Condensing CO2 droplets

Condensed contaminant centrifugal separation is an emerging mechanical method to separate carbon dioxide (CO2) out of contaminated natural gas and combustion effluent. Cooling is achieved by using an expansion device and/or a heat exchanger. CO2 rich droplets are formed in the cooled gas stream. These CO2 rich droplets are then separated out of the gas stream by a rotational particle separator. The droplet size is a key parameter for the efficient design of the centrifugal separation unit. This thesis predicts and measures the behavior of condensing droplets formed by continuous expansion cooling. This study also describes the effect of relevant process parameters like feed composition, expansion rate, pressure and temperature. Measurements of CO2 rich droplets in CH4 or N2 are done for various feed compositions of CO2 (20-60 mole %) at the experimental test rig in Shell Amsterdam. A novel optical reflectance method is employed under high pressure (up to 30 bar) and semi-cryogenic conditions (down to -60 oC). A scheme to obtain the real droplet diameters from the measured chord lengths is also described and validated. CO2 rich droplet diameters down to 2 µm are measured. Measurements give the range of condensed CO2 rich droplets between 2 and 200 µm. 50-80 % of the liquid volume consists of droplets between 2 and 20 µm. Measured number densities are on the order of 1014 m-3. A theoretical model characterizes the effect of various process parameters on the droplet size. The number density of droplets depends on the expansion rate and surface tension. Droplet size depends on the condensed liquid amount for the given number density. For the typical expansion rates of turbines and valves (102-103 s-1), the calculated number densities lie between 1012-1016 m-3. The corresponding droplet size range is 0.5-10 µm for a typical liquid volume fraction of 1 %. The droplet growth model is validated in a steam/air condensation test rig. Laser diffraction measurements of condensed water droplet sizes in air are done for a range of steam in air compositions and flow rates. It is also found that condensation droplet growth is completed within 1 ms for the typical expansion rates of 102 s-1. Thus, the induction time can be shorter than the 0.2 s used in experiments. This enables design of more compact separation units. We show that CO2 droplets in a full scale unit will be as small as 1 µm. Thus, a separator needs to be designed to separate all the droplets down to this value in order to achieve 100 % droplet separation efficiency