Special report83a7383a5e33475eed0e-e819cda5edf0a946af164bb0b2f2ae3c.r0.cf1.rackcdn.com/...Special report Transforming Subsurface ... Mr Holdaway was followed by a talk on seabed seismic

Special reportTransforming Subsurface InterpretationFinding Petroleum forum in London - Apr 13 2015

SAS - data mining on subsurface data

Seabed Geo Solutions - why seabed seismic recording gives better images

Dynamic Graphics - using fibre optic welldata with reservoir models

Geodirk - an automated system for usinggeology with seismic processing

Official publication of Finding Petroleum

Special event report - April 2015

Contact us [email protected]

For ultimate seismic data, in any marine environment, we deliver the most versatile seabed solutions.

se seitamitlor uFe mhr teivle dew

e eniray mnn a, iataic dmsd soebae selitarset vsoe m

, tnemonrivne e.siontuld so

s utactonC nfi onf @ mco.sgbs

Data mining in the oil andgas industry is about work-ing out which dependentand independent data rela-tionships are identified andmapped to your businessobjectives, said KeithHoldaway, advisory indus-try consultant with SAS In-stitute, speaking at the Apr13 Finding Petroleum Lon-don forum “TransformingSubsurface Interpretation”.

It usually involves both data scientists anddomain (oil and gas) experts, he said.

Mr Holdaway worked as a geophysicist atShell for 15 years, and subsequently has beenworking on ways to develop ‘machine learn-ing’ systems to try to complement traditionalways of looking at oil and gas data.

He is also the author of a book, “Harness Oiland Gas Big Data with Analytics: OptimizeExploration and Production with Data DrivenModels”, published in May 2014 by Wiley.

Data mining techniques can be used to try towork out ways to make models with the data,and generating hypotheses worth modeling,

to help you achieve your results.

“We try to integrate some of these analyticalworkflows and see if the data tells you astory,” he said.

All of the work needs to be geared aroundworking out an ‘objective function’, the rela-tionship between data elements which willhelp you achieve the objective you are look-ing for (usually, more oil).

A common trap is that people get very ex-cited with the relationships they discover, butthe relationships don’t actually help achievethe business objective, he said.

“Putting these [models] together from a datascientist perspective is one thing, but youmust be able to operationalise these modelsin an existing architecture. If you can't oper-ationalise you can't really gain,” he said.

Deterministic and probabilistic

Subsurface oil and gas data has many vari-ables, and works best with probabilities, tohelp people make the best decisions, he said. You need to work out which variables will

help you work out the probabilities in a highdimension input space.

The oil and gas industry has traditionallyworked in a very deterministic way, lookingfor the right answers, using ‘first principles’based on traditional scientific calculationsand equations, he said.

Oil companies have got to a situation wherethey are only using 20 per cent of the datathey have, because they only use data whichenable these equations.

Keith Holdaway – how to do oil and gas data mining

This special edition of Digital Energy Jornal is a report from the Finding Petroleum forum in London on Apr 13 2015, Transforming Subsurface Interpretation.

Data mining in the oil and gas industry is about trying to work out which hidden trends and relationshipsin the data coupled with appropriate data-driven models will lead you to the right answer, with datascientists working together with domain experts, said Keith Holdaway of SAS

Transforming Subsurface Interpretation

In this example, the computer has found a way ofgrouping wells according to their characteristics. If thegrouping makes sense to a reservoir or production en-gineer, it might point to where other successful wellscould be found, if they have the same characteristicsas a group of successful wells.

3Digital Energy Journal - Special report, Apr 13 2015 Finding Petroleum forum "Transforming Subsurface Interpretation"

Event websitewww.findingpetroleum.com/60143.aspx

Report written by Karl Jeffery, editor of Digital Energy [email protected] Tel 44 208 150 5292

