝛽𝑖𝑒𝑥𝑝−𝐸𝑖𝑅𝑇

𝑘

10𝜂𝑘𝑖Θ𝑘 𝑒𝑥𝑝−𝜀𝑘𝑖Θ𝑘𝑅𝑇



Slide 20

DRM with heat transfer through solid spheres

3D Sim.

Exp.

3D Sim.

Temperature Methan

Wehinger et al. (2016) AIChE J., 62, 12, 4436-4452

Due tothermo-

dynamics ofmicrokinetics

Interplay between

thermo andspecies

development.



Slide 21

Lessons learned: DRM on Ni

1. Constant temperature over radius?

• Catalyst color is function of T and c

• CFD shows large gradients

2. Uncertainty of kinetics

• Thermodynamic consistency

• Surface coverage and coking

• Characterization of the catalyst

➢ For the future: improved experimental conditions

and analysis

➢ Still: complexity of the system remains

Third layer of catalyst after DRM

Side view after DRM



Slide 22

What’s happening inside fixed-bed reactors?

Heat transfer


✓

✓✘Local velocities

✓

Morphology

✓



Slide 23

CFD

The future of CFD and fixed-bed reactors!?

Pellet shapes Reactor parts

3D

pri

nte

rOptimization Detailed insights

Additive manufacturing

Lumped models



Slide 24

Thank you for your attention and DAAD for the Doktorandenstipendium for staying at Brown

University, and DFG in the framework of the cluster of excellence “Unifying concepts in

catalysis (Unicat)” for financial support.

Unifying Concepts in CatalysisResearch area: D1/E1 and D2/E2

Unifying Concepts

Thank you for your attention!



Slide 25

Backup



Slide 26

Application: Design exploration

Spheres Cylinders Raschig rings

➢ Quantitative comparison of RTD, local transport phenomena and local kinetics



Slide 27

Investigated cases

Case 1:

Fixed temperature from experiment

• Temperature profile from experiments

• Tcat, surf(z) = Texp(z) ≠ f(r)

• Isothermal particles

• Only gas phase and particle surface

considered

➢ Testing microkinetics

Case 2:

Heat transfer through solid spheres

• Energy exchange between gas phase and

solid phase

• Tcat, surf(z) = f(r, 𝜙, z)

• Twall and Tinlet are fixed

• Gas phase and solid phase considered.

➢ Testing particle-resolved approach



Slide 28

Case 1: fixed temperature

Tuning the kinetics with one parameter necessary

Original kinetics

• Small differences between 2D and 3D• Calculation time:

• 2D: ~3 minutes• 3D: ~3 days

3D Sim.

Tuned kinetics

Inlet

Outlet



Slide 29

Potential and weaknesses

Potentials:

• Exploration of novel particle shapes or foam structures

• No transport correlations needed

• Quantifications of local interactions between transport phenomena and kinetics

Weaknesses:

• High-tech tool, where conventional models fail

• Calculation time: several days, even on cluster

• Microkinetics valid over wide range of T, p, c and catalyst and support

composition

➢ Particle-resolved CFD simulations helps to understand catalytic flow

reactors.

what‘s happening inside fixed-bed reactors? berlin –chemical and process engineering wehinger,...

Documents