Numerical and Experimental Investigation of Biodiesel/Glycerin Separation

Show full item record

Title: Numerical and Experimental Investigation of Biodiesel/Glycerin Separation
Author: Garner, John Dail
Advisors: Richard Johnson, Committee Member
Herbert Eckerlin, Committee Member
Alexander Hobbs, Committee Chair
Abstract: During the production of biodiesel, oils and fats are vigorously mixed with an alcohol to produce biodiesel and glycerin. These two fluids need to be separated in order to obtain a pure stream of biodiesel. In continuous flow biodiesel production facilities, centrifuges are used to aid in this separation; however, centrifuges are expensive to purchase and operate, especially for small scale producers. For this reason, an alternative to centrifuges is needed for continuous production by small scale producers. In this biodiesel/glycerin mixture, small droplets of glycerin are dispersed within the biodiesel. If this mixture is placed into a tank where fluid motion is absent, gravity causes the droplets to settle out and form a layer of glycerin below the biodiesel. After the separation has occurred, the glycerin and biodiesel can be drained off separately. This is the technique that batch production facilities use. Using this technique, it is possible to use a series of tanks as a continuous flow separation device. While one tank was being filled, others would be settling. Once settled, the phases could be drained off and the tanks could be refilled. In order to predict whether the separation times are acceptable for biodiesel producers, a simple model is needed. In 1997, Parichay K. Das proposed such a model in his paper entitled 'Prediction of Settling Velocities of Drops in a Concentrated Batch Liquid-Liquid Dispersion.' In the present work, the estimated separation times from this model are compared to experimentally determined settling times. It was hoped that this model could accurately predict the separation time for this liquid-liquid dispersion; however, after comparison, it was determined that the model didn't accurately predict these settling times. Possible sources of error are discussed; however, no one reason for the model discrepancy could be determined. In the present work, a continuous separation process was designed to mate onto a continuous flow production facility that was developed by researcher at the University of Idaho. A simple bank of 5 small tanks was designed to replace the centrifuge that the Idaho researcher proposed. This design was based on Das's model, even though the model was suspected of overestimating the settling times.
Date: 2005-05-02
Degree: MS
Discipline: Mechanical Engineering
URI: http://www.lib.ncsu.edu/resolver/1840.16/232


Files in this item

Files Size Format View
etd.pdf 975.3Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record