Sales manager Richard [email protected] Tel 44 208 150 5291

Conference produced by David Bamford

Layout by Laura Jones, Very Vermilion Ltd

Cover art by Alexandra Mckenzie

Digital Energy Journalwww.d-e-j.com

Future Energy Publishing, 39-41 North Road, London, N7 9DP, UK www.fuenp.com

Transforming Subsurface InterpretationFinding Petroleum's forum in London on April 13, 2015, "Transforming Subsurface Interpretation", looked at ways that subsurface interpretationcould be improved by using state of the art technology, tradition geophysics / geology skills, and structured work processes.The opening talk was from Keith Holdaway, advisory industry consultant with SAS Institute, and author of a book, “Harness Oil and Gas BigData with Analytics: Optimize Exploration and Production with Data Driven Models”, published in May 2014 by Wiley.Mr Holdaway was followed by a talk on seabed seismic recording with John Moses, Sales Director with Seabed Geosolutions.A session on data integration included a talk on gravity gradiometry recording and data integration with Claire Husband, Senior Geophysicistwith ARKeX, and John Brennan, Analytics and Data Management Strategy Lead, Oil & Gas, Hewlett-Packard.A session on data visualisation and analysis included talks from Mike Leach, workstation technologist with Lenovo, on using high performancecomputing; Ken Armitage, managing director of Geodirk, on integrating traditional geological skills with seismic data processing; and JaneWheelwright, technical application specialist with Dynamic Graphics, on integrating fibre optic data from wells with other subsurface data.This special Digital Energy Journal report includes an outline of all of the talks. For most of the speakers, the videos and slides are availablefree of charge online - you can see how to access from the links at the end of the article.Karl Jeffery, editor, Digital Energy Journal

Explaining how todo data mining in oil and gas: KeithHoldaway, advisoryindustry consultantwith SAS Institute


4

But there is a big move towards probabilisticapproaches, which most ‘data mining’ tech-niques include, he said.

The Society of Petroleum Engineers recentlycalculated that the number of papers with theterm ‘data mining’ in it had increased by 20xin the past 7-8 years, he said.

The data modelling approach can be usedalongside the more traditional calculations,so you are using traditional and non-tradi-tional ways of looking at seismic data to-gether.

These techniques can be applied to any E+Pdata set, he said.

Reservoir characterisation

A typical oil and gas business objective couldbe reservoir characterisation, using subsur-face data to try to determine a higher fidelitygeologic model.

Geophysicists calculate “seismic attributes”,a range of different data properties calculatedfrom the seismic, such as amplitudes, coher-ence and variance.

It would be useful if you could map seismicattributes to reservoir properties, but to do thisyou need to know which attributes are themost useful from a pre- and post- stack per-spective.

Automated data mining sensitivity techniquescan help you work out which attributes mightbe reflecting something useful, and so whichones are worth studying more closely.

As an example, you can get the data outputin the form of ‘self-organising maps’, whichcan help you spot patterns and relationshipswhich might show you direct hydrocarbon in-dicators.

To map seismic attributes to reservoir prop-erties, you need both data scientists and reser-voir experts.

“You can't just go off and talk to PhDs in sta-tistics, they come up with these algorithmswhich aren't totally applicable to your busi-ness,” he said.

“We need some kind of visualisation alongthe roadmap, so experienced people can say,yes you're going in the right direction here.We may need to tweak something to improvethat process,” he said.

Trying to find the right answer is an iterativeprocess, not a process where you can enteryour data into a ‘black box’ and get the an-swer out.

In one example, Mr Holdaway gave subsur-face data for a field in Oman, which con-tained a known oilfield, to someone with datamining expertise in the medical sector, but nogeophysics expertise, and without telling himwhere the oil was. Deep Learning workflowswere implemented to identify features indica-tive of DHIs based on a recursive set of neu-ral networks.

“I said, do your thing as you did in the otherindustry,” Mr Holdaway said.

“He identified other areas close to a produc-ing well. It worked very well.”

Understanding production

In another example, you might have one wellwhich delivers good oil production, and awell nearby which doesn’t deliver anything,and you can’t understand why.

You can apply data mining techniques to tryto identify the variables which are consistentin both regions, and then use those as a basisto understand the second well.

You could use this to come up with a betterfracturing strategy for the second well, per-haps with a different proppant, based on whatyou have identified as the more importantgeomechanics of the field.

In one big oil company example, the com-pany had an employee who was one of theworld’s biggest experts on reservoir waterlevel. He claimed that he knew everythingthere was to know about water level in thecompany’s reservoirs, and no computer sys-tem would be able to say why water cuts insome wells are higher than others.

The SAS data mining techniques showed thathe might be wrong in how he was interpretingthe fracture network.

“His ego wasn't so big that he entertained theidea,” Mr Holdaway said.

“He decided to change the location of an in-jector well and producer well and it obviatedsome of the problems.”

