Experimental Investigations of Single Oil Droplet Breakage in a Turbulent Water Flow

Abstract

Multiphase flow and the transient behavior of the dispersed phase is important to numerous industrial applications. Predictive modeling of the dispersed phase would be beneficial in designing, for example, phase separation equipment. A possible framework for modeling of the dispersed phase is the population balance equation. As the dispersed phase may undergo breakage, the breakage phenomena must be sufficiently understood to formulate universally predictive models describing the breakage processes. However, additional data from experimental investigations is required as the current understanding of the breakage phenomena is not at the required level.

In this work, the breakage phenomena are investigated experimentally by high-speed imaging of single octanol droplets in a turbulent water flow. A new experimental facility has been designed and constructed to perform the investigation. To determine the design criteria of the experimental facility the derivation of available turbulent breakage models was examined. In addition, a review of previous experimental setups and a review of isotropic turbulence facilities were performed. Four criteria were identified as critical for experimental investigation of turbulent droplet breakage. One, single droplets should be considered. Two, the entire breakage event must be observed by high-speed imaging and the procedure for extracting data must be transparent and well defined. Three, the experiments must be repeatable and reproducible as several experiments under the same conditions are required. Four, the region of breakage should be defined by known local flow conditions exhibiting low gradients in the turbulence level. To fulfill the determined criteria, a facility utilizing channel flow was constructed.

A LDV investigation was performed to characterize the continuous flow conditions. The resulting instantaneous velocity measurements were used to obtain the turbulent kinetic energy. Taylor's frozen hypothesis was used for estimating two-point correlations, which were used to obtain the turbulent kinetic energy dissipation rate.

A well-defined image analysis procedure was defined, elucidating the procedure of interpreting individual videos of breakage. Two interpretations of the breakage event, the initial breakage event definition and the cascade breakage event definition, were considered in the analysis.

To combine the information obtained from several videos, a clearly defined statistical analysis procedure was provided. In the procedure a quantitative precision of the measured quantities were obtained using 95% confidence intervals. Based on the statistical procedure it was determined that the number of investigations required for statistically relevant results were ~30.

Single octanol droplet experiments were performed and investigated to elucidate the breakage phenomena. The impact of both the mother drop size and the turbulence characteristics could be investigated, as each breakage event was associated with known local flow conditions from the LDV investigation. Known model concepts could be fitted to the data for the breakage time and the breakage probability with reasonable accuracy. However, the model coefficients were different from previous investigations, thus the models can not be considered to be universal. The average number of daughters and the daughter size distribution function exhibits behaviors which are not in agreement with available model concepts.