Brandbrief 2015 Clean Combustion Concepts

Transcription

1 Brandbrief 2015 Clean Combustion Concepts Nr 21 Schone en Zuinige Verbranding februari 2015 Technologiestichting STW

2 Colofon Brandbrief De brandbrief, met een oplage van 450 exemplaren, is een uitgave van het Platform van het STWtechnologieprogramma Schone en Zuinige Verbranding (SZV). De Brandbrief bericht over lopende zaken in het programma, vorderingen van onderzoek en ander nieuws van participerende bedrijven en universi-teiten De uitvoering is in handen van Technologiestichting STW. Redactie Prof.dr.ir. Th.H. van der Meer Universiteit Twente Dr.L.J. Korstanje Technologiestichting STW A.M. van der Stroom Technologiestichting STW Voor een exemplaar van de Brandbrief kunt u zich aanmelden bij het programmabureau. Dit is de laatste uitgave van De Brandbrief. Programmabureau Schone en Zuinige Verbranding Postadres Technologiestichting STW Linda de Groot Schone en Zuinige Verbranding Postbus GA UTRECHT Internet clean-combustion-concepts Concept en uitgave Technologiestichting STW, Utrecht Ontwerp Room for ID s, Nieuwegein Realisatie Argante Argante Amsterdam Fotografie Betrokken instellingen Druk Repro-afdeling FOM/STW-bureau Niets uit deze uitgave mag worden overgenomen of vermenigvuldigd zonder uitdrukkelijke toestemming van de redactie. STW-nummer 2015/00765/STW ISBN Samenstelling Platform Schone en Zuinige Verbranding Prof.dr.ir. Th.H. van der Meer, voorzitter Universiteit Twente Prof.dr.ir. R.S.G. Baert TNO Automotive Dr.ir. M.F.G. Cremers DNV GL Energy Prof.dr. L.P.H. de Goey Technische Universiteit Eindhoven Ir. B. Hakstege DAF Trucks NV Dr.ir. J.H.A. Kiel ECN Dr.ir. W. de Jong Technische Universiteit Delft Prof.dr. H.B. Levinsky DNV GL Energy Dr.ir. L. Post Shell Global Solutions International Dr.ir. P. Pronk Tata Steel Dr.ir. C.J.A. Pulles KIWA Technology Prof.dr. D.J.E.M. Roekaerts Technische Universiteit Delft Ir. J.N.A. Koomen Stork Thermeq Dr. L.J. Korstanje, secretaris Technologiestichting STW 2 Brandbrief STW 2015

3 Brandbrief 2015 Nr 21 Schone en Zuinige Verbranding februari 2015 Technologiestichting STW Schone en zuinige verbranding 3

4 Inhoud 06 Voorwoord 07 1 / Projects 08 MILDNOX: Fuel flexibility and NO formation in dilute combustion 12 BIOxyFuel: Torrefied Biomass Combustion under Oxy-fuel Conditions in Coal Fired Power Plants 14 XCiDE: Crossing the Combustion modes in Diesel Engines 17 HiTAC: Heavy Fuel-Oil Combustion in a HiTAC Boiler 20 ULRICO: Ultra Rich Combustion of Natural Gas to Syngas 23 MoST: Multi-scale modification of swirling combustion for optimized gas turbines 26 ALTAS: Advanced Low NO x Flexible Fuel Gas Turbine Combustion, Aero and Stationary 28 flexflox: Flameless combustion conditions and efficiency improvement of single- and multiburner-flox TM furnaces in relation to changes in fuel and oxidizer composition 31 2 / Promotions 32 Dr. L. Zhou 30 September Dr. S. Ayyaoureddi 9 January Dr. P.G.M. Hoeijmakers 28 January Dr. M. Shahi 24 September Ir. N. Speelman September Brandbrief STW 2015

5 Schone en zuinige verbranding 5

6 Voorwoord Beste Brandbrief lezer, Met gemengde gevoelens presenteren we hier Brandbrief no. 21. Met een tevreden en trots gevoel presenteren we in deze Brandbrief de resultaten van de 8 projecten die intussen zijn afgerond binnen het Perspectief programma Clean Combustion Concepts. Met de resultaten van deze projecten zijn stappen gezet naar schonere en zuinigere verbrandingstechnologiën, waar de betrokken industriële deelnemers mee verder kunnen. Nu het CCC programma succesvol is afgesloten komt ook het Platform Schone en Zuinige Verbranding tot een einde. Een platform waarbinnen gedurende de afgelopen 18 jaar de universitaire groepen met een aantal belangrijke bedrijven en instituten het verbrandingsonderzoek in Nederland hebben afgestemd. We kunnen met een tevreden gevoel terugkijken op een zeer vruchtbare samenwerking binnen het platform. Aan de andere kant is het natuurlijk jammer dat aan dit platform nu een einde komt. Voor de toekomst zijn we echter bijzonder positief gestemd, omdat de functie van het platform SZV vanaf nu wordt overgenomen door de Nederlandse Vereniging voor Vlamonderzoek, de NVV. Ook de jaarlijkse nationale conferentie COMBURA zal in de toekomst door de NVV worden georganiseerd. We zijn ervan overtuigd dat hiermee de overlegstructuur van de academische groepen en de contacten met het bedrijfsleven van deze groepen gewaarborgd zijn. Het is nog niet duidelijk wat de toekomst zal zijn van deze Brandbrief. Zeker is dat dit de laatste aflevering is in deze vorm. Theo van der Meer (Voorzitter Platform Schone en Zuinige Verbranding) Philip de Goey (Voorzitter Programmacommissie Clean Combustion Concepts) 6 Brandbrief STW 2015

