Modelling and efficiency improvement of a plasma-arc gasification reactor quench probe

(1)

Modelling and efficiency improvement of a plasma-arc

gasification reactor quench probe

J. van Rooyen

23425652

ORCid ID: https://orcid.org/0000-0002-3931-8548

Dissertation submitted in fulfilment of the requirements of the degree Master of Engineering in

Chemical Engineering at the Potchefstroom campus of North West University

School of Chemical and Minerals Engineering

North-West University Potchefstroom Campus

Supervisors: Mr. A.F. van der Merwe Prof. K.R. Uren

(2)

(3)

_____________________________________________________________________________ ABSTRACT

Page | i

ABSTRACT

A steady state computational fluid dynamics (CFD) model was developed to describe the temperature and flow characteristics inside a laboratory scale plasma-arc gasification reactor quench probe with the aim of efficiency improvement. “Thermal efficiency” was mainly described according to the thermal and fluid profiles and characteristics inside the quench probe, while the chemical efficiency, which contributed to a minor part of the study, was based on an experimental approach.

The CFD model was constructed using the STAR-CCM+ CFD simulation software in the 3-dimensional Eulerian-Lagrangian framework. The realisable k-ε turbulence two-layer model was applied to describe the physics properties of the simulation and water droplets were considered spherical and of constant density and size. An unreactive model was developed as temperatures were too low to deem the dominant reaction, i.e. the water gas shift reaction, active.

Model data was validated through comparison thereof with experimental thermocouple measurements positioned inside the laboratory scale quench probe. U-tube measurements described the chemical gas composition downstream to the quench probe.

Experimental data was categorised into phase 1 (non-reactive approach) and phase 2 (reactive approach), of which the feed of the latter included organic material as opposed to phase 1, which consisted of only a carrier gas feed. A variety of experimental cases were investigated by inducing changes to the experimental setup with regards to the water spray rate, gas feed and type, and supplied electrical current. Phase 1 was used to construct the base model, of which parameter analyses were conducted to refine the model for phase 2. The final model could describe quench probe conditions fairly accurate, with an average error of 9.87 % and a root mean square error of 6.96 °C.

It was found that the temperature distribution inside the quench probe was strongly dependant on the velocity profile. The development of a recirculation zone inside the quench resulted in a longer residence time, increasing the cooling effect of the spray water. Furthermore, temperatures within range of full quenching were achieved relatively early after the first spray injection, indicating redundant water spray.

Through use of the dimensionless temperature gradient and H2/CO ratio, the thermal and chemical

efficiencies could respectively be investigated. Generally, the dimensionless temperature gradient averaged 0.8 towards the exit of the quench probe, indicating adequate quenching. Contrarywise, the H2/CO ratio ranged between 0.5 and 0.7 when ideally ratios of 1.0 to 2.5 are preferred for

industrial application. It is therefore cardinal to improve chemical efficiency whilst not sacrificing the integrity of the thermal efficiency.

(4)

Page | ii

Lastly, the model was used to investigate improvements to the current quench probe design with regards to the water flow rate, nozzle placement, number of nozzles and geometry. It was concluded that the water flow rate could be reduced in addition to lessening the number of nozzles to effectively achieve the same quenching results. Additionally, a larger diameter quench probe would achieve faster quenching rates, but due to the redundant water spray nozzles in the current application, similar results were achieved for smaller diameter cases.

(5)

___________________________________________________________________ ACKNOWLEDGEMENTS

Page | iii

ACKNOWLEDGEMENTS

“Knowledge is knowing a tomato is a fruit; wisdom is not putting it in a fruit salad”

– Miles Kington

Writing a Master’s dissertation is no small feat and I would therefore like to extend my gratitude towards those that did not only help me improve my knowledge, but added spoons full of wisdom into the mix:

 Mr. Frikkie van der Merwe and Prof. Kenny Uren who guided me with patience, diplomacy and persistence;

 My family, who always reminded me of the importance to take breaks and treat myself;  Nico Bijzet for his unwavering patience, motivation and unlimited supply of chocolate.  Mr. Tobie Loftus for being a good minion and helping immensely with my experimental work.

Furthermore, without the aid of Dr. IJ van der Walt and Mr. P Scheepers at the Nuclear Energy Corporation of South Africa (NECSA) in the use of their facilities and continuous guidance in the laboratory, this would not be possible.

Lastly, I would like to thank my Heavenly Father for giving me the opportunity, talent and endurance to complete this research dissertation and for paving the way forward.

(6)

Page | iv

_____________________________________________________________ TABLE OF CONTENTS

Page | v 3.2 Material ... 57 3.3 Measuring equipment ... 57 3.3.1 Type-R thermocouple ... 57 3.3.2 Type-K thermocouple ... 57

3.3.3 Picolog® TC-08 USB thermocouple data logger ... 57

3.3.4 Gilian® Gilibrator-2 calibrator ... 57

3.4 Experimental procedure ... 58

3.4.1 Phase 1: non-reactive experimental procedure ... 58

3.4.2 Phase 2: reactive experimental procedure ... 59

Bibliography 61 CHAPTER 4: MODEL DEVELOPMENT ... 62

4.1 CFD modelling assumptions ... 63

4.2 Generation of the quench probe geometry for the CFD model ... 64

4.3 Mesh continua ... 66

4.4 Physics continua ... 69

4.5 Boundary conditions ... 80

4.6 Summary of final conditions used in model ... 83

Bibliography 85 RESULTS AND DISCUSSION ... 88

4.7 CFD Modelling results ... 89

4.7.1 Velocity profile ... 89

4.7.2 General temperature profile ... 91

4.8 Validation of modelled results ... 95

4.8.1 Phase 1 results ... 96

4.8.2 Phase 2 results ... 102

4.9 Quenching efficiency ... 104

4.10 Chemical efficiency ... 107

4.11 Quench probe improvement study ... 108

4.11.1 Nozzle flow rate ... 109

4.11.2 Nozzle positioning ... 109

4.11.3 Efficiency of nozzle clusters ... 110

4.11.4 Geometric variations... 112

Bibliography ... 117

(8)

Page | vi

5.1 Conclusion ... 120

5.1.1 Flow and temperature profile ... 120

5.1.2 Experimental data ... 120

5.1.3 Chemical efficiency ... 121

5.1.4 Improvement considerations ... 121

5.1.5 Geometric considerations ... 121

5.2 Recommendations ... 122

Appendix A: Previous conducted research ... 123

Bibliography 125 Appendix B: Description of physics models ... 127

B.1. List of definitions ... 127

Bibliography 131 Appendix C: Calibration of testing equipment ... 132

C.1. Calibration of feed ... 132

C.2. Nitrogen flow meter ... 132

Appendix D: Nozzle specifications ... 134

D.1. Determining parameters ... 134

D.2. Summary of water nozzle (injector) parameters... 137

Bibliography 138 Appendix E: Ultimate analysis for pine wood chips ... 139

E.1. Ultimate analysis for pine wood chips ... 139

Appendix F: Quench probe technical drawing ... 140

F.1. Quench probe technical drawing ... 140

Appendix G: SIMPLE algorithm ... 144

G.1. SIMPLE algorithm ... 144

Appendix H: Temperature profile influence of individual nozzles ... 145

H.1. Nozzle temperature profile ... 145

Appendix I: Experimental data ... 148

I.1. Phase 1: 32.0 ℓ/min N2 flow rate ... 148

I.2. Phase 1: 32.4 ℓ/min N2 and 65 ℓ/min air flow rate ... 149

I.3. Phase 2: 35.6 ℓ/min air flow rate, 150 A power supply and low (4.8 kg/h) wood chip feed 150

(9)

_____________________________________________________________ TABLE OF CONTENTS

Page | vii I.4. Phase 2: 35.6 ℓ/min air flow rate, 150 A power supply and medium (6 kg/h) wood chip feed 151

