The size distribution of oil droplets formed in subsea oil and gas blowouts is known to have a strong impact on their subsequent fate in the environment. Small droplets have low rising velocities, are more influenced by oceanographic turbulence and have larger potential for natural biodegradation. Subsea Dispersant Injection (SSDI) is an established method for achieving this goal, lowering the interfacial tension between the oil and water and significantly reducing oil droplet size. However, despite its many advantages, the use of SSDI could be limited both by logistical constraints and legislative restrictions. Adding to the toolkit a method to achieve subsea dispersion, without the use of chemicals, would therefore enhance oil spill response capability. This option is called Subsea Mechanical Dispersion (SSMD).

An extensive feasibility study on SSMD has been performed and the main findings are reported in this paper. The work was initiated by BP in 2015 and later followed up by a consortium of Equinor, Total Norge, Aker BP and Lundin. The first phase explored multiple principles of generating subsea dispersions (ultrasonic, mechanical shear forces and water jetting) through both laboratory experiments and modelling. These studies clearly indicate that SSMD has an operational potential to significantly reduce oil droplet sizes from a subsea release and influence the fate and behaviour of the released oil volume.

The recent work reported in this paper on operationalisation, upscaling and large-scale testing of subsea water jetting. This work is performed by SINTEF in close cooperation with Exponent (computational fluid dynamics and shear stress modelling) and Oceaneering (operationalisation and full-scale prototyping).

This paper describes the results from a program exploring the potential of using a mechanical device for creating mechanical dispersions in response to a subsea oil release. This response option is called subsea mechanical dispersion (SSMD) and the main objective is to influence the fate of the released oil in the marine ecosystem by significantly reducing the droplet sizes of the released oil. The initial feasibility study (Phase-I) performed in 2013–15 focused different technologies for mechanical subsea dispersion; Mechanical Shear, Ultrasound and Water jetting (Davies at al., 2015). The main conclusions from this study were that the SSMD concept was promising and should be further evaluated (Brandvik et al., 2016).

The follow-up study (Phase-II) was performed in 2016–17 and focused on both down-scaled experimental work to verify the principle of water jetting (SINTEF), Computational Fluid Dynamics (CFD) modelling to study the fundamentals (Exponent) and a review of available equipment suitable for a subsea water jetting operation (Oceaneering). Both small-scale laboratory testing and modelling indicated high effectiveness using water jetting as a method for SSMD. This together with a positive outcome of the equipment survey pointed at water jetting as an operationally viable method for SSMD.

Phase-III (2018–19) aimed at reducing some of the uncertainties regarding effectiveness with mixed releases (oil & gas), optimizing nozzle designs and feasibility/costs of a prototype. This phase was mainly performed in 2018 focusing on:

  1. Small-scale testing of combined releases of oil & gas,

  2. Modelling of different nozzle configurations using turbulent dissipation rate as a performance metric

  3. Conceptual design of a full-scale prototype (subsea pump & nozzles operated by a ROV).

The results show only a minor reduction in effectiveness on combined releases of oil and gas compared to oil alone, increase the knowledge regarding nozzle design (one vs. multiple nozzles) and showed that a full-scale prototype can be based on existing submersible pumps and ROVs.

Phase-IV of this program includes large-scale testing of combined releases (oil & gas) at the Bureau of Safety and Environmental Enforcement's (BSEE) Ohmsett facilities in the U.S. This testing will be performed in December 2019 and this study will be reported in a separate paper.

Why do we study the potential of mechanical subsea dispersion on subsea oil releases? Do we need an additional or alternative response technique? This question has been asked both during the previous feasibility studies and when the findings from these studies have been presented at conferences (Brandvik et al., 2016). The main strategy for applying SSMD is to significantly reduce the oil droplet sizes from a subsea oil release, which changes their fate and behaviour in the marine environment. This can also be achieved by subsea dispersant injecting (SSDI), which is a well-established technology in many regions. However, in many cases it is beneficial to have a wide range of response technologies to choose among. SSMD can be beneficial in some scenarios where SSDI cannot used, for example due to regional legislation or in release scenarios were SSDI is expected to have very low effectiveness, for example in releases with very low turbulence. There is also a significant operational advantage to avoid transport large quantities of dispersants.