7 1 Projects Schone en zuinige verbranding 7

8 MILDNOX: Fuel flexibility and NO formation in dilute combustion NO species using Laser-Induced Fluorescence. The mathematical description of the model is governed by a set of conservation equations for mass, momentum, energy and species in the cylindrical coordinate. The GRI Mech. 3.0 chemical mechanism is used to obtain the required thermodynamic and transport data involved. Mixture-Averaged transport is used to calculate diffusion velocities of each species. Radiation effects were also added to calculation using optically-thin approximation. Projectleaders: prof.dr. H.B. Levinsky, prof.dr. L.P.H. de Goey, dr. A.V. Mohkov, dr.ir. J.A. van Oijen 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 NO x emissions, and offer the possibility of substantially homogenizing the temperature field in furnaces. To permit the optimization of NO x control, and to provide insight into the ultimate low-no x potential of these methods, in the proposed research we investigate the paths to NO formation in dilute, high temperature combustion. Towards this end, we have performed quantitative laser diagnostic measurements of flame structure, using Raman and LIF in the laminar coflow geometry (see Fig. 1), combined with detailed numerical simulations of the structure of the reaction zone (see Fig. 2). An important part of this research is the analysis of the preheating and dilution of the fuel and/or oxidizer on spatial structure and NO formation. The predictive power of the detailed simulations made using the GRI-Mech 3.0 chemical mechanism is tested by comparison of the measured and calculated distributions of temperature and major species fractions. Laminar diffusion flames with different degree of preheating of the coflow and fuel were studied. The structures of a normal non-preheated diffusion flame (Case NP) and a MILD flame with preheated and diluted reactants (Case M) are compared here. Calculated temperature distributions of these flames are shown in Fig. 2. The flame temperature and major species (CO, CO 2, N 2, H 2, H 2 O, CH 4 and O 2 ) were also measured using spontaneous Raman scattering and Measurements of the in-house developed diffusion burner are compared against computations and a good agreement was found for major species and temperature (see Fig 3). NO concentration obtained by Laser Induced Fluorescence is compared with computations as shown in Fig. 4. It is seen that amount of NO is predicted with a reasonable accuracy (see Fig 4). Additionally, measurements have been performed using a LJHC burner for Case M. In this burner the diluted oxidizer coflow is generated by a lean premixed ceramic burner. In essence, this geometry is a diluted laminar jet in hot coflow. Computations of this flame have also been performed using detailed chemistry of GRI 3.0 and Mixture-Averaged transport. Comparison of computations and measurements of temperature for this flame is shown in Fig. 5 at three different heights above the fuel jet exit. The mild increase in temperature in the mixing layer (~200 K) is indicative of MILD combustion under these conditions. NO concentrations of this burner are also compared with computations (see Fig. 6). It is seen that NO fractions are below 10 ppm in which majority of the NO is formed in the coflow. To study flame stabilization of this combustion regime, we perform numerical study of the the H 2 -enriched Delft Jet-in-Hot Coflow (DJHC) burner which is shown in Fig.7. This burner mimics conditions of Mild combustion in which a fuel jet is ignited due to being issued into hot burned gases of coflow. Base fuel in the experiments is Dutch Natural Gas (DNG) and very recently it has been mixed with various amounts of H 2. It has been observed that addition of H 2 has a significant effect on the flame structure and stabilization mechanism of the lifted turbulent non-premixed flame. The present study also reports on the numerical investigation of preferential diffusion effects in autoignition of H 2 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 CH 4 /H 2 flames in a hot environment. Various amounts of H 2 ranging from 0 to 25 percent of fuel volume have been added to the base fuel and a significant change in lift-off height and stability of 8 Brandbrief STW 2015

9 Measurement of a MILD flame in the laminar jet-inhot-coflow (LJHC) burner. 2 Temperature computations of (left) Case NP and (right) Case M. 3 The measured (symbols) and calculated temperature (lines) in Case NP at three different heights as a function of the radial distance. 4 The measured (symbols) and calculated NO in Case NP at three different heights as a function of the radial distance. 5 The measured (symbols) and calculated temperature in Case M at three different heights as a function of the radial distance. 6 The measured (symbols) and calculated NO in Case M at three different heights as a function of the radial distance. Schone en zuinige verbranding 9

10 the flames has been observed. 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. Comparison of computed mean streamwise velocities against measurements is shown in Fig.8 for case DJHC-00H2 (pure DNG). It can be seen that the mean velocity field agrees very well with the measurements. Fig. 9 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 applica- tion of the random noise generator has been successfully reproduced the experimental inflow turbulent fluctuations. The temperature field has been computed by application of the developed LES-FGM-PDF model employing non-unity Lewis numbers. Instantaneous snapshots of the temperature field are shown in Fig. 10. In these snapshots the formation of ignition kernels can be observed. This observation corresponds to experimentally observed ignition kernels. A comparison of Favre-averaged distributions of predicted OH mass fraction for all studied cases with unity Lewis and non-unity Lewis numbers (not shown here) indicates that the concentration of OH increase significantly by addition of hydrogen. 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 (5 and 10% H2 addition, respectively). By comparison with the most probable flame Schematic of Delft Jet-in- Hot Coflow (DJHC) burner of Oldenhof et al. 8 Comparison of computed radial profiles of mean streamwise velocity at heights Z = 15, 60 and 90 (solid lines) against measurements (open symbols) for Case 00H Comparison of computed centerline RMS values of streamwise and spanwise velocity and turbulent kinetic energy (solid lines) against measurements (symbols) for Case 00H2. 10 Brandbrief STW 2015