I.5. Phase 2: 35.6 ℓ/min air flow rate, 150 A power supply and high (7.63 kg/h) wood chip feed 151

Appendix J: Phase 1 model results ... 153

J.1. 10 parcel streams model results ... 153

J.2. 30 parcel streams model results ... 155

J.3. Statistical significance ... 156

Appendix K: Phase 2 modelling results ... 158

Appendix L: Dimensionless temperature gradient ... 162

Appendix M: Flow rate influence on temperature profile ... 166

Appendix N: Configurations of moved nozzles ... 167

Appendix O: Temperature profile images of moved nozzles ... 169

Appendix P: Influence of nozzle position on temperature distribution ... 171

Appendix Q: Temperature profile for clusters of nozzles ... 172

(10)

Page | viii

TABLE OF FIGURES

CHAPTER 1

Figure 1-1: A direct water quench used in syngas cooling applications ... 4 Figure 1-2: Simplified scheme of the laboratory scale plasma gasification setup ... 5 Figure 1-3: Workflow and chapter contents of the research study ... 8

CHAPTER 2

Figure 2-1: The dimensionless temperature gradient as a function of residence time at Tinlet of 850

°C (Adapted from Hawboldt et al., 1999) ... 14 Figure 2-2: Comparison between quenching rates of different quenching methods at typical quenching conditions (Adapted from Sundstrom & DeMichiell, 1971) ... 16 Figure 2-3: Syngas heat recovery systems as investigated by Uebel et al. (2014) (Taken from Uebel et al., 2014) ... 17 Figure 2-4: Syngas cooling configurations as tested by Ni et al. (2011) where (a) direct water quenching (b) radiant syngas cooler (RSC) (c) RSC and (d) gas quenching (Taken from Ni et al., 2011) ... 18 Figure 2-5: Exiting CO conversion for different reactor temperatures and H2O/CO ratios at 1 atm and

present catalyst (Taken from Choi & Stenger, 2003) ... 20 Figure 2-6: Influence of steam flow of (a) 20 kg/h, (b) 50 kg/h, (c) 100 kg/h and (d) 300 kg/h in a quench reactor on the formation of different chemical species (Taken from Yan et al., 2012)... 22 Figure 2-7: The effect of water injection with (i) one nozzle, (ii) four nozzles and (iii) five nozzles on (a) the CO conversion and (b) the reaction equilibrium (Taken from Kiso & Matsuo, 2011) ... 23 Figure 2-8: Optimisation study on the influence of (a) outlet temperature, (b) h/d ratio, (c) volume, (d) wall cooling capability and (e) steam flow as led by Uebel et al. (2016b) ... 32 Figure 2-9: Primary and secondary injection points in the study conducted by Wang et al. (2011) 36 Figure 2-10: CO conversion in relation to steam mass flow and H2O/CO ratio (Adapted from Uebel

(11)

_______________________________________________________________ TABLE OF FIGURES

Page | ix Figure 2-11: Coherence between H2/CO ratio and syngas exit temperature (Adapted from Uebel et

al., 2016a) ... 40 Figure 2-12: Temperature profiles in radial syngas coolers (adapted from (a) Ghouse et al. (2015), (b) Risberg & Gebart (2013) and (c) Johansson & Westerlund (2011)) ... 41 Figure 2-13: Velocity and temperature profiles of spray applications inside a wind tunnel (Taken from Montazeri et al., 2015) ... 42 Figure 2-14: Longitudinal velocity and temperature profiles of synthetic gas quenched with recycled gas (Taken from Ye et al., 2013) ... 43 Figure 2-15: Cross sections of the temperature profile inside a synthesis gas quench (Taken from Ye et al., 2013) ... 44 Figure 2-16: Longitudinal cuts of the flow and temperature distribution inside and RSC (Taken from Yu et al., 2009) ... 45 Figure 2-17: Flow profile in the connection between a gasifier and a RSC (Taken from Ni et al., 2009)

... 45 Figure 2-18: RSC with additional quench setup (Taken from Li et al., 2016b) ... 46 Figure 2-19: Temperature profile for (a) no spray and (b) spray and velocity profile inside an RSC with additional quench (Taken from Li et al., 2016b) ... 47 Figure 2-20: Cross sectional temperature distribution inside an RSC with additional water spray (Taken from Li et al., 2016b) ... 48

CHAPTER 3

Figure 3-1: Experimental investigation summary ... 58

CHAPTER 4

Figure 4-1: Simulated quench probe shell in the CFD platform (b) according to the technical drawing (a)... 64 Figure 4-2: Spray nozzle injection cones in STAR-CCM+ ... 65 Figure 4-3: Thermocouple measuring points in STAR-CCM+ ... 66

(12)

Page | x

Figure 4-4: Mesh type model analysis ... 67

Figure 4-5: Polyhedral mesh in phase 2 – (a) surface mesh and (b) volume mesh ... 68

Figure 4-6: Mesh cell size analysis ... 69

Figure 4-7: Mean absolute error analysis on two different EoS models ... 77

Figure 4-8: Mean absolute errors analysis based on different turbulence models ... 77

Figure 4-9: Influence of the amount of parcel streams on the absolute mean absolute error for different cases ... 80

Figure 4-10: Boundary condition locations specified in the STAR-CCM+ model ... 81

Figure 4-11: Summary of models used in the CFD direct quench probe simulation ... 84

CHAPTER 5 Figure 5-1: Laboratory-scale quench probe velocity profile ... 90

Figure 5-2: Laboratory-scale quench probe temperature profile with all nozzles open ... 91

Figure 5-3: Individual spray nozzle influence on the temperature profile ... 93

Figure 5-4: Radial temperature profile at (a) TC1, (b) TC2, (c) TC3 and (d) TC4 ... 95

Figure 5-5: Phase 1 - model values vs. experimental values for (a) 32.4 ℓ/min N2 and (b) 32.4 ℓ/min N2 and 65 ℓ/min air at different currents and number of nozzles open ... 100

Figure 5-6: Model values vs. experimental values for (a) all nozzles open, (b) nozzle 2 closed and (c) nozzles 2 and 3 closed at different wood chip feed rates (150 A) ... 103

Figure 5-7: Dimensionless temperature gradient for (a) no nozzles closed, (b) nozzle 1 closed and (c) nozzles 1 and 2 closed ... 106

Figure 5-8: Effect of flow rate in spray nozzles on quench probe temperature profile ... 109

Figure 5-9: Effect of nozzle position in the temperature profile of the quench probe ... 110

Figure 5-10: Effect of clusters of spray nozzles on the temperature profile of the laboratory scale quench probe ... 111

Figure 5-11: Effect of geometry changes on the flow profile inside a laboratory scale quench probe ... 113

(13)

_______________________________________________________________ TABLE OF FIGURES

Page | xi Figure 5-13: Comparison of modelled results for different geometrical configurations of the quench

probe at a low feed rate ... 115

APPENDICES Figure C-1: Calibration curve of the organic waste feeder ... 132

Figure C-2: Calibration curve of the nitrogen flow (ℓ/min) ... 132

Figure C-3: Calibration curve of the air flow (ℓ/min) ... 133

Figure D-1: Fulljet® spray nozzle specifications (Spraying Systems co., 2016) ... 134

Figure D-2: Theoretical orifice velocity of Fulljet water nozzles (Spraying Systems co., 2016) .... 135

Figure D-3: Diameter of sprayed particles (Spraying Systems co., 2016) ... 136

Figure F-1: Full quench probe technical drawing ... 141

Figure F-2: Quench probe bend technical drawing ... 142

Figure F-3: Quench probe quenching section technical drawing ... 143

Figure H-1: Temperature profiles for nozzles spraying in the negative x-direction for (a) nozzle 1, (b) nozzle 2 and (c) nozzle 3 ... 146

