Utilization of a Three Dimensional Model in Designing and Tuning of Large Scale Boilers

Transcription

1 Utilization of a Three Dimensional Model in Designing and Tuning of Large Scale Boilers Marko Lyytikäinen 1, Ari Kettunen 1, Kari Myöhänen 2, Timo Hyppänen 2 1 Research and Development Department, Foster Wheeler Energia Oy, FI Varkaus, Finland 2 Lappeenranta University of Technology, FI Lappeenranta, Finland Presented at 11th International Conference on Fluidized Bed Technology May 14-17, 2014 Beijing, China

2 UTILIZATION OF A THREE DIMENSIONAL MODEL IN DESIGNING AND TUNING OF LARGE SCALE BOILERS Marko Lyytikäinen 1, Ari Kettunen 1, Kari Myöhänen 2, Timo Hyppänen 2 1 Research and Development Department, Foster Wheeler Energia Oy, FI Varkaus, Finland 2 Lappeenranta University of Technology, FI Lappeenranta, Finland Abstract In this paper, a general overview about the utilization of the comprehensive threedimensional steady state model called CFB3D for the CFB boiler engineering and operation optimization is given. It is shown as an example of how to apply a three-dimensional model for temperature and oxygen profile tuning in a commercial scale CFB boiler. Tuning is completed by changing fuel and air feed distributions. That affects formation and degradation of emissions, which have been investigated in this paper as well. INTRODUCTION The complexity of the CFB furnace process and the large dimensions of commercial circulating fluidized bed (CFB) boilers set challenges for understanding and controlling of combustion phenomena in furnace. Combustion phenomena are connected to formed gas and temperature profiles. They have a big role in the designing sense as they have direct impacts on heat flux distribution, combustion efficiency, and formation of emissions. In a large-scale CFB furnace, the local feeding of fuel, air, and other input materials, as well as the limited mixing rate of different reactants produce inhomogeneous process conditions. Thus, the temperature and gas profiles can be affected by solid and gas feed locations as well as changing feed shares. To understand local process conditions inside the furnace and how to affect to them is crucial for optimizing the whole system. There are different approaches to increase the understanding of combustion phenomena inside the furnace. Direct method could be measurements. However, comprehensive measurements inside the commercial scale furnace are practically impossible because of large dimensions. Thus, many modeling approaches and levels have been developed. Furnace can be modeled as 0D, 1D, 1.5D, 2D or 3D. Modeling methods from zero to two dimensions are suitable for fast modeling of effectiveness of boilers. They are normally tuned to work well with certain types of boilers and configurations. The limitations come from understanding the detailed phenomena inside the furnace and how the different configurations affect to phenomena and to outputs. Three-dimensional modeling is able to fill this gap. Even if measurements have limitations they can be used for validating modeling tools. Three dimensional modeling tool can be validated using limited number of measurements and after that the model can be used for predicting phenomena all over the furnace. At Foster Wheeler, three dimensional model called CFB3D is in use. At the moment CFB3D model is widely utilized in CFB engineering and research and development (R&D) purposes. The model is used especially in large scale CFB designing for the determination of the optimal fuel and secondary air feeding locations. Also the optimal locations of the internal heat transfer surfaces are designed with the CFB3D model. In R&D, the CFB3D model is important and widely applied tool for further development of large scale CFB performance, like emissions and low load operation In this paper, CFB3D is used in simulating combustion processes inside a furnace. In the first example, fuel feed distribution has been optimized in a sense of temperature and heat flux profiles. In the second example, the knowledge gained from the modeling is utilized for emission reduction. SOx emissions have been reduced by optimizing air feed shares between the furnace walls. In addition, optimal ammonia feeding locations have been found out for reduction of NOx emissions.