11 10 10 Computed instantaneous distributions of temperature field using FGM-LES-PDF model with non-unity Lewis numbers for Case DJHC-00H2. These snapshots show the localized temperature rise corresponding to formation of ignition kernels. luminescence line, 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. Concluding Remarks We have developed an efficient and reliable numerical model to predict MILD combustion of natural gas and hydrogen mixtures. With this model, new furnaces for the high-temperature process industry can be developed. In real furnaces, however, the conditions might be quite different from those in the JHC burners. For instance, Reynolds numbers might be larger which can increase turbulence intensities. Furthermore, entrainment of burned gas into the fuel stream shifts most reactive mixture fraction toward the fuel stream. In this condition, turbulent structures have a larger impact on ignition events resulting in an increased role of turbulence transport with respect to molecular diffusion. In the future research, experimental and numerical investigations of these conditions are indispensable in order to move toward more practical situations. Publications [1] [2] [3] [4] [5] [6] S.E. Abtahizadeh, PhD Thesis Numerical study of Mild combustion from laminar flames to Large Eddy Simulation of turbulent flames with Flamelet Generated Manifolds, Eindhoven University of Technology (2014). Advisers: Prof. Dr. Philip de Goey and Dr. Ir. Jeroen van Oijen Sepman, A., Abtahizadeh, E., Mokhov, A., van Oijen, J., Levinsky, H., de Goey, P. Experimental and numerical studies of the effects of hydrogen addition on the structure of a laminar methane nitrogen jet in hot coflow under MILD conditions. Int. J. Hydrogen Energy, Vol. 38, (2013). Abtahizadeh, S.E., Sepman, A.V., Hernandez-Perez, F.E., van Oijen, J., Mokhov, A.V., de Goey, P. and Levinsky, H.B. Numerical and experimental investigations on the influence of preheating and dilution on transition of laminar coflow diffusion flames to Mild combustion regime. Combust. Flame Vol. 160, (2013). Abtahizadeh, S.E., Oijen, J.A. van, Goey, L.P.H. de (2012). Numerical study of Mild combustion with entrainment of burned gas into oxidizer and/or fuel streams. Combustion and Flame, 159(6), Sepman, A.V., Abtahizadeh, S.E., Mokhov, A.V., Levinsky, H.B. and de Goey, P. Numerical and experimental studies of the NO formation in laminar coflow diffusion flames on their transition to MILD combustion regime. Combust. Flame. Vol. 160, (2013). Mokhov, A.V., Smirnov, B.M., Dutka, M., Vainchtein, D., Levinsky, H.B. and De Hosson, J. Th.M. Formation of chain aggregates in external electric field. Chem. Phys. Lett. Vol. 570, (2013). [7] [8] [9] Sepman, A.V., Toro, V., Mokhov, A.V. and Levinsky, H.B. Determination of temperature and concentrations of main components in flames by fitting measured Raman spectra. J. Appl. Phys. B, Vol. 112, (2013). Sepman, A.V., Mokhov, A.V., and Levinsky, H.B. Spatial structure and NO formation of a laminar methane nitrogen jet in hot coflow under MILD conditions: A spontaneous Raman and LIF study. Fuel, 103, pp (2013). Sepman, A.V., Mokhov, A.V. and Levinsky H.B. The effects of the hydrogen addition on the HCN profiles in fuel-rich-premixed, burner-stabilized C1-C3 alkane flames. Int. J. Hydrogen Energy, vol. 36, no. 21, pp (2011). [10] Sepman, A.V., Mokhov, A.V. and Levinsky H.B. Extending the predictions of chemical mechanisms for hydrogen combustion: comparison of predicted and measured flame temperatures in burner-stabilized 1D flames. Int. J. Hydrogen Energy, Vol. 36, pp (2011). [11] Sepman, A.V., Mokhov, A.V., and Levinsky, H.B. The effects of hydrogen addition on NO formation in atmospheric-pressure, fuel-rich-premixed, burner-stabilized methane, ethane and propane flames. Int. J. Hydrogen Acknowledgement: The authors would like to thank STW for sponsoring this project under the CCC program project number: Schone en zuinige verbranding 11

12 BIOxyFuel: Torrefied Biomass Combustion under Oxy-fuel Conditions in Coal Fired Power Plants Research method The scientific results of the research program are the development of new and unique experimental facilities in The Netherlands, advanced mathematical models for the key processes at different scales (particle, flow, furnace), and model validation using data from full-scale plants. Results In the experimental phase the reactivity, burnout, and emissions of raw and torrefied biomass are measured. In this way more insight has been obtained in the dominating combustion mechanisms of coal/biomass-mixtures under both air-blown and oxyfuel conditions. Projectleaders: prof.dr.ir. G. Brem, dr. C.W.M. van der Geld, prof.dr.ir. B.J. Geurts, prof.dr. L.P.H. de Goey, prof.dr. J.G.M. Kuerten, prof.dr.ir. Th.H. van der Meer, dr.ir. J.A. van Oijen The objective of the BIOxyFuel project was to increase understanding and predictive capabilities of torrefied biomass combustion at high co-firing rates under oxy-fuel conditions in coal fired power plants. The combination of biomass co-firing and oxy-fuel power plants will have a double effect on the reduction of CO 2. In fact, the combination of oxy-fuel combustion and biomass could be used as a sink for CO 2. The research program is carried out for different types of biomass and torrefied biomass as from an economic perspective fuel flexibility is essential because of fluctuating availability and prices of the different biomass streams. Industrial partners involved in the project were NVV, KEMA, and TSA (electricity power companies). Parallel to the experimental work different models have been developed, ranging from a single particle model, to a particle-laden turbulent flow model and a furnace or full-scale model. The single particle biomass combustion model has been validated using experimental findings and integrated in the particle-laden turbulent flow simulations under pyrolysis conditions. These models have been integrated in a furnace model, that is validated against experimental data obtained from measurement campaigns at a full-scale power plant while co-firing biomass. The BIOxyFuel project has contributed to the CCC programme by developing a new combustion concept, reducing unwanted emissions (CO 2, NO x ), giving insight in a higher fuel flexibility (different types of torrefied biomass/coal mixtures), improving the potential use of sustainable fuels (biomass) Brandbrief STW 2015

