The paper presents the outline of the Spill Detection and Recognition system – SpiDeR and its application to underwater oil and gas detection, classification and source characterization demonstrated in the remote-sensing survey of Mississippi Canyon area in the Gulf of Mexico founded by BSEE in 2017. The main objective of the operation was to deploy sensor package from a remotely-operated vehicle (ROV) to survey, detect, and map the location(s) of hydrocarbon emissions that are responsible for the surface oil spill and sheen footprint in the Mississippi Canyon Area. The objectives have been accomplished by conducting a multi-day, three-part survey mapping the area of interest, generation of georeferenced charts and 3D visualizations with detected oil active spills, all supported by a ROV intervention outfitted with oil spill detection and recognition system SpiDeR.

SpiDeR is a modular sensor suite capable of detecting, recognizing the source and classifying the hydrocarbon underwater leaks. The sensor suit with selectable configuration can be installed on any type of ROV vehicle and interfaces to the ROV with a single cable conducting the power and data. The presented here and used during the mission complete sensor suite consist of two 3D, broad band, electronically scanning multibeam sonar systems NORBIT WBMS STX, one Forward Looking Sonar NORBIT WBMS FLS, fluorescent oil classifier LIF – Laser Induced Fluorescence detection unit and the video camera with lights.

The most useful capability of the SpiDeR is the ability to generate 3D imagery (georeferenced bathymetry) even when the ROV is not moving. That combined with time gives 4D observable capabilities of the oil spill. The 4D capabilities have been proven useful during the u-bathymetry part in Phase 2 and forward-looking 3D in Phase 3 of this mission.

The system has been deployed from the ROV in the area where it has been known for the last decade that the leak of hydrocarbons is coming from. The real task at hand was to recognize the leak source and that source contain hydrocarbons and accurately document the source location and provide measurable documentation of its character.

The particular project where the SpiDeR was deployed was to conduct a multi-day, three-part survey covering the area of interest (1,000 ft. x 1,000 ft.) supported by a ROV intervention outfitted with oil spill detection and recognition system.

The leakage detection mission was divided into three phases. Phase 1 - General Bathymetric Survey, Phase 2 - Precision Bathymetric Survey and Phase 3 - Source Recognition Survey.

An outline of these three phases is as follow.

1.1. Phase 1 - General Bathymetric Survey.

The operation to identify the location(s) of the sheen source(s) starts with the high-resolution bathymetric survey of the area under consideration using 3D scanning sonar looking down. During the survey the bathymetry data, as well as full water column data were collected. The objective was to perform general survey and to locate any potential hazards in the area, either suspended in the water column or on the seabed.

1.2. Phase 2 - Precision Bathymetric Survey.

After the preliminary processing was done and all potential hazards were identified, the secondary bathymetry survey was performed from lower altitude to obtain a higher resolution of the mapped area.

1.3. Phase 3 - Source Recognition Survey.

Once the area(s) of interest was located, the source recognition component started with the ROV navigating to that particular area and use the sensor packages to classify the sheen source and verify it contains hydrocarbons.

2. Mission preparation

To satisfy the mission objectives, NORBIT Subsea has developed a modular sensor suite to detect, recognize and classify underwater oil leakages called SpiDeR (Spill Detection and Recognition). The following chapters outline the system characterization in the view of the goals of the mission.

3. SpiDeR – Spill Detection and Recognition system

SpiDeR is a modular sensor suite capable of detecting, recognizing the source and classifying the hydrocarbon underwater leaks. The module can be installed on any type of ROV vehicle and interfaces to the ROV with a single cable conducting the power and data. The sensors being part of SpiDeR are listed in the Fig. 1 and consists of three sonar systems, lights, video camera, pan and tilt mechanism and the fluorescent laser based unit LIF – Laser Induced Fluorescence detection unit.

Fig. 1

SpiDeR modular system and its components

Fig. 1

SpiDeR modular system and its components

Close modal

All devices are connected to the pressure vessel which contains necessary power conditioning and data connections. It also passes the trigger and timing information from SpiDeR to ROV over the same cable.

SpiDeR is powered with 48V with 6A current capabilities. The power consumption of the entire module depends on the use scenario and varies between 100W and 250W.

