57
A possible explanation for this unexpected size increase could be the increased
compressibility of the “live oil” (Satter and Iqbal 2016) which would reduce the
TKE of the jet by elastic deformation of individual drops rather than breakup.
In addition, the size distribution of the “live oil” was significantly broader compared to the dead oil. While both “live” and “dead” oil discharges followed a lognormal distribution function, the spreading factor σ varied widely for the different
oils and increased from “dead” to “live” oil by over 30%.
4.4.2 Influence of Rapid Pressure Loss at the Wellhead
and Phase Changes of the Oil
According to Lehr et al. (2010), the pressure of the oil at the DWH spill site changed
rapidly from 241 bar measured inside the BOP to 154 bar ambient hydrostatic pressure after exiting the wellhead. This change took place over a height of only 16.4 m.
As a result from this rapid pressure loss and the previous, slow decompression along
the borehole, a multiphase flow of “live oil” and gas is discharged from the wellhead. Due to the pressure-dependent solubility of gaseous components in the crude
oil (Chap. 3 and Sect. 4.4.1), such a massive pressure drop can lead to an oversaturation of the oil and therefore outgassing of C 1 to C 5 in addition to the expansion of
the already existing gas phase. This outgassing and gas expansion might affect the
drop formation in different ways. First, the expansion of a separate gas phase will
add to the overall TKE of the multiphase plume by its expansion energy. Secondly,
outgassing from the oversaturated oil will lead to the formation of a gas phase
within the oil drops, which will significantly alter and destabilize the drop. The gas
microbubbles might either leave the oil drop at its surface, thereby removing parts
of the oil from the “mother drop” or expand within the drop, thereby forming a twophase particle with different breakup and rising characteristics (see also Pesch et al.
2020 on the rise velocity of live oil droplets). Both effects are depicted in Fig. 4.2.
To assess the effect of such a pressure drop on the DSD, oil jets of both “live” and
“dead” oil were generated at Hamburg University of Technology with a defined
pressure difference between oil reservoir pressure and hydrostatic pressure of the
seawater. For this purpose, a throttle valve was added in front of the nozzle of the
experimental setup described in Seemann et al. (2014) and Malone et al. (2018) to
generate the required pressure drop. The oil reservoir was pressurized to 161 bar,
while the seawater basin was kept at 151 bar pressure like in the earlier experiments.
The DSD was determined from manually evaluated images of an endoscopic camera system (Malone et al. 2018).
With this pressure drop of Δp = 10 bar, the “dead oil” droplet sizes were distributed log-normally as expected, whereas the “live oil” was distributed bimodally.
This bimodal distribution could very well be described by superposing two lognormal distributions. It is assumed that this bimodality is caused by the processes
hypothesized above and depicted in Fig. 4.2. Under this assumption, the larger
4 Jet Formation at the Spill Site and Resulting Droplet Size Distributions
A possible explanation for this unexpected size increase could be the increased
compressibility of the “live oil” (Satter and Iqbal 2016) which would reduce the
TKE of the jet by elastic deformation of individual drops rather than breakup.
In addition, the size distribution of the “live oil” was significantly broader compared to the dead oil. While both “live” and “dead” oil discharges followed a lognormal distribution function, the spreading factor σ varied widely for the different
oils and increased from “dead” to “live” oil by over 30%.
4.4.2 Influence of Rapid Pressure Loss at the Wellhead
and Phase Changes of the Oil
According to Lehr et al. (2010), the pressure of the oil at the DWH spill site changed
rapidly from 241 bar measured inside the BOP to 154 bar ambient hydrostatic pressure after exiting the wellhead. This change took place over a height of only 16.4 m.
As a result from this rapid pressure loss and the previous, slow decompression along
the borehole, a multiphase flow of “live oil” and gas is discharged from the wellhead. Due to the pressure-dependent solubility of gaseous components in the crude
oil (Chap. 3 and Sect. 4.4.1), such a massive pressure drop can lead to an oversaturation of the oil and therefore outgassing of C 1 to C 5 in addition to the expansion of
the already existing gas phase. This outgassing and gas expansion might affect the
drop formation in different ways. First, the expansion of a separate gas phase will
add to the overall TKE of the multiphase plume by its expansion energy. Secondly,
outgassing from the oversaturated oil will lead to the formation of a gas phase
within the oil drops, which will significantly alter and destabilize the drop. The gas
microbubbles might either leave the oil drop at its surface, thereby removing parts
of the oil from the “mother drop” or expand within the drop, thereby forming a twophase particle with different breakup and rising characteristics (see also Pesch et al.
2020 on the rise velocity of live oil droplets). Both effects are depicted in Fig. 4.2.
To assess the effect of such a pressure drop on the DSD, oil jets of both “live” and
“dead” oil were generated at Hamburg University of Technology with a defined
pressure difference between oil reservoir pressure and hydrostatic pressure of the
seawater. For this purpose, a throttle valve was added in front of the nozzle of the
experimental setup described in Seemann et al. (2014) and Malone et al. (2018) to
generate the required pressure drop. The oil reservoir was pressurized to 161 bar,
while the seawater basin was kept at 151 bar pressure like in the earlier experiments.
The DSD was determined from manually evaluated images of an endoscopic camera system (Malone et al. 2018).
With this pressure drop of Δp = 10 bar, the “dead oil” droplet sizes were distributed log-normally as expected, whereas the “live oil” was distributed bimodally.
This bimodal distribution could very well be described by superposing two lognormal distributions. It is assumed that this bimodality is caused by the processes
hypothesized above and depicted in Fig. 4.2. Under this assumption, the larger
4 Jet Formation at the Spill Site and Resulting Droplet Size Distributions