Figure H-2: Temperature profiles for nozzles spraying in the negative z-direction for (a) nozzle 4 and (b) nozzle 5 ... 147

Figure J-1: Model values vs. experimental values for (a) 32.4 ℓ/min N2 (b) 32.4 ℓ/min N2 and 65 ℓ/min O2 at different currents and amounts of nozzles open in phase 1 for the 10 parcel streams simulation ... 154

Figure J-2: Model values vs. experimental values for (a) 32.4 ℓ/min N2 (b) 32.4 ℓ/min N2 and 65 ℓ/min O2 at different currents and amounts of nozzles open in phase 1 for the 30 parcel streams simulation ... 156

Figure J-3: Modelled temperatures vs. experimental temperatures for (a) 10 parcel streams, (b) 20 parcel streams and (c) 30 parcel streams ... 157

Figure K-1: Phase 2 modelling results for different experimental conditions ... 161

Figure L-1: Dimensionless temperature gradient for different experimental conditions ... 165

(14)

Page | xii

Figure N-1: Spray nozzle configurations for (a) top and bottom, (b) 5 in a row (downwards), (c) 5 in

a row (sideward), (d) 3 spraying inwards and (e) 5 spraying inwards ... 168

Figure O-1: Temperature profiles for different spray configurations ... 170

Figure P-1: Graphs showing the influence of moved nozzles inside the quench probe ... 171

Figure Q-1: Graphs showing the influence of moved nozzles inside the quench probe ... 172

(15)

__________________________________________________________________ LIST OF TABLES

Page | xiii

LIST OF TABLES

CHAPTER 2

Table 2-1: Kinetic expressions for the WGSR ... 19

Table 2-2: Ideal distance water injection points in the 5-stage optimised case (Taken from Kiso & Matsuo, 2011) ... 24

Table 2-3: Summary of CFD models used in studies conducted on syngas cooling ... 27

Table 2-4: Base case syngas entering the reactor quench probe from Uebel et al. (2016b) ... 31

Table 2-5: Wang et al. (2011) CFD model reaction prediction (Adapted from Wang et al. 2011) ... 35

Table 2-6: Kinetic parameters as taken from Wang et al. (2011) ... 35

Table 2-7: Kinetic parameters as adapted by Wang et al. (2011) ... 36

Table 2-8: Gas characteristics in a study conducted by Uebel et al. (Adapted from Uebel et al. 2016a) ... 38

CHAPTER 3 Table 3-1: Operating conditions investigated in non-reactive experimental procedure (phase 1) .. 59

Table 3-2: Operating conditions investigated in reactive experimental procedure (phase 2) ... 60

CHAPTER 4 Table 4-1: Dimensionless coefficients for the realisable k-ε model ... 72

Table 4-2: Absolute error of individual thermocouples with different turbulence models ... 78

Table 4-3: Boundary conditions specified for the CFD model ... 81

Table 4-4: Mole fractions as specified in phase 2 model ... 82

CHAPTER 5 Table 5-1: Experimental values for the temperature profile (°C) of the quench probe at a constant current and process gas flow rates ... 96

(16)

Page | xiv

Table 5-2: Gaseous species concentration after quenching process ... 108

APPENDICES

Table A-1: Synthesis gas quenching studies as conducted by a variety of authors ... 123 Table D-1: Summary of injector values in STAR-CCM+ model ... 137 Table E-1: Wood chip ultimate analysis ... 139 Table I-1: Phase 1 experimental results for different power supplies and flow rates at a constant gas flow rate of 32.4 ℓ/min N2 (no solid feed) ... 148

Table I-2: Phase 1 experimental results for different power supplies and flow rates at a constant gas flow rate of 32.4 ℓ/min N2 and 65.0 ℓ/min air (no solid feed) ... 149

Table I-3: Phase 2 experimental results for different water flow rates and a wood chip feed of 4.8 kg/h and air flow rate of 35.6 ℓ/min ... 150 Table I-4: Phase 2 experimental results for different water flow rates and a wood chip feed of 6 kg/hand air flow rate of 35.6 ℓ/min... 151 Table I-5: Phase 2 experimental results for different water flow rates and a wood chip feed of 7.63 kg/h and air flow rate of 35.6 ℓ/min ... 151

(17)

___________________________________________________________ LIST OF ABBREVIATIONS

Page | xv

LIST OF ABBREVIATIONS

AFROX - African Oxygen limited

CFD - Computational fluid dynamics

DO - Discrete ordinate

DRW - Discrete random walk

EDC - Eddy dissipation concept

EoS - Equation of state

IGCC - Integrated gasification combined cycle

ISAT - In-situ adaptive tabulation

MAE - Mean absolute error

RANS - Reynolds averaged Navier Stokes

RNG - Renormalisation group

RR - Rosin Rammler

RSC - Radiant syngas cooler

SA - Spalart Allmaras

SIMPLE - Semi-implicit method for pressure linked equations

SST - Shear stress transport

Syngas - Synthesis gas

TC - Thermocouple

UDF - User defined function

WGS - Water gas shift

(18)

Page | xvi

NOMENCLATURE

Symbol Representation Value

𝜌 Density 𝑉 Volume 𝐴 Area 𝑡 Time 𝒗 Kinematic velocity 𝒗_𝑔 Grid velocity

𝑻 Viscous stress tensor

𝒇_𝑟 Rotational force

𝒇_𝑔 Gravitational force

𝒇_𝜔 Vorticity confinement force

𝜒 Porosity 𝑰 Identity matrix 𝑛 Mole amount 𝑇 Temperature 𝑅 Gas constant _8.314 𝐽 𝑚𝑜𝑙 𝐾

𝑘 Turbulent kinetic energy

𝜀 Turbulent dissipation rate

𝜇 Dynamic viscosity

𝜇_𝑡 Turbulent viscosity

𝜎𝑘 Dimensionless constant 1.0

𝑓_𝑐 Curvature correction factor

𝐺_𝑘 Turbulent production

𝐺_𝑏 Buoyancy production

𝛶𝑀 Compressibility modification

𝑆_𝑘 User-specified source term

𝑆_𝑢 User-specified source term

𝜎_𝜀 Dimensionless constant 1.2

𝑆_𝜀 User-specified source term

𝑣′ _{Root mean square of local velocity fluctuations}

𝑣̅ Mean velocity

𝐼 Turbulence intensity

(19)

________________________________________________________________ NOMENCLATURE

Page | xvii

𝐴𝑜 Dimensionless constant 4.0

𝑈(∗) _𝜇_𝑡_{model coefficient}

𝑺 Mean strain rate tensor

𝑾 Mean vorticity tensor

𝐴𝑠 𝐶𝜇 model coefficient

𝑆 Modulus of the mean strain rate tensor

𝑊 Modulus of the mean vorticity tensor

𝛽 Thermal expansion coefficient

𝒈 Gravitational vector

𝑐 Speed of sound

𝐶_𝑀 Dimensionless constant 2.0

𝜂 𝐶_𝜀₃ model coefficient

𝒗_𝒃 Velocity component parallel to 𝒈

𝒖𝒃 Velocity component perpendicular to 𝒈

𝐶𝑚𝑎𝑥 Dimensionless constant 1.25

𝑙_𝜀 Length scale

𝑅𝑒_𝑦 Wall distance Reynolds number

𝑅𝑒𝑦∗ Wall distance Reynolds number limit 60.0

𝐴𝜆 𝜆 model coefficient

𝜆 Wall proximity indicator

𝑦 Distance from wall

𝑛_𝑖 Iteration number

𝑐_𝑙 𝑙_𝜀 model coefficient

𝐴𝜀 𝑙𝜀 model coefficient

𝜅 Dimensionless constant 0.42

𝐴_𝜇 Dimensionless constant 70.0

𝐶𝜀3 Realisable k-ε model coefficient 1.9

𝐶𝜀2 Dimensionless constant

𝐶𝑟2 Dimensionless constant 0.25

𝐶𝜀1 Realisable k-ε model coefficient

𝐶_𝑟₁ Dimensionless constant 0.04645

(20)