There are 2D multibeam, camera, lights and LIF detector unit which are mounted on the pan and tilt mechanism allowing free rotation in horizontal and vertical planes. The video camera is capable of recording the video but also periodically taking still high-resolution images. This feature has been found very useful in analyzing the plumes. The 2D Forward Looking Sonar (FLS) is mounted vertically allowing for wide vertical coverage (180deg) and offers excellent navigation and visualization aid. It is also equipped with the object detection capabilities useful for autonomous missions.

The LIF (Laser Induced Fluorescence) detector is used as classifier whether the leak contains the hydrocarbons or not. It is a point source and needs to be positioned in a close vicinity of the plume to perform detection. The range will depend on the turbidity of the water.

There are two 3D scanning systems, which are affixed to the base plate. One of the 3D scanning sonars is pointing down and is used to perform bathymetry operation similarly to other NORBIT bathymetry systems.

The STX 3D&4D imaging sonar is shown in Fig. 2. is based on a proven WBMS platform but the transmitting antenna has a new capability of electronically changing the direction of the emitted sound wave. That process is called beam steering or similarly to the receiving beamforming called transmit beamforming operation. The transmit and receive beamforming processes are outlined in Fig. 3.

Fig. 2,

NORBIT STX - 3D&4D multibeam

Fig. 2,

NORBIT STX - 3D&4D multibeam

Close modal
Fig. 3,

Transmit and receive beamforming in STX

Fig. 3,

Transmit and receive beamforming in STX

Close modal

The most useful capabilities of the NORBIT STX sonar system is the ability to generate 3D imagery (bathymetry or raster image) even when the vessel is not moving. That combined with time gives 4D observable capabilities of the STX. The 4D capabilities have been demonstrated during the u-bathymetry part in Phase 2 and forward looking 3D in Phase 3 of this mission.

Perhaps the most challenging problem for reconnaissance and recovery operation using underwater acoustic equipment is the inability for the acoustic wave to penetrate hard structures and consequent shadows causing gaps in the coverage. While for some application it is desired to see shadows (e.g. mine-hunting and side-scan operation), for bathymetry mission the goal is to cover as much of the sea bottom as possible and eliminate the unwanted holidays in coverage. In some cases, it is possible to go around the structures with the vessel and fill up the gaps, but in other cases, it may deem impossible due to the presence of structures or lack of access or time. In such cases, the mission operators do not need to live with gaps in the data from these areas and can utilize the NORBIT STX.

The STX with its unique capabilities to steer the transmit beam can also “look” behind the structures in a way that standard multibeam cannot. It also allows for a more accurate representation of the underwater structures as it increases the number of observable angles at which the object is seen. It is intuitively very simple when one passes an object which can only be seen in one direction, then it will be visualized from that side only. If, however, one can look forward and backward at the same object, the number of observables is much increased and captures more details of the underwater structures.

The above has been demonstrated by surveying close to the jacket area in u-bathymetry Phase 2 of this project.

3.1. Integration with the ROV.

The multidisciplinary sensor suite SpiDeR facilitates the integration operation due to simplistic mechanical and electrical interfacing. The mechanics are limited to fastening the mounting frame of the module onto the ROV frame and ballasting for weight. The electrical interfacing consists of a single cable which interfaces directly to the main multiplexer of the ROV. This time Matrix MK II from Innova technology was used. This multiplexer offers two Gb Ethernet connections one was used for the SpiDeR and one was used for the government instruments running two single beam echo sounder transducers. The multiplexer also provided timing pulse (PPS). The SpideR also triggered the acoustic transmission of the DVL attached to the INS system to avoid interference. The trigger signal is available at the same connection as power and data.

Fig. 4

SpiDeR integration onto ROV

Fig. 4

SpiDeR integration onto ROV

Close modal

The SpiDeR module was integrated in a short time and tested for operation in the test tank. Special care had to be taken to allow the bathymetry system to have unobstructed view through the ROV's base plate this was accomplished by cutting out a square in the plate, However, this limited the view of the down looking sonar to roughly 120deg. which was not a big problem for this application.

In addition, the ROV was also mobilized with transponder beacons,INS and 1200kHz DVL System for navigation purposes.

3.2. Testing in the test tank

Testing in the test tank comprises of buoyancy test and checking of all on board systems, such as propelling system, lights, cameras, powering of the devices and telemetry of the data over the fiber.