3 MODEL DESCRIPTION AND VALIDATION CFB3D is an important tool in increasing knowledge about the CFB furnace process and nowadays it is widely utilized in CFB furnace concept development and operation tuning. The model solves the threedimensional CFB furnace process in steady state conditions. The furnace model is linked to separate submodels that model the other hot loop processes: separators, return legs, and possible external heat exchangers. The model frame is illustrated in Fig. 1. (Myöhänen and Hyppänen, 2011) Three dimensional furnace volume is discretized by hexahedral control volumes. The grid size is normally up to some hundreds of thousands of cells and cell dimensions are typically m. The model solves fluid dynamics of solids and gases, combustion and gasification reactions, comminution of solids, homogeneous reactions, heat transfer, and forming of emissions in each cell. Formation of different NOxspecies can be solved by a post-solver. The boundary conditions of the model include the different gas and solid feeds as local volumetric source terms and temperature profile of cooling fluid. The model is semiempirical: it combines fundamental balance equations with empirical correlations, which enables practical calculations of large scale furnaces. For example, the solid concentration profiles are set by empirical correlations, the solution of solid flow fields are based on potential flow approach, and the heat transfer to walls is determined as a combination of convective and radiative heat flux from dilute phase and convective heat flux from cluster phase (or wall layer) near the walls (approach resembling the cluster renewal model, Dutta and Basu, 2003). The model theory has been described by Myöhänen and Hyppänen (2011) and Myöhänen (2011). This kind of tool makes it possible to study combustion phenomena locally inside a furnace. All gas and solid feedings are modeled in actual places and those locations affect the calculation results. Thus, it is possible to investigate the effect of different geometries and operation parameters to combustion process, for example, the effect of oxygen-fired conditions (Myöhänen et al., 2009). The estimation capability of the model has been widely verified against process data collected from various commercial scale boilers (see Myöhänen and Hyppänen, 2011; Vepsäläinen et al., 2009; Myöhänen et al., 2005). Validations cover different operation conditions with full and lower load operation - unbalancing tests, air staging tests and co-combustion tests. Fig. 2 presents an example of a validation study, in which the fuel feed was intentionally set higher to the right side of the furnace. Because of the relatively slow lateral mixing of gases inside the furnace, the unbalance, which has been created at the bottom of the furnace, can be seen at the upper part of the furnace as well. The oxygen concentration is clearly smaller on the right side and by comparing the measured and modeled values, the different model parameters, which affect the reaction rates and mixing, can be tuned. Knowledge gathered by model validation studies and up-to-date parameter databases provides solid basis for performance prediction and design development of CFB boiler furnaces. Fig. 1. CFB3D model frame (Myöhänen and Hyppänen, 2011). Fig. 2. Modeled (contours) and measured (dots) oxygen contents in a case where fuel feed is higher on the right side than on the left side.

4 OPTIMIZING OF FUEL FEED DISTRIBUTION The design of the fuel feeding system is usually limited by economical and engineering aspects. The number of feed points should be minimal, but still provide uniform distribution of fuel feed and thermal load at the bottom of the furnace. Moreover, the feed points cannot be freely placed, because the bottom section of the furnace needs to accommodate e.g. return legs from separators, start-up burners, and secondary air ports. In most cases, the fuel feed is distributed evenly to different feed points, but this is not necessarily optimum for the furnace process. The following presents a calculation study of a 300 MWe oxygen-fired CFB with once-through steam cycle. This is an updated calculation of the cases presented by Myöhänen et al., The fuel is fed to nine feeding ports located at the bottom of the furnace. With a uniform fuel feed distribution, the temperature near the centerline of the furnace tends to be higher than near the side walls, which is due to cooling effect of the side walls. In this study, the fuel feed was decreased to the three ports, which are located near the centerline and increased to the other ports. The relative changes were in the order of 20% or less. Fig. 3 presents the distribution of the total heat from reactions at the bottom of the furnace before and after the modification. The modification affects the oxygen distribution: with non-uniform fuel feed, the oxygen concentration is slightly higher near the centerline, but the change is quite small (Fig. 4). Fig. 3. Distribution of total heat from reactions at the bottom of the furnace with uniform and nonuniform fuel feed. The locations of fuel feeding ports indicated by arrows. Fig. 4. Modeled oxygen concentration fields with uniform and nonuniform fuel feed distribution. When the fuel feed is reduced to the center ports, the temperature in the core as well as the maximum temperature inside the furnace are reduced (Fig. 5). After the modification, the heat flux distribution at the furnace walls is clearly more uniform (Fig. 6). Although in this case, the original heat flux profile is uniform enough for successful operation of the boiler, the 3D modeling provides means to optimize the thermal design. This is especially important for future oxygen fired units with higher inlet oxygen concentration and higher thermal load. Fig. 5. Modeled temperature fields with uniform and nonuniform fuel feed distribution. Fig. 6. Modeled heat flux profiles with uniform and nonuniform fuel feed distribution.

