Degree Name

Doctor of Philosophy


Department of Materials Engineering


A detailed knowledge of the flow distribution is required for further studies of rate processes and their mechanism taking place in the packed beds. A better understanding of the rate processes is essential for proper process design and process optimisation of packed bed systems to maximise the comparative advantage and safety factor in industrial operations.

A new mathematical model of velocity distribution of single phase fluid flow in packed beds was developed by assuming the flow characteristic is a combination of a continuous and a discontinuous systems of fluids between voids in the bed. In order to allow a comparison with data measured at the downstream of the bed, the model was completed by a new mathematical model of a developing flow profile in an empty pipe. The model can be applied for both compressible and incompressible fluid.

The validity of the model has been checked using previous experimental data and new measurement results. The agreement between measured data and results predicted by the mathematical model is good. The new model favourably compares with previous models in terms of accuracy and its simplicity does not require new empirical constants.

It is clearly demonstrated that the fluid flow distribution in a packed bed is influenced by the Reynolds number and the bed characteristics. However, when the Reynolds number is higher than 500, the flow profile is mostly determined by the bed characteristics. Moreover, it is also demonstrated that the disagreement of previous investigators in the effect of Reynolds number and particle diameter on fluid flow distribution in packed beds is mostly due to limitation experimental.

Similar to the macroscopic view of fluid flow in packed beds, it can be shown at the microscopic level that models based on discontinuous systems are only successful for local porosity less than 0.5. In conditions when the local porosity is higher than 0.5, especially in the vicinity of the wall, a model based on the continuous systems approach, as used in the present case, is more accurate.

The restriction for the flat flow profile assumption for packed bed systems was investigated by using the present model. It is clearly shown that the deviation of flat profile condition not only depends on the D/DP ratio but also depends on the Reynolds number. The deviation of flat flow profile condition was also used to investigate the possibility of generating a distorted physical model in terms of D/DP ratio and L/D ratio.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.