13 2 3 Publications [1] [2] [3] [4] [5] [6] Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2013). A quasisteady analysis of oxy-fuel combustion of a wood char particle. Combustion Science and Technology, 185(4), in Web of Science Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2013). Reduced model for combustion of a small biomass particle at high operating temperatures. Bioresource Technology, 131, ; in Web of Science Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2012). Predicting the pyrolysis of single biomass particles based on a time and space integral method. Journal of Analytical and Applied Pyrolysis, 96(July), ; in Web of Science Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2012). A simplified pyrolysis model of a biomass particle based on infinitesimally thin reaction front approximation. Energy & Fuels, 26(6), ; in Web of Science Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2012). Analytical solutions for prediction of the ignition time of wood particles based on a time and space integral method. Thermochimica Acta, 548, 65-75; in Web of Science Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2011). A detailed one-dimensional model of combustion of a woody biomass particle. Bioresource Technology, 102(20), ; in Web of Science [8] [9] Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2011). Numerical study of the conversion time of single pyrolyzing biomass particles at high heating conditions. Chemical Engineering Journal, 169(1-3), ; in Web of Science E. Russo, J.G.M. Kuerten, B.J. Geurts, Delay of biomass pyrolysis by gas-particle interaction, J. Anal. Appl. Pyrolysis, 110, (2014) [10] E. Russo, J.G.M. Kuerten, B.J. Geurts. C.W.M. van der Geld, Water droplet condensation and evaporation in turbulent channel flow, J. Fluid Mech., 749, (2014) [11] E.M.Gucho, E.A.Bramer and G.Brem, Experimental studies of torrefied biomass co-firing with coal in drop tube furnace, June 06, 2011, 19th European Biomass Conference and Exhibition, Berlin [12] E.M.Gucho, K.Shazhad, E.A.Bramer and G.Brem, Parametric study on the torrefaction of beech wood and miscanthus for co-firing application., ToTeM 37, September 2011, Technical University of Wroclaw, Poland [13] E.M.Gucho, E.A.Bramer and G.Brem, Áir and oxyfuel combustion of torrefied biomass in new spiral combustion reactor, 3rd Oxyfuel Combustion Conference, 9-13 September 2013, Ponferrada, Spain [7] Haseli, Y., Oijen, J.A. van & Goey, L.P.H. de (2011). Modeling biomass particle pyrolysis with temperature-dependent heat of reactions. Journal of Analytical and Applied Pyrolysis, 90(2), ; in Web of Science Schone en zuinige verbranding 13

14 XCiDE: Crossing the Combustion modes in Diesel Engines Projectleaders: dr.ir. L.M.T. Somers, dr. N.J. Dam, prof.dr. L.P.H. de Goey Ever increasing demands from legislation forces OEM s of HD Diesel engines to apply EGR (Exhaust Gas Recirculation) often in combination with after treatment systems (SCR, DPF). This will induce a so-called fuel penalty and increases the cost of the powertrain. Hence the active research in intrinsically clean combustion concepts that apply a more premixed type of combustion (HCCI, PCCI). Unequivocally these concepts try to create a more or less homogeneous charge but still rely on auto-ignition. To achieve this, injection of fuel has to be separated from the ignition event allowing ample mixing time. Principally separation can be obtained by lowering the temperature and/or the reactivity of the fuel. The goal of the project is to RCCI Determine the load range for PCCI with conventional diesel fuels Similar but with alternative (high-octane) fuels Develop a combustion model in a CFD setting that naturally takes the fuel reactivity into account Build a knowledge base on new combustion concepts Research method To adequately determine concept boundaries a special engine set-up is used. It is based on a HD diesel engine which is adapted such that one-cylinder is separated from the rest. The test-cylinder has a separate intake and exhaust allowing for flexible setting of temperature, boost pressure and EGR percentage (up to 75%). The fuel injection equipment applies a modern common-rail system able to deliver pressures up to 3500 bar (mostly limited by the injector) and is freely programmable. All exhaust emission are measured. In figure 1 a photograph of the engine is depicted. The numerical approach is based on the efficient FGM methodology. This methodology is a chemical reduction method that is based on a so-called flamelet approach in combination with a tabulation method. In this project the method is extended for modelling engine combustion, including features like ignition for diffusive and pre-mixed combustion of large alkane (i.e. automotive) fuels. Results First experiments on PCCI applying regular diesel fuel revealed that the applicable load range is small and introduces a fuel penalty. The compression needs to be lowered and high levels of EGR are inevitable to reach the necessary separation between injection and ignition. This is mainly due to the low resistance of diesel fuel, as it is designed to be, against auto-ignition. A promising path to a PCCI concept is proven to be a change in fuel reactivity. Two prevailing implementations to reach that currently exist: RCCI and PPC. RCCI is a dual-fuel concept and PPC applies a single fuel (blend) with a moderate but much higher octane number than diesel (see figure 2). Dual-fuel PPC In contrast to normal dual fuel applications in RCCI the low octane fuel (diesel) is injected early to ensure separation between ignition and injection. Initial experiments show the potential of the approach for a reduction of NO x by one order of magnitude compared to regular diesel at similar EGR levels (figure 3). In fact the soot and NO x emissions Single-fuel The single cylinder HD diesel engine at the TU/e (Cyclops). 2 Hi-octane PCCI concepts. 14 Brandbrief STW 2015