“The data told him a story and he tried to giveit some credence, and managed to increaseproduction in that asset.”

The systems have also been used on produc-tion data, by one oil company which nearlydestroyed part of its reservoir with waterflood.

The data mining techniques came up with aprobabilistic system which could forecastproduction for a particular well with 90 percent confidence, he said.

This helped the company work out whichstages of the well to leave open and which toclose.

“All those kind of answers came through in-tegrated production data, PLT data, and someof the geologic, petrophysical data,” he said.

Expertise

A company in Oklahoma City in the US hasput together an ‘analytical centre of excel-lence’, where it uses many different softwaretechniques to try to help upstream engineers,he said.

The company hired data scientists who havesome understanding of the oil and gas indus-try.

“You have to be able to steer these guys sothey are not trying to create algorithms, tryingto be Newton and Einstein, and come up withsomething incredible which isn't useful to thebusiness,” he said.

Sometimes you need a lot of expertise to de-velop the right methods, but once they havebeen developed, you have a repetitivemethodology which does not need so muchskill to use.

“There's a lot of interaction between data sci-entists and geoscientists,” he said.

“The output from predictive models is fedback into the traditional tools. It will give youa much more robust visualisation.”

Managing data

There are three different systems in upstreamoil and gas – the reservoir, wellbore and fa-cilities. All of these generate many differentsorts of data.

The data has many different forms, includingstructured, unstructured, “big” data, high ve-locity data, real time data, batch data, spatial(related to space) and temporal (related totime).

Compared to other industries, the volumes ofoil and gas data are not particularly large, but the variety of data can be very large, hesaid.

Digital Energy Journal - Special report, Apr 13 2015 Finding Petroleum forum "Transforming Subsurface Interpretation"

The oil and gas has so much different data, there aremore possible relationships and correlations than ahuman being could figure out.


Seabed Geosolutions – acquiring seismic on the seabedThere is growing interest in recording seismic data with nodes on the seabed, particularly with areas hard toaccess with streamer vessels, said John Moses of Seabed Geosolutions

There is a growing interest in recording seismicdata with “nodes”, or small recording devices,placed on the seabed, as an alternative to tradi-tional streamers towed behind vessels, said JohnMoses, Sales Director with Seabed Geosolutions.

Seabed seismic can offer a much clearer picture,and give you more flexibility in how you doyour survey, than towed streamer, he said, speak-ing at the Finding Petroleum London conferenceon April 13, “Transforming Subsurface Interpre-tation”.

The cost per km2 of seabed acquisition is still rel-atively high compared to towed streamer meth-ods for anything but reservoir scale studies, butthe added value of information combined withevolving operational methods has been enoughto persuade many oil companies to take theplunge to the ocean bottom.

The company is a joint venture between Fugroand CGG.

In the current oil price environment, spendingmoney on good data can be a cost cutting measure, Mr Moses said, if it enables drilling tohit the reservoirs more efficiently. “It is aboutde-risking the [drilling] decisions,” he said.Good decisions have to be based on the best in-formation.

There are many areas of the world which havebeen surveyed many times by towed streamer,such as the North Sea. Using seabed seismicrecording could illuminate the fields in a newway, he said, which helps understanding of thereservoirs and optimises the way they can be ex-ploited.

Seabed seismic recording is not a new technol-ogy, but in the past has mainly only been usedfor areas which could not be accessed with con-ventional streamers, Mr Moses said. This wasbecause seabed seismic used to be relatively ex-pensive and had technical limitations. The latestadvances have removed the barriers that previ-ously existed and are also starting to close thecost gap. At the same time the uplift in the valueof information is now being recognised.

The quality of seismic data improves withseabed recording. The receivers are able torecord both P and S waves thanks to their 4 com-ponent sensors. The sources and receivers arecompletely decoupled which gives the oil com-pany complete freedom to design full offset andfull azimuth data sets, recorded in a quiet envi-ronment with full frequency bandwidth. Re-ceivers can be placed in obstructed and shallowareas which are inaccessible by conventionalmeans. Both imaging and reservoir attributes aredramatically enhanced and offshore field infra-structure is no longer a barrier to data acquisi-tion

“We are seeing a quantum step change in qualityof seismic data, using receivers on the seabed,”he said.

The Dan Field showing the reservoir and wells - there are 58 oil producers and 50 water injectors.