5 OPTIMIZING OF GAS FEED SHARES In this example, the controlling of emissions by adjusting the secondary air feed shares is presented. The example has been made for the boiler which generates 150 MW of thermal energy. The fuel used is Polish coal. Fuel is supplied through the front wall. In the basic situation, without any feed share optimization, 43% of secondary air is supplied through the front wall and 38% through the rear wall. 19% of secondary air is supplied through the side walls. Air shares for the basic and modified cases are presented in Table 1. Modeled temperature field for the basic case is shown in Fig. 7. Temperature field is quite uniform even if fuel is supplied through the front wall. However, oxygen content is not uniform due to combined impact of evenly distributed secondary air and unevenly distributed fuel. Oxygen content is lowest on the front wall side as shown in Fig. 8. Because of low oxygen content areas, sulfur capture is not optimal and high SO 2 concentration can be detected on the front wall side (see Fig. 9). In contrast to sulfur emission, high NOx emissions are located at the oxygen-rich region. This can be seen in Fig. 10. Table.1: Air shares for the basic and modified cases. Front Rear Sides Basic 43% 38% 19 % Modified 66% 30% 4% Fig. 7. Temperature profile in the basic case. Fig. 8. O2 profile in the basic case.

6 Fig. 9. SO 2 profile in the basic case. Fig. 10. NO profile in the basic case. Because there is a low oxygen content area on the front wall side, the air flow should be weighted more to that side. High oxygen areas on the back corners are unnecessary, thus, the side wall air flows can be reduced. In the modified case, the air flow is distributed so that front wall share is 66%, rear wall share is 30% and the rest of the air is fed through the side walls. In the following figures, the color scales are the same as in the above figures of the basic case. Temperature field of the modified case is shown in Fig. 11 and the oxygen contours are shown in Fig. 12. The modified case produces almost identical temperature profile as the original one. The low oxygen content area on the front wall side is decreased because of improved air share weighting. More even oxygen content improves sulfur capture and thus, high sulfur areas on the front wall side have been vanished. In the modified case, the SO 2 levels are smoother and lower as shown in Fig. 13. However, NO concentration has increased a little bit because there is more oxygen available in a wider area. NO concentration of the modified case is shown in Fig. 14. In this case, it is possible to add ammonia feeds to reduce NOx concentration. Modeling helps to find out best possible locations for the nozzles. Based on these results, good places for the nozzles are on the rear wall side at the places of high NO levels. In this way to examine the phenomena inside the furnace CFB3D tool can be used in the design and tuning of large boilers. Fig. 11. Temperature profile in the modified case. Fig. 12. O2 profile in the modified case.

7 Fig. 13. SO 2 profile in the modified case. Fig. 14. NO profile in the modified case. SUMMARY When building large CFB boilers, it is important to understand the formation of temperature and gas profiles inside the furnace and how they are affected by fuel and air feed distributions. They have a direct impact for example in heat flux distribution, combustion efficiency, and formation and degradation of emissions. The presented modeling tool together with measurements can be utilized for this purpose. The benefits of a valid three-dimensional CFB furnace model are clear: with a support of such a model, the placement of the feeding points and heat transfer surfaces can be designed optimally, and the model can be applied for further optimization of the fuel and air feed settings. The approach introduced in this paper provides considerable advantages in designing and tuning of new boilers. REFERENCES Dutta, A., Basu, P. (2003). An improvement of cluster-renewal model for estimation of heat transfer on the water-walls of commercial CFB boilers. In: Pisupati, S. ed. Proceedings of the 17th International Conference on Fluidized Bed Combustion, ASME, New York. Myöhänen, K., Hyppänen, T., Loschkin, M. (2005). Converting measurement data to process knowledge by using three-dimensional CFB furnace model. In: Cen, K., ed. Proceedings of the 8th International Conference on Circulating Fluidized Beds, International Academic Publishers, Beijng, Myöhänen, K., Hyppänen, T., Pikkarainen, T., Eriksson, T., and Hotta, A Near zero CO2 emissions in coal firing with oxy-fuel CFB boiler. Chemical Engineering & Technology, 32(3), Myöhänen, K., Hyppänen, T A three-dimensional model frame for modelling combustion and gasification in circulating fluidized bed furnaces. International Journal of Chemical Reactor Engineering, 9, Article A25. Myöhänen, K., Modelling of combustion and sorbent reactions in three-dimensional flow environment of a circulating fluidized bed furnace, Doctor thesis, Lappeenranta University of Technology. Myöhänen, K., Eriksson, T., Kuivalainen, R., Hyppänen, T Design and modelling of a 330 MWe Flexib-Burn CFB for oxygen-fired and air-fired combustion. In: Arena, U., Chirone, R., Miccio, M., and Salatino, P., eds. Proceedings of the 21 st International Conference on Fluidized Bed Combustion, 2012, EnzoAlbanoEditore, Vepsäläinen, A., Myöhänen, K., Hyppänen, T., Leino, T., Tourunen, A Development and validation of a 3-dimensional CFB furnace model. In: Yue, G., Zhang, H., Zhao, C., and Luo, Z., eds. Proceedings of the 20 th International Conference on Fluidized Bed Combustion, 2009, Tsinghua University Press, Beijing,