15 are below EUROVI limits without a DPF and after treatment system. The fuel economy has improved by nearly 10% which shows its potential towards CO 2 reduction. The RCCI concept also has its drawbacks mainly due to trapped fuel in crevices and overleaning. This results in relatively high CO and UHC (Unburned HydroCarbons) emissions compared to conventional diesel combustion (CDC). The PPC concepts does not suffer from high UHC and CO emissions because of the fact the fuel can be targeted better towards the piston bowl. It was found that the specific composition of the fuel blend is not really important but the performance is mainly determined by the Fuel octane number. A such even Naphtha fuels and low injection pressures can be used as shown in figure 4. Consequently cheaper injection equipment can be applied and the demand for high grade fuels can be minimized. Note that this on itself can lead to CO 2 reduction at the production side of the fuel as well. As these concepts heavily rely details of the mixing process and combustion computational fluid dynamics (CFD) will be a necessary tool to optimize these concepts in relation to the fuel composition. As fuel details are important chemical kinetic schemes need to be incorporated in an efficient way. In this project the FGM methodology is extended to CDC and RCCI/PPC combustion. To validate the approach it has been extensively compared to the detailed database of the Engine Combustion Network ( 3 Specific NOx emissions (logarithmic scale!) at medium load at different timings. The diamonds are results from an RCCI experiment applying 90% gasoline/10% diesel. 4 Effect of fuel pressure on soot emissions. Three naphtha blends in the PPC concept compared to regular diesel in a CDC concept. 5 Prediction of 50% heat release (CA50) as a function of gasoline percentage in RCCI experiment. 6 CO mass-fraction of a PCCI case running on regular diesel depicting highest levels in the crevice region. After that it has been applied to a RCCI study with varying gasoline content (figure 5). The results are promising as the method predicts the same trend as the experim-ents show. The big advantage of the numerical app-roach is that it is now possible to investigate where exactly the emissions are created. Detailed information as presented in figure 6 can pinpoint exact measures to deal with the specific problems of the new concepts. The program has come to the conclusion that these new concepts pro-vide a way to develop clean and efficient engines. A larger follow-up of the project is currently formulated in which DAF, TNO and Shell will participate actively. Clearly the optimization of the fuel together with engine technology offers a win-win situation to reduce the carbon footprint of the transport sector considerably. Schone en zuinige verbranding 15

16 Journal Publications [1] [2] [3] [4] [5] [6] [7] [8] [9] Egüz, U., Leermakers, C.A.J., Somers, L.M.T. & Goey, L.P.H. de (2014). Modeling of PCCI combustion with FGM tabulated chemistry. Fuel, 118, Egüz, U., Ayyapureddi, S., Bekdemir, C., Somers, L.M.T. & Goey, L.P.H. de (2013). Manifold resolution study of the FGM method for an igniting diesel spray. Fuel, 113, Egüz, U., Maes, N.C.J., Leermakers, C.A.J., Somers, L.M.T. & Goey, L.P.H. de (2013). Predicting auto-ignition characteristics of RCCI combustion using a multi-zone model. International Journal of Automotive Technology, 14(5), Egüz, U., Leermakers, C.A.J., Somers, L.M.T. & Goey, L.P.H. de (2013). Premixed charge compression ignition combustion modeling with a multi-zone approach including inter-zonal mixing. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 227(9), Egüz, U., Ayyapureddi, S., Bekdemir, C., Somers, L.M.T. & Goey, L.P.H. de (2012). Modeling fuel spray auto-ignition using the FGM approach: effect of tabulation method. SAE International Journal of Engines: Egüz, U., Somers, L.M.T., Leermakers, C.A.J. & Goey, L.P.H. de (2011). Multizone modelling of PCCI combustion. International Journal of Vehicle Design, 55(1), U. Egüz, L.M.T. Somers, C.A.J. Leermakers, L.P.H. de Goey, Multi-zone modelling of PCCI combustion, Int. J. of Vehicle Design, 55(1), 76-90, (2011) C.A.J. Leermakers, M.P.B. Musculus, In-cylinder soot precursor growth in a low-temperature combustion diesel engine: Laser-induced fluorescence of polycyclic aromatic hydrocarbons, Proceedings of the Combustion Institute, Available online 19 July 2014, ISSN , org/ /j.proci C.A.J. Leermakers, P.C. Bakker, B.C.W. Nijssen, L.M.T. Somers, B.H. Johansson, Low octane fuel composition effects on the load range capability of partially premixed combustion, Fuel, Volume 135, 1 November 2014, Pages , ISSN , fuel [10] Leermakers, C.A.J., Luijten, C.C.M., Somers, L.M.T., Goey, L.P.H. de & Albrecht, B.A. (2013). Experimental study on the impact of operating conditions on PCCI combustion. International Journal of Vehicle Design, 62(1), [11] Leermakers, C.A.J., Bakker, P.C., Somers, L.M.T., Goey, L.P.H. de & Johansson, B.H. (2013). Commercial Naphtha blends for partially premixed combustion. SAE International Journal of Fuels and Lubricants, 6(1): [12] Leermakers, C.A.J., Bakker, P.C., Somers, L.M.T., Goey, L.P.H. de & Johansson, B.H. (2013). Butanol-diesel blends for partially premixed combustion. SAE International Journal of Fuels and Lubricants, 6(1): [13] Leermakers, C.A.J., Somers, L.M.T. & Johansson, B.H. (2012). Combustion phasing controllability with dual fuel injection timings. SAE International Journal of Engines, 2012: [14] Leermakers, C.A.J., Luijten, C.C.M., Somers, L.M.T., Kalghatgi, G.T. & Albrecht, B.A. (2011). Experimental study of fuel composition impact on PCCI combustion in a heavy-duty diesel engine. SAE International Journal of Engines, /20. [15] Leermakers, C.A.J., Berge, B. van den, Luijten, C.C.M., Somers, L.M.T., Goey, L.P.H. de & Albrecht, B.A. (2011). Gasoline diesel dual fuel : effect of injection timing and fuel balance. SAE International Journal of Engines, 4(3): [16] C.A.J. Leermakers, C.C.M. Luijten, L.M.T. Somers, G.T. Kalghatgi, B.A. Albrecht, Experimental Study of Fuel Composition Impact on PCCI Combustion in a Heavy-Duty Diesel Engine, SAE Technical Papers, -, , (2011) [17] M.D. Boot, C.C.M. Luijten, L.M.T. Somers, U. Egüz, D.D.T.M. van Erp, B.A. Albrecht and R.S.G. Baert, Uncooled EGR as a Means of Limiting Wall-Wetting under Early DI Conditions, SAE Technical Papers, 2009, [18] M.D. Boot, C.C.M. Luijten, L.M.T. Somers, U. Egüz, D.D.T.M. van Erp, B.A. Albrecht, R.S.G. Baert, Uncooled EGR as a Means of Limiting Wall-Wetting under Early Direct Injection Conditions, in Homogeneous Charge Compression Ignition Engines, 2009; Editors: SAE, / , SAE International, Book Chapter ISBN (2009) Conference papers [19] U. Egüz, C.A.J. Leermakers, L.M.T. Somers and L.P.H. de Goey (2011), Preliminary analysis of soot and UHC emissions under PCCI conditions, Proceedings of European Combustion Meeting (ECM2011), 28 June- 1 July 2011, Cardiff, Wales. [20] U. Egüz, C. Bekdemir, L.M.T. Somers and L.P.H. de Goey (2011), Study of PCCI modeling with the FGM approach, Proceedings of Towards Clean Diesel Engines (TCDE), 8-9 June 2011, Chester, United Kingdom. [21] U. Egüz and L.M.T. Somers (2011), Modeling of PCCI Combustion with FGM Approach, Oral Presentation, International Conference on Numerical Combustion (ICNC), April, 2011, Corfu, Greece. [22] U. Egüz, C.A.J. Leermakers, L.M.T. Somers and L.P.H. de Goey (2010) Multi-zone Modelling of PCCI Combustion with CFD Coupling for Stratification, Proceedings of Towards Sustainable Combustion (Speic2010), June 2010, Tenerife, Spain. Conference posters [23] L.M.T Somers, C.A.J. Leermakers and U. Egüz (2010), Crossing the Combustion Modes in Diesel Engines, Oral Presentation, Combura 2010, October 2010, Maastricht, the Netherlands. [24] C.A.J. Leermakers, B.A. Albrecht, L.M.T. Somers, C.C.M. Luijten, Euro VI without fuel penalty?, in 7th International Automotive Congress.nl; Helmond, Netherlands, Conference Poster (2010) [25] B. Berge, van den, C.A.J. Leermakers, L.M.T. Somers, C.C.M. Luijten, L.P.H. de Goey, Impact of fuels with lower reactivity on PCCI combustion in a heavy-duty engine, in Combura; Maastricht, Netherlands, Conference Poster (2010) [26] C.A.J. Leermakers, L.M.T. Somers, C.C.M. Luijten, L.P.H. de Goey, Euro VI without fuel penalty?, in Combura; Maastricht, Netherlands, Conference Poster (2010) [27] C.A.J. Leermakers, R.P.C. Zegers, L.M.T. Somers, C.C.M. Luijten, N.J. Dam, L.P.H. de Goey, High speed combustion imaging, in Combura; Maastricht, Netherlands, Conference Poster (2010) Acknowledgement: The authors would like to thank STW for sponsoring this project under the CCC program project number: and DAF, Delphi, Shell and Avantium for their contributions. 16 Brandbrief STW 2015