The main objectives for the work described in this paper have been to:

  1. Verify that the high SSMD effectiveness measured earlier also are valid for combined releases (oil & gas).

  2. Study different nozzle designs and the effect of upscaling by CFD modelling.

  3. Perform a concept design study for a full-size SSMD based on existing submersible pumps and ROVs.

The experimental work was performed in the SINTEF MiniTower and Titanium high pressure tank (TiTank) in Trondheim, Norway. This paper presents only a part of the experimental work and further details are available in the full technical report from this study (Brandvik et al., 2019a). Since these experiments include gas (air), the LISST instrument couldn't be used for quantifying particle sizes. The LISST is not capable of distinguishing between oil droplets & gas bubbles. The SINTEF Silhouette Camera (SilCam) has been successfully used for this purpose in multiple projects (Davies et al., 2017 and Ahnell et al., 2018). Both the MiniTower and the TiTank were modified for using the SilCam to detect particle sizes (oil droplets and gas bubbles).

Table 2-1:

Experimental conditions

Experimental conditions
Experimental conditions
Experimental conditions
Experimental conditions

These conditions (Table 2-1) gave the following oil droplet sizes for untreated oil, expressed as the median diameter from a lognormal volume distribution (d50);

  1. Oil alone experiments: 740 μm

  2. Combined experiments with oil and gas (1:1 volume): 530 μm

This particle sizes are slightly larger compared to standardised condition for this test system (Brandvik et al., 2019b), but were needed for the SilCam to effectively discriminate between oil droplets and gas bubbles.

In addition to a single horizontal nozzle, two additional nozzle configurations were tested (see Table 2-1). The “Star” configuration consisted of three nozzles configured axially pointing (inward) horizontally towards the released oil. The “Grid” configuration consists of multiple nozzles located in the rising plume with the jetting nozzles pointing downward against the rising oil & gas plume. In full-scale experiments this design could consist of a high number of nozzles (50 – 150). Due to the limited dimensions of the released oil & gas plume in the laboratory experiments, the number of nozzles that could be used is limited. In these experiments three downward pointing nozzles are used to simulate the “grid” configuration.

Figure 2.1:

Star nozzle configuration (3 x 0.17 mm ID) with oil and gas release (HP TiTank). Upper part: Oil and gas untreated (left), with water jetting – SSMD (right) Central part: Image showing oil droplets and gas bubbles (from SilCam) Lower part: Oil droplet distributions quantified from SilCam images (30 sec average)

Figure 2.1:

Star nozzle configuration (3 x 0.17 mm ID) with oil and gas release (HP TiTank). Upper part: Oil and gas untreated (left), with water jetting – SSMD (right) Central part: Image showing oil droplets and gas bubbles (from SilCam) Lower part: Oil droplet distributions quantified from SilCam images (30 sec average)

Close modal

2.1 Low pressure experiments (MiniTower)

The bench-scale laboratory system used for this testing (80 L) was developed by SINTEF and is a scaled-down version of SINTEF Basin Tower (40 000 L) used for studies of subsurface releases of oil and gas and the effectiveness of dispersant injection (Brandvik et al., 2013). The effectiveness of a dispersion technique is quantified as the relative shift in median volume droplet size (MVD or d50) compared to the untreated oil. The tank has a flow-through system for sea water and the laminar flow makes it possible to perform continuous experiments. A typical experiment series starts with background measurements of any natural particles in the sea water and then often a period with untreated oil, before different types of treatment is started. One example could be water jetting with different water velocities. Usually is each treatment continued for 60 seconds to allow sufficient data collection (30 seconds) to obtain data with sufficient statistical quality. Further details regarding the SINTEF MiniTower and its operation are available in Brandvik et al., 2019a.

Cavitation due to high water velocities limits the water jetting velocity used during the MiniTower experiments. Cavitation contribute to the oil droplet splitting process and produce smaller droplets. However, it is an artificial laboratory process since it is not expected to be present at operational hydrostatic pressures (> 100 meter). To be able to perform high velocity experiments under “non-cavitation” conditions, experiments were performed under pressure (20 bars), see next chapter.

