Improved multi-stage cross-flow fluidized bed classifier
Peer reviewed, Journal article
Accepted version
Permanent lenke
https://hdl.handle.net/11250/2688678Utgivelsesdato
2019Metadata
Vis full innførselSamlinger
- Publikasjoner fra CRIStin - SINTEF AS [6009]
- SINTEF Industri [1650]
Sammendrag
In the present work, improvements to a novel fluidized bed solids classification system previously published by Jayarathna et al. are discussed (Jayarathna et al., 2018 [1]). The system is designed for a high temperature application, namely solids classification in a CO2 capture plant incorporated with a coal-fired power plant. The capture plant is proposed to be equipped with the latest calcium looping technology, fully integrated calcium looping (FICaL). The classifier will be fed with a mixture of sorbent and heat transfer (inert) particles in the real plant at the operating temperature in between 910 °C and 600 °C. Due to the higher temperature it is not possible to use any sensitive process equipment such as filters or metal screens or any mechanically moving parts. Stepwise development of a such a system is explained in the previous publication by the authors (Jayarathna et al., 2018 [1]). In this work, the system is down-scaled into a smaller cold flow system, and zirconia and steel particles are used to mimic the sorbent and heat transfer solids particles in the hot flow system, respectively. Experiments and CFD simulations are carried out for the cold flow classification system. The previous fluidized bed design (Jayarathna et al., 2018 [1]) reached an efficiency of 90% and 99% recovery (10% and 1% loss) of the lighter (zirconia) and heavier (steel) components, respectively, but a 10% loss of sorbent particles in the real system would not be economical. The new modified system reached recovery efficiencies of 97% and 98% (3% zirconia loss and 2% steel loss) of the lighter and heavier component, respectively. CFD is used as a supporting tool in the design process, and also in making improvements to the design. Improvements are made by modifying the classifier geometry such as shape, height and internal design features. The experimental results from the improved classifier are compared with the CFD predictions made by the commercially available CFD software Barracuda® 17.1. Discrepancies between the experimental and simulated results are discussed, and the Barracuda CFD model is recommended as a good simulation tool for such studies and for upscaling the simulations for the hot flow system. The improved solid classifier design obtained good enough classification performances to proceed with upscaling and continue to make further improvements in the hot flow system. The iterative design and modeling effort from this research work has produced a functional, high-efficiency classification concept that can be used for the 1000 μm HT-solids particles and the 175 μm sorbent particles in the full-scale hot-flow system.