tekatesfayemengesha

OCTO

BER

&910

Kasteel VaesharteltMaastrichtThe Netherlands

COMBURA13COMBUSTION RESEARCH AND APPLICATION

i

COMBURA 2013 was organized on behalf of STW-Platform Clean and Efficient Combustion Technology Foundation STW NVV (Nederlandse Vlamvereniging) Dutch Section of the Combustion Institute and was organized by Theo van der Meer, University of Twente Rob Bastiaans, Eindhoven University of Technology Liselotte Vogels-Verhoeven, Eindhoven University of Technology Joris Koomen, Stork Technical Services and Secretary of the NVV Leo Korstanje, Technology Foundation STW

Platform Clean and Efficient Combustion Prof.dr.ir. Th.H. van der Meer (chairman) Universiteit of Twente Prof.dr.ir. R.S.G. Baert TNO Automotive Dr.ir. M.F.G. Cremers KEMA Nederland BV Dr.ir. M.A.F. Derksen Stork Thermeq BV Prof.dr. L.P.H. de Goey Eindhoven University of Technology Ir. B. Hakstede DAF Trucks Dr.ir. W. de Jong Delft University of Technology Dr.ir. J.H.A. Kiel ECN Prof.dr. H.B. Levinsky University of Groningen and

DNV KEMA Dr.ir. L. Post Shell Global Solutions Int. BV Dr.ir. P. Pronk Tata Steel Dr.ir. C.J.A. Pulles KIWA Gas Technology Prof.dr. D.J.E.M. Roekaerts Delft University of Technology Dr. L.J. Korstanje (secretary) Technology Foundation STW Secretariat Astrid van der Stroom Technology Foundation STW P.O. Box 3021 3502 GA Utrecht The Netherlands Tel: +31 (0)30 600 1 297 Fax:+31 (0)30 601 44 08 E-mail: [email protected]

Photo front cover: Room for ID's, Nieuwegein

ii

Abstracts of CCC projects, 9 October 2013 1

1. flexFLOX - Flameless combustion conditions and efficiency improvement of 3 single- and multi-burner-FLOXTM furnaces in relation to changes in fuel and oxidizer composition

Luis Arteaga-Mendez et al., Delft University of Technology 2. HiTAC Boiler - Heavy fuel-oil combustion in a HiTAC boiler 9 Th.H. van der Meer et al, University of Twente 3. MILDNOx - Fuel flexibility and NO Formation in dilute combustion 12 Ebrahim Abtahizadeh et al., Eindhoven University of Technology 4. MoST - Multi-scale modification of swirling combustion for optimized gas turbines 14 A.A. Verbeek et al., University of Twente 5. XCiDE - Crossing the Combustion modes in Diesel Engines 16 Ula Egz et al.,Eindhoven University of Technology 6. ALTAS - Advanced low NOx flexible fuel gas turbine combustion, aero 18

and stationary Andrea Donini et al., Eindhoven University of Technology

7. BiOxyFuel - Torrefied biomass combustion under oxy-fuel conditions in coal fired 19

power plants Eyerusalem Gucho, University of Twente

Combustion of torrefied beech wood in new spiral combustion reactor E. Russo, Eindhoven University of Technology Biomass pyrolysis in DNS of turbulent particle-laden flow

8. ULRICO - Ultra rich combustion of hydrocarbons and soot formation 25 Mark Woolderink et al., University of Twente

iii

Abstracts of presentations, 10 October 2013 27

1. Pepijn Pronk (Tata Steel) Application of model predictive control for steel heating furnaces 29

2. Pierre Ploumen (Ploumen Energy Consultancy) 30 Pulse combustor applied in integrated paper sludge conversion process with

drying and sequential combustion 3. Bart Venneker (Stork Thermeq) 32

Towards HiTAC conditions in a 9 MW oil fired industrial boiler 4. Likun Ma (Delft University of Technology) 34 Modeling of Delft ethanol jet-in-hot-coflow flame with tabulated chemistry method 5. David Diarra (OWI Oel-Waerme-Institut) 35 Characterizing ignition behavior of primary reference fuels by numerical stability

analysis of detailed reaction mechanisms 6. Moresh Wankhede (Dacolt) --

Unsteady gas turbine model combustor simulations with FGM 7. Ferry Tap (Dacolt) --

Advanced Diesel engine combustion modeling with Dacolt PSR+PDF 8. Peter-Christian Bakker (Eindhoven University of Technology) 37

Enabling partially premixed combustion using ON70 fuels in the low load regime 9. Bas Nijssen (Eindhoven University of Technology) 39 Experimental investigation of low octane fuel composition effects on load range

capacity in partially premixed combustion (PPC) 10. Andreas Ortwein (DBFZ Deutsches Biomasseforschungszentrum gemeinntzige 41

GmbH, Leipzig) CFD in biomass combustion research activities and challenges 11. Jim Kok (University of Twente) 43 Limit cycles of combustion and pressure oscillations in gas turbine engines:

highlights of results by 18 PhD students in the LIMOUSINE project 12. Robin Doddema (Eindhoven University of Technology) 47 How premixed is partially premixed combustion: optical engine experiments to study fuel stratification

iv

Abstracts of posters 49 1. A.G.Iyer, Eindhoven University of Technology 51

Unsteady stretch effects in an oscillating premixed counterflow flat flame 2. A. Donini, Eindhoven University of Technology 52

The application of flamelet generated manifolds in the modelling of partially premixed cooled flames

3. M.U. Gktolga Eindhoven University of Technology 54

The effect of initial turbulence level and temperature profile on autoignition timing of turbulent mixing layers

4. R. Haas-Wittm, RWTH Aachen 56 Determination of the effective activation energy of ethanol by measuring the adiabatic laminar burning velocity 5. J.D. Herdman, University of Groningen 58

Dual-channel angle-dependent light scattering measurements of silica aggregate growth in premixed flames

6. F.E. Hernndez-Prez, Eindhoven University of Technology 60 Steady closed reacting fronts under terrestrial gravity conditions:

their numerical prediction and computation 7. M. Meijer, Eindhoven University of Technology 62

Engine combustion network: cracking diesel spray fundamentals together 8. N. Speelman, Eindhoven University of Technology 64 Development of a flame Ionization model for the prediction of electric currents

in methane-air flames List of participants COMBURA 2012 67

Abstracts of CCC-projects

1

Combura13 symposium, Kasteel Vaeshartelt, Maastricht, The Netherlands, October 9-10, 2013

Flameless combustion conditions and efficiency improvement of single- and multiburner- FLOXTM furnaces in relation to changes in fuel and oxidizer composition

(10428)

L.D. Arteaga Mendez1, G. Sarras2, E.-S. Cho2 , S.Y. Mahmoudi Larimi1, J. Lu2, W. de Jong2 , M.J. Tummers1, E.H. van Veen1 and D.J.E.M. Roekaerts1,2,*

1Department of Process and Energy, Section Fluid Mechanics, Delft University of Technology, The Netherlands 2Department of Process and Energy, Section Energy Technology, Delft University of Technology, The Netherlands

*[email protected]

Key words: Flameless combustion, furnaces, fuel flexibility, jet-in-coflow, tabulated chemistry

Flameless combustion is a clean combustion concept leading to strongly reduced pollutant emissions compared to traditional combustion processes. To reach the flameless combustion regime the air (and/or fuel) streams are diluted with hot combustion products with a temperature sufficiently high for the combustion process to be stable and occurring in a distributed reaction zone. In furnace applications it leads to a higher efficiency because of the more extensive use of air preheat. To widen the range of applicability of the this technology, the impact of changes in fuel and oxidizer composition on the combustion process in furnaces operated using flameless combustion is investigated. Dutch natural gas (DNG) being the base-line fuel, also the combustion of biogas (represented as mixture of DNG and CO2) and mixtures of DNG and H2 are investigated. By combining detailed measurements and modeling, in single- and multi-burner furnaces insight is gained in the flame structure, heat transfer enhancement and the emission reduction. The relevant turbulent reactive mixing processes are examined using laser diagnostic methods and computed in detail using detailed and reduced chemical models and using statistical models for turbulence and turbulence chemistry interaction.

SINGLE BURNER SYSTEMS

In an earlier STW project the Delft jet-in-hot-coflow (DJHC) burner was built to study the fundamentals of flameless combustion for traditional gaseous hydrocarbon fuels, mainly DNG [1,2]. In the flexFLOX project, the DJHC was upgraded so that the base fuel can be blended with hydrogen and carbon dioxide to extend the research to fuels that resemble nontraditional gaseous fuels like biogas, coke oven gas or refinery gas.

EXPERIMENTS

For ordinary turbulent lifted flames in cold air, flame stabilization is governed by flame propagation. The stabilization mechanism changes to auto-ignition kernels when DNG is oxidized in a hot coflow of lean

combustion products [1]. Flame luminescence recordings show that addition of hydrogen to the fuel results in a shift in upstream direction of the stabilization region regardless the coflow conditions. Addition of hydrogen to DNG fuel (Fig. 1) gradually changes the flame structure at the stabilization position until auto-ignition kernels can no longer be detected and a connected lower flame boundary appears as in ordinary lifted flames in cold air.