5

“Data management is one of the key issues,”he said. “We have to aggregate the data, in-tegrate in an analytical data warehouse.”

Data mining techniques

There are two basic data mining techniques,supervised and unsupervised.

With “supervised” data mining, you split thevariables into explanatory (or independent)variables and dependent variables. The aimis to try to find a relationship between inde-pendent and dependent variables. With “unsupervised” data mining, all vari-

ables are treated in the same way (there is nodistinction between explanatory and depend-ent variables) and you are trying to spot pat-terns and trends that cluster into characteristicprofiles.

The data mining methods will look at manydifferent models which could be used to giveyou the answer you are looking for.

Examples of data mining techniques includefuzzy logic, cluster analysis and neural net-works. “These are all approaches to let thedata do the talking,” he said.

“We have to quantify the uncertainty in thosevariables and evaluate the value inherent inthe patterns,” he said.

The process looks for trends and hidden pat-terns.

Principal component analysis is a methodol-ogy to reduce the high ‘dimension’ of thedata, if you can figure out which componentsare worth paying most attention to.

Watch Mr Holdaway’s talk on video atwww.findingpetroleum.com/video/1336.aspx


Helping you get a clearer subsurface picture by usingseabed data acquisition - John Moses, sales director,Seabed Geosolutions

Towing streamers in the vicinity of platforms isa risky undertaking. It can only be done whencurrent conditions are pushing the cables awayfrom the obstruction which inevitably leads to alarge gap in data coverage. Usually in the mostcritical area.

Technology

The technology, put simply, is to record seismicdata using devices on the seabed, rather thanfrom streamers towed behind a vessel.

The sensors (hydrophones or geophones) forrecording the data can either be fixed to cableslaid on the seafloor, or installed in standalone‘nodes’ which are laid by a remote operated ve-hicle (ROV)or attached to robust connectingropes.

The jargon is “Ocean Bottom Cable” (OBC) and“Ocean Bottom Nodes” (OBN).

Azimuth

With conventional towed streamer survey, theacoustic source is also towed behind the record-ing vessel. The sound waves echoing back fromthe sub-surface are only registered or sampledin a narrow azimuthal field.

Sometimes, a narrow azimuth is all you need,but a wider or full azimuth will give you muchmore ability to understand what you are lookingat, he said.

“If your geophysical challenges are harder, youneed more accurate information,” he said.

You can get multiple azimuth with a conven-tional towed streamer survey, if the vessel passesover the same area of subsurface three times indifferent directions. “But this multiplies yourcost by three,” he said.

You could also do it by having a number ofsource vessels making seismic waves from dif-

ferent directions, but this also means more costs.

But with the seismic data being recorded withreceivers on the seabed, you can record in anygeometry you like. You can create seismicwaves in different positions on the water surfaceand the nodes record everything they hear.

To illustrate the importance of multi-azimuth,consider if you were trying to work out the lo-cation of an object at the bottom of a swimmingpool in the night, using a torch.

The torch light bends (refracts) as it goes throughwater. But since you don’t know the depth of theswimming pool or the degree to which the lightrefracts, you can’t calculate the location by shin-ing the beam from one position.

The only way to do it is to shine the torch in theswimming pool from many directions and many

distances, and then you will have enough data tomake a calculation, he said.

Similarly, with seismic recording, if you canlook at a subsurface object from many differentazimuths and offsets, you can make a more ac-curate calculation and have confidence in locat-ing your sunken treasure.

Geophones and hydrophones

Another benefit of recording on the seabed isthat you can use geophones, which directlyrecord ground movement directly as well as hy-drophones.

Hydrophones can only record plane (P) waves,where the wave ups and downs are in the samedirection as the direction of travel of the wave,as you would get from hitting the surface ofwater.

But geophones can also record shear (S) waves,where the ups and downs of the wave are per-pendicular to the direction of travel of the wave(like waves on the surface of water). Shearwaves are extremely interesting as their propa-gation in rocks charged with hydrocarbons ismarkedly different to their propagation throughnon charged rocks. It means that reservoir attrib-utes can be more reliably estimated.