In the Fig. 5 below one can see the LIF laser beam bouncing off the water surface. It is interesting to zoom a little bit at the beam cross-section at the water surface to see some shining fluorescing particles of motor oil present in the test tank.

Fig. 5

ROV in the test tank (left), Lifting ROV off the test tank facility in Robert, LA (right)

Fig. 5

ROV in the test tank (left), Lifting ROV off the test tank facility in Robert, LA (right)

Close modal

3.3. Integration of the ROV on the vessel

The vessel used for operation was a 170 ft x 36 ft DP1 mini- supply vessel M/V Gerry Bordelon built in 2011 with a gross tonnage just below 500 tons. The vessel offers accommodations for up 22 passengers and has a usable deck area of 110 ft. X 30 ft. The vessel comes with no infrastructure on the deck and all equipment needed to be secured on the deck before the trial could take place.

Fig. 6

M/V Gerry Bordelon - vessel for operation

Fig. 6

M/V Gerry Bordelon - vessel for operation

Close modal
Fig. 7

ROV integration on the vessel

Fig. 7

ROV integration on the vessel

Close modal

4. Mission execution

4.1. Location of the survey site

The project area is located around 95 miles South-East of New Orleans.

Fig. 8

The location of the survey site (Source: Google Earth) and historical photo before tower collapse.

Fig. 8

The location of the survey site (Source: Google Earth) and historical photo before tower collapse.

Close modal

5. Phase 1 Bathymetry survey

Before the general bathymetry survey starts, the ROV was prepared and checked for all subsystems.

The USBL system is visible in Fig. 9. The picture on the right-hand side shows the pole mount and several devices mounted on that pole.

Fig. 9

ROV being launched from the starboard side (left), the USBL positioning system is mounted on the pole (right).

Fig. 9

ROV being launched from the starboard side (left), the USBL positioning system is mounted on the pole (right).

Close modal
Fig. 10

The area of 1000 x 1000ft covering the location of the jacket and the original location.

Fig. 10

The area of 1000 x 1000ft covering the location of the jacket and the original location.

Close modal

In the center part of the pole, the USBL system is mounted. The USBL system with internal motion sensor with the beacon being installed on the ROV side providing real-time transponder to the unit.

The general bathymetry was conducted from the ROV surveying between 50m-90m altitude using the bathymetry 3D mapping sonar. This was accomplished by running parallel lines until the area of interest was covered (Fig. 12).

Along with bathymetry acquisition, water column data was collected. Water column data allows suspended plumes to be observed as well as to detect any hazard in the water column (see Fig. 11). Real-time maps are generated while surveying allowing the surveyors to quickly assess the insonified area. After the survey, post-processing will take place to clean the data and to identify the area(s) of interest (potential sheen source location(s)).

Fig. 11

Different captures of the main plume character

Fig. 11

Different captures of the main plume character

Close modal

As indicated on the Fig. 11 (left) the main plume has been detected to extend from the bottom up to the surveying altitude, most likely proceeding up to the water surface. The plume grows in size and changes directions. Sometimes observed to reverse two times before reaching the top due to currents.

It could also be seen that the plume separates and more buoyant bubbles, having higher vertical speed, reach the surface while the less buoyant (perhaps dissolved hydrocarbons) separate and stay in the water column. Taking a different view at the plume from different angle (right-hand side of Fig. 11) it can be seen that there are actually two plumes, not one as initially believed. The exact location and recognition of the source will be done in Phase 2.

During the initial survey, it was also possible to see the hazard suspended roughly 35 m above the jacket (right-hand side of Fig. 12). Initially, it was not clear what this hazard was but then later using the forward looking sonars it was shown to be some sort of a structure attached to the jacket with metal rods (left two images in Fig. 12).

Fig. 12

Images showing the hazard located in the area of the main leak

Fig. 12

Images showing the hazard located in the area of the main leak

Close modal

The hazard is believed to be a collapsed collector dome structure left from the past. The remains are located at roughly 80m below the surface and do not pose a danger to general navigation, but it may cause trouble for any towed objects such as fishnets if they happen to be in that area.

5.1. General bathymetry survey results

During the data processing of Phase 1 it was possible to capture the scene from various survey lines and various directions. An instructive illustration of this phase effort is presented in the figure Fig. 13.