2.2 High pressure experiments (TiTank)

To simulate conditions on the Norwegian continental shelf, a pressurised titanium is established at SINTEF. This tank can attain a pressure of 30 bar (300 meters depth) with a volume of 1.4 m3. The tank is equipped with a decompression chamber, various sample holders, circulating pump etc. However, most of this equipment was sealed off during our experiments, to ease cleaning of the facility at the end of the experiments. The operating procedure for the TiTank is very similar to what described earlier for the MiniTower. The distance from the release nozzle to the SilCam was comparable to the MiniTower. After the tank was closed and pressurized video monitoring of the nozzle, plume and instrumentation ensured necessary documentation. Further details can be found in Brandvik et al., 2019a.

3.1 Combined releases of Oil and Gas

This section presents only selected parts of this study, more details are found in Brandvik et al., 2019a. The experiments presented in Figure 3.1 show the reduction in oil droplet sizes directing a singular water jet (0.17 mm nozzle) at the released oil (and gas) to simulate subsea mechanical dispersion (SSMD). The water jetting was performed at a height of ten release diameters (15 mm) above the release nozzle. This allows the initial oil droplet formation, due to the turbulence in the jet, to finish before we mechanically reduce the oil droplet sizes by water jetting. Water jetting is performed with a velocity ranging from 22 – 92 m/s. All experiments performed under non-cavitating conditions and the water flow rates used represent 7.5 – 31% of the oil release rate.

Figure 3.1:

SSMD simulated by a Singular Horizontal Water Jet (0.17 mm ID) for oil alone and combined experiments: Relative droplet size distribution (volume %) for oil alone (black lines) and as a function of increasing water velocities (coloured lines). The droplet sizes in the boxes are d50 quantified from the distributions. All experiments are performed under non-cavitating conditions.

Figure 3.1:

SSMD simulated by a Singular Horizontal Water Jet (0.17 mm ID) for oil alone and combined experiments: Relative droplet size distribution (volume %) for oil alone (black lines) and as a function of increasing water velocities (coloured lines). The droplet sizes in the boxes are d50 quantified from the distributions. All experiments are performed under non-cavitating conditions.

Close modal

From Figure 3.1, we observe only minor differences between the oil alone experiments (upper part) and the oil and gas experiments (lower part), and there is no systematic trend showing larger oil droplets in the experiments with oil & gas compared to the oil alone experiments. The oil droplet sizes for the highest water jetting velocities varies in the 120–130 μm range for both types of experiments. However, since the untreated oil droplets are smaller for the combined releases, the effectiveness (relative droplet reduction) is slightly smaller for the combined releases. See oil reduction factors in summary conclusions for more details (Table 6-1).

Table 6-1:

Subsea dispersion effectiveness in small-scale laboratory experiments (oil droplet size reduction factor). Only for experiments without cavitation.

Subsea dispersion effectiveness in small-scale laboratory experiments (oil droplet size reduction factor). Only for experiments without cavitation.
Subsea dispersion effectiveness in small-scale laboratory experiments (oil droplet size reduction factor). Only for experiments without cavitation.

3.2 Different Nozzle Designs

As presented earlier, the “Grid” nozzle configuration consists of multiple nozzles with the water jetting pointing downward against the rising oil & gas plume. In full-scale experiments this design could consist of a high number of nozzles (50 – 150). Due to the small plume dimensions in the down-scaled laboratory experiments, the number of nozzles that can be used are limited. In these experiments three downward pointing nozzles are used to simulate the “Grid” configuration. In the experimental set-up used in this study only the low to medium water velocities could be used, since the higher velocities caused the released oil to interact with the release arrangement and the bottom section of the MiniTower, due to the downward momentum of the nozzle grid. Results comparing the three different nozzle configurations are presented in Figure 3.2.

Figure 3.2:

Both Star- , Grid- and Single nozzle configuration: Comparison of reduction in oil droplet sizes (SSMD effectiveness) with both combined- and oil only experiments. Water jetting velocities of 86–92 m/s for Star and single-nozzle and 49 m/s for Grid-configuration.