Figure: 1 Instantaneous snapshots of the studied H2/DNG flames oxidized in hot coflow. The red cross denotes the location of the fuel pipe exit. Cases (a),(b),(c) and (d) respectively are for pure DNG, and mixtures with 5% H2 ,10% H2 and 25% H2 .

Analysis of the luminescence images shows that the mechanism governing the lift-off process of DNG/CO2 flames is similar to that of natural gas flame in which ignition kernel generation by autoignition is responsible for flame stabilization. Luminescence measurements show that compared to the lift-off height of the DNG flame, addition of 30% carbon dioxide to the fuel increases the lift-off height by less than 20%. In summary, in hot coflow burner mimicking the MILD regime, the stabilization mechanism and the flame lift-off height for biogas are very similar to those of natural gas. The effect of hydrogen and carbon dioxide addition on the temperature of turbulent DNG jet flames oxidized in cold air and hot diluted coflow was

3


studied using Coherent anti-Stokes Raman Scattering Spectroscopy (CARS).

Figure: 2 Temperature PDFs at points where flame luminescence is always present for the 25% H2 fuel jet flame (a) oxidized in cold air and (b) in hot coflow.

For flames oxidized in cold air, the temperatures measured in the stabilization region present a bimodal distribution; further downstream the mean temperature gradually increases up to the region where high temperatures are predominant. For flames oxidized in hot coflow, hydrogen has a strong effect. In Fig. 2 temperature PDFs for the 25% H2 flame case oxidized in cold and hot coflow in a point where the flame luminescence is present most of the time, show a considerable peak temperature reduction typical of flameless oxidation combustion.

Figure: 3 Radial profile of mean and 95% upper confidence limit of the temperature for DNG and DNG+H2 flames in hot coflow at an axial position in the stabilization region.

When the amount of hydrogen in the fuel is low, the mean value of the temperature in the stabilization region is not a good indicator of the presence of reactions (ignition) (Fig. 3). Then the tail of the PDF has to be taken into account. At higher hydrogen concentration in the fuel, a clear and sharp increase in the mean temperature is indicating the start of combustion reactions. The high peak temperatures measured are attributed to the effect of hydrogen preferential diffusion. Carbon dioxide addition to the fuel does not produce significant changes in the temperature of the studied flames (Fig. 4). The result shows that the MILD combustion of the studied fuel systems is a promising technique for stable combustion of hydrogen containing and biogas fuels. A single burner furnace allowing the study of flameless operation in a situation with aerodynamic flue gas recirculation was designed and built. However, the completion of its construction is ongoing.

Figure: 4 Radial profiles of mean temperature for the DNG and 30% CO2 flames oxidized in hot coflow.

MODELING A new computational model for the Delft jet-in-hot coflow burner using hybrid finite-volume transported PDF methods has been developed. The results of the application of this model have been compared with experiments and with a more standard model available in a commercial CFD code.

4


PDF-MODELING The transported PDF model combines an underlying Reynolds stress model with the Monte Carlo simulation of velocity and scalar statistics. In the current work a joint scalar PDF model was used. A 3D reduced chemistry model using two mixture fractions and a progress variable is used. A second mixture fraction is needed to represent the dependence of coflow composition and temperature on the radial direction. The dynamics of the progress variable is governed by a reaction rate retrieved from a flamelet generated manifold (FGM) based on unsteady non-premixed flamelets. Igniting flamelets are used to construct the 3D-FGM in order to represent the auto-ignition and lift off observed in the jet-in-coflow burner. In a subsequent step the representation of enthalpy (temperature) fluctuations has been refined by the addition of enthalpy deficit as independent variable, leading to a 4D-FGM For computational efficiency, the FGM-library is re-tabulated using the Delft FLAME code, using adaptive refinement in the regions of composition space where large gradients are present. The model has first been used for the simulation of the DNG flame DJHC-I [1,2]. The composition is given by the two mixture fractions, progress variable and enthalpy deficit and all other scalar variables (temperature, species mass fractions, density, etc.) are retrieved from the look-up table. Figure 5 shows the radial profiles of the mean temperature for four axial locations. In the inner shear layer, r < 12mm, the mean temperature is in good agreement with experimental results for all axial positions. In the outer region, an over prediction is observed at downstream positions. This may be attributed to the use of only a small number of flamelets (Z2 values) to represent the outer region of the coflow and a too narrow range of enthalpy deficit. Furthermore, in the present simulation the temperature fluctuations at the coflow inlet having a value of about 120 K, have not been taken into account. This explains the deviation in Trms at z=30mm (Fig.6). In view of the strong dependency of autoignition on temperature, it is clear that including these fluctuations is expected to lead to more accurate results. The simulation with the suggested improvements are in progress. COMPARISON OF EDC AND PDF MODELS Numerical simulations of the experiments in the jet-in-hot-coflow burner with DNG and with DNG+30%CO2 fuel, representing biogas, were conducted by solving the RANS equations using Reynolds stress model (RSM) as turbulence model in combination with EDC (Eddy Dissipation Concept)

for turbulence-chemistry interaction and also with the transported PDF/3D-FGM method described above. The DRM19 reduced mechanism was used as chemical kinetics with the EDC model. The performance of EDC and PDF models in predicting the flame structure was then evaluated against the available experimental measurements. The results show that the EDC/DRM19 and PDF/3D-FGM models predict the experimentally observed decreasing trend of lift-off height with decrease of the coflow temperature. Further, although more detailed chemistry is used with EDC, the PDF model in combination with an FGM leads to results in better agreement with experiments [3].

Figure 5: Radial profiles of the mean temperature at four axial locations. The solid line denotes simulation results and the circles denote experimental results.

5


Figure 6: Radial profiles of the rms of the temperature at four axial locations. The solid line denotes simulation results and the circles denote experimental results. MULTI-BURNER FURNACE STUDY Figure 7 shows a schematic diagram of the TU Delft multi-burner FLOXTM combustion furnace (MEEC furnace [4]) and the burner configuration. At full load it has three pairs of regenerative FLOXTM burners, each with a rated thermal power of 100 kWth. Thermal load is simulated by a cooling system consisting of eight single ended concentric tubes, four placed at the bottom of the furnace and four at the top. EXPERIMENTS Recently experimental studies using two burner pairs have been made to resolve some issues raised by the flue gas measurements in previous experiments.

AirBlower

Exhaust GasInduce Fan

Dutch NG

Data Acquisition

O2, CO2, CO, NO

Gas Analyzer

Cooling Tube Lower

Cooling Tube Upper

S-Type TC

FlowmeterTemp.

1

2

Stack

Velocity

Temperature

Figure 7: Schematic diagram of the MEEC furnace and of the burner configuration during operation with two burner pairs in parallel firing mode. Firstly, it has been noticed repeatedly that the CO emission in the regenerator flue gas has a periodicity, with time between wave peaks about 2.5 times the cycle time of MEEC burner firing cycle. And until now this phenomenon remained unexplained. The reliability of the peaks was in doubt since it could be also caused by the inaccuracy of gas analyzer. For further investigation of this phenomenon, a series of experiments have been carried out with natural gas as fuel. Flue gas measurements were made using two gas analyzers and related equipment provided by Tata Steel and the gas analyzer used previously. Figure 8 shows the comparison of CO emission for 3 cycle times (10s and 30s) measured with two different gas analyzers (MRU, provided by Tata Steel and S710) but with exactly same gas sample. For the case of a cycle time of 30s, the results of the two gas agree, but for the cases of 10s cycle time, only the MRU result is correct. The S710 is not as sensitive as the MRU, which is the reason of the previously found 2.5 times cycle time periodicity. The main conclusion is that fluctuations in the CO emission are really present and the periodicity of the CO peak coincides with the burner cycle time. To find out more about the cause of the incomplete combustion signaled by the CO, the CH4 concentration in regenerator flue gas and in the exhaust gas from different burner was measured.

6


Figure 8: Comparison of CO emission for 2 cycle times (10s and 30s) measured with two different gas analyzers

Figure 9 CH4 concentration in flue gas streams coming from regenerators of three different burners (E5, E6, and E8) and regenerator flue gas. Figure 9 shows the comparison of CH4 concentration measured with a CH4 gas analyzer (provided by Tata Steel) in flue gas streams coming from regenerators

of three different burners (E5, E6, and E8) and in the combined stream of all flue gas leaving the furnace via the regenerators. It is clear that the CH4 emission is burner dependent. The reasons for this are under investigation. Emission measurements were made during a period of about 40 minutes, including the transition from Flame mode to FLOX mode, during start up of the furnace. The measurements were including CH4 data, as Figure 10 shows. The transition starts automatically at furnace temperature 850 C which is a set point of the LabView MEEC control program. After the transition the CO emission becomes more fluctuating. CO2 concentrations measured increase a little bit with the same fluctuation range and O2 concentrations analyzed decrease slightly with the same fluctuation range. The advantage of the FLOX operation is again shown clearly by the sharp NO decrease at the transition.

Figure 10 Time evolution of emissions during a period of about 40 minutes, including the Flame to FLOX transition A new measurement campaign for operation with DNG and biogas, employing an improved gas analyzer and additional measurement probes for composition and temperature is ongoing and so is the effort to obtain accurate CFD predictions of the composition and temperature fields in the furnace. CONCLUSION The highlights of the past year are threefold: detailed measurements of flame structure in the Delft jet-in-hot-coflow (DJHC) burner, detailed comparison of CFD predictions for DJHC flames using EDC and a new PDF model, upgrade of the MEEC furnace and the measurement equipment to compare operation with natural gas and biogas fuel.