Shear waves are very useful when sending seis-mic data through an area with lots of gas bubbles(as often found over a reservoir). P waves willundergo interference as they pass through gas,which makes them very hard to interpret. If thereis a gas cloud above a reservoir it can be verydifficult to see the reservoir beneath using P-wave data recorded by hydrophones, but shearwaves recorded by the geophones allow a clear

The Ocean Bottom Node survey layout - with nodes spaced 225m x 225m. Note line L1 corresponds to the seis-mic image shown on the previous page


6 Digital Energy Journal - Special report, Apr 13 2015 Finding Petroleum forum "Transforming Subsurface Interpretation"

Comparing seismic images gathered using streamer survey (left) and ocean bottom seismic (right) on Maersk'sDan field. Note the much clearer imaging of the zone of interest (on the right). Image courtesy Maersk Oil, orig-inally presented at EAGE conference in Amsterdam, June 2014

image to be achieved. Shear waves do not prop-agate at all in a liquid medium so cannot berecorded by receiver arrays towed into the watercolumn

Case studies

Mr Moses presented five examples of how thesystems have been used.

In the first example, the system was used toimage a reservoir which had two surface ob-structions (offshore platforms).

A towed streamer survey had been shot over thearea in 2012, but was not able to image the areabelow the platforms.

Mr Moses displayed a subsurface image show-ing what was possible with streamers. “You cansee the holes left by the obstructions which aredirectly above the reservoir and damage the re-sulting image,” he said.

By being able to safely put receivers near the ob-structions on the seabed you can complete themissing data and have a much improved imageof the reservoir

In a second example, the area to be imaged wasoffshore West Africa.

The area had very strong water currents, whichmade it impossible for a streamer vessel to ap-proach the tension leg platform located abovethe reservoir.

Seabed Geosolutions delivered nodes to the re-gion in containers, then using in-field ROV fa-cilities a carpet of nodes was positioned on theseabed centred on the platform in water depthgreater than 400m.

The inevitable huge hole in the streamer deriveddata was seamlessly filled by the node data,which also showed improved detail thanks to thefull azimuth characteristics.

A third example from the North Sea, looking forshallow gas below the platforms, which couldnot be imaged using streamers.

A fourth example was from South East Asia,where the oil company had a problem with shal-low gas cloud disturbing the seismic image ofthe reservoir. The P-wave data even recorded onthe seabed showed a lot of disturbance above thereservoir, but the S-wave data was much cleaner.

A fifth example from North West Australia,where the seabed data was used to help de-riskdrilling and injection decisions.

The company had tried to work out where theoil-water contact was from their towed streamerseismic data, but it did not correlate with the welllog data. Also the injection wells did not seemto be having any impact on oil-water contact.

Once a new dataset was acquired using oceanbottom cable, the oil water contact point was re-calculated, and did correlate with the well logs,and it was possible to see the effect of water in-jection.

The company calculated that the ocean bottomseismic data added 3 years to the life of its field, he said.

Equipment

The latest node system from Seabed Geosolu-tions, MantaTM , uses standard rechargeablelithium ion batteries that can last for 90 days. Ordisposable batteries can be offered which canlast for 200 days. Advances in battery technol-ogy are dramatically changing the power densityof the cells and their capacity increases about10% every year.

The nodes are sized about 1 ft (30cm) diameter,and two thirds of the space is taken up by alithium battery.

The company has a new ocean bottom nodetechnology called “Manta”, which will be intesting from June 2015. It records four compo-nents (3 geophones and 1 hydrophone), and op

erates seamlessly through all water depths from0 to 3,000 m.

The Seabed Geosolutions’ sensors, both nodesand cables typically contain 4 components, hesaid.

The unit has an ‘inclinometer’ which can recordthe angle of the node once it is on the seafloor.

In the future we may see other types of sensorsin nodes.“Once you have a node on the seabed, it is up toyou what receiver you put in it,” he said.

There is no limit to how many nodes you canuse. One vessel could carry 10,000 nodes atonce.

The Manta units are stacked in modules on theback deck of a supply vessel. They can be de-ployed by rope and cable, as well as from ROV. The data cannot be analysed until the node isphysically recovered after the survey. Thismeans that it is not possible to do real time qual-ity control on the survey (as you can with towedstreamer surveys). However improvements inelectronics mean that clients are increasinglyconfident about the integrity of their data.

Node receiver carpets are optimised in densityfor the objective of the survey. Shallower targetsrequire more dense sampling. Deep targets canbe imaged with sparse receiver carpets.