(21)

_______________________________________________________ CHAPTER 1: INTRODUCTION

Page | 1

CHAPTER 1:

INTRODUCTION

Chapter 1 will provide a background on waste management, plasma gasification and hot-gas quenching in the broad context. In sections 1.2, 1.3 and 1.4, the research study focus, aim and objectives will be discussed respectively. The chapter will be brought to a close with the scope of

(22)

Page | 2

1.1 Background and motivation 1.1.1 Waste management prospect

Effective waste treatment and management is becoming a more urgent matter in the current age as a constant increase in population and the use of less biodegradable products call for alternative waste treatment options (Heberlein & Murphy, 2008; Ruj & Ghosh, 2014). Currently, landfilling, incineration, biological treatment and recycling are the primary methods used to dispose waste (Heberlein & Murphy, 2008; Eriksson et al., 2005). Due to the heterogeneous nature of waste, it is difficult to sort and dispose of its constituents, which typically result in water, soil or air pollution (Ruj & Ghosh, 2014; Giusti, 2009).

In 2015 waste generation in South Africa totalled approximately 3.31 kg waste per person per day (SAWIC, 2016). Here, mining waste represents the majority, while domestic, trade and sewage waste contribute to only a small percentage (Greben & Oelofse, 2009). Landfilling is currently the primary waste treatment method, but is becoming less attractive due to leachate and methane emissions from waste and the excessive surface area that is consumed by this implementation (Greben & Oelofse, 2009; Byun et al., 2012). Simultaneously, an increasing need exists for alternative power generation due to the elevating demand for energy (Greben & Oelofse, 2009). Various thermal processes regarding the vitrification, melting and gasification of wastes are currently being studied; their aim to destroy the organic fraction in waste and convert the inorganic portion to harmless silica glass (Gomez et al., 2009). An attractive option is the use of gasification to process waste material. Gasification is defined as a non-incineration process in which carbon containing compounds are decomposed into simple molecules at high temperatures (Lemmens et al., 2007; Rajasekhar et al., 2015). These products include synthesis gas (syngas), i.e. carbon monoxide and hydrogen, which is a valuable source for alternative energy applications (e.g. fuel) and the production of chemicals (i.e. methanol, ammonia and di-methyl ether) and an inert vitrified slag that can be implemented as a building material additive (Bratsev et al., 2006; Galeno et al., 2011; Lemmens et

al., 2007). Plasma gasification, moreover, utilises a plasma (i.e. ionised gas) for the destruction of

compounds at temperatures exceeding 5 000 °C for cleaner and enhanced synthesis gas generation, which releases a controllable amount of air pollutants (Moustakas et al., 2008; Cho et al., 2015). 1.1.2 Plasma gasification vs. traditional gasification

Plasma gasification is a non-incineration thermal process in which feed material is broken down into their constituent molecules through the use of exceptionally high temperatures in an oxygen deprived environment (Mountouris et al., 2008). This technology can be implemented for a variety of waste feedstocks, but is especially advantageous in the management of hazardous (i.e. radio-active and toxic) waste as it converts it into inert, non-leachable products that does not require controlled disposal (Tang et al., 2013; Janajreh et al., 2013; Fabry et al., 2013). The non-leachable

(23)

_______________________________________________________ CHAPTER 1: INTRODUCTION

Page | 3 characteristic of the slag makes plasma gasification an appropriate pre-treatment for landfilling in addition to a waste-to-energy solution (Tendler et al., 2005).

Plasma technology is receiving increased attention due to several advantages in comparison to traditional gasification. While traditional gasification also occurs in an oxygen deprived environment, operational temperatures for traditional gasification range from 450 to 850 °C and is dependent on the exothermic nature of the decomposition process itself to sustain reactions (Mountouris et al., 2008; Lemmens et al., 2007; Byun et al., 2012). Typically, most of the carbon in the reaction is consumed by sustaining the exothermal reaction with sub-stoichiometric oxygen (Mountouris et al., 2008; Lemmens et al., 2007). The exit gas from standard gasification is similar to the plasma alternative, but is more contaminated due to the presence of tars, soot and char (Mountouris et al., 2008). In addition, waste feed to traditional gasifiers must be sorted prior to the gasification process as opposed to plasma gasification which can process heterogeneous feeds (Mountouris et al., 2008). Plasma gasification utilises temperatures exceeding 5000 °C from a plasma burner (or torch), an independent heat source, to the effect that the resultant gas is not dependent on self-sustaining reactions (Lemmens et al., 2007). The application of high temperatures has the advantageous effect that unwanted dioxins, furans, soot, char and tar are broken down into smaller molecular chains, resulting in cleaner syngas (Galeno et al., 2011; Materazzi et al., 2015). The syngas received from the plasma can be used to completely recover the energy used for its creation in addition to contributing to other energy associated activities (Bratsev et al., 2006). Among the benefits mentioned above, other advantages include:

i. A low concentration of hazardous emissions and reduction in waste volume makes it an environmentally friendly process (Minutillo et al., 2009; Zhang et al., 2012);

ii. Syngas with a higher calorific value is produced (Minutillo et al., 2009; Cho et al., 2015); iii. Enhanced control in the process environment due to the presence of an external heat source

(Heberlein & Murphy, 2008);

iv. Heavy metals are trapped in the molten slag (Zhang et al., 2013); v. Easy process start-up (Cho et al., 2015; Heberlein & Murphy, 2008);

vi. A small reactor geometry with increased product flow rates (Gomez et al., 2009; Byun et al., 2012);

vii. Low gas flow rates resulting in cost-effective and easier gas treatment (Gomez et al., 2009; Byun et al., 2012);

viii. High residence times (Fabry et al., 2013).

Plasma gasification, however, has the disadvantage that its external energy source is quite energy intensive due to high electricity consumption, and is resultantly expensive to operate (Galeno et al., 2011; Heberlein & Murphy, 2008).

(24)

Page | 4

Gomez et al. (2009) concluded that plasma gasification will be economically feasible if implemented on a long-term basis, with the commercial viability even further improved by the avoidance of landfill and emission taxes.

1.1.3 Quenching of syngas

Gaseous hydrocarbons exit the plasma-arc reactor at extremely high temperatures (1100 – 1400 °C) after gasification and subsequently needs to be cooled (Wang et al., 2011; Hawboldt et al., 1999). Syngas cooling should occur rapidly to ensure that equilibrium is reached and that no unwanted side reactions occur downstream in the quench probe (Hawboldt et al., 1999). The quenching process ensures that downstream equipment is not damaged by slag deposits while simultaneously satisfying the low temperature requirements of the water-gas shift reaction and desulphurisation processes, ultimately ensuring the formation of appropriate products (Wang et al., 2011; Ye et al., 2013). Current quenching technologies entail the use of radial cooling, cold syngas, chemical quenching or the direct spraying of water as quench medium (Ye et al., 2013). An example of the latter, which forms the main quenching focus of this study, is shown in Figure 1-1.

Figure 1-1: A direct water quench used in syngas cooling applications

Syngas in Water inlet Syngas out Thermocouple 1 2 3 4 5 1 2 3 4 Elbow

(25)

_______________________________________________________ CHAPTER 1: INTRODUCTION

Page | 5 As indicated in Figure 1-1, syngas enters at the left side of the probe from the gasification reactor and exits at the right side en route to a syngas cleaning system. The syngas is directly quenched with water droplets entering via nozzles positioned axially on the surface of the quench probe. As the quenching step in the process plays such an important role in the production of syngas, it is important to develop a thorough understanding of the flow regimes, temperature profile, particle behaviour, reactivity and residence time inside the quench probe.