7


ACKNOWLEDGEMENT The authors would like to thank Technology Foundation STW, Tata Steel, Shell, Numeca Int, NVV, Celsian BV, and WS Gmbh for supporting this project. REFERENCES [1] E. Oldenhof, M.J. Tummers, E.H. van Veen, and D.J.E.M. Roekaerts, Ignition kernel formation and lift-off behaviour of jet-in-hot-coflow flames, Combustion and Flame, 157(6) 1167 1178, 2010. [2] E. Oldenhof, M.J. Tummers, E.H. van Veen, and D.J.E.M. Roekaerts. Role of entrainment in the stabilisation of jet-in-hot-coflow flames. Combustion and Flame, 158(8), 1553 1563, 2011 [3] Y. Mahmoudi, G. Sarras, L.D. Arteaga Mendez, E.H. van Veen, M.J. Tummers, D.J.E.M. Roekaerts, Numerical and experimental investigation of turbulent flame structure in a Jet-in-Hot-Coflow burner. On the effect of coflow settings and fuel composition, submitted for publication [4] E.-S. Cho, D. Shin, J. Lu, W. de Jong, and D.J.E.M. Roekaerts, Configuration effects of natural gas fired multi-pair regenerative burners in a flameless oxidation furnace on efficiency and emissions, Applied Energy, 107, 25-32, 2013

8


HiTAC: Heavy Fuel-oil combustion in a HiTAC boiler

Th.H. van der Meer1, S. Zhu1, H. Rodrigues2, E.H. van Veen2, D.J.E.M. Roekaerts2, M.J. Tummers2,

B.C.H.Venneker3

1Thermal Engineering Group, University of Twente, P.O.Box 217, 7500 AE Enschede, The Netherlands 2 Department of Process and Energy, Section Fluid Mechanics, Delft University of Technology, The Netherlands

3Stork Technical Services, Langelermaatweg 12, 7553 Hengelo, The Netherlands

Key words: High temperature air combustion, spray combustion, boiler.

This project concerns the extension of the application of High Temperature Air Combustion (HiTAC) to heavy-oil combustion processes in a boiler. To generate the knowledge needed to be able to develop and design such a boiler, experimental and computational investigations will be made of turbulent spray flames under HiTAC conditions. OBJECTIVES: An experimental study of spray flames of light fuel oil burning in a co-flow of hot air diluted with combustion products will provide detailed knowledge of the relations between atomization process, ignition, entrainment and burnout. A spray combustion model for the HiTAC regime will be developed for heavy fuel combustion, including the prediction of emissions under HiTAC conditions (NOx, CO and soot). This will provide a tool to assist in burner and boiler development. The spray model will be embedded in the commercial CFD code ANSYS-Fluent and used in combination with appropriated models for turbulence, soot formation and radiative heat transfer. This tool will be used to support the preparation of a HiTAC boiler field test. This field test will be done at 9 MW scale by Stork Technical Services, using heavy fuel oil characterized and delivered by Shell. Stork will also develop water-steam cycles optimized for application in combination with HiTAC combustion. The principle of HiTAC lends itself ideally for application in boilers. The key features of this high-efficiency combustion process can be utilized to lead to simpler, cheaper and more reliable designs of boilers, with very low emissions of harmful species such as Nitric Oxides (NOx) and Carbon Monoxide (CO). Gains could be significant especially for heavy fuel-oils, whereas the HiTAC combustion process lends itself ideally for the combustion of all sorts of difficult fuels, ranging from low-calorific gases such as waste-gases, to heavy fuel-oils. The resulting optimized boiler design should be cheaper and more reliable than the present standard. Combining this with a better environmental performance, the economic and market perspective of the technology and its products

is very healthy. Moreover, these systems also lend themselves ideally for application of Oxyfuel combustion, which is one of the investigated routes to Carbon Capture and Storage (CCS). This means that these systems will be at least capture-ready, or could, with further research effort, also be further developed as Carbon Capturing. RESULTS-WP1: EXPERIMENTS ON LIGHT FUEL SPRAY FLAMES Figure 1 shows the burner facility for the study of light fuel oil flames at Delft University of Technology. It consists of a pressure-swirl atomizer producing a spray of fine fuel droplets issuing in a hot coflow of combustion products. These products are produced by the secondary burner in which air and Dutch natural gas (DNG) mix and generate a matrix of lean flames. The air/DNG ratio in combination with the effects of two perforated plates along the pipe length determines the temperature, oxygen and turbulence levels.

Figure 1 - Experimental setup and photographs of a spray flame in hot-dilute coflow and the secondary burner. The field of view of the spray flame photograph has dimensions 170x160mm2.

Several ethanol and acetone sprays have been studied. Here we consider one illustrative example, an ethanol flame. Simultaneous measurements of droplet velocity and diameter were made using Phase Doppler anemometry (PDA) in the spray region. Results are presented in Figure 2. Gas phase velocity statistics were derived selecting from the full record only

9

Combura13 symposium, Kasteel Vaeshartelt, Maastricht, The Netherlands, October 9-10, 2013 droplets that were suitable tracers of the gas phase (d

Combura13 symposium, Kasteel Vaeshartelt, Maastricht, The Netherlands, October 9-10, 2013 z=10 to 50 mm under similar co-flow inlet conditions as presented in WP1. Good agreement was observed between the experimental data and simulation results. The simulation results also showed that an increased temperature of the co-flow leads to a higher peak temperature in the flame, while a lowered O2 concentration in the co-flow leads to a lower peak temperature. With the low O2 concentration of 6%, the temperature difference between the peak temperature and co-flow temperature can even drop to about 200K, which is a significant reduction from conventional flames that often have a temperature difference above 1000K. NOX formation was also reduced considerably with the reduced peak temperature. Thus we achieved the HiTAC condition in ethanol spray combustion.

800

1200

1600

2000

2400

T (K

)

Coflow inlet conditionsXo2=9% Ucf=3.5 m/s Tcf= 1500K

z=50mm

z=60mm

Figure 3 Predicted radial profiles of mean temperature at z=50 and 60 mm above the nozzle exit (top) and predicted SMD (d32) of droplets at z=10 to 50 mm (bottom) under coflow inlet conditions close to the ones presented in WP1. We then went further to numerical investigation of heavy fuel-oil combustion towards HiTAC conditions in a 9MW boiler. The Euler-Lagrange method and the Eddy Dissipation model with a two-step global reaction mechanism were employed. The standard k- model was used for the turbulence, and an empirical droplet size distribution was used for the spray injected by a steam blast atomizer. Simulation results with the existing burner at Stork Technical Services showed that increasing of the temperature of the combustion air from 373 K to 746 K leads to a higher peak temperature from 2240 K to 2390 K, while reducing O2 concentration of the combustion air from 23.065 wt% to a half results in more uniform temperature distribution with a peak temperature of 1510 K. Further investigation was done with recycling various

ratios of flue gas into the primary and secondary air respectively for introducing various O2 concentration conditions for the combustion air flow. Four cases which are possible for application in the field test were then numerically studied and compared with the base case. The predicted temperature difference between the average temperature and the peak temperature showed that the case with the lowest O2 concentration in the primary air has the least temperature difference in the boiler. It was also shown that besides thermal NOX, fuel NOX is also one of the dominant contributions to NOX formation in heavy fuel-oil combustion. By introducing flue gas recirculation, thermal NOX can be reduced to a very low level, leaving the fuel NOX playing the dominant role. The interaction between soot and radiation also showed considerable influence on the predicted temperature profiles. In the case with hot combustion air, the peak temperature was reduced by 140 K and the NOX emission was reduced to about one fourth. As a result, for heavy fuel-oil combustion, a more uniform temperature distribution in the boiler can be achieved by diluting the primary and secondary air flow with flue gas recirculation and thermal NOX can be effectively reduced, while the remained fuel NOX formation is mainly dependent on the local combustion characteristics and the initial concentration of nitrogen-bound compounds. Realization of HiTAC condition in heavy fuel-oil combustion depends on the possibility to guarantee a feasible and sufficiently high level of entrainment of flue gas into the evaporating spray jet. Knowledge of these simulations will be used to get a better understanding of spray combustion towards HiTAC conditions and to help with the design of the HiTAC burner. ACKNOWLEDGEMENT The authors would like to thank STW, NVV, Stork Technical Services and Shell for sponsoring this project.

0

10

20

30

40

50

60

0 5 10 15 20 25 30 35 40

d 32

(um

)

Radial coordinate (mm)

z=10mmz=20mm

11


MILDNOx: Fuel flexibility and NO formation in dilute combustion

S.E. Abtahizadeh1, J.A. van Oijen

1, L.P.H. de Goey

1

A.V. Sepman2, A.V. Mokhov

2 and H.B. Levinsky

2,3

1Combustion Technology, Eindhoven University of Technology, P.O. Box 513 5600 MB Eindhoven, the Netherlands 2Combustion technology, University of Groningen, 9747 AG Groningen, The Netherlands

3KEMA, P.O. Box 2029, 9704 CA Groningen, The Netherlands

Key words: Large-eddy simulation, flamelet-generated manifolds, jet-in-hot-coflow, hydrogen addition.

