Name: MURILO ZUCATELLI ELIAS
Publication date: 20/03/2025
Examining board:
Name |
Role |
|---|---|
| PAULO LARANJEIRA DA CUNHA LAGE | Examinador Externo |
| RENATO DO NASCIMENTO SIQUEIRA | Examinador Interno |
| ROGERIO RAMOS | Presidente |
Summary: During oil production, shear and turbulence in the ow progressively break water droplets, leading to the formation of stable water-in-oil (W/O) emulsions. This study investigated the droplet breakup and coalescence dynamics in W/O emulsions through laboratory experiments and population balance modeling. The research was conducted using the NEMOG/UFES ow circuit, focusing on the impact of gate valve restrictions on the droplet size distribution (DSD). The Coulaloglou and Tavlarides (CT) and Mitre et al.(CEM) published models were applied to simulate the observed phenomena, with parameter adjustments to optimize predictions against experimental data. Results showed that the MV-01 valve, located at the circuitŠs inlet, predominantly induced droplet breakup, causing a moderate reduction in the mean droplet size. In contrast, the downstream MV-02 valve exhibited greater variability in DSD due to a signicant increase in coalescence events, particularly at dispersed phase concentrations exceeding 15% v/v. The implemented models effectively predicted the DSD for water volume fractions of 5%, 10%, and 15% v/v. The CEM model demonstrated superior overall performance, even with its original parameters, although it underestimated coalescence effects at high dispersed-phase concentrations. With mean errors of approximately 40% for MV-01 and 16% for MV-02 when considering the distribution. Predictions of the mean droplet diameter, De brouckere, yielded mean errors below 23% for MV-01 and 9% for MV-02. The CT model showed limitations in fully describing coalescence phenomena, requiring additional adjustments under severe conditions. Factors such as water volume fractions and the heterogeneous distribution of turbulent energy were identied as critical for improving predictions. Sensitivity analysis revealed that breakup and collision parameters ( and ) had the most signicant inuence on prediction errors, whereas the coalescence efficiency parameter () had a negligible impact. Experimental results partially corroborated model predictions, with average and maximum errors comparable to those reported in the literature. The maximum error in the predicted volume-weighted mean droplet diameter ([4,3]) was 25.6%, highlighting the need for advancements in submodels for coalescence and energy dissipation to enhance simulation accuracy.