1.2 Research focus

With a waste-to-energy solution in mind, a laboratory scale plasma-arc gasification reactor system was constructed for research purposes and possible scale-up application in the future. Figure 1-2 shows a schematic representation of the setup:

Figure 1-2: Simplified scheme of the laboratory scale plasma gasification setup

As shown in Figure 1-2, shredded waste material is fed into the reactor and is decomposed through the heat supplied by a plasma burner (torch). The argon and nitrogen is used as a plasma gas, aiding in the operation of the plasma torch, while air is used as process gas and is fed simultaneously with the solid feed. The gaseous products formed in the reactor is transported through a quench probe (i.e. Figure 1-1) where rapid cooling takes place to effectively stop the formation of side reactions,

Screw feeder Filter Argon/nitrogen feed Air compressor _Slag Water supply External power source Plasma reactor Quench probe Torch

(26)

Page | 6

including the formation of unwanted hydrocarbons (e.g. methane, ethane etc.). Afterwards, the gas is transported to filters connected in series where further cleaning takes place.

The reactor and quench probe are, however, both estimated to operate inefficiently necessitating an investigation into the improvement of the system to ensure optimum H2/CO yield and cleaner syngas

recovery. The improvement of this system will finally aim to lay the foundation for an industrial scale plasma-arc gasification reactor setup.

This study will therefore concentrate on the improvement of the laboratory scale quench probe and, although emphasis will only be given to this unit, the study will run in parallel with similar investigations focusing on the improvement of other components in the reactor setup.

1.3 Aim

The research study will aim to improve the current water-spray quench probe connected to a laboratory-scale plasma-arc gasification reactor. The term ‘improve’ will be based largely on the emerging H2/CO gas ratio at the exit of the quench probe as well as the temperature profile inside

the quench probe.

1.4 Objectives

The primary objectives can therefore be defined accordingly:

i. Model and validate the design of the quench probe connected to a laboratory scale plasma-arc gasification reactor using computational fluid dynamics (CFD);

The secondary objectives are stated as follows:

i. Suggest an improved quench probe design based on the modelled temperature profiles. ii. Analyse the chemical efficiency of the quench probe.

1.5 Investigation outline 1.5.1 Method of investigation

In order to successfully represent heat transfer, chemical equilibrium and flow characteristics in the quench probe, it is required to perform an analysis based on a CFD modelling approach of the quench probe. The CFD model will give a clear indication of flow patterns within the quench probe, being able to accurately solve evaporation equations, chemical equilibria equations and flow models.

(27)

_______________________________________________________ CHAPTER 1: INTRODUCTION

Page | 7 The model will be validated through experimental runs conducted on the laboratory scale quench probe in which temperatures, pressures, gas- and feed compositions and flow rates will be measured. A homogeneous feed (pine wood chips) will be used to ensure consistency of results. A variety of scenarios will be investigated to ultimately construct an applicable model.

Suggestions on the improvement of the quench probe will be motivated by the constructed model according to the quenching efficiency as well as the measured chemical yields. As the mentioned synthesis gas will be mainly used for application in alternative fuel, a H2/CO ratio applicable to the

electricity generation is necessary. A qualitative analysis will be conducted based on the expected chemical reaction equilibria at the determined temperature and pressure. In addition, efficiency will also be based on the temperature profile which will signify the effectiveness of the cooling medium. 1.5.2 Scope of the study

The study will be outlined by an in-depth literature study, which will include sources on current synthesis gas cooling configurations, flow- and temperature profiles inside syngas coolers, reactivity of quench coolers and existing CFD configurations.

The experimental analysis will follow, in which temperature and specie yields will be measured according to different experimental situations. A CFD model will be constructed according to this and therefore a chapter on model development will describe the procedure followed to construct the model.

Experimental data will be validated on the CFD model in a chapter titled “Results and discussion”. Further results to be discussed in this chapter will include the chemical efficiency of the quench probe as well as possible improvements on the current installation according to changes implemented in the model. The report will be drawn to a close by the insertion of a conclusion chapter as the final chapter. Figure 1-3 shows the workflow of the research study and highlights the contents of each chapter.

(28)

Page | 8

Figure 1-3: Workflow and chapter contents of the research study

CHAPTER 1: INTRODUCTION

1.1 Background and motivation

1.1.1 Waste management prospect

1.1.2 Plasma gasification vs. traditional gasification 1.1.3 Quenching of syngas 1.2 Research focus 1.3 Aim 1.4 Objectives 1.5 Investigation outline 1.5.1 Method of investigation 1.5.2 Scope of the study

CHAPTER 2: LITERATURE STUDY

2.1 Synthesis gas quenching

2.1.1 Quenching methods 2.1.2 The direct quenching method

2.2 Computational fluid dynamics (CFD) modelling of synthesis gas cooling systems

2.2.1 Approaches in CFD modelling for synthesis gas coolers

2.2.2 CFD modelling applicable to direct water quenching – reactivity

2.2.3 CFD modelling applicable to direct water quenching – temperature and flow profiles

2.3 Concluding review

CHAPTER 3: EXPERIMENTAL METHOD

3.1 Laboratory-scale experimental setup 3.2 Material

3.3 Measuring equipment

3.3.1 Type-R thermocouple 3.3.2 Type-K thermocouple

3.3.3 Picolog® TC-08 USB thermocouple data logger 3.3.4 Gilian® Gilibrator-2 calibrator

3.4 Experimental procedure

3.4.1 Phase 1: non-reactive experimental procedure 3.4.2 Phase 2: reactive experimental procedure

CHAPTER 4: MODEL DEVELOPMENT

4.1 CFD modelling assumptions

4.2 Generation of the quench probe geometry for the CFD model

4.3 Mesh continua 4.4 Physics continua 4.5 Boundary conditions

4.6 Summary of final conditions used in the model

CHAPTER 5: RESULTS AND DISCUSSION

5.1 CFD modelling results

5.1.1 Velocity profile

5.1.2 General temperature profile

5.2 Validation of modelled results

5.2.1 Phase 1 results 5.2.2 Phase 2 results

5.3 Quenching efficiency 5.4 Chemical efficiency

5.5 Quench probe improvement study

5.5.1 Nozzle flow rate 5.5.2 Nozzle positioning 5.5.3 Efficiency of nozzle clusters 5.5.4 Geometric variations CHAPTER 6: CONCLUSION AND RECOMMENDATIONS

6.1 Conclusion

6.1.1 Flow and temperature profile 6.1.2 Experimental data

6.1.3 Chemical efficiency 6.1.4 Improvement considerations 6.1.5 Geometric considerations

6.2 Recommendations

Do results agree with objectives, aim and focus?

Compare literature and

results

Consolidate literature and model development

(29)

__________________________________________________________________ BIBLIOGRAPHY

Page | 9

Bibliography

B

Bratsev, A. N., Popov, V. E., Rutberg, A. F. & Shtengel, S. V., 2006. A facility for plasma gasification of waste of various types. High Temperature, Issue 44, pp. 823-828.

Byun, Y., Chung, J., Cho, M. & Hwang, S. M., 2012. Thermal Plasma Gasification of Municipal Solid Waste (MSW). INTECH Open Access Publisher.

C

Cho, I. J., Park, H. W., Park, D. W. & Choi, S., 2015. Enhancement of synthesis gas production using gasification-plasma hybrid system. International Journal of Hydrogen Energy, Issue 40, pp. 1709-1716.

E

Eriksson, O. et al., 2005. Munisipal solid waste management from a systems perspective. Journal of Cleaner Production, Issue 13, pp. 421-439.

F

Fabry, F., Rehmet, C., Rohani, V. & Fulcheri, L., 2013. Waste gasification by thermal plasma: a review. Waste and Biomass Valorization, Issue 4, pp. 421-439.