Figure 3.2:

Both Star- , Grid- and Single nozzle configuration: Comparison of reduction in oil droplet sizes (SSMD effectiveness) with both combined- and oil only experiments. Water jetting velocities of 86–92 m/s for Star and single-nozzle and 49 m/s for Grid-configuration.

Close modal

Comparing the experiments with only oil, “Single” and “Star” perform very similar (reduction in droplet size versus jetting velocity), while “Grid” shows a significant lower ability to reduce oil droplet size. However, the comparison in slightly unfavourable for the “Grid” approach since we could not use the highest jetting velocities because we forced the released oil down towards the bottom of the minitower where the oil coated the release system. However, if we compare the similar velocities used for all experiments (49 – 51 m/s), we observe a d50 of; “Single nozzle”: 297 μm, “Star”: 308 μm and “Grid”: 480 μm.

To obtain comparable jetting velocities with the “Star” configuration, flow rates three times larger than for a “Single” configuration are needed. This imply that the water flow rates vary between 22 – 32% in the “Single Nozzle” experiments and 38 – 88% in the “Star” experiments (compared to oil). From an operational perspective, with an objective of keeping water rates < 50%, the “Single Nozzle” would be a more favourable configuration. However, in a full-scale scenario the difference between the diameter of the release (25 – 500 cm) on the area covered by a single nozzle (5 – 32 mm) will be larger. In such cases using multiple nozzle could be favourable to cover the entire cross section of the release. Comparing “Single” nozzle and “Star” configuration on a large-scale release with realistic oil droplet sizes will be an important objective for the planned large-scale testing at Ohmsett.

The objective of this work has been to use Computational Fluid Dynamics (CFD) to assess the viability of the subsea mechanical dispersion concept for conditions that have not yet been the subject of experimental testing. In this work it is shown that enhanced turbulent dissipation rates can be used as a metric for improved water jet nozzle performance. This modelling effort has been performed by Exponent. Only selected parts of this work are described in this section and the full report can be found in the technical report (Brandvik et al., 2019a).

In the earlier CFD work the focus was to numerically predict droplet sizes (d50) to justify that SSMD is feasible. Using current computational methodologies, this is highly computationally intensive and not practical to do at full scale. A more feasible use of CFD within the scope of this project is to predict the interaction of the oil plume and water jet without attempting to resolve droplet size distributions. This type of modelling is less computationally intensive and can be performed at full scale in 3-D. This approach provides an efficient way to make relative comparisons between different nozzle configurations and also provides a way to assess how the SSMD concept is expected to perform with more practical oil release sizes.

In order to make quantitative comparisons between different water nozzle configurations, a novel methodology that involves computing statistics on the turbulent dissipation rate history experienced by tracer particles released with the oil plume is utilized. The classical theory of droplet splitting in a turbulent flow field predicts a maximum stable droplet size (dmax) given by (Hinze, 1955):
formula

Here a is constant of proportionality, σ is the interfacial surface tension (oil-water), ρ is the density of the continuous phase, and ε is the stationary turbulent dissipation rate. The goal of SSMD is to use water jets to reduce oil droplet sizes, and to achieve this the water jet nozzles should be designed to maximize the turbulent dissipation rates experienced by the released oil.

To make this evaluation in the computational models, tracer particles are released at the location of the oil plume, and their trajectories, as dictated by the surrounding flow-field, are computed. Along each particle trajectory, the turbulent dissipation rate experienced by that particle is recorded as a function of time. In particular, it is shown that if the maximum turbulent dissipation rate (TDR) is recorded for each tracer particle, and then the median value of the maximum TDR for the whole tracer particle population is computed, this value correlates well with the observed droplet size distribution (d50). This is demonstrated through modelling of the existing experimental measurements performed by SINTEF.

The distribution of this TDR metric throughout the tracer particle population is also postulated to correlate with the shape of the droplet size distribution. If the TDR metric is relatively uniform across the tracer particles then this would indicate a log-normal size distribution with a prominent peak, whereas a non-uniform TDR metric may indicate a broader droplet size distribution.

