(104g) Redrop - An Efficient Monte-Carlo Method for Simulating Extraction Columns

The usual approach to design extraction columns is mostly based on pilot-plant experiments. These experiments are time- and material-consuming. Therefore an alternative method is to simulate the behaviour of technical equipment. This can be achieved by solving the drop-population balances. In this case the mathematical complexity of solving the population balances limits the number of considered drop properties. A meaningful simulation in a reasonable time is possible by instead using a Monte-Carlo method to solve the drop-population balances. ReDrop is such a Monte-Carlo method. The idea of the ReDrop model is to follow individual drops during their total life time in the column. Thus individual values like concentrations, diameter and life-time are stored for every single drop. If a new property of the drops has to be regarded, a single value for each drop is added.

The models for describing the individual drop behaviour are determined from measuring cells that characterize individual drop behaviour like sedimentation and mass transfer. Thus to perform ReDrop simulations, still experiments are necessary. But these experiments can be carried out in lab-scale cells and the material consumption is by far less than in experiments on pilot-plant scale, only a few liters are required of the original system. The experiments have to be carried out with the original system because even small amounts of impurities can change the drop behaviour dramatically. In these experiments sedimentation, mass transfer and breakage behaviour of single drops is investigated. Single-drop models are fitted to the experimental values to obtain system-specific model parameters. The parameters are then input to the ReDrop simulations.

ReDrop simulations have been successfully validated for different physical systems, including technical systems. ReDrop is able to predict the separation efficiency of a column as well as operating limits like flooding. Another advantageous characteristic of ReDrop is the column designer can simulate different types of internals. Thus the selection of the optimal column internals can be the last step in the column design based on the simulations and is no longer required as experienced first step which was the case for the design based on pilot-plant experiments.

The ReDrop algorithm has first been developed for physical extraction in pulsed sieve-tray columns and was then extended to columns with regular or random packings. The next step is currently to extend the application of ReDrop to reactive extraction. Therefore, modules were added to the ReDrop algorithm in order to calculate reaction kinetics and equilibria. To include non-ideality in both phases, different activity coefficient models were added.

In the presentation the concept of the ReDrop algorithm and the necessary single-drop experiments will be presented. First results of ReDrop for reactive systems will also be a central topic of the presentation.