Back to Table of Contents

43857-AC9
An Integrated Numerical and Experimental Study of Scaling Effects in Stirred Tank Reactors

David F. Hill, Pennsylvania State University and S. Balachandar, University of Florida

First Year Report

-------------------------

1. Overview

This project proposed to conduct an integrated numerical and experimental study of the turbulent flow in stirred tank reactors. These flows are widely used in the chemical, petroleum, and environmental engineering industries. Knowledge of the mean and turbulent velocity fields in these flows is important as the large and small scale flow structures help to control reaction and blending rates. The work is motivated by the need to be able to scale up small laboratory-scale experimental results to full prototype scale. The funded project seeks to conduct the most extensive set of scaling experiments to date. In addition, computation simulations of the flow, using the experimental results, will be performed.

2. Project Management

The funded project spans a period of two years. The first year (covered by this annual report) was conducted at the Pennsylvania State University, under the direction of Dr. D. Hill. The second year (to be covered by the next annual report) will be conducted at the University of Florida, under the direction of Dr. S. Balachandar. The work during the first year was experimental in nature, with the goal of investigating the scaling of the flow. The work during the second year will be computational in nature, using the results from the first year as a foundation.

3. Work Accomplished

3.1 Development of Experimental Facility

The first task accomplished during this year was the construction of the experimental facility. Since the goal was to investigate the geometric scaling of the flow, a progressive series of four mixing tanks was constructed. The tanks ranged in size from (inner diameters) approximately 6 inches to 24 inches. The general arrangement was an unbaffled cylindrical tank with a fluid depth equal to the inner diameter. A Rushton turbine was positioned on the tank axis and was located at the vertical midpoint. Following the usual convention for Rushton impeller driven flows, the impeller diameter was 1/3 of the tank inner diameter. The impeller was connected to a programmable variable speed mixer via a shaft. The shaft passed through the impeller body and extended to the bottom of the tank, where it was supported by a bearing. This had the effect of reducing shaft wobble. Given the emphasis on scaling effects in the present study, extreme care was taken to ensure that the four tank / impeller sets were as geometrically similar as possible. This attention to geometric scale extended to to thickness of the stock material used to construct the impellers and the diameter of the shaft used to support the impeller.

3.2 Experimental Matrix

For each tank, the goal was to conduct experiments at eight Reynolds numbers, spanning the range from approximately 10,000 to 150,000. This leads to a total of 32 experimental trials. Due to constraints on minimum / maximum mixer speed and maximum mixer power output, not all 32 trials were possible. In the end, 28 trials were conducted.

3.3 Velocity Data Acquisition

For each experimental trial, two-dimensional velocity data were obtained using the particle image velocimetry (PIV) method. The field of view was positioned to be very near to, and include, one of the impeller blades. An optical trigger was used to initiate the data acquisition process. This allowed for the collection of 'phase locked' data, or data at the same relative angle of rotation of the impeller shaft. For each trial, an ensemble of 1000 realizations was obtained, allowing for the computation of mean and turbulent velocity statistics. The raw data were scaled (non-dimensionalized) by parameters such as blade tip velocity and impeller diameters, allowing for study of how the flow field depends upon Reynolds number and geometric scale.

3.3.1 Three Dimensional Velocity Data Acquisition

In an exciting collaboration between the PIs and TSI, Inc. (one of the global leaders in velocity laser diagnostics), the experimental facilities developed with this grant were used to test the V3V (Volumetric 3-dimensional Velocity) system. This experimental system allows for the instantaneous acquisition of velocity data over a three-dimensional volume (cube). Within this cube, approximately 20,000 points of data are acquired with each realization. The PIs and scientists from TSI collaborated to obtain data sets for three of the conditions in the experimental matrix. These data are completely unique and have never been obtained before for flows in stirred tank reactors. The chief benefit of these data is that they provide a fully three-dimensional and instantaneous look at the extremely complex flow field (tip vortices, radial jet, etc.) in a stirred tank.

3.4 Future Plans

The PIs are presently analyzing their data. One manuscript has been submitted, and several others will be forthcoming. Additionally, the work is being presented at the Penn State Research Symposium (presentation by the funded graduate student).

Back to top