Combustion using highly preheated air, together with diluted air and/or fuel, is a clean combustion concept that combines high efficiency and low pollutant emissions in industrial heating processes. Having names such as flameless oxidation, high efficiency combustion and MILD combustion, these methods allow the use of recuperated heat in high-temperature processes without the penalty of increased NOx emissions, and offer the possibility of substantially homogenizing the temperature field in furnaces.

In the MILDNOx project, researchers from TU/e and RuG investigate the structure, stabilization and NO chemistry of mild flames. The obtained knowledge is transferred to the industrial partners, Tata Steel and Numeca Int., in the form of validated numerical models. In the previous years, the transition from normal to MILD combustion in a laminar coflow burner was investigated both experimentally and numerically. The effect of hydrogen addition to the fuel (methane) on flame stabilization and NO formation was studied. In general, a very good agreement was obtained between numerical predictions and experimental measurements. In the last phase of this project, the numerical model is further developed for the simulation of turbulent MILD combustion.

To study the stability of flames in turbulent MILD combustion, we perform a numerical study of H2-enriched flames on the Delft Jet-in-Hot Coflow (DJHC) burner which is shown in Fig.1.

Figure 1: Schematic of Delft Jet-in-Hot Coflow (DJHC) burner of Oldenhof et al.

This burner mimics the conditions of MILD combustion in which a fuel jet is ignited by being issued into a coflow of air mixed with hot burned gases. The base fuel in the experiments is Dutch Natural Gas (DNG) and very recently it has been

mixed with various amounts of H2. It has been observed by Roekaerts and coworkers that addition of H2 has a large effect on the flame structure and stabilization mechanism of the lifted turbulent flame. This numerical study reports on the investigation of preferential diffusion effects in autoignition of H2 containing fuels. These effects are implemented in the FGM technique for LES of MILD combustion. For this purpose, a flamelet-based combustion model has been developed based on non-unity Lewis mixing layers for LES of the turbulent igniting CH4/H2 flames in a hot environment. Various amounts of H2 ranging from 0 to 25 percent of the fuel volume is added to the base fuel and a significant change in lift-off height and stability of the flames is observed in agreement with the experimentally observed trend. The goal of this research is not to provide a comprehensive validation of all cases against experimental data (which is not available) but to illustrate effect of preferential diffusion in complex interactions of mixing and kinetic on the flame's stability.

The LES has been performed by taking into account variances of controlling variables that have been computed by presumed beta-PDF approach. Turbulent inflow conditions are generated using a random noise generator.

Figure 2: Comparison of computed radial profiles of mean streamwise velocity at heights Z = 15, 60 and 90 (solid lines) against measurements (open symbols) for Case 00H2. Bottom figure shows centerline RMS values of streamwise and spanwise velocity and centerline turbulent kinetic energy.

A comparison of computed mean stream wise velocities against the measurements is shown in

12


Figure 2. It is seen that the mean velocity field agrees very well with measurements. Figure 3 shows a comparison of the computed and measured RMS values of streamwise and spanwise velocity and the resulting turbulent kinetic energy. It is clear that application of the random noise generator successfully reproduces the experimental inflow turbulent fluctuations and accordingly the mixing field.

Figure 3: Comparison of computed centerline RMS values of streamwise and spanwise velocity and turbulent kinetic energy (solid lines) against measurements (open symbols) for Case 00H2.

The temperature field has been computed by application of the developed LES-FGM-PDF model with non-unity Lewis combustion model. Instantaneous snapshots of the temperature field have been shown in Fig. 4. In these snapshots formation of ignition kernels can be observed. Such an observation corresponds to the experimentally observed ignition kernels by flame chemiluminescence.

Figure 4: Computed instantaneous distributions of temperature field using FGM-LES-PDF model with Non-unity Lewis combustion model for Case DJHC-00H2. These snapshots show the localized temperature rise corresponding to formation of ignition kernels.

Figure 5 displays Favre-averaged distribution of predicted OH mass fraction for all studied cases with unity Lewis numbers (top figures) and non-unity Lewis numbers combustion model. It can be seen that average quantities of OH increase significantly by addition of hydrogen which may affect stabilization mechanisms of these flames. Inclusion of preferential diffusion in the combustion model affects the stabilization and lift-off height of the predicted flames significantly especially for DJHC-05H2 and DJHC-10H2.

By comparison with the most probable flame luminescence lift-off height (red lines in Fig. 5), it is indicated that hydrogen-enriched cases require inclusion of preferential diffusion effects in the combustion model for an accurate prediction of lift-off height especially for cases DJHC-05H2 and DJHC-10H2. Quantitative agreement requires further improvement of the LES-FGM model, in particular the inclusion of heat loss.

Figure 5: Favre-averaged distribution of computed OH field for different cases (from left to right DJHC-00H2, DJHC-05H2, DJHC-10H2 and DJHC-25H2). Computations of FGM-LES-PDF model with unity Lewis numbers (top) and non-unity Lewis numbers (bottom). Red lines correspond to the heights where 50% probability of flame luminescence has been measured.

ACKNOWLEDGEMENT

The authors would like to thank STW, TATA Steel and the NVV for financial support of this project under the CCC program.

13


MoST: MULTI-SCALE MODIFICATION OF SWIRLING COMBUSTION FOR OPTIMIZED GAS

TURBINES (10425)

A.A. Verbeek1, T. Cardoso de Souza2, T.H.W.M. Bouten1, G.G.M. Stoffels1, R.J.M. Bastiaans2, B.J. Geurts3,4, Th. Van der Meer1, and L.P.H. de Goey2

1Laboratory of Thermal Engineering, University of Twente, P.O.Box 217, 7500 AE Enschede, The Netherlands 2Mechanical Engineering, Eindhoven University of Technology, P.O. Box 513 5600 MB Eindhoven, The Netherlands

3MMS, Applied Mathematics, Faculty EEMCS, PO Box 217, 7500AE Enschede, The Netherlands 4Anisotropic Turbulence, Dept. of Appl. Phys., Eindhoven Univ. of Tech., P.O. Box 513, 5600MB Eindhoven, The Netherlands

Key words: premixed combustion, low swirl burner, fractal grids, large-scale modulated turbulence

A clean and efficient way of generating heat and momentum by combustion is to operate in the lean premixed regime. While lean combustion results in low NOx emissions, low CO and low un-burnt hydrocarbons, it also reduces the stability of the combustion. A newly invented stabilization mechanism, the low swirl burner (LSB), is very promising in further reduction of NOx emissions. The combustion in such burners is very much dependent on the turbulence of the upstream flow. In the MoST-project different methods are investigated which either generate turbulence efficiently or generate specific turbulence that is efficient for combustion.

Using fractals to enhance Low Swirl Combustion

Fractal grids [1] provide an efficient way of generating turbulence. The multi-scale geometry creates a flow field of which the turbulence is more intense and decays in a slower fashion compared with regular grids. It is demonstrated that these grids can be used to increase the flame surface density (FSD) of the low NOx LSB.

The standard central blocking grid of a LSB, with perforations in a hexagonal pattern, is replaced with a circular cropped fractal grid of the cross-family. This grid contains an iterative pattern of crosses as illustrated by fig.1. The level of fractality is defined by the parameter = +1/ .The blockage is kept constant between the two classes of disks. While the mean flow profile is not influenced much when changing the central blocking disk, the level of turbulent fluctuations is enhanced significantly. Fig 2

shows as function of the downstream distance. Using a fractal grid, i.e., < 1, results in an increase of roughly a factor of two compared to the classical case. The grid with = 1, which is not fractal, does not show an increase in .

The effect of the turbulence on the flame is assessed by means of planar OH-LIF images post-processed with an edge-detection algorithm. In fig 3 the resulting FSD is shown for the central part of the flame, which is affected most by the turbulence generated by the grid. The maximum FSD is increased only 15%, but the flame brush thickness increased as well. The averaged consumption speed [2] is increased up to 95% for the central region and is a function of . No adverse effect on the emissions was observed. For both cases low NO and

Figure 1: (a) Description of fractal cross pattern. (b) Circular cropped grid for use in combination with LSB.

Figure 2: Description of fractal grid

Figure 3: Flame surface density of the flame corresponding to original LSB (left) and LSB with fractal grid (right).

14


CO emissions were measured that were in agreement with literature [3]. The use of multi-scale grids enables one to efficiently increase the power density of a low swirl flame, which may lead to a better applicability of LSB in gas turbine application, where such high power density is desirable.

Modulated Turbulence for Premixed Combustion

As a different strategy for turbulence enhancement in premixed combustion, in the project MoST numerical experiments focusing on the use of spatially-modulated turbulence were additionally considered.

The effects of a spatial time-modulated forcing acting on a stoichiometric turbulent Bunsen flame were investigated using direct numerical simulations of the Navier-Stokes equations coupled with a tabulated chemistry technique, namely the flamelet generated manifold (FGM). The premixed flame was agitated in space and time by a hybrid forcing consisting on filtered small-scale random perturbations and a coherent large-scale spatially periodic modulation imposed at the inflow plane. The forcing in this case consisted on a set of coherent vortical structures, with specific wave-number, periodically distributed over the periodic x-direction at the inflow. In these simulations, the large-scale modulation, defined as a Beltrami flow type, has action limited to the region where the cold unburnt mixture emerges. Fig 4 shows an example of the spatial pattern associated with the large-scale modulation imposed.

