In this work, the effect of the configuration on the collection efficiency and pressure drop in small diameter cyclones is evaluated based on neural networks (Functional Link Networks). The experiences were conducted at LSP/COPPE Laboratory in a Stairmand high efficiency prototype (Dc = 5 cm) with variable overflow diameter. Three different configurations were tested, and it is possible to observe a significant increase in the collection efficiency with the reduction of the overflow diameter.
small diameter cyclone; neural networks; radioactive wastes
