Numerical model of flue gas desulfurization spray tower in power plant

Flue gas desulphurization system is designed by many companies for power plant system due to the environmental regulations. As coal is the main energy source, many power plants were built in coal-reserve areas of Turkey. Besides the plants were built, many FGD systems were also built by different global companies. Every company has its own FGD technology to clean the gas before releasing from the flue. These technologies are developed and protected by patent applications. The Babcock & Wilcox Company, Lenthes Bischoff GMBH, Mitsubishi Heavy industries and Noell-KRC Umwelttechnik GMBH are well-known FGD companies which they were built Kemerköy, Yatağan, Afşin and Orhaneli power plants in the 1990s. These plants powers are various from 210 MW to 360 MW. FGD systems have been installed in as many power plants as ever, but there is still no method to accurately predict their behavior. It is desired to study the flow in the column in detail to achieve optimum column design, which is its difficulty. Among of FGD systems design requirements, cleaning of flue gas at high (90% and above) rates, maximum desulphurization and minimum pressure drop, is a priority. In this thesis numerical model of the flue gas tower is applied to a real scale power plant and investigate the performance of the column. Some design factors are selected to compare the results. These are (1) adding spray level, (2) integrate perforated plate, like Babcock & Wilcox patented tray and (3) putting a ring to reduce leaks either flue gas or slurry liquid due to the column wall. Numerical analyses in FGD system designs have become one of the most widely used methods today due to the development of computational fluid dynamics (CFD) techniques and increase in computational power. As a result of the CFD analysis, detailed flow field estimates are obtained, and this provides a better understanding of dynamics. In the literature review given in the introduction part of the thesis, besides numerical modeling studies, laboratory scale testing and mathematical modeling studies were included. Studies on FGD systems in the literature have been investigated and referenced in the thesis content. In the section on physical systems and simplifications, the modeling steps of the FGD system are given in titles. To summarize, the modeling steps are respectively solved by creating the geometry, extracting the flow volume, dividing the flow volume into the finite elements, meshing, and the geometry that is ready for model definition. Modeling steps of the flue gas system have also been given in this order. The FGD system, the column geometry, and piping in the spray levels were primarily created with Siemens NX cad program. Spray levels are modeled as surface due to simplification of the geometry. The prepared geometry was transferred to Ansys-SpaceClaim program and necessary definitions were made and sent to Ansys-Meshing program. Appropriate size meshing was created and necessary surfaces in the boundary definitions were also named. In Fluent program, it was solved by making appropriate definitions and results were obtained. The information obtained from literature and with some simplifications, the flue gas desulphurization system was modeled on the power plant scale. These simplifications are that the density of the sprayed droplets is much greater than the flue gas, and as a result, the forces can be neglected other than drag force. In the calculation of the drag force acting on the droplets, it is assumed that the droplets are in spherical form. At the same time, it was considered two-way interaction between the gas and liquid phases in the flue-gas column. Eulerian and Lagrangian approaches generally used for mathematical modeling of two-phase flow. The advantage of Lagrangian approach is that droplet physical properties (diameter, density, etc.) can differ in. In this thesis, Euler-Lagrange approach is used for these reasons. When the gas phase is defined as Euler, the droplets are modeled as a discrete phase by Lagrange trajectory approach. The modeling of the discrete phase is defined detailly. Especially since working FGD model has too many nozzles, a total of 460 nozzles, must be defined together with the position informations and the properties of each injection. For this process, MATLAB program and Excel file that define the injector properties are used. How the droplets will behave after hitting the column wall is governed by the added UDF code. Reflection angles and velocities are calculated by using the velocity and incidence angle of each droplet that hit the column wall. Numerical and test results of Weiss and Wieltsch were taken as reference in these calculations. First, FGD column was analysed without the injectors to provide convergence for the gas phase, and the results are given as a comparison. After the gas phase converges, the discrete phase is activated with unsteady particle tracking. While the quencher level and the second spray level are operated in all analyses, the first spray level, which usually reserves in power plants, is operated in a single analysis condition. When the tray is modeled, the porous-jump boundary condition is applied on the surface where the tray is located. The coefficients necessary for porous-jumping were found out from the literature with perforated-plate values of 50% porosity and estimated the height of slurry flow values to be accumulated on the plate were entered. If FGD column has a ring geometry, a ring surface was created in the model geometry, and wall boundary condition was defined for this surface. The results and interpretations are given in the last two sections, with each analysis of the resulting velocity and pressure figures, and the effects of comparative graphics and design factors are summarized. From these sections, it is seen that the pressure drop in the flow of the flue gas is observed as a corrective effect of flue gas distributions. However, this pressure drop will increase operating costs. Efficiency reduction of the desulfurization systems is mainly caused by colliding gases in the column walls. The inability to place a nozzle near the wall areas and the flow of the sprayed slurry from the wall cause inefficiency. Therefore, the design proposed by the ring addition analysis clearly shows that the flue gas escaping without desulfurization accelerates to the center of the column and helps to meet with the slurry. In the appendix, simple implementations that can be done before the modeling of the flue gas systems are given. UDF code for controlling droplet-wall interaction and some numerical results which were used to create figures were given in the appendix, too. In addition, since the study requires for defining a hollow-cone nozzle, the parameters to be defined for the hollow-cone nozzle, have been summarized from the Ansys Fluent User's Guide document along with explanations.