17 HiTAC: Heavy fuel-oil combustion in a HiTAC boiler specifically for this project. Numerical modeling of both the Delft laboratory scale flame as well as the Stork industrial test boiler were done at the University of Twente with the aim of coupling both experiments and understanding the underlying processes. Finally Stork developed water-steam cycles optimized for the application in combination with HiTAC combustion Projectleaders: prof.dr.ir. Th.H. van der Meer, prof.dr. D.J.E.M. Roekaerts, dr.ir. M.J. Tummers PhD s: S.L. Zhu, H.R. Correia Rodrigues The aim of this project was to improve the efficiency and reduce NOx and CO emissions of heavy fuel-oil combustion in industrial boilers by applying High Temperature Air Combustion (HiTAC). HiTAC relies on rapid dilution of fuel and combustion air with combustion products before the combustion reactions take place. In the case of liquid fuels this leads to the question whether the entrainment rate of an evaporating fuel spray can be high enough to reach sufficiently dilute conditions of the fuel. A very detailed experimental study was performed at Delft University of Technology of spray flames of light fuel-oils (ethanol and acetone) in hot-diluted co-flow conditions. In parallel field tests were performed at Stork Thermeq in a 9 MW test boiler with spray flames of heavy fuel-oil with hotdiluted combustion air. The fuel-oil for these experiments was provided and characterized by Shell Global solutions Research method A laboratory test burner was developed for a spray flame in hot diluted co-flow. Figure 1 shows an ethanol flame from this burner in a co-flow with a temperature of 1300 K and an oxygen concentration of 9.3%. Several numerical methods were used for detailed in-flame measurements, such as: high speed visualization of the liquid break-up process; Phase Doppler anemometry (PDA) for simultaneous measurements of droplet velocity and size statistics; coherent anti-stokes Raman spectroscopy (CARS) for gas-phase temperature statistics and laser Doppler anemometry for co-flow velocity measurements. A numerical model was developed within Ansys-Fluent with a pressure-swirl atomizer model including coalescence, secondary break-up and evaporation of the droplets, a laminar flamelet model for combustion, the discrete ordinate models for radiation and the k- model for turbulence. This model was first used to simulate a welldocumented spray flame from literature, the so-called NIST flame. Then the model was used for simulations of the Delft laboratory flame and finally for the Stork test boiler. For the last simulations the Eddy Dissipation model was used as the combustion model in stead of the flamelet model. 1 Ethanol spray flame in hot diluted co-flow. 1 Schone en zuinige verbranding 17