4.1 Turbulent Dissipation Rate Metric Validation Study

To test the validity of the TDR metric, the approach was first tested on the results from the oil-only experiments conducted by SINTEF as part of the Phase II study. This was originally planned to be performed for the single horizontal nozzle and for the Star configuration for cases where cavitation was not observed. Cavitation is expected to create an additional mechanism for droplet breakup that would not be captured by the TDR metric. However, since experiments were also performed under high pressure in the SINTEF TiTank as part of Phase III, results from high water velocity experiments with a combined release of oil and air were also included in the evaluation of the method. The combined oil and air release cases were also used to determine if the modelling methodology could identify the best performing nozzle configuration.

The predicted TDR for the two selected horizontal nozzle cases is illustrated in Figure 4.1. This figure shows the region of high TDR values is clearly larger for the high water flow rate nozzle (135 mL/min, 91 m/s) compared to the low flow rate nozzle (70 mL/min, 51 m/s). This larger region of high TDR values increases the chances that an oil droplet will experience these values.

Figure 4.1:

Illustration of TDR for the two selected experiments from Phase-III experimental work (Horizontal single nozzle, 1a and 1c, 70 and 135 ml//min or 51 and 99 m/s.), see Brandvik et al., 2019a for further details.

Figure 4.1:

Illustration of TDR for the two selected experiments from Phase-III experimental work (Horizontal single nozzle, 1a and 1c, 70 and 135 ml//min or 51 and 99 m/s.), see Brandvik et al., 2019a for further details.

Close modal

The tracer particle trajectories with associated TDR values are plotted in Figure 4.2. Here the more penetrating effect of the higher velocity water jet is clearly illustrated. It should be noted that in Figure 4.2 and Figure 4.1 the red colour indicates the TDR is greater than or equal to the maximum value indicated on the colour bar. This was done for more equitable visual comparison between cases, but it should be noted that the max TDR values in Case 1c are much higher than in Case 1a, even though they are both red in the figures.

Figure 4.2:

Tracer particle TDR history for the two selected experiments from Phase-III experimental work (Horizontal single nozzle, 1a and 1c, 70 and 135 ml//min or 51 and 99 m/s.

Figure 4.2:

Tracer particle TDR history for the two selected experiments from Phase-III experimental work (Horizontal single nozzle, 1a and 1c, 70 and 135 ml//min or 51 and 99 m/s.

Close modal

To isolate the contribution of the water jet to droplet breakup a normalized turbulent dissipation rate metric can be defined. This is defined by dividing the traditional median of max TDR above the vertical threshold by the median of max TDR defined below the vertical threshold illustrated in Figure 4.2. The vertical threshold is selected to capture the max TDR values just downstream from the oil release point, but not capture any impact of the water jet. This normalized metric indicates the TDR contribution of the water jet relative to the TDR contribution that naturally occurs as the oil-air jet is released. Thus, normalized values greater than unity signify greater relative contribution of the water jet, whereas values less than unity indicate the water jet is likely not contributing significantly to droplet breakup.

The complete validation study, as discussed in the full technical report (Brandvik et al., 2019a), reveals this normalized metric is the best metric to use when comparing the relative performance of various nozzle configurations. Using this metric, it is shown that a single horizontal water jet is predicted to generate the best droplet breakup while using the least amount of water, as compared to the Star and Grid nozzle configurations. This conclusion is affirmed by the experiments conducted at SINTEF.

4.2 Modelling SSMD Effectiveness at Larger Scales

Additional modelling was performed to evaluate expected SSMD effectiveness for the larger scale experiments planned to be performed at the Ohmsett facility and an even larger scale field release planned performed in open water field experiment. While the Lab-scale cases consider release diameters of approximately 1 mm, the Ohmsett-scale cases consider a 25 mm release and the Field-scale cases consider a 250 mm release. In the planned Ohmsett experiments, the release point will be towed at a steady velocity to simulate a side current and this effect is also included in the Ohmsett-scale models. As illustrated in Figure 4.3, the side current causes increased oil jet dispersion, leading to the water jet influencing a smaller fraction of the oil than in the Field Case which does not include a side current.

Figure 4.3

Comparison of tracer particle trajectories and TDR levels in the water jet for corresponding Ohmsett and Field-scale cases.

Figure 4.3

Comparison of tracer particle trajectories and TDR levels in the water jet for corresponding Ohmsett and Field-scale cases.