Fig. 13

The general bathymetry survey already reveled the plume character

Fig. 13

The general bathymetry survey already reveled the plume character

Close modal

During the processing, a couple of new leaks were found. However, they were smaller than the main leak. All findings are outlined in the figure Fig. 14. In total four potential leaks were located and one navigation hazard.

Fig. 14

Location of potential leaks in the area

Fig. 14

Location of potential leaks in the area

Close modal

The leak depicted as Leak 2 which is located in the original installation angle of the jacket before the collapse is believed to be gas only.

6. Phase 2 - u-Bathymetry survey

Once the general bathymetry was performed, and the hazards, as well as the main plume, was identified then the u-bathymetry phase was conducted. The u-bathymetry consisted of a high-resolution survey in the given locations where the leak was thought to originate. The detailed, high-resolution bathymetric survey requires the ROV to be operated at a lower altitude which has been determined to be roughly 20m above the seafloor. The bathymetry 3D mapping sonar will be used to collect bathymetry and water column data to map the exact location of the source and prepare for the classification phase.

A sample frame of the u-bathymetry is presented in Fig. 15 showing the complicated structure just below the ROV and two leaks on the left-hand side.

Fig. 15

The u-bathymetry snap shot using 3D scanning down-looking sonar.

Fig. 15

The u-bathymetry snap shot using 3D scanning down-looking sonar.

Close modal

The online processing was followed by onboard processing of the collected data and identification of the potential sheen source area(s). The result of the processing is presented in Fig. 16. The u-bathymetry aims to heavily oversample the space and statistically, average the false data to map the features with more details and with better coverage.

Fig. 16

The resulting point cloud of the u-bathymetry using 3D scanning down-looking sonar. Structure (jacket) is colored in blue. The leaks are colored in yellow.

Fig. 16

The resulting point cloud of the u-bathymetry using 3D scanning down-looking sonar. Structure (jacket) is colored in blue. The leaks are colored in yellow.

Close modal

As one can imagine a standard method of surveying of this type of environment results in many gaps behind the steel legs of the underwater structure as sound cannot penetrate hard objects. It may take many attempts to position the ROV at different angles to try to hit the target with different angles and sometimes it's an impossible mission.

But when the STX sweeping capabilities are used, a single line covers all looking angles and allows a quick and efficient cover of the area without gaps and positioning errors.

For the visualization purposes, it is instructive to overlay the raster sonar image onto the digital point cloud to obtain the confidence in the detection scheme (Fig. 17).

Fig. 17

The u-bathymetry using 3D scanning down-looking sonar with a single snap shot overlaid on the digital point cloud data. Structure (jacket) is colored in blue. The detected leaks are colored in yellow.

Fig. 17

The u-bathymetry using 3D scanning down-looking sonar with a single snap shot overlaid on the digital point cloud data. Structure (jacket) is colored in blue. The detected leaks are colored in yellow.

Close modal

The result of u-bathymetry and Phases 1 & 2 is the exact location of the leaks and the visualization allowing decisions on the last phase of the plan – the recognition and classification phase 3.

7. Phase 3 - Recognition and classification

Once the Phase 1 and Phase 2 are done, the exact locations of the leaks are identified as well as any hazardous object which needs to be avoided. The chart in Fig. 14 showed the locations of the above.

Leak source recognition consists of navigating the ROV in close vicinity of the areas of interest identified in the bathymetry surveys from previous days. The forward-looking 4D sonar and FLS sonar will now be used for acoustic sampling. The ROV operator has safely navigated to a proper location of the ROV to ensure best possible visibility of the source sites for the cameras and laser unit. The LIF (Laser Induced Fluorescence) detector unit has been used along with the video camera to classify the leak to contain the hydrocarbons.

Fig. 18

Image from ROV camera showing LIF beam (center) and lights (left) for video camera (right)

Fig. 18

Image from ROV camera showing LIF beam (center) and lights (left) for video camera (right)

Close modal

The LIF detector unit contains specialized optics which filters only the fluorescent response corresponding to the hydrocarbon emission. The device has been manufactured by partner Spectra Solutions and has been successfully used in the past to classify the hydrocarbon plumes. However, in order for the LIF detector to be able to operate it needs to be put in a close vicinity of the plume, so the laser light has a chance to reach it. The performance will greatly depend on the turbidity of the water and the closer one could get to the plume the better the chances for success. To navigate the ROV and LIF unit in the close vicinity of the plume the two forward-looking sonars were used.