18 In the Stork 9 MW test facility experiments were conducted under conditions of elevated combustion air temperature, high amount of flue gas recirculation and fuel staging, aiming at more uniform temperature distributions throughout the combustion chamber. Conventional gas analysers were used to monitor NO x, CO and O 2 in the flue gases. Results When comparing conventional ethanol spray flames with ethanol spray flames in hot-diluted co-flow the results of this project show that for the flames in hot-diluted co-flows: 1. The mean flame temperatures are more uniform with less steep temperature gradients. With an oxygen concentration of only 6% the numerical simulations showed that the difference between peak temperature and co-flow temperature drops to about 200 K. 2. The maxima of the mean flame temperature are similar or considerably lower, depending on temperature and O 2 concentration of the co-flow. Also the NO x concentrations were reduced considerably. Figure 2 shows computed results of the peak mean temperature under various co-flow conditions. 3. The turbulent temperature fluctuations are much lower. The measurements showed maximal temperature fluctuations of 700 K in the conventional ethanol flame and maximal temperature fluctuations of 200 K in the flame with a co-flow temperature of 1200 K and an oxygen concentration of 9.2%. 4. From the experiments on the lab scale burner it was concluded that small droplets are quickly vaporized and the combustion process is mainly depending on the mixing of the fuel vapor with the entrained flow and the ignition delay time. Higher co-flow temperatures leads do a reduction of the flame lift-off height and an earlier formation of intermediate species leading to an increase of the peak temperature in lower axial position. From these observations it is clear that the conditions in the flames in a hot-diluted co-flow are close to HiTAC conditions. Although experimental restrictions did not make it possible to reach HiTAC conditions in the field tests with the 9 MW boiler, the results from the field test and the results from the numerical simulations on the same boiler are promising. The experiments showed that flue gas recirculation as well as fuel staging led to a decrease in NO x by 20%. The combination of flue gas recirculation and fuel staging decreased NO x by 30%. Simulation results in the test boiler showed that an increase of the temperature of the combustion air from 373 K to 746 K leads to a higher peak temperature in the combustion chamber from 2240 K to 2390 K. A reduction of the O2 concentration of the combustion air from wt% to wt% results in more uniform temperature distribution with a peak temperature of 1510 K. Further numerical investigation was done with recycling various ratios of flue gas into the primary and secondary air respectively for introducing various O 2 concentration conditions for the combustion air flow. The predicted temperature difference between the average 2 2 Model predictions of peak temperature as a function of co-flow conditions. 18 Brandbrief STW 2015

19 temperature and the peak temperature showed that the case with the lowest O 2 concentration in the primary air has the smallest temperature difference in the boiler. It was also shown that besides thermal NO x, fuel NO x is an important contributor to NO x formation in heavy fuel-oil combustion. By introducing flue gas recirculation, thermal NO x can be reduced to a very low level, leaving the fuel NO x 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 NO x emission was reduced to about one fourth. The results from this project are promising with respect to HiTAC combustion of heavy-fuel oil in boilers. The next step is to further develop this boiler concept by realizing optimal internal recirculation in the combustion chamber. For this reason Stork will use numerical simulations, which will be validated with experiments in a new test boiler, which is currently under development Publications [1] [2] [3] [4] S. Zhu, D.J.E.M. Roekaerts, and T.H. van der Meer, Numerical study of a methanol spray flame. 5th European Combustion Meeting, T. Griffiths (Ed.), Cardiff, UK, 2011, paper 067, 6 pages S. Zhu, D.J.E.M. Roekaerts, and T.H. van der Meer, Numerical simulation of a turbulent methanol spray flame using the Euler-Lagrange method and the steady laminar flamelet model. In Proceedings of the Mediterranean Combustion Symposium. Chia Laguna, Sardinia, Italy, 2011 H. Rodrigues, M.J. Tummers, D.J.E.M. Roekaerts, Experiments on turbulent ethanol reacting sprays in HiTAC conditions, 12th International Conference on Liquid Atomization and Spray Systems, Heidelberg, September, 2-6, 2012 S. Zhu, D.J.E.M. Roekaerts, A.K.Pozarlik, B. Venneker, T.H. van der Meer, Numerical investigation towards a HiTAC condition in a 9MW heavy fuel-oil boiler, 6th European Combustion Meeting, Lund, Sweden, 25-28th June, [9] Hugo Correia Rodrigues, Mark J. Tummers, Eric H. van Veen, Dirk J.E.M. Roekaerts, Spray flame structure in conventional and hot-diluted combustion regime, Combustion and Flame, [10] L. Ma, H.R. Correia Rodrigues, S. Zhu, M.J. Tummers, D.J.E.M. Roekaerts, Modelling of Delft Spray-in-Hot-Coflow flame with steady flamelet and FGM, in Book of Abstracts of the 23rd Biennial Meeting of the Belgian Section of the Combustion Institute, Brussels, May, 2014, 2 pages [11] Shanglong Zhu, Dirk Roekaerts, Artur Pozarlik, Theo van der Meer, Eulerian-Lagrngian RANS model simulations of the NIST turbulent methanol spray flame, submitted to Combustion Science and Technology [12] Shanglong Zhu, Dirk Roekaerts, Artur Pozarlik, Hugo Rodriguez, Theo van der Meer, Numerical investigation towards HiTAC conditions in laboratory-scale ethanol spray combustion, in preparation [5] L. Ma, S. Zhu, H.R.C. Rodrigues, M.J. Tummers, T.H. van der Meer and D.J.E.M. Roekaerts, Numerical investigation of ethanol spray-inhot-coflow flame using steady flamelet model, 8th Mediterranean Combustion Symposium, Çesme Izmir, Turkey, September 8-13, 2013, Paper EGTSC-13, 13 pages, Editors: Nevin Selcuk, Federico Beretta, Mohy S. Mansour, and Andrea d Anna. Publisher: International Centre For Heat and Mass Transfer, METU, Ankara, Turkey [13] Hugo Rodrigues, Spray combustion in moderate and intense low-oxygen conditions. An experimental study, PhD thesis, 2015, TU Delft. [6] H.R.C.Rodrigues, M.J. Tummers, D.J.E.M. Roekaerts, Turbulent Spray Combustion in hot-diluted co-flow, 9th Asia-Pacific Conference on Combustion, Gyeongju Hilton, Gyeongju, Korea, May 2013, 4 pp [7] H.R.C. Rodrigues, M.J. Tummers and D.J.E.M. Roekaerts, Turbulent Spray Combustion of ethanol and acetone flames in flameless conditions, In: European Combustion Meeting 2013 Paper P3-80, 6 pp, June 25-28, 2013, Lund, Sweden, ISBN [8] H. Correia Rodrigues, M.J. Tummers, E.H. van Veen, D.J.E.M. Roekaerts, Effects of coflow temperature and composition on ethanol spray flames in hot-diluted coflow, Int. J. Heat Fluid Flow, 2014, Schone en zuinige verbranding 19