G

Galeno, G. K., Minutillo, M. & Perna, A., 2011. From waste to electricity through integrated plasma gasification/fuel cell (IPGFC) system. International Journal of Hydrogen energy, Issue 36, pp. 1692-1701.

Giusti, L., 2009. A review of waste management practices and their impact on human health. Waste Management, Issue 29, pp. 2227-2239.

Gomez, E. et al., 2009. Thermal plasma technology for the treatment of wastes: a critical review. Journal of Hazardous Materials, Issue 161, pp. 614-626.

Greben, H. A. & Oelofse, S. H. H., 2009. Unlocking the resource potential of organic waste: a South African perspective. Waste Management & Research, Issue 27, pp. 676-684.

H

Hawboldt, K. A., Monnery, W. D. & Svreck, W. Y., 1999. A study of the effect of quench design on experimental data. Industrial & Engineering Chemistry Research, Issue 38, pp. 2260-2263. Heberlein, J. & Murphy, A. B., 2008. Thermal Plasma Waste Treatment. Journal of physics D:

Applied Pyrolysis, Issue 41, pp. 53002-53021. J

(30)

Page | 10

Janajreh, I., Raza, S. S. & Valmudsson, A. S., 2013. Plasma gasification process: Modelling, simulation and comparison with conventional gasification. Energy Conversion and Management, Issue 65, pp. 801-809.

L

Lemmens, B. et al., 2007. Assessment of plasma gasification of high calorific waste streams. Waste Management, Issue 27, pp. 1562-1569.

M

Materazzi, M. et al., 2015. Reforming of tars and organic sulphur compounds in a plasma-assisted process for waste gasification. Fuel Processing Technology, Issue 137, pp. 259-268.

Minutillo, M., Perna, A. & Di Bona, D., 2008. Modelling and Performance analysis of an integrated plasma gasification combined cycle (IPGCC) power plant. Energy Conversion and Management, Issue 50, pp. 2264-2271.

Mountouri, A., Voutsas, E. & Tassios, D., 2008. Plasma gasification of sewage sludge: Process development and energy optimization. Energy Conversion and Management, Issue 49, pp. 2264-2271.

Moustakas, K. et al., 2008. Analysis of results from the operation of a pilot plasma gasification/vitrification unit for optimizing its performance. Journa; of Hazardous materials, Issue 151, pp. 473-480.

R

Rajasekhar, M. et al., 2015. Energy GEneration from Municipal Solid waste by Innovative Technologies - Plasma Gasification. Procedia Materials Science, Issue 10, pp. 513-518.

Ruj, B. & Ghosh, S., 2014. Technological aspects of thermal plasma treatment of municipal solid waste - a review. Fuel Processing Technology, Issue 126, pp. 298-308.

S

SAWIC (South African Waste Information Centre), 2016. Tonnage reports. [Online] Available at: http://sawic.environment.gov.za/index.php?menu=15 [Accessed 15 August 2016].

T

Tang, L., Huang, H., Hao, H. & Zhao, K., 2013. Development of plasma pyrolysis/gasification systems for energy effiecient and environmrntally sound waste disposal. Journal of Electrostatics, Issue 71, pp. 839-847.

Tendler, M., Rutberg, P. & Van Oost, G., 2005. Plasma based waste treatment and energy production. Plasma physics and controlled fusion, Issue 47, pp. A219-A230.

W

Wang, T., Lu, X., Hsu, H. W. & Shen, C. H., 2011. Investigation of the performance of a syngas quench cooling design in a downdraft entrained flow gasifier. Pittsburgh, 28th international coal conference.

(31)

__________________________________________________________________ BIBLIOGRAPHY

Page | 11 Y

Ye, I., Park, S., Ryu, C. & Park, S. K., 2013. Flow and heat transfer characteristics in the syngas quench system of a 300 MWe IGCC process. Applied thermal Engineering, Issue 58, pp. 11-21. Z

Zhang, Q. et al., 2012. Performance analysis of municipal solid waste gasification with steam in a Plasma gasification melting reactor. Applied Energy, Issue 98, pp. 219-229.

Zhang, Q. et al., 2013. A thermodynamic analysis of solid waste gasofocation in the Plasma Gasification Melting Process. Applied Energy, Issue 112, pp. 405-143.

(32)

Page | 12

CHAPTER 2:

LITERATURE STUDY

Chapter 1 outlined the background and scope of the current investigation. The aim, focus and objectives of the study were determined within this context. The purpose of chapter 2 will be to investigate current and previous literature with regards to quenching, quenching technology and

syngas cooling. This will be followed by a detailed section on syngas modelling in the CFD environment, which will include previous studies on both general syngas cooling models as well as

(33)

____________________________________________________ CHAPTER 2: LITERATURE STUDY

Page | 13 2.1 Synthesis gas quenching

During plasma gasification, reactions occur at temperatures ranging in excess of 1 200 °C (Huang & Tang, 2007; Byun et al., 2012) resulting in high temperature raw synthesis gas formation, which consists mainly of H2 and CO along with H2O, CO2, CH4 and H2S in lesser amounts (Ye et al., 2013;

Wiinikka et al., 2012). It is essential that this hydrocarbon gaseous product is cooled down rapidly to avoid thermal attack, slag formation and the damaging of downstream equipment (Ye et al., 2013). The quenching vessel plays a distinctive role in freezing chemical reaction products to ensure the correct pure gas consistency according to further processing requirements, usually characterised by the H2/CO ratio in the gas (Wiinikka et al., 2012; Sundstrom & DeMichiell, 1971; Uebel et al., 2016a).

Synthesis gas quality can also be defined by the synthesis gas module, as defined by Equation 2.1.

𝑀 = 𝑋𝐻2− 𝑋𝐶𝑂2

𝑋_𝐶𝑂− 𝑋_𝐶𝑂₂

2.1

Here 𝑋 denotes the molar fractions for the mentioned species (Uebel et al., 2016a).

In a poorly designed quench system, unwanted side-reactions occur over the length of the probe, forming longer chain molecules (i.e. methane, acetylene and ethylene) which result in an altered equilibrium composition followed by a consequential shift in the reaction rate. The correct design of such quenching systems is therefore essential to prevent the formation of these species (Hawboldt

et al., 1999). The choice of an appropriate synthesis gas cooling concept is usually subject to gas

impurities, slag formation tendency and gas exit temperature (Uebel et al., 2016a).

The gaseous product exiting from the reactor is typically quenched in the industry though either direct or indirect methods. Direct methods entail cooling through immediate contact with either water, chemical species or recycled hydrocarbon gas, while radiant synthesis gas cooling is classified as an indirect method (Ye et al., 2013; Oh et al., 2014). The various syngas cooling methods will be discussed in detail in section 2.1.1.

According to Hawboldt et al. (1999) a system can be deemed quenched when the dimensionless temperature gradient, given by Equation 2.2, is equal to one.

𝜎 = 𝑇𝑖𝑛𝑙𝑒𝑡− 𝑇𝑔 𝑇𝑖𝑛𝑙𝑒𝑡− 𝑇5%

2.2

The dimensionless temperature gradient parameter is equal to the difference between the inlet temperature (𝑇_{𝑖𝑛𝑙𝑒𝑡}) and the current temperature at the measuring point (𝑇_𝑔) divided by the difference between the inlet temperature and the temperature at which the reaction is within 5% of its final conversion, represented by 𝑇5% (Hawboldt et al., 1999). A value larger than one would therefore

(34)

Page | 14

mean that the system is “over-quenched” and that the quenching is not viable anymore as it would not account for the insignificance of the reaction. Likewise, a constant gradient shows that no more significant heat exchange is taking place as the system is saturated.