Close modal

The actual towing speed (to induce a side current) used during the Ohmsett experiments ended up in the 0.18 – 0.25 m/s range and gave higher effectiveness' than indicated by the modelling in Figure 4.4.

Figure 4.4.

Normalized median of max TDR versus water usage across the Field, Ohmsett, and Laboratory scales for a single horizontal nozzle. The water jet velocity is indicated above each data point and the oil-air mixture release velocity is denoted by Voil.

Figure 4.4.

Normalized median of max TDR versus water usage across the Field, Ohmsett, and Laboratory scales for a single horizontal nozzle. The water jet velocity is indicated above each data point and the oil-air mixture release velocity is denoted by Voil.

Close modal

Summary results for the scaling study are presented in Figure 4.4. The chart illustrates that for the same water jet velocity the relative impact of the water jet, compared to naturally occurring TDR levels, is expected to increase when moving from the laboratory scale to the field scale. If the effect of the side current were removed, the Ohmsett cases are expected to be shifted upward in performance relative to the values plotted (see explanation above). These results optimistically suggest that provided the water jet(s) influence a significant fraction of the release, then meaningful enhancements in droplet breakup can be achieved even without having to go to the extreme (99 m/s) velocities that were required at the laboratory scale. The full-scale prototype illustrated in Figure 5.1 (Star configuration) is designed to operate with a water jetting velocity in the 30 – 40 m/s range.

Figure 5.1:

Water jet system setup (pump skid and nozzle manifold) with three 31 mm nozzles in a Star-configuration and a 4 m3/min water flow rate. Estimated capacity to treat a 12 000 m3/day (50% water treatment).

Figure 5.1:

Water jet system setup (pump skid and nozzle manifold) with three 31 mm nozzles in a Star-configuration and a 4 m3/min water flow rate. Estimated capacity to treat a 12 000 m3/day (50% water treatment).

Close modal

The main objective with this part of the study was to design a full-scale prototype for SSMD based on water jetting. This includes a subsea pump and nozzle system all operated by a ROV.

For water jetting to become an operational method for SSMD, high capacity pumps together with suitable nozzles and ROV arrangements are needed. As a part of the phase-II Oceaneering performed a market survey and identified submersible pumps as the most suitable options for subsea water jetting. An alternative to submersible pumps could be a topside approach, but the operational challenges with pumps located on a response vessel, bringing the water down with flexible tube are significant. Both pressure-drop, and handling of the hose will limit the operational water depth for a topside approach.

Conceptional designs of equipment, procedures needed to perform SSMD and tentative costs was worked out by Oceaneering. Only a summary figure from this work (Figure 5.1) is presented and further details can be found in the technical report (Brandvik et al., 2019a).

6.1 The Effect of Combined Releases on SSMD Effectiveness

The results from the non-cavitating and experiments with combined releases are summarized and compared to results obtained by simulated subsea dispersant injection (SSDI) in the table below. The droplet size reduction factor is the ratio between the untreated and treated oil droplets.

6.2 Different Nozzle Configurations

  1. Single horizontal nozzle and Star-configurations gave comparable results in the small-scale testing, with the Star-configuration performing slightly better, but at a cost of higher water rates. The Grid-configuration showed a lower effectiveness.

  2. The high efficiency at reduced water consumption for the single nozzle approach is clearly operationally favourable.

  3. Multiple nozzles could be needed to treat a larger cross section of a release and could be more applicable in a full-scale situation.

  4. The modelling showed less favourable results for the Star configuration compared to the single nozzle, probably due to a perfect alignment of the jetting beams, cancelling out each other in the modelling.

6.3 Modelling of SSMD Effectiveness for Large-Scale Releases

  1. The modelling of the planned Ohmsett testing gave reduced effectiveness due to spreading of the released oil prior to water jetting. The widening of the oil plume was caused by the high (max) towing speed used in the modelling (0.5 m/s). However, the towing speeds used at Ohmsett will wary in the 0.18 – 0.25 m/s range.

  2. The full-scale modelling (Macondo conditions) showed very promising results compared to the small-scale modelling and lab-results.