20 ULRICO: Ultra Rich Combustion of Natural Gas to Syngas pressure and CFD modeling will be employed. The work is carried out by PhD student Marc Woolderink at the University of Twente and PostDoc Michael Stoellinger at the Technical University of Delft, supervised by Prof. Dirk Roekaerts (TUD) and Dr. Jim Kok (UT). The work was supported by the industrial partners Shell Global Solutions (financial support and engineering guidance) and ANSYS Europe (CFD modeling support). Projectleaders: dr.ir. J.B.W. Kok, Prof.dr. D.J.E.M. Roekaerts A major issue in Partial Combustion plants is to achieve an optimal syngas output composition, with low soot content and small combustor volume. These are conflicting demands and the design needs to be optimized with a view to the downstream process. To achieve high hydrogen and carbon monoxide concentrations a high combustor volume is required. But this will result in high soot content of the product gas. This will lead to fouling of downstream heat exchanger systems and hence to loss of reliability and high costs of maintenance. Gasification plant users and manufacturers will use the knowledge from this project to improve existing plants and optimize designs for new plants. With the expected increasing demand for new fuels like synthetic Diesel, syngas and hydrogen the proposed project will render crucial information and design tools on chemical reaction processes and soot formation in ultra rich conditions. Both experimental tests at elevated Research method A rich combustion test rig at elevated pressure was realized at the UT. Numerical modelling of the turbulent rich combustion process of perfectly premixed natural gas and oxidizer to syngas, and of a nonpremixed system with product gas recirculation was performed at both UT and TUD. AT the UT The gaseous chemistry is described by a reaction progress variable based combustion model with tabulated detailed chemistry. The soot formation and radiative heat loss of the gases and the soot particles is taken into account by 2 extra transport equations. All models are implemented in the commercially available CFD package ANSYS-CFX and applied on a premixed reactor design and a nonpremixed design with product gas back mixing. At the TUD a simple semi-empirical soot model based on the soot number density and soot mass concentration is adopted in a transported PDF method for turbulent diffusion flames. The gas phase chemistry is reduced by a flamelet generated manifold (FGM) based on the mixture 1 Schematic cross section of reactor. 2 Partial oxidation reactor experimental setup Brandbrief STW 2015

21 fraction, progress variable and enthalpy loss. To account for the radiative heat transfer, the Reynolds averaged radiative transfer equation (RTE) is solved by means of a discrete transfer method. The proposed modeling approach is validated in simulations of two turbulent non-premixed methane-air flames at 1 bar and 3 bar pressure. Results Experiments In this research measurements were done on the setup with a reactor with swirl stabilized flame (Figs 1,2) at several pressures and at equivalence ratios 2 to 4. A sample flow was extracted from the reactor downstream the oxidation front. This flow was quenched and diluted rapidly in order to avoid changes in chemical composition and soot particle growth and coagulation. For this application a special dilution system with dilution factor 104 was developed in the project. Upstream of the dilution step the gas composition was analysed with a gas chromatograph and downstream the dilution step, the soot concentration and size distribution (2-200 nm)was measured with a Scanning Mobility Particle Sizer. Previous experiments with a very similar setup and under the same conditions were done by Albrecht et al. [1] but without measurement of soot particle concentration. In figure 3 the experimental concentration measurements are shown. All CO concentration measurements show a generally decreasing trend with increasing equivalence ratios from F=2.5 to F=4.0. Table 1 gives the measured species concentrations at an equivalence ratio F of 3 and their values at 50 ms in Chemkin PSR, Chemkin Premix and chemical equilibrium. It can be concluded that the dry mole fraction of CO measured in the produced syngas is approximately 0.16, which is significantly lower than the equilibrium value of This is a clear indication that the mixture would move more in the direction of equilibrium if the residence time in the reactor would have been longer. This also explains the decreasing CO concentration at high equivalence ratio s, which is contradictory with equilibrium data. At high equivalence ratio the chemical reactions slow down, necessitating an increased residence time to reach equilibrium. This emphasizes the need of an accurate chemical activity prediction far from equilibrium. Simulations The CFI model was applied on a premixed natural gas flame, Nitrox40 as oxidizer, a preheat temperature of 573 K, equivalence ratio of 2.5 and a pressure of 4.0 bar. The axial velocity flow field shows a central recirculation area. The reaction progress variable shows a delay by chemical reaction kinetics after a steep increase in the flame front, steadily rising throughout the entire reactor until it reaches CCFI =0.88 at the outlet. The soot particle number density increases the most in the flame front, keeps increasing for a certain length before decreasing towards the outlet. In Figure 5 the axial profiles of CO, H 2, C 2 H 2 and the temperature of the CFI simulation have been plot. It can be seen that the CFI model is able to predict the endothermic reforming processes that are taking place downstream of the flame front in ultra-rich combustion. The decrease in temperature after the flame front due to reforming reactions is visible, as well as the steady increase of CO and H 2 and decrease of C 2 H 2. 3 Comparison of measured CO concentration in syngas produces on basis of natural gas/nitrox. table Measured and predicted species concentrations [vol %] in dried produced syngas. 3 Schone en zuinige verbranding 21