Quenching efficiency can be influenced by a number of factors in which the temperature profile, pressure drop and residence time are included (Hawboldt et al., 1999). In a study conducted by Hawboldt et al. (1999), it was concluded that the residence time and temperature profile are the most prominent of these parameters as shown in Figure 2-1. Figure 2-1 introduces the concept of the dimensionless temperature gradient in addition to how it was used in the study to determine quenching efficiency.

Figure 2-1: The dimensionless temperature gradient as a function of residence time at Tinlet of 850 °C (Adapted from Hawboldt et al., 1999)

In Figure 2-1, the effect of forced (water) quenching in comparison to free (air) quenching can clearly be differentiated. Even though the flow rate of air was increased from 1.5 to 8 ℓ/min, there was almost no discernible change in the quenching propensity. The authors therefore concluded that the most effective way to improve an air quenching system is to modify the geometry to successfully decrease residence time.

2.1.1 Quenching methods

Due to a diverse availability of synthesis gas quenching techniques, a wide variety of studies have been conducted regarding these methods. To evaluate the efficiency of the geometry, it is necessary

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 5 10 15 σ Residence time (ms) Water quench (8 ℓpm) Air quench (8 ℓpm) Air quench (1,5 ℓpm)

(35)

____________________________________________________ CHAPTER 2: LITERATURE STUDY

Page | 15 to analyse all possible syngas quenching options. Appendix A summarises syngas quenching studies conducted by different authors on various syngas quenching configurations.

Quenching methods include radial syngas cooling, direct water quenching and quenching through recycled syngas, although radiant syngas cooling and direct water quenching are more popular technologies (Li et al., 2016b; Ni et al., 2011). Chemical quenching is not conventionally used as it could lead to formation of by-products that might be expensive to separate at later stages and is therefore deemed economically unviable (Pylilo, 2012). Each of these quenching methods are subsequently discussed in more detail.

2.1.1.1 Radiant syngas cooling

During radiant syngas cooling, hot gas enters a cooled tube annulus that is of constant diameter (Sundstrom & DiMichiell, 1971). Even though an RSC is a more expensive application than water quenching, it is valued as sensible heat can be recovered from the syngas and slag and redirected for application elsewhere (Ni et al., 2010; Ni et al., 2011). RSCs can therefore increase the energy efficiency of the entire system between 4-5 % through sensible heat recovery and can substantially reduce operational cost in the long term (Li et al., 2016a; Li et al., 2016b). Direct quenching systems are less bulky and easier to maintain and are typically installed to reduce space, cost and maintenance (Wang et al., 2011).

2.1.1.2 Gas recycle quenching

Gas recycle quenching entails quenching of the hot syngas with recycled, cold synthesis gas redirected from downstream processes (Sundstrom & DeMichiell, 1971). To obtain the sought quenching effect, it is necessary that both the cold and hot streams are thoroughly mixed (Pylilo, 2012). This results in molecular diffusion and convective heat transport (Pylilo, 2012). Cold syngas cooling is therefore dependant on temperature, gas velocities, turbulent mixing, gas densities and quench geometry for efficient cooling (Pylilo, 2012).

2.1.1.3 Chemical quenching

A process in which the quenching medium undergoes transformation due to chemical reactions is defined as a chemical quench (Pylilo, 2012). Hydrocarbons are generally used in this type of application (Pylilo, 2012). A potential benefit of this type of quenching system is the formation and recovery of valuable products (other than syngas) and could, in combination with steam, result in a higher processing efficiency (Pylilo, 2012). Chemical quenching can for example be used as reactive slag removal stage during the formation of liquid slag and ash in a gasifier (Pylilo, 2012).

2.1.1.4 Direct water quenching

In a direct water quenching system, water is fed or sprayed into the quenching section and comes into direct contact with the hydrocarbon gas stream (Wiinikka et al., 2012). The syngas stream is

(36)

Page | 16

cooled via heat and mass transfer occurring between the two streams (Sundstrom & DeMichiell, 1971) which may induce side reactions.

A drawback may be that the large amounts of water that is used for cooling purposes could become contaminated with slag and would require additional water treatment facilities (Li et al., 2016a). It does, however, assist with scrubbing of ash and other particulate matter, but can sometimes sweep along chars which can reduce overall carbon conversion (Wang et al., 2011). Partial and full quench systems are typically employed to exceed the sticky slag temperature zone (Uebel, 2016a) – a term describing the temperature at which slag fines tend to become adhesive to vessel walls in the phase change between solid particle and fluid. This quenching method would therefore benefit from organic feed as there would be no slag formation from inorganic components.

As quenching needs to occur rapidly in quenching equipment, it is needed to investigate the quenching rate of the different technologies (Sundstrom & DeMichiell, 1971). Figure 2-2 compares the quenching rate between radial (cold wall), and liquid spray (direct water) and gas mix (recycling of syngas) quenching as investigated by Sundstrom & DeMichiell (1971).

Figure 2-2: Comparison between quenching rates of different quenching methods at typical quenching conditions (Adapted from Sundstrom & DeMichiell, 1971)

Sundstrom & DeMichiell (1971) found that gas mixing resulted in the quickest quench time, which was followed by liquid spray and cold wall quenching. Gas mixing resulted in an increasing quench rate, due to turbulent mixing that occurs when applying this method. During liquid spray quenching, however, drops accelerated due to the faster moving carrier gas. As the difference between the velocities of the moving gas and spray drops decrease, the quenching rate slows due to a decrease in the heat and mass transfer coefficients. This is due to a decrease in droplet velocity. For all the techniques, it was found that the quenching rates were in the order of 106_ºR/sec.

0 500 1000 1500 2000 2500 3000 0 2 4 6 8 10 12 Te m pe ratu re (K) Residence time (ms) Gas mixing Liquid spray Cold wall

(37)

____________________________________________________ CHAPTER 2: LITERATURE STUDY

Page | 17 2.1.1.5 Combined quenching technology

Quenching technologies are regularly combined to make the process more efficient depending on gas exit temperatures and impurities (Uebel et al., 2016a).

A study directed by Uebel et al. (2014) on syngas heat recovery systems tested different syngas cooling configurations in an integrated gasification combined cycle (IGCC) plant based on thermodynamic and technical considerations. The different concepts were integrated with a process model to find the most efficient design. Figure 2-3 presents the different configurations that were tested, taking into account both convective and radiative heat recovery.

Figure 2-3: Syngas heat recovery systems as investigated by Uebel et al. (2014) (Taken from Uebel et al., 2014)

The technical evaluation found that convective coolers are more sensitive regarding fouling and slag deposits than RSCs. The combination of a partial water quench with a convective cooler is questionable due to difficulty in controlling a constant outlet temperature upstream in the convective cooler. This could result in challenging process control setups due to varying slag properties. Furthermore, the radiant system shows more robustness against fouling and is easier to clean. Combining the technical and thermodynamic preferences, the verdict is in favour of radiant coolers. For the thermodynamic analyses, designs A and D presented higher efficiencies with 1.42 and 1.68 % respectively, while concepts B and E showed lower efficiency with 1.49 and 1.19 % respectively. It should be noted that the thermodynamic analysis was quite fundamental and that more detailed analyses proved that the efficiency could improve even further, despite the original low percentages. Concept D proved to exhibit more technical difficulty than A, regardless of a higher efficiency value.

(38)

Page | 18

Therefore, the authors concluded concept A as being the most promising design due to relatively high efficiency and limited technical complexity.

In a similar investigation steered by Ni et al. (2011), the flow and temperature fields in syngas coolers of an entrained flow gasifier were compared. Figure 2-4 shows the different configurations that were investigated.