  3. Water jetting effectiveness is generally expected to increase at larger scales, since higher water rates, probably will compensate for reduced water velocities (20–40 m/s) compared to laboratory testing (40 – 90 m/s). This is indicated by very high normalized TDR-metrics (modelling) and very high Reynolds numbers (nozzle flow calculations) for large- and full-scale water jetting.

6.4 Summary Conclusions

  1. Small-scale experiments indicate a SSMD effectiveness (water jetting) comparable with SSDI.

  2. Modelling indicate an increase in effectiveness at larger scale (wider release diameters and higher oil flow rates).

  3. Designed prototype shows that full-scale SSMD (water jetting) can be performed with the main components based on available subsea equipment (ROVs and subsea pumps).

This study has received funding from multiple energy companies and this long-term support is highly appreciated by the authors. The contact persons at these companies are thanked for valuable discussions and comments; P.A. Beynet; M. Agrawal and P. Evans (BP UK), T. McKeever and K. Heitnes Hofstad (Equinor), G. Kjeilen-Eilertsen, T. Merzi and A. Cramer (Total) and A. Kelley (Lundin).

Ahnell,
A.,
Davies
E.J.,
Leirvik
F.,
and
Brandvik
P.J.,
2018
.
“Optical Monitoring of Subsea Blowout Droplets and Subsea Dispersant Efficacy”
,
Proceedings of the Forty-first AMOP Technical Seminar
,
Environment and Climate Change Canada
,
Ottawa, ON, Canada
,
pp.
90
111
,
2018.
Brandvik,
P.J.,
Johansen,
Ø.,
Leirvik,
F.,
Farooq,
U.,
and
Daling,
P.S.
2013
.
Droplet breakup in sub-surface oil releases – Part 1: Experimental study of droplet breakup and effectiveness of dispersant injection
.
Mar. Pollut. Bull. 2013
Volume 73
,
Issue 1
,
15 2013, pp.
319
326
.
Brandvik,
P.J.,
Davies
E.J.,
Krause
D.F.,
Beynet
P.A.;
Agrawal
M.
and
Evans
P.
2016
.
Subsea mechanical dispersion, adding to the toolkit of oil spill response technology
.
In
proceedings from Society of Petroleum Engineers - SPE International Conference and Exhibition on Health, Safety, Security, Environment, and Social Responsibility
,
Stavanger, Norway
,
April. 2016.
Brandvik
P.J.,
Krause
D.F.,
Owens,
Z.,
Gilman,
L.,
Yun,
M.,
Carlsen,
T.,
Koranek,
M.
2019a
.
Subsea Mechanical Dispersions – a Feasibility Study Phase-III
.
Technical Report. SINTEF Technical report,
pp.
70
.
Trondheim, Norway.
Brandvik,
P.J.,
Daling,
P.S.,
Dunnebier,
D.
and
Makatounis,
P.,
Leirvik,
F.
and
Krause,
D.F
2019b
.
A Proposed New Laboratory Protocol for Dispersant Effectiveness Testing Adapted for Subsea Dispersant Injection
.
Journal of Environmental Protection
.
Volume 10
,
May 2019 pp.
694
709
.
Davies
E.J.,
Brandvik
P.J.,
Krause
D.F.
and
Johansen
Ø.
2015
.
Properties of mechanical dispersions from subsurface releases – an experimental study in the SINTEF Tower Basin
,
SINTEF Report A27123 (unrestricted), Trondheim Norway.
Davies,
Brandvik
and
Leirvik
2017
:
The use of spectral transmittance imaging to size and classify suspended particulate matter in seawater
.
Mar. Pollut. Bull. 2017
,
vol. 115
Issue 1–2
pp.
402
410
.
Hinze,
J.O.,
1955
.
Fundamentals of the hydrodynamic mechanism of splitting in dispersion processes
.
A.I.Ch.E. Journal
,
Vol. 1
,
pp.
289
295
.

Author notes

Contact person: Per Johan Brandvik, PhD., Senior Scientist/Professor, SINTEF Ocean AS, Marine Environmental Technology, 4762 Trondheim, Norway. Email: per.brandvik@sintef.no. Phone: +47 9095 8576