This thesis belongs to the research field of applied computational fluid dynamics (CFD). The research focus is on utilization of existing state-of-the-art CFD modeling and mathematical optimization methods for very large and multiscale design optimization problems. The application under study is the recovery boiler, which is used to combust black liquor, a by-product of the pulp making process. The objective is to bring new understanding to the complex physical and chemical processes inside recovery boilers. In this thesis, the effects of various design choices on these processes are systematically and quantitatively investigated for the first time. The first part focuses on the combustion performance in the furnace. A surrogate-based optimization method and CFD modeling are combined to understand and quantify the connection between the furnace dimensions and combustion processes. As a result of the optimization, a set of Pareto-optimal geometries is obtained and the reasons for the improved combustion performance are investigated. Thereafter, a highly systematic CFD study is performed to understand the effects of typical design choices regarding the secondary air system on the mixing, penetration, and uniformity of the vertical velocity field. Several implications for the optimal design of the air system are formulated. The second part focuses on the flow field and heat transfer in the superheater region. A surrogate-based optimization method and CFD modeling are integrated to investigate the effects of the superheater region geometry on the flow field and heat transfer. The numerical results are analyzed to explain the physical mechanisms for the performance improvements and the linkage between the geometry, flow field, and heat transfer. After this, a new fully three-dimensional CFD model is developed to simulate the complex three-dimensional flow and heat transfer phenomena in the superheater region. Two sets of industrial full-scale measurements are reported and utilized to validate the simulation results. The added-value and new implications of the three-dimensional results are demonstrated. The main added-value of this thesis is considered to consist of the following factors. Primarily, it is one of the first extensive numerical studies into the internal processes of recovery boilers. Significant new knowledge is obtained in the following areas: 1) geometrical design of the furnace, 2) geometrical design of the superheater region, and 3) design of the combustion air system. In addition, an exceptional validation study is done between the numerical simulations and experiments in the superheater region. Finally, substantial new knowledge is obtained concerning the utilization of optimization methods in the context of recovery boilers and similar very large-scale applications.
|Publication status||Published - 2019|
|MoE publication type||G5 Doctoral dissertation (article)|
- computational fluid dynamics
- recovery boiler
- surrogate modeling
- heat transfer