To visualize the mutual relation of the vertical and horizontal sonar the following image is presented.

Fig. 19

Images from vertical and horizontal forward lookers give full picture

Fig. 19

Images from vertical and horizontal forward lookers give full picture

Close modal

The combination of the two sonars has proven to be very efficient in navigating the ROV and helpful visualizing the plume. With the horizontally looking sonar it was easy to instruct the ROV navigator to perform the ROV positioning in the required manner, and with the vertical sonar, it was possible to lower the ROV to the save altitude without the risk of hitting any objects.

Fig. 20

Two plumes coming from two sources visible on vertical (left) 2D and horizontal (right) 3D-FLS sonar imagery

Fig. 20

Two plumes coming from two sources visible on vertical (left) 2D and horizontal (right) 3D-FLS sonar imagery

Close modal

Superimposing the sonar single pings onto the grid model and the point cloud gives a good visual interpretation of the environment.

Fig. 21

Vertical and horizontal sonar images overlaid with the DTM and point cloud at the site.

Fig. 21

Vertical and horizontal sonar images overlaid with the DTM and point cloud at the site.

Close modal

With such presentation, it is possible to see the precise navigation needed to conduct safe operation with the ROV. The horizontal sonar cuts the space in the horizontal plane, yet scanning the jacket up and down building its three-dimensional point cloud. When at the same time the vertically looking FLS sonar offers the full display of the entire scenery in the active sonar sector. Both of them give a full picture of the situation developing in the dynamic underwater environment where the video cameras do not offer any aid due to low visibility.

Fig. 22

Vertical and horizontal sonar images overlaid with the DTM and point cloud at the site.

Fig. 22

Vertical and horizontal sonar images overlaid with the DTM and point cloud at the site.

Close modal

The entire positioning procedure allowed getting the ROV to a very close proximity of the plumes and to enter the classification part of Phase 3 with the LIF detector.

The LIF operates using laser light and as such is subject to the law of physics of optical light. That means its range will depend on the turbidity of the medium. However, provided it gets to the detection range, which in this case we found to be 2–4m, it offers unsurpassed classification capabilities by being able to distinguish between a gas leak and an oil leak.

In Fig. 23 one can see raw data coming from the LIF detector. The first part of the data is collected when the ROV was navigating toward the plume. Then by the end of the data plot the laser cuts through the plume and finds the large content of hydrocarbons.

Fig. 23

LIF classification of the plume (top) and video camera (bottom) showing shining fluorescent particles passing by the laser beam

Fig. 23

LIF classification of the plume (top) and video camera (bottom) showing shining fluorescent particles passing by the laser beam

Close modal
Fig. 24

Oil samples collection using Niskin bottle and robotic arms to release the lock-in mechanism.

Fig. 24

Oil samples collection using Niskin bottle and robotic arms to release the lock-in mechanism.

Close modal

Further inspection of the video camera also confirms the shining fluorescent objects passing by the laser beam. These are the oil blobs which contribute to the large spikes on the laser display. The general increase in the background of the LIF readings is attributed to the increased content of dissolved hydrocarbons in the plume.

After the leak is recognized the oil sampling has been taken with Niskin Bottles already mounted on the ROV.

With the help of the FLS sonars (as before) the bottles have been positioned as close to the source as possible to capture samples of the oil. The bottles were closed using the robotic arms on the ROV and securely delivered to further lab analyses which were confirmed to contain volatile compounds, crude oil, and petroleum hydrocarbons.

Conclusions

The paper outlines the use of the Spill Detection and Recognition module SpiDeR integrated on the ROV used for the operation.

The location of the leak of interest along with recognition of the source and classification of the leak have been successfully performed. It was shown that the leak contained petroleum hydrocarbons in the close vicinity of the jacket in the location outlined in this document.

[1]
Scott Pegau
W.,
et al,
Detection of oil on-in-and-under ice., Report from Joint Industry Programme on oil spill detection and mapping in low visibility and ice,
September
2015
[2]
Bello
Job,
Eriksen
Peter,
and
Pocwiardowski
Pawel,
Oil Leak Detections with a Combined Telescopic Fluorescence Sensor and a Wide Band Multibeam Sonar, 2017
International Oil Spill Conference