Droplet separators for evaporative towers: efficiency estimation by PDA

Abstract A Phase Doppler Anemometry system has been set up to characterize the behaviour of different arrangements of static impactors that are used as drift eliminators to intercept and remove residual water droplets entrained in the hot air flow released by an evaporative tower. The investigated evaporative tower has a square section of 60x60 cm, that is the standard size for modular separation elements; it was tested while working with or without the droplet separators; for each test condition, the residual water droplets entrained by the air flow and expulsed by the tower are detected few centimetres above the tower exhaust and are characterized by the PDA system that can measure the velocity and diameter of each droplet. Many parameters describing the droplet population over the whole exit section, (number, mean velocity and diameter) can be obtained. The distribution of the droplets is calculated as a function of their diameter, and represented as a percentage of their total number and of their total volume. The velocity-diameter plot of the same droplets shows other aspect of the population. Droplets with same diameter show a spread velocity distribution due to the high turbulence of the exhaust air flow. Droplets with larger diameters have smaller mean velocity, since the gravitational downward force is not negligible compared to the aerodynamic upward drag from the air flow. Few very large droplets have even negative velocity: they are interpreted as droplets that, after the expulsion, are falling back downward. The use of an LDV-PDA system allows to detect such droplets and to discard them from the separation efficiency calculations. The efficiency can be calculated by direct comparison of the number of water droplets that are detected in the exhaust air flow, both globally or for any specific class of droplet size. Global results can be calculated on the basis of the number or of the volume of the droplets. The accuracy of the efficiency estimation is also studied. Two main aspects are considered. The first is that the PDA measurement volume dimension variation with the detected droplet size: it has negligible effect on the result per classes of diameters, but the effect on the global efficiency is present and will be discussed. The second aspect is the presence of droplets that are falling downward in the upward air flow: their presence should be considered and corrected for. The results of this paper are useful when comparing them to other measurements obtained with techniques that are not able to detect the droplet velocity.