Statistical techniques for correlating the shape and performance for new random packings.

摘要

Separation processes are the heart of the chemical industry and are used for product purification. Distillation, absorption and stripping are some of the basic types of separation processes encountered in the industry. The operations are carried out in columns containing trays or packings. Packed columns have become popular in the industry because of the advantages associated with them. There are many types of packings available in the market and research is on going to find better packings. In the present study, we deal exclusively with random packings.; The existing models calculate the mass transfer efficiency and other performance parameters depending on the properties of the fluids being handled, the geometrical area, porosity and the shape of the random packing. The shape of the packing is generally characterized using a packing specific shape constant. Though these models cover a wide variety of random packings, one cannot predict the performance of newly developed packing elements without experimentation. In this work, we investigate the relation between the performance and the shape of the random packings.; Since geometric parameters like surface area and porosity do not define the shape of the packing uniquely, we attempt to define the shape of the packing using the porosity distribution characteristics. The porosity distribution was computed using an algorithm developed for packing random packings. Higher order statistics were used to define the porosity distribution and Partial Least Squares (PLS) modeling was used to develop a linear relationship between the shape defining parameters and the performance characteristics of random packings. New packings were developed by selecting various shapes available in geometry and the mass transfer efficiency and pressure drop were predicted for these packings using the model.