Ocean Bottom Cables are well suited to rela-tively shallow water where dense sampling isbeneficial, but the electrical connecting cablesthemselves limit the number of receivers that canbe placed on the seabed and the depth fromwhich they may be recovered. The cable sensorsuse accelerometers in place of geophones whichhave a very good low frequency response –down to DC. Accelerometers are rather powerhungry and are not suitable for use in batterypowered nodes. However, technology is alwayson the advance and new low powered ac-celerometers are now entering the market.

Technology development

In partnership with Saudi Aramco, SeabedGeosolutions has a research project under waydeveloping fully autonomous nodes, which canfly themselves through the water to the desiredlocation.

New sources are being developed by others inthe industry. In particular the marine vibratorwill be very well suited to seabed seismic acqui-sition methods.

Watch Mr Moses’ talk on video and downloadslideswww.findingpetroleum.com/video/1285.aspx

The difficulty of imaging through gas. The left image shows Prestack Depth Migration (PSDM) with just p waves(left) and with shear waves (right). Image courtesy PETRONAS and PETRONAS Carigali, originally presented atEAGE 2014.



Full Tensor Gravity Gradiometry (FTG) surveys areincreasingly used for oil and gas exploration.

The technique provides a broadband gravity datasetthat can be integrated with existing seismic to addan increased understanding of the subsurface.

As a dataset that can be used throughout the explo-ration phase, it is cost effective to acquire early in theproject life cycle in order to gain maximum value,said Claire Husband, Senior Geophysicist withARKeX, speaking at the Finding Petroleum forumin London on April 13, “Transforming SubsurfaceInterpretation”.

During the presentation Claire Husband explainedthe difference between gravity and gravity gradiom-etry then went on to show gravity gradiometry andseismic integration examples from various stages ofthe exploration workflow from regional reconnais-sance to final interpretation.

Gravity and Gravity Gradiometry?

Gravitational force or acceleration due to gravity onthe Earth’s surface can vary spatially for geologicalreasons.

For example if you are standing above a high densitybody, you will experience a slightly higher acceler-ation due to gravity than at surrounding locations.The converse is also true, if you stand above a geo-logical body which is less dense that the surround-ings, you will experience a slightly loweracceleration due to gravity.

Lithologies vary in density depending on composi-tion. This varies from igneous rocks, which tend tobe the densest, to coal and salt which are commonlythe least dense materials in the subsurface. So bymeasuring gravity, or how gravity changes, you canget an understanding of the subsurface structure.

The instruments used to acquire conventional gravityand broadband gravity/ gravity gradient data are fun-damentally different.

A conventional gravimeter can be thought of con-ceptually as a mass on a spring. As the mass on aspring is moved over the ground, the ‘pull’ on themass will vary with the mass in the subsurface.

However, the gravity gradiometer has conceptually,although not in practice, two masses on springs lo-cated one above the other.

This measures variation in gravity between twopoints otherwise known as the ‘gravity gradient’.Modern gravity gradiometers, such as an FTG in-strument, conceptually comprise of an assembly of‘masses on springs’ oriented in different directionsto allow the measurement of the horizon components of the gravity gradient as wellas the vertical component.

Gravity gradient surveys and conventional gravity surveys can be conducted from moving platforms,

such as a boat or aircraft, which can be highly ad-vantageous as large areas can be covered quickly. However, the conventional gravimeter is hamperedas it cannot distinguish between acceleration causedby turbulence motion and the acceleration caused bya geological origin.

In order to overcome this, conventional gravity dataare filtered to remove the high frequency turbulencerelated component. This high frequency filteringprocess also removes geological signal and de-creases the resolution of the data.

There is no way round this. However, gravity gra-diometry does not suffer from this issue and does notneed to undergo the same high frequency filtering.

A common question which is asked is “What is thesmallest feature that gravity gradiometry/ broadbandgravity can resolve?” The resolution of the gravitygradiometry data (and size of the smallest object de-tectable) primarily depends on the magnitude of thedensity contrast between the target and the surround-ing lithology, the target depth and survey line con-figuration, she said.

Regional Reconnaissance – Northeast Greenland

The first example was from North East Greenlandwhere in April to June 2012 ARKeX conducted a50,000sqkm broadband gravity survey in conjunc-tion with Ion Geophysical’s 2D seismic acquisitioncampaign “Northeast GreenlandSPAN.”