Figure 4: 3D view of the rectangular domain showing 2D contours of the modulation streamwise velocity component for the case K=4/Lx. Color lines correspond to the downstream trajectory evolution of particles initialized at different inflow locations. The dashed lines in the inflow denote the cold mixture region.

A parametric variation using different length-scales, i.e., wave-numbers, and distinct forcing amplitudes was considered to investigate the flame response. Fig 5 provides a qualitative view of the instantaneous effects that the imposed modulations have on the flame front.

Figure 5: 3D snapshots view of the flame front after 8 flow-through times colored with vorticity in the z-direction, for different modulation wave-numbers: a) K=24/Lx, b) K=4/Lx.

Next the flame response is obtained in terms of the time-averaged total flame surface density, FSD. Fig 5 show results of FSD obtained for several cases.

Figure 5: a) time-averaged FSD, , as a function of the spatial modes K.L/ (circle symbols) b) time-averaged FSD as a function of the normalized modulation amplitude, A/A0, where A0 is the amplitude of the forcing used for the cases shown in a). In b) the symbols correspond to the cases K=2/L (square symbols), K=24/L (square symbols) and the unmodulated flame (triangle symbols)

Fig.5 (a) shows no significant maximum in the global flame response compared to the unmodulated case, i.e., K.L/=0, despite of the broad range of modulation scales used to force the flame. However, results in Fig.5 (b) show that the flame response becomes selective to the size of the length-scales imposed when the amplitude of the inflow modulation is sufficiently large. It can be observed, for instance an increase up to 20% in the time-averaged FSD when the flame is agitated by a set of modulation scales with wave-number K=2/Lx and twice the amplitude of cases shown in Fig.5 (a).

ACKNOWLEDGEMENT The authors would like to thanks STW/AnsaldoThomassen for sponsoring this project. REFERENCE [1] D. Hurst, J.C. Vassilicos, Phys. Fluids 19 (2007) [2] J. Driscoll, Prog. Energ. Combust. 34 (2008) [3] M.R. Johnson et al, Proc Combust. Inst. 30 (2005)

15


XCiDE: CROSSING THE COMBUSTION MODES IN DIESEL ENGINES MODELLING LOW TEMPERATURE COMBUSTION WITH THE FGM METHOD

EFFICIENT FUELS FOR FUTURE ENGINES

U. Egz, C.A.J. Leermakers, L.M.T. Somers, N.J. Dam, L.P.H. de Goey, B.H. Johansson Combustion Technology Group, Eindhoven University of Technology, Den Dolech 2 5612AZ, Eindhoven, The Netherlands

Key words: Low temperature combustion, tabulated chemistry, ignition delay, fuels, partially premixed combustion, laser diagnostics.

Apart from the presumable challenge of future crude oil shortage, internal combustion engines are associated with problems regarding local emissions, like Nitrogen Oxides (NOx) and Particulate Matter (PM), as well as carbon dioxide (CO2), which is generally thought to have a global impact. For local emissions many solutions exist. However, these solutions come with either additional investment costs, a fuel consumption penalty (higher in-use costs), or both. Therefore, often a choice has to made between local or global emissions. To avoid this trade-off, a simultaneous reduction is preferred.

The reduction of carbon dioxide in road transport depends, starting upstream, on the CO2 footprint of the fuel, the efficiency of energy conversion, and the energy used (i.e. vehicle efficiency). The CO2 footprint of the fuel depends on the energy used in producing and refining. Both bio-fuels and crude-based fuels can reduce the impact significantly, depending on production methods, and the upgrading necessary. Furthermore, the fuel can act as an enabler for certain high efficiency combustion regimes. The combustion regime can have an intrinsic benefit in the energy conversion efficiency. Both the combustion regime and the fuels impact the local emissions and, whichever has lower PM or NOx emissions can save fuel and CO2 emissions, by avoiding the formerly mentioned trade-off.

In the experimental part of this project [1], both various alternative combustion concepts are investigated, as well as a variety of fuels (or fuel combinations) to use in either conventional or these alternative combustion concepts.

As a first investigation, a short term solution for using Premixed Charge Compression Ignition combustion was researched [2]. In such a short term scenario, engine hardware and fuels would be conventional, and the combustion concept would be made possible by a smart choice of operating conditions. The impact of the composition of conventional diesel fuels in this short term scenario has also been looked at [3]. In a somewhat longer

term scenario, fuel reactivity can be tailored to the demands of the combustion concept, of which a first investigation has also been shown in [3]

Besides tailoring a single fuels reactivity, other researchers have shown remarkable results using the Reactivity Controlled Compression Ignition concept. This concept uses in-cylinder blending two fuels with different reactivity to a desired value. This concept has been evaluated briefly first [4] after which also the controllability, required for real-life use, has been shown [5].

Commercially available naphtha blends can also help to make the transition from fully mixing controlled combustion (as in a Diesel engine) to more premixed regimes [6]. The application of such naphtha blends as well as a number of n-butanol/diesel blends, used in Partially Premixed Combustion mode, has been presented [7]. Such bio-fuels (be it blended with Diesel fuel) can enable the further use of aforementioned concepts. Given an optimum reactivity for the latter concept, a number of fuels have been blended both from refinery streams and from surrogate fuel components. The differences between these components is presented in [8].

Further improvements of the combustion regimes described earlier rely on detailed knowledge and understanding of the underlying processes. High speed LIF techniques for detection of both the CH [9] and the OH radical have been developed, and an attempt has been made to apply these techniques to both conventional and advanced combustion regimes, to gain the aforementioned understanding.

Because the application of the two former techniques has proven to be limited under engine conditions, a third technique for studying early PAH and soot formation has been developed. While this technique does not work at high-speed, it still provided a significant addition to the understanding of both the effects of operating conditions and fuel effects.

16


The operation of a conventional diesel engine suffers from the trade-off between NOx and soot emissions. In general, a decrease in NOx emissions is achieved at the expense of higher particulate emissions and vice versa. New combustion concepts target a simultaneous reduction of both emissions. Low temperature combustion (LTC) emphasize obtaining a more homogeneous mixture and limiting local in-cylinder temperatures. During the moment of injection, the in-cylinder conditions are low to trigger auto-ignition. Combustion starts after the end of injection and the process is primarily governed by chemical kinetics, not so much by diffusive mixing anymore. As there is no more overlap between combustion and injection, no direct control over ignition is possible, which is a common issue for new combustion concepts. Dual fuel Reactivity Controlled Compression Ignition (RCCI) combustion concept is introduced to have a better control over combustion phasing. Generally, port injection of a high octane fuel is combined with early direct injection of a low octane fuel. RCCI combustion targets to control combustion by tuning fuel blends and injection strategies.

A direct application of detailed reaction mechanisms within the CFD framework for complex systems like engines is not practical because of the number of additional equations and associated length and time scales. Therefore several reduction methods exist for the reacting flows. In this work, the Flamelet Generated Manifold (FGM) method is applied to model combustion chemistry. The first part of the FGM application is creating a representative manifold. For this, relevant canonical systems need to be identified. Here, Homogeneous Reactors (HR) are applied to store all thermo chemical properties in a look-up table which is characterized by the mixture fraction and a reaction progress variable.

The simulations in the HR canonical system are performed at constant pressure. However, the pressure inside the cylinder during an engine operation is not constant but it changes during the (compression/expansion) strokes. Therefore, a set of pressure levels needs to be defined. It is observed that LTC is extremely sensitive to the chosen pressure levels due to the broad pressure range, long duration between injection and ignition events, and the dual fuel operation. Especially the selection of high temperature/pressure levels is very critical. Once the correct pressure levels are set, it is possible to predict ignition phasing accurately with the FGM method.

ACKNOWLEDGEMENT

This project is funded by the Dutch Technology Foundation STW, which is the applied science division of NWO, and the Technology Programme of the Ministry of Economic Affairs. The project number is 10417: Crossing the Combustion modes in Diesel Engines (XCiDE). DAF Trucks N.V., Shell Global Solutions, Avantium Technologies B.V. and Delphi are also acknowledged for their contributions to the project. REFERENCES [1] Leermakers, C.A.J., (2014) Efficient fuels for future

engines,PhD thesis, Eindhoven University of Technology.

[2] Leermakers, et al., (2013) Int. J. of Vehicle Design 62(1):1-20.

[3] Leermakers, et al., (2011) SAE Technical Papers 2011-01-1351

[4] Leermakers, et al., (2011) SAE Technical Papers 2011-01-2437

[5] Leermakers, et al., (2012) SAE Technical Paper 2012-01-1575

[6] Leermakers, et al., (2013) SAE Int. J. of Fuels and Lubr. 6(1):2013-01-1681

[7] Leermakers, et al., (2013) SAE Int. J. of Fuels and Lubr. 6(1):2013-01-1683.

[8] Nijssen, et al., (2013) Experimental investigation of low octane fuel composition effects on load range capacity in Partially Premixed Combustion (PPC), Proceedings of the 6th European Combustion Meeting, Lund Sweden.