Figure 2-4: Syngas cooling configurations as tested by Ni et al. (2011) where (a) direct water quenching (b) radiant syngas cooler (RSC) (c) RSC and (d) gas quenching (Taken from Ni et al., 2011)

In conclusion, it was found that either an IGCC, a RSC or a RSC with syngas quench would be the preferred option due the recovery of sensible heat from syngas that can produce higher quality water vapour which is advantageous for use elsewhere in the process. This is in agreement with the findings of Uebel et al. (2014). It is therefore not only important to consider the quenching efficiency when designing a syngas cooling setup, but to take into account the application of the setup as well. 2.1.2 The direct quenching method

The core focus of this study will be on a direct water quench probe for cooling of the plasma gasifier’s product gas stream. The reactivity, thermodynamic aspects and quenching configuration of such systems will subsequently be discussed.

2.1.2.1 Reactivity and thermodynamic considerations in the direct quench system

The direct water quench is associated with the exothermic water gas shift (WGS) reaction, governed by Equation 2.3 (Uebel et al., 2016b).

(39)

____________________________________________________ CHAPTER 2: LITERATURE STUDY

Page | 19 𝐶𝑂(𝑔) + 𝐻2𝑂(𝑔) ⇌ 𝐶𝑂2(𝑔) + 𝐻2(𝑔) ∆𝑅𝐻° = −41.1 𝑘𝐽/𝑚𝑜𝑙 2.3

The low temperature steam that is formed during water quenching shifts the reaction towards CO2

and H2 formation and can be advantageous for the H2/CO yield in the system when employed

correctly (Wang et al., 2011; Uebel, 2016b). The quench probe will thus, in this case, serve as a reactor. Decreased temperatures favour the conversion of carbon monoxide and steam to carbon dioxide and hydrogen (Bustamante et al., n.d.; Rhodes et al., 1995) and due to the low temperature requirements, the reaction is typically carried out in a catalytic environment to effectively increase reaction rate (Bustamante et al., 2004). The current setup is non-catalytic and catalysts are therefore negligible in the current study.

Several studies were conducted on the kinetics of the reverse uncatalysed WGS reaction, with the forward reaction not studied as extensively due to low equilibrium conversions at elevated temperatures (Bustamante et al., n.d.). All of these investigations were carried out in quartz reactors, and the main findings are summarised in Table 2-1.

Table 2-1: Kinetic expressions for the WGSR

Author P (MPa) T (K) Ea (kJ/mol) ko (ℓ/mol)0.5s-1

Reverse reaction

Bustamante et al. (2004) 0.1 1148-1198 222.2 1.09 × 107

Bustamante et al. (2004) 1.6 1148-1198 218.4 5.99 × 108

Graven & Long (1953) 0.1 1148-1323 234.3 2.90 × 109

Karim & Mohindra (1974) 0.1 <2500 397.5 2.30 × 1016

Kochubei & Moin (1969) 0.1 1023-1523 326.4 6.40 × 1012

Tingey (1965) 0.1 673-1073 164.2 7.60 × 104

Tingey (1965) 0.1 1073-1323 318.0 1.20 × 1013

Forward Reaction

Bustamante et al. (n.d.) 0.1 1070-1134 304.6 7.68 × 1010

Bustamante et al. (n.d.) 1.6 1070-1134 288.3 7.40 × 108

Graven & Long (1953) 0.1 1173 274.1 7.97 × 109

 Reverse reaction

Karim & Mohindra (1974), Kochubei & Moin (1969) and Tingey (1965) received reaction rates at high temperatures which were comparable with each other. Graven & Long (1953) and Bustamante et al. (2004) both obtained faster reaction rates which were in agreement. In a further study conducted by Bustamante et al. (2004), a CFD analysis was used to investigate the influence that the different reactor geometries could have on the reaction rate of the WGS reaction. It was concluded that this

(40)

Page | 20

had no influence on the reaction rate, but the influence on the quenching rate is found to be noteworthy.

 Forward reaction

Bustamante et al. (n.d.) and Graven & Long’s (1953) results were synonymous with each other. The slightly lower reaction rate found by Bustamante et al. (n.d.) was allotted to nickel deposits found in the reactor which formed nickel carbonyl with the CO species. It is noteworthy considering low temperature catalytic data as there are limited non-catalytic studies available.

In a study conducted by Choi & Stenger (2003), the exiting conversion of CO was reported for different reactor temperatures in the presence of a catalyst. This is graphed by Figure 2-5.

Figure 2-5: Exiting CO conversion for different reactor temperatures and H2O/CO ratios at 1 atm and present catalyst (Taken from Choi & Stenger, 2003)

Choi & Stenger (2003) concluded that higher conversions of CO are achieved with higher temperature conditions as well as elevated H2O/CO ratios in the presence of a catalyst. In general,

lower temperatures had low conversion, a phenomenon that will be even more crippled when the catalyst is removed.

 Pressure dependency

According to Rhodes et al. (1995) the selectivities of the products and reactants are not determined by the position of the forward and reverse reactions, concluding that the equilibrium constant of the WGS reaction is deemed independent of pressure. This is evident through the studies conducted by Bustamante et al. (2004) and Bustamante et al. (n.d.) in Table 2-1. Sato et al. (2004) reported this phenomenon in a study on the kinetics of the WGS reaction on supercritical water. In all three cases, an increase in pressure proved to have a lesser influence on the reaction rate constant, indicating that the reaction will reach equilibrium.

(41)

____________________________________________________ CHAPTER 2: LITERATURE STUDY

Page | 21 The WGS reaction is nevertheless not the only reaction that could occur in the quench probe. In the presence of sulphur compounds, the hydrogenation reaction (Equation 2.4) and the hydrolysis reaction (Equation 2.5) may also occur (Wiinikka, 2012):

𝐶𝑂𝑆(𝑔) + 𝐻₂(𝑔) ↔ 𝐻2𝑆(𝑔) + 𝐶𝑂(𝑔) ∆𝑅𝐻° = 7 𝑘𝐽/𝑚𝑜𝑙 2.4

𝐶𝑂𝑆(𝑔) + 𝐻₂𝑂(𝑔) ↔ 𝐻₂𝑆(𝑔) + 𝐶𝑂₂(𝑔) ∆_𝑅𝐻° = −34 𝑘𝐽/𝑚𝑜𝑙 2.5

Wood chip ultimate and proximate analyses, as conducted by Koukouzas et al. (2007), Shi et al. (2016) and Varol et al. (2010), indicated the presence of sulphur, albeit it is a small fraction (0 – 0.14 wt% dry basis). Equation 2.4 – 2.5 may therefore be included in the reactive study if the influence thereof is of note.

An investigation led by Yan et al. (2012) evaluated the product formation in a quench reactor according to the quenching temperature. A coal gasification system was used in which the primary objective was to increase acetylene production. The inlet gas temperature to the quench was estimated to be 1 630 K and a variety of chemical species were assumed to be forming within the quench reactor. The formation of hydrocarbon species was predicted in a range of 1 000 – 4 000 K at various steam feed rates. The results are summarised in Figure 2-6.

In Figure 2-6 it can be observed that an increase in steam flow rate causes an increase in the temperature at which most of the products form, although it simultaneously decreases the rate of coke formation on the reactor wall (Yan et al., 2012). The formation of C2H2 is more dominant at

lower temperatures. For the purpose of this study, it is important to note that a higher steam flow rate leans toward the formation of a greater number of useable products (CO and H2) in relation to

unwanted products. Likewise, lower temperature operation tends to inhibit the formation of longer chain molecules, which would finally increase the synthesis gas yield. Optimal operating conditions will therefore tend to be at relatively low temperatures (1000 – 1400 K) and high steam flow rates. Yan et al. (2012) concluded that reactor performance is influenced by two dominant factors which include the C/H mass ratio in the feed gas as well as the quench temperature. Increased quench temperature and a higher mass ratio of C/H would effectively increase the formation of C2H2.

Additional water/steam injection would cause a decrease in C/H ratio, but would also lessen slag formation on the quench wall. It is therefore necessary to optimise both of these factors and simultaneously observing the influence of the parameters on one another.