Differences in interpretation before and after the in-tegration of broadband gravity data with seismicwere presented.

Fundamentally, however good the seismic data ac-quisition and processing, a 2D seismic campaign is,by definition, spatially limited. In the case of theNortheast GreenlandSPAN, the seismic lines areregularly and sensibly spaced throughout the surveyarea. Dr Husband showed examples from two blockswhere the seismic had not traversed over the salt ormajor faults which were both clearly visible in thebroadband gravity data. The broadband gravity sur-vey revealed a significant area of salt in both blockswhich had not been covered by the 2D seismic ac-quisition.

As well as gaining an increased understanding of thesubsurface, this information can also be used to planfuture seismic surveys. Both salt and faults can causemajor challenges to seismic processing and imagingworkflows.

But if you know in advance, you might plan the seis-mic acquisition more effectively which might entailthe use of more advanced seismic acquisition con-figurations, such as MAZ, WAZ or coil, to overcomepotential seismic illumination issues, she said.

Aiding seismic processing – offshore Gabon

The next examples showed the integration of gravitygradiometry with 2D seismic data from offshoreGabon at both the reconnaissance and the seismicdepth imaging stage.

In contrast to the North East Greenland area, highdensity carbonates as opposed to low density salt,were one of the potential exploration targets. Claire Husband showed an example of the broad-band gravity used with the 2D seismic data to mapthe lateral extent and spatial distribution of continu-ous carbonate bodies in the survey area.

The processing example started by showing the in-tegration of legacy data and broadband gravity datato build a density model of the subsurface. In short,this model was converted to density using velocity-density relationships derived from well data.

The density derived velocity model was used as astarting model for the PSDM (pre stack depth mi-gration) workflow. The standard top-down velocitymodel building approach was adopted.

The velocity of the first layer was updated using re-flection tomography, the density was then adjustedaccordingly in this layer and the other layers as ap-propriate. This was then converted to velocity. Thenew velocity forms the input to the next layer in thetomography. The initial results look promising.

Aiding seismic interpretation – on-shore Gabon

The final example presented by Claire Husbandlooked at how FTG data had been successfullyutilised to understand why an exploration well in on-shore Gabon was dry.

The exploration well target was a reservoir locatedon a structural high beneath a salt body. The basis ofthe reservoir seal was that the salt above the targetwas one intact salt body.

The clients drilled the well expecting to drill througha thick, intact layer of salt. However little to no saltor hydrocarbons were discovered in the well.

Presumably, with no salt acting as a seal, the hydro-carbons (if once present) had migrated away. Theclient undertook a high grade gravitygradiometry/Broadband gravity survey to help un-derstand the spatial distribution of salt.

The gravity gradiometry/Broadband gravity dataclearly indicated that the well had been drilled in asaddle area between two salt bodies.

Broadband gravity data was compared to the re-gional wells. The correlation between producing pre-salt wells and the salt as shown by the broadbandgravity data was 100%.


ARKeX – integrating gravity and seismic dataClaire Husband, Senior Geophysicist with ARKeX, explains the benefits of integrating broadband gravity datawith seismic at different stages throughout the exploration work cycle.


HP – data scientists meet subject matter expertsThe best way to do analytics in the oil and gas industry is to bring subject matter experts together with datascientists, said John Brennan of Hewlett PackardWhen it comes to data analytics, “we think thekey is to bring subject matter experts togetherwith data scientists who can provide real insighton what you can extract from the information,”said John Brennan, Analytics & Data Manage-ment Strategy Lead, Oil and Gas, with Hewlett-Packard.

He was speaking at the Apr 13 Finding Petro-leum London forum, “Transforming Sub-Sur-face Interpretation”.

“You get a team of business people, IT people,and put your foot to the floor and say, what isthe data, what is the hypotheses, and test in avery rapid manner.”

“Sometimes you find nothing at all and that's avalid answer - you can park that particular prob-lem.”

You need to make sure you start off with a spe-cific question which you are trying to answer.

90 per cent of the effort of doing a data miningproject is often gathering the data together in thefirst place, he said.

In the oil and gas industry, “we want to be ableto learn from the data that we've got,” Mr Bren-nan said.

Examples

Many companies around the world are findingnew ways to work with large data sets, he said.