[9] Leermakers, et al., Cinematographic imaging of methylidine (CH) radicals by planar laser-induced fluorescence for combustion applications, submitted to Appl. Phys. B: Laser and Optics.

17


ALTAS Project: Advanced Low NOx Flexible Fuel Gas Turbine Combustion, Aero and Stationary

A. Donini, S. Mukhopadhyay, R.J.M. Bastiaans, J.A. van Oijen and L.P.H. de Goey Combustion Technology Group, Eindhoven University of Technology, Postal Address P.O. Box 513 5600 MB Eindhoven.

Key words: Gas turbine combustor, fuel flexibility, fuel composition, NOx reduction, turbulent combustion, Computational Fluid Dynamics (CFD), Large-Eddy simulation (LES), Reynolds Averaged Navier-Stokes (RANS), chemical reduction techniques, Flamelet-Generated Manifolds (FGM), combined combustion modes, combustor design.

Gas turbine combustion is the most important energy conversion method in the world today. Using gas turbines, large scale, low emission energy production is possible. Nitrogen oxide emissions are one of the most important technology drivers for combustion systems today. For land based engines, low NOx emissions can be achieved by very lean premixed combustion (dry low NOx). For most aero applications, gas turbines are the only option to achieve the required thrust. Conventional combustion systems are based on diffusion mode combustion, but to be able to meet emission requirements, the next generation of aero gas turbines will be based on lean premixed combustion technology. LPP (lean pre-vaporized premixed) combustion is not possible, due to the short ignition delay times of kerosene, but premixing can be achieved by suitable fuel injection systems (e.g. LDI: lean direct injection).

It is expected that fuel composition will change in the future. This is due to depletion and independency of oil reserves and the greenhouse effect of burning fossil fuels. New fuels will be generated from biomass, landfill, waste and coal. Coal gasification with CO2 sequestration looks very promising. This results in gaseous fuels with high hydrogen and carbon monoxide content with various fractions of diluents. Liquid fuels for aero applications will contain larger fractions of synthetic fuel, such as generated through the Fischer-Tropsch synthesis.

In the field of stationary land based engines, Siemens Power Generation (SPG) is very active and would like to increase their turbulent combustion knowledge to predict engine performance. This is mainly in connection with techniques to include high levels of hydrogen in the fuel. Also, Rolls-Royce faces challenges with their gas turbines for aero-planes in relation to new synthetic fuels and combustion concepts. Both companies are interested in carrying out research together with TU/e to improve their predictive capability.

In the current project detailed knowledge for modeling of combustion with alternative fuels will be developed. This is done by means of detailed descriptions in the framework of Computational Fluid Dynamics (CFD). The ultimate goal is to predict the combustion process of gas turbines, including complex physical real fuel phenomena (temperature-traverse, NOx, preferential diffusion, thermo-diffusive effects, soot, ignition, extinction etc.). This requires a method in which different combustion modes can be captured, different fuels can be covered and extinction, ignition, heat loss and slow chemistry effects can be included. To that end the promising technique of flamelet-generated manifolds (FGM) will be extended in this project. The technique is developed in its basic form at TU/e and has been continuously tested and extended to more general situations over the last years.

Figure 1: Siemens stationary gas turbine.

The current project is heavily supported by Siemens Power Generation, SPG, and Rolls-Royce Deutschland, RRD. The vision of these companies is that a joint project performed in collaboration with the Combustion Technology group of Eindhoven University of Technology, can create added value to their individual competitiveness in CFD based gas

18


turbine design. It has to be mentioned that this is a unique development dictated by the challenging changes with which gas turbine developers are confronted in the near future. The benefit in a competitive situation of a joint project becomes significant with the existence of common fundamental problems, which can be tackled within a unified framework. The methods developed by the Combustion Technology group of TU/e will provide such an approach.

Currently at Siemens Power Generation, SPG, Reynolds Averaged Navier-Stokes (RANS) based CFD is used with simple combustion models. One of the aims of the current project is to enhance the CFD capabilities to support the design of future combustors. It is in the interest of SPG to gain the knowledge to be able to use CFD as a tool to optimize designs as well as to obtain fundamental knowledge to arrive at possible solutions for performance and emission issues. Here different types of RANS and LES (Large-eddy simulation) techniques become an important item. Since SPG is particularly interested in combustion of hydrogen enriched fuels, this issue will guide the research. Although the fuel itself is somewhat simpler, complex effects arise due to the high diffusivity of hydrogen, resulting in a more complex combustion behavior. The goal is to develop models that are capable of representing these effects at high temperatures and pressures. For SPG, methods will be developed for commercial CFD codes, e.g. CFX (RANS and LES versions).

At Rolls-Royce the research is oriented towards the implementation and application of more advanced chemistry models for complex combustion concepts. The interaction between turbulence and chemistry will be studied in more detail using a partially premixed fuel and air stream configuration, in which extinction and re-ignition occurs. Such a configuration requires very advanced modeling approaches. The reduction of complex fuel kinetics (such as for kerosene) adds another dimension to the problem. Can the main effects of the fuel be represented in CFD (by a model fuel) and if yes, would it be possible to predict slow chemical processes associated to final emissions at the exhaust level?

The current project is embedded in a long term, large scale research effort that is ongoing in the combustion group of TU/e.. Many other universities and research institutes meanwhile follow these developments. Also several industrial partners, both international and Dutch, are involved. All these

partners will benefit (and already did) from the developments towards a more general applicability of the method. This is a real challenge and requires a large investigation effort.

ACKNOWLEDGEMENT The authors would like to thank STW, Siemens Power Generation (SPG), and Rolls-Royce Deutschland (RRD) for sponsoring this project.

19

Combustion of torrefied beech wood in new spiral combustion reactor Eyerusalem M.Gucho,Eddy A.Bramer, Gerrit Brem

Department of Energy Technology, Faculty of Engineering Technology, University of Twente, Drienerlolaan 5, 7500 AE Enschede, The Netherlands

Biomass fuel has a great potential in reducing the greenhouse gas emissions during electricity generation. One way to integrate biomass in electricity production is via co-firing it with coal in the current coal fired power plants. The main advantage of biomass co-firing is that it can be used in the existing power plants with little or no alteration. This makes the biomass co-firing a relative economical technology which can be made operative in a short term. However, some of the properties of the biomass hampers its utilization at high cofiring rate. Torrefaction, which is a thermal pre-treatment of a biomass helps to produce a material more similar to coal.

Several experimental techniques have been used to study the combustion characteristics of solid fuels like biomass or coal. Thermogravimetric analyser (TGA) , which is commonly used technique to study the reactivity, kinetics of the fuel using mass change as indicator. However, the heating rate is quite low to compare to actual boiler conditions. In contrast, drop tube furnace (DTF) with high heating rate close to actual boiler condition is used to study the ignition and burnout characteristics of solid fuel. However in DTF, it is difficult to measure the residence time of the particle experimentally. Most of the time CFD modelling are used to determine the residence time of the particle or assuming as the gas residence time, which leads to some errors. These days some optical access on the DTF are used to visualize the particles. However, this optical access location are fixed to one place not the whole reactor makes difficult to follow the particle for the whole journey during its decomposition.

In this work, a new experimental reactor with spiral shape was designed that facilitates the visual observation of whole life of single particle combustion at various process conditions. This paper will present the preliminary experimental results from the new spiral reactor designed in our Thermal Engineering laboratory in University of Twente.

1. Experimental section

1.1. Material properties

For this investigation severe beech wood torrefied at torrefaction temperature between 513-623 K and residence time of 10 min were considered.

1.2. Experimental setup

A new spiral combustion reactor was developed to study visually the volatile and char ignition, and burnout behaviour of solid fuels. The reactor is made from quartz glass that facilitates the visual combustion study. The spiral shape makes the setup compact and easy to access during maintenance. The spiral reactor has a total length of 3.5 m in which the first 1.5 m from the outside diameter used as the gas preheater.The gas flow can be adjusted in different range depending on the requirement of the study. As it is shown in Fig.2 the setup consists of different parts, fuel injection with pressurized gas that is activated by electric signal from the computer. For visual observation, high speed camera with specification frame rate of up to 1000 frame/sec and with image quality of 460 x 460 pixel is used. The heating rate of the particle is ~1000 oC/s.

20

Figure.1 Schematic diagram and top view of Spiral combustion reactor

A single-particle experiment of (STBW) with particle diameter of 500-600 m was carried out in reactor temperature range of 500-950 oC. The experiment helps us to determine the effect of temperature on the ignition and combustion time. The ignition time is the time between the time at which the particle mixed with the hot gas till flash of the particle happens. The ignition time is the parameter which describes the reactivity of the fuel. The ignition time of the fuel depends on many process parameter like process temperature, initial volatile matter and gas composition. On other hand, the combustion time is the time between the flash occurred and till the flash disappears. In case of lower temperature the combustion time could not been seen as the residence time was not enough to be fully converted. The longer the combustion time , the char burnout efficiency in actual case will be low and vice versa for short combustion time.

