Residence Time Distribution modeling of combustors through Chemical Reactor Network

Abstract

The growing globalization and continuous industrial and urban development result in an ever-increasing energy demand for transportation and power generation. Therefore, providing sustainable and clean conversion processes is one of the most crucial challenges worldwide. By virtue of the high energy density of chemical fuels and the capability of satisfying strongly fluctuating energy demand, combustion dominates the global energy scenario. According to the crucial reasons listed above, its importance is not diminishing in the foreseeable future, despite the increasing share of renewable conversion sources. Hence, a continuous improvement of combustion devices, to increase their efficiency and reduce pollutants emission, is mandatory, and new combustion concepts must be explored and tested. Lean combustion fulfills the requirements of an efficient and clean combustion process, applicable to aviation gas turbines. As it results in higher temperatures of the combustor wall, this process is often coupled with effusion cooling of the combustor liner. This poses the problem of the interaction between a hot reactive environment and cold cooling air, which is not yet fully understood. This strategy needs further investigation because it could lead to high carbon monoxide emissions due to rapid and inhomogeneous cooling of pockets of reacting fluid. Oxyfuel combustion, instead, is a strategy applicable to stationary power plants. In this case, the substitution of air with a mixture comprising carbon dioxide and oxygen takes place. Such an approach represents a point of novelty with respect to standard combustion in air. Therefore, an improved understanding of the effect of excess CO2 in a reaction environment is crucial. Both outlined technologies present undeniable advantages, yet their possible drawbacks need to be understood and avoided, and a deeper knowledge is required to properly exploit their potential. To this aim, thorough experimental investigation in suitably designed close-to-reality configuration is mandatory, and ideally complemented by the modeling of the observed phenomena. A modeling strategy that well fits the idea of a strong synergy between experiments and modeling, is referred to as Chemical Reactor Network (CRN) modeling. This strategy proposes a simplified version of the flow field based on two extreme mixing possibilities. Such models give insight into the mixing and reactive features of a complex flow, as the ones encountered in practical combustion devices. To achieve a proper CRN model, it is possible to design and size it against the Residence Time Distribution (RTD) of a certain configuration. Even without further modeling, RTD data alone yields precious information on the mixing characteristic of the system under investigation. In the current work, Chemical Reactor Network modeling based on experimental Residence Time Distribution data is applied to two suitably designed close-to-reality configurations. They are representative of an aviation gas turbine combustor and a power generation furnace. These systems were designed to better understand the underlying phenomena while investigating new combustion concepts, such as lean combustion and oxyfuel combustion. CRN models are designed and tested in both situations. These models are developed based on zonal modeling of the flow field and on the Residence Time Distribution of the systems. In both cases, their performances are tested against experimental data available for both test-rigs regarding pollutants emissions. Additionally, they are employed to understand the impact of the operating conditions on the combustion process. This work states the importance of such simple and flexible tools in combustion research.