Numerical modeling and experimental study of a plasma lift reactor

Abstract

The plasma lift reactor (PLR) is a new type of reactor used to degrade organic materials in an aqueous medium. This work focuses on improving the understanding of how this PLR works. Important contributions to science were made including: the first laboratory validated Computational Fluid Dynamics (CFD) model of the PLR, new experimental data for gas holdup, bubble velocity, and temperature, and improvments to the conductive probe measurement technique. A literature review was performed to better understand relevant technologies for both the experimental and numerical work. Conductive needle probes were selected from the literature review as a potential candidate for flow measurements in the PLR. However, in preliminary tests it became evident that improvements were possible to this technique. Four probes of different design were built and tested to determine which was the most effective at measuring gas holdup and bubble velocity. The largest of the probes exhibited the best signal to noise ratio and was effective in measuring bubble velocity. This style of probe was chosen for further experimental measurements in the PLR. Additionally a new technique was devised to process the probe signals which improves over existing technology. Software was written to demonstrate the new capabilities. The importance of the signal to noise ratio in conductive needle probe measurements and the new data processing algorithm were new and unique advances in the technology. As such, they have been published in the Review of Scientific Instruments. Computational fluid dynamics modeling was also performed. Results from these models will be the first ever to appear in a peer reviewed journal. A 2D axis-symmetric model was used to limit computational expense while providing reasonable accuracy. A customized drag model which accounts for bubble-bubble interactions was also implemented. Predictions for the liquid phase recirculation in the reactor agreed well with available laboratory measurements. Predictions for the gas holdup and velocity were better for lower flow rates than higher gas flow rates. Improved modeling techniques to be implemented in the future will undoubtedly bring numeric predictions closer to laboratory measurements.