1. Result and discussion

The statistical distribution is formulated from 50 test at each temperature. As it is shown in Fig2.(a) as expected the ignition delay moves towards to shorter time as the gas temperature increases to 950 oC. For temperatures between 500 to 600 oC the ignition time distribution decreases sharply and becomes narrow , which is the boundary temperature region for the ignition to commence. The ignition time also has shown small standard deviation compared to combustion time as ignition mainly depends on the availability of oxygen and the gas temperature around.

21

Biomass pyrolysis in DNS of turbulent particle-laden flow E. Russo1, J.G.M. Kuerten 1,2, and B.J. Geurts1,2 1 Eindhoven University of Technology, Eindhoven, The Netherlands, 2 Faculty EEMCS, University of Twente, Enschede, The Netherlands

Biomass is important for co-firing in coal power plants thereby reducing CO2 emissions. Modeling the combustion of biomass involves various physical and chemical processes, which take place successively and even simultaneously [1, 2]. An important step in biomass combustion is pyrolysis, in which virgin biomass is converted into char. The simplified model for biomass pyrolysis provided by Haseli [3] is used to write a Matlab code that solves the equations of a single biomass particle immersed in a gas of uniform temperature. Next, this model is adapted and implemented in a compressible code that was developed in our group. The code simulates the turbulent flow of gas in a channel with heated walls, using two-way coupling of mass, momentum and energy. We do not model combustion and gasification of biomass, but focus on pyrolysis, in particular on the effect of particle-gas interaction on the conversion time, i.e. the time needed to convert biomass into char. This is the first attempt in modeling the pyrolysis of biomass in a 3D flow framework. The gas of the channel flow consists of Nitrogen. The compressible Navier-Stokes equations are solved with a second-order accurate finite volume method. Because of the two-way coupling, some extra terms are present in the Navier-Stokes equations. In this way the contribution of a particle to the gas equations is added to the control volume in which the particle is present. The particle-gas interaction does not change the total mass, momentum and energy of the system. Integration in time is performed using the same four-stage second-order accurate Runge-Kutta method for the gas phase and the particles. The channel geometry has a size of 4H in streamwise direction and 2H in the spanwise direction, where H is half the channel height. The mesh is uniform in the periodic directions (streamwise and spanwise) and clustering is adopted near the walls. At the walls the no-slip condition is adopted and the temperature is kept fixed. For the particles, periodic conditions are applied in the homogeneous flow directions. If a particle approaches a wall within a distance of its radius, an elastic collision without heat transfer is applied. Simulations are performed at Reynolds number Re = 150, based on friction velocity and half the channel height. The flow is initialized by a fully developed turbulent velocity and temperature field obtained from a previous simulation without particles. The initial particle distribution is random and homogeneous. The initial particle velocity is equal to the gas velocity at the particle location. The initial particle temperature is 300 K.

By means of the numerical simulation method an analysis of the dependency of the conversion time on the particle volume fraction and the particle size has been performed. We found that gas-particle interaction affects the conversion time, which shows a characteristic dependence on particle size and concentration as shown in the results presented in the following.

We performed simulations with different numbers of particles in order to analyze the pyrolysis conversion time as a function of the particle volume fraction. The diameter of all particles was 0.7 mm. The initial gas temperature was 1400 K. The particle Stokes number changes from 40 to 5 during pyrolysis. The results show that the heating-up time increases with increasing volume fraction. The two-way particle-gas interaction affects the gas temperature because of the convective heat exchange: more particles extract more heat from the gas compared to simulations with fewer particles. This results in a slower particle heating up and increases the mean particle pyrolysis conversion time, as shown in Figure 1.

In order to investigate the effect of particle size on the conversion time, we performed simulations with constant volume fraction and varying particle diameter. The volume fraction (2.55x10-4) is in

22

the range where particle collisions can be ignored but two-way coupling is relevant. The initial gas temperature in these simulations was taken 1200 K. Without two-way coupling the gas temperature would remain constant and the conversion time would increase linearly with the diameter of the particles. Due to two-way coupling, the gas temperature is time dependent and the particle conversion time shows a more than linear dependence on particle diameter, see Figure 2.

Figure 1 Conversion time for different volume Figure 2 Conversion time versus particle fraction with particle diameter 0.7 mm. diameter at constant volume

fraction.

In practical applications of co-firing of biomass and coal, the probability density function of biomass particle diameter follows a Rosin-Rammler distribution. In Figure 3 the distribution of torrefied beech wood is shown. We used this size distribution but truncated at 3 mm. For larger particles the point-particle assumption is no longer valid. It can be observed that only few particles are larger than 3 mm. Very small particles require a very small time step in our explicit time-integration method. Therefore, we also bounded the size distribution from below. We performed four simulations in which the cut-off of the particle size pdf was set at 0.5, 0.3, 0.2 and 0.15 mm in order to investigate whether it is possible to disregard very small particles without noticeably influencing the results. The volume fraction of biomass is 5.3*10-5.

In the absence of very small particles the pyrolysis and therefore the particle mass loss starts somewhat later, as Figure 4 shows for an initial gas temperature of 1200 K. The results show that a cut-off at 0.2 mm or less hardly influences the results. The volume fraction used is in the one-way coupling range, where the coupling terms in the momentum equation can be ignored. The results in Figure 5 demonstrate that the same holds for the coupling terms in the gas temperature equation, as the differences between one-way coupled and two-way coupled simulations are small. The two-way coupling becomes important at higher volume fractions. If we consider the gas to be oxygen with the same properties, the volume fraction needed for stoichiometric combustion is more than four times larger than the one when the gas is air. However, in our simulations the oxygen does not react because we do not have a combustion model. In such a case, for a volume fraction of 3.7x10-4 the comparison between one-way and two-way coupling in Figure 6 shows a noticeable difference. In the future we will add a model for coal particles and a combustion model for the particles and the volatiles as well in order to investigate and optimize the co-firing.

23

Figure 3 Rosin-Rammler size distribution Figure 4 Particle mass versus time at of torrefied beech wood. different cut-off values.

Figure 5 Particle mass loss: 1-way and 2-way Figure 6 Particle mass loss: 1-way and 2-way coupling for volume fraction ~6x10-5 coupling for volume fraction

~3.7x10-4

References [1] Haseli, Y., van Oijen, J.A., de Goey, L.P.H.: Modeling biomass particle pyrolysis with temperature-dependent heat of reactions. Journal of Analytical and Applied Pyrolysis 90, 140154 (2011) [2] Haseli, Y., van Oijen, J.A., de Goey, L.P.H.: Numerical study of the conversion time of single pyrolyzing biomass particles at high heating conditions. Chemical Engineering Journal 169, 299312 (2011) [3] Haseli, Y., van Oijen, J.A., de Goey, L.P.H.: Predicting the pyrolysis of single biomass particles based on a time and space integral method. Journal of Analytical and Applied Pyrolysis 96, 126138 (2012)

24


ULRICO: ULTRA RICH COMBUSTION OF HYDROCARBONS AND SOOT FORMATION

SOOT MEASUREMENTS IN A PREMIXED REACTOR

M.H.F.Woolderink1, J.B.W.Kok1, M.Stllinger2 and D.J.E.M. Roekaerts2

1Laboratory of Thermal Engineering, University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands

2Department of Process and Energy, Delft University of Technology,

Mekelweg, 2, 2628 CD Delft, The Netherlands [email protected]

Key words: Ultra rich combustion, natural gas, syngas, soot, experiments

Introduction Ultra rich combustion (partial oxidation) of natural gas is a process applied to produce synthesis gas (or syngas). This gas is composed primarily of hydrogen and carbon monoxide. Syngas represents the intermediary step from hydrocarbons to bulk chemicals and synthetic fuels. The reactor design and the operating conditions have to ensure a high conversion of natural gas to hydrogen and carbon monoxide. In addition to this, the hydrogen to carbon monoxide ratio in the syngas is relevant for the downstream application of the syngas produced. The syngas soot content is also of concern, in view of the syngas fouling the reactor system and to minimize the downstream effort of soot removal. The operating conditions for the large scale application of the partial oxidation process are characterized by turbulent flow and a high fuel to oxidizer ratio (rich combustion), far beyond the stoichiometric ratio. The goal of this research project is twofold: development of computational models that can be used as a design tool for the partial oxidation reactor and the development and use of a reliable measurement system to quantify relevant soot properties such as the size distribution. Experimental setup In order to validate the numerical results, measurements are performed in a reactor operating at pressures from 1 to 6 bar and at

equivalence ratios ranging from 2 to 4. To investigate the gas composition and soot production during ultra-rich combustion conditions, an experimental setup is available. This setup has been made suitable to collect in-situ flame samples continuously to examine the local chemical composition as well as the soot content and properties. The flame samples are chemically quenched and diluted with a custom designed and built sample system. With this system it is possible to decrease the temperature and pressure of the flame samples from 1400 C and 6 bar to atmospheric pressure and room temperature. The sample can be diluted with a factor of 10.000. This is necessary because the soot concentration in the flame can significantly exceed the upper concentration limit of the Scanning Mobility Particle Sizer (SMPS) spectrometer that is used to measure the soot properties. The gas composition is measured before the dilution step with a by-pass that leads to a gas chromatograph. Results Experiments were performed where different parameters were varied to investigate their influence on the soot particle size distribution and concentration. The varied parameters were equivalence ratio, oxidizer composition and pressure. In all experiments the soot particle concentration was in the order of 107 109 particles/cm3. Increasing the equivalence ratio leads to higher soot particle concentrations and larger mean particle diameters. These effects are

25


also seen when the oxygen concentration in the oxidizer is decreased from 40% to 21%. All measured soot concentration profiles have a log-normal distribution. An interesting observation was that the logarithmic standard deviation has approximately the same value for all measurements. Conclusion The gas composition and soot properties measurements give valuable validation data for the performance of the combustion and soot formation models at ultra rich conditions that are not yet available in the literature. The discovery that the logarithmic standard deviation of the log-normal soot concentration distribution has approximately a constant value simplifies the numerical modeling of the soot formation significantly. With a constant logarithmic standard deviation, it is possible to describe the particle size distribution by only solving the transport equations for the soot mass fraction and the soot number density. From these, the mean particle diameter and the total particle concentration can be calculated. These two variables define the position and the height of the log-normal distribution. The width of the log-normal distribution is constant because of the constant logarithmic standard deviation. ACKNOWLEDGMENT This project is supported by Technology Foundation STW and Shell.

26

Abstracts of presentations

27

Combura 2013 symposium, Vaeshartelt, Maastricht, The Netherlands, October 9-10, 2013

Application of model predictive control for steel heating furnaces

P. Pronk, R. Speets, H. Wu Tata Steel, Research & Development, P.O.Box 10.000, 1970 CA IJmuiden, The Netherlands

Key words: Furnaces, modelling, control

Tata Steel is a top ten global steel maker with a total steel making capacity of 27 million tones. IJmuiden in the Netherlands is one of its major production sites, which annually produces 7 million tonnes of flat steel products. These flat products, also referred to as steel strip, are produced from slabs with a thickness of 225 millimetres. Via hot and cold rolling processes, the slabs thickness is reduced to create a strip with a gauge ranging from 0.2 to 10 millimetres depending on customer requirements.

The strip production process contains two major heating steps, namely reheating prior to hot rolling and annealing after cold rolling. Reheating furnaces are gas-fired and heat up slabs to a temperature of approximately 1200C, while the slabs are in direct contact with the combustion process. Annealing furnaces heat up cold-rolled strip to temperatures between 600 and 800C. The strip is indirectly heated to avoid oxidation of the steel strip. In continuous annealing furnaces, this indirect heating is realised via radiant tubes (see Figure 1). These tubes transfer the heat released by combustion inside the tubes to the steel strip outside the tubes by means of radiation. A continuous annealing line consists of multiple radiant tubes through which strip travels in a protective atmosphere of hydrogen and nitrogen as shown in Figure 2.

Figure 1: Radiant tube with recuperator

Most steel heating furnaces are currently controlled via conventional PID control. Each product, either slab or strip, has a target temperature evolution throughout the furnace. The actual product

temperature obtained via a measurement or a model is compared to this target temperature. If necessary, the furnace temperature and/or the product speed are adjusted to meet the product temperature target. The furnace temperatures are subsequently controlled via atmosphere temperature measurements and the fuel supply to the burners.

Strip chargeStrip discharge

Figure 2: Continuous annealing line

There is a strong drive to increase the production capacity and energy efficiency of these heating furnaces, while (temperature) tolerances get tighter and product batches become smaller resulting in bigger and more frequent transitions from strip to strip. These requirements make the conventional control systems inadequate, since these do not take furnace dynamics into account, do not anticipate on batch transitions, and have no interaction between speed and furnace temperature control. Tata Steel sees model predictive control based on state-space descriptions, as a control method that does meet the control requirements of today.

Model predictive control has recently successfully been implemented in continuous annealing line #12 at the Tata Steels site in IJmuiden. The challenge of this line is to heat cold-rolled strip to temperatures between 600 and 700C with a tolerance of 10C. Due to strip transitions and enforced speed changes, the conventional control has difficulties with meeting these targets. The recent implementation of an in-house developed model predictive control system has increased the on-spec length from 75 to 98%, while increasing the production capacity with 3%, and reducing the natural gas consumption with 6% and operator intervention to minimum.

In the near future, the model predictive control system will also be implemented at Tata Steels other continuous annealing lines. Next to that, a similar model predictive control system will be also developed for and implemented at reheating furnaces.

29

Pulse combustor applied in integrated paper sludge conversion process with drying and sequential combustionby PhD P.J. Ploumen PLOUMEN ENERGY CONSULTANCY and M.Sc M.A. van Praag PULSED HEAT B.V.

Abstract of presentation COMBURA 2013 on 9 & 10 October 2013 Maastricht

PulsedHeat BV has developed applications based on pulse combustion. With pulse combustion not all the chemical energy in the fuel is converted to heat but a part (to 10%) will be converted to kinetic energy of the flue gas molecules. The amplitude of the pressure wave can reach values of 0.2 bar. The frequency of the wave depends on the geometry of the burner, air valve and tailpipe and the composition of the fuel but will be in the range of 180 to 200 Hz.

Due to the very frequent chances in velocity and direction of the molecules the combustion of the dried sludge in the combuster will be much more intensive as has been proven in Russian boilers with pulsating burners using coal. With pulse combustion the drying process is much more intensive, which will result in a smaller dryer. This is also demonstrated in practice.

In the first approach of Pulsed Heat an integrated concept was realised in which the drying process was integrated in the combustion process providing the heat for the drying process. This is illustrated in figure 1, where all the main components are put in line and are placed in a container. Tests have shown that the integrated concept works.

Figure 1: Pilot plant installed at IWE Eerbeek

With subsidary of EFRO and with a contribution of the province Gelderland, the CalciPulse project was started, with the aim to design and built a demo installation to handle all the paper sludge of Industrie Water Eerbeek (IWE). In this project Pulsed Heat is working together with the TU/e, IWE and the manufacturer RhineTech to realise the project. Ploumen (P.E.C.) is assisting Pulsed Heat in optimizing the process. Based on literature, old drawings and some test results of a previous model of the pulsating burner, carried out by TU/e, a parametric model was built and a redesign of the pulsating burner was realized. The new burner is build. The burner is a combination of two burners, the first one is a pulsating burner, with an aerovalve and a tailpipe. The consequence is that the operation range of this burner is limited. In the second burner, the possibility to reduce the amount of

30

air is added. The capacity of the second burner, also indicated as booster, will be 10 times the capacity of the first burner. In the first burner the pulsating flow will be created. With test already carried out with pulse combustor and booster the impression was, that the booster got the same frequency and amplitude. This will be checked with additional tests to be carried out in September 2013.

To maximize the benefit of the pulsating flow, the concept of the pilot plant was changed for the demo installation. All the necessary main equipment was put in line, to avoid the reduction of pulsating flow in pipes and benches. Also a new design for the converter was realised based on data of Awerbuch, and a new type of dryer where the sludge will fall in the dryer like a curtain and will met the pulsating hot flow. Pulsed Heat has taken a patent on the design of the dryer. With CFD modelling the TU/e will analyse the design of the converter and the dryer in the period September to December 2013. At the TU/e also some prediction of CO and NOX emissions of pulse combustors are carried out.

The design of the demo plant is illustrated in figure 2. With IWE a modification of the feeding system of the sludge will be carried out. It turned out the behaviour of the mixed sludge and the secundary sludge are so different that two separate feedings systems are necessary to supply the sludge to the feeder of the dryer.Another complication is the large amount of Calcium Carbonate in the sludge. In the converter the Calcium Carbonate can react if the combustion temperature is above 840 C. If this occurs, some heat is required for the exothermal reaction, which heat will not be available for the drying process.

Figure 2: Design demo installation

In the presentation more details will be given of the results of the measurement carried out in September 2013.It is the aim to build the demo installation in the last quarter of this year. The ultimate project goal is to develop a broad range of driers and combustors which make it possible to convert several materials locally to more value materials without the large transportation costs and without the disadvantage of the general burning in centralized municipal solid waste incineraters.

31

Combura13 symposium, Kasteel Vaeshartelt, Maastricht, the Netherlands, October 9&10, 2013

Towards HiTAC conditions in a 9MW oil fired industrial boiler B.C.H. Venneker, J.N.A. Koomen, M.P. Morskate, W. Jansen, M.A.F. Derksen

Stork Technical Services Process Equipment, P.O.Box 33, 7500AA Hengelo, the Netherlands

High Temperature Air Combustion (HiTAC) is a combustion process in which fuels are burned at high air temperatures with reduced oxygen-concentrations. The principle of HiTAC lends itself ideally for application in boilers and especially for all kinds of difficult fuels, ranging from low-calorific gases to heavy fuel-oils. For heavy fuel-oils, the expectations are that in combination with HiTAC, these can be utilized for steam generation with very low harmful emissions such as NOx, CO, and particulates. If in the design of boilers the differences in heat release between conventional combustion and HiTAC are truly accounted for, it is even foreseen that the resulting optimized boiler design will be cheaper and more reliable than the present standard. As a first step taken on the route towards an optimized HiTAC boiler, some tests were performed with a Stork Double Register Burner (DRB) with heavy fuel-oil a

Date post:	04-Mar-2016
Category:	Documents
Upload:	erin-walker
View:	21 times
Download:	0 times

tekatesfayemengesha

Documents