It built an analytics system was for US stock carracing company NASCAR, to help the companysort through all the comments made on socialmedia while races are going on, so companystaff can get a better understanding of what thefans are thinking about.

NASCAR has a "Fan Engagement Centre"where all the data is gathered together andsorted. The solution includes hardware (power-ful servers to store and process data) and soft-ware, such as Autonomy, to extract meaningfrom the data. The data is then displayed to staffon video walls.

It receives data from Twitter, other social mediasites, and media outlets. The company can trackwhat the fans are thinking about as major an-nouncements are being made, such as a teamproducing a new car.

Another example is the UK meteorological of-fice, which has ‘big data’ systems which cangenerate a weather forecast for a 200x200m areaof land.

Data mining has also been used by supermar-kets, to calculate which product sells the mostafter a hurricane warning (which turns out to bebeer, Mr Brennan said).

Some companies are using automated tools toanalyse Twitter messages to see what the publicis saying about them.

“As the volume of data in the world grows, ourability to extract value from it also going up,” hesaid. “We can pull in multiple data sources fromthousands of feeds.”

HP’s offering

HP has a pre-packaged system for big data calledHAVEn, which stands for Hadoop, AutonomyIDOL, HP Vertica, HP Enterprise Security, andthe capability for building “n” Apps.

It can be provided as software or run in thecloud.

You can pull in information from a whole varietyof different systems, including video, text, callcentre data and social media feeds.

“HAVEn allows you to import the data into acentral area and run a range of analyticalprocesses on it,” he said.

“You can chew through the data and see whatthe correlation and causation is and what the in-sight is,” he said.

This software was used in the LondonOlympics, where video from CCTV cameraswas continually analysed, comparing faces withthe faces of known terrorists.

HP offers analytics as a service, or provides con-sulting for the best way to do it.

It helps companies gather their data together soit can be analysed, and share the results throughthe company.

See John’s talk on video atwww.findingpetroleum.com/video/1293.aspx

The HP stand at the Finding Petroleum London forum “Transforming Sub-Surface Interpretation”



Computer manufacturer Lenovo has developed a workstation especially for subsurface work in the oiland gas industry

Lenovo – a workstation for oil and gas subsurface

Computer manufacturer Lenovo has put to-gether a workstation computer package specif-ically for subsurface work in the oil and gasindustry, said Mike Leach, Workstation Technologist with Lenovo.

He was speaking at the Finding Petroleum con-ference in London on Apr 13 2015, “Transform-ing Subsurface Interpretation”.

In general, workstations have improved per-formance by about 15 per cent over the past 12months, he said.

The aim is to make subsurface data faster toanalyse and interpret, he said. Its processingspeed should be particularly appreciated forlong, iterative workflows, avoiding lengthywaits while computer processing takes place.

The system was tested for use in a fracture de-tection workflow. The work started with a 610

MB subsurface data set. The interpreter neededto remove noise, identify the direction of slopeof the rock layers, then four further processes -working out the fault structure, identifying thefracture, analysing fracture brittleness, lookingat microseismic data.

The fracture identification work included look-ing at fracture curvature, understanding the frac-ture network, fracture services, fracture azimuth(direction), and fracture density.

“It is a repetitive and tedious workflow, involv-ing many iterations,” he said.

Sometimes geoscientists go through the sameworkflow with 12 iterations to clean up animage, he said.

The workstation

The Lenovo workstation has been put togetherto give you the computational and visualisationtechnologies subsurface interpreters need, hesaid.

It uses Lenovo’s ThinkStation P900 worksta-tion, configured with 2 Intel Xeon E5-2697v3CPUs, 256 GB of 2133MHz DDR4 memory;delivering a total of 28x CPU cores (14x coresper CPU)

By adding more graphic processing units(GPUs) to your computer you can make it evenfaster. You can use the GPU for computing aswell as graphics display. “The more GPUs youadd, the faster the performance.”

You can also add high performance data storagesystems to the workstation, with up to 40 ter-abytes of storage in one workstation. “You canload in data very quickly, analyse it veryquickly,” he said.

“To get the performance you need doesn't haveto cost as much as you might think.”

The basic ThinkStation P900 computer costsunder £1600, but the cost can be as much as£30,000 depending on the configuration, hesaid. Oil and gas users will typically spend£5,000 to £7,000, but the final price is ulti-mately based on end user workflow and re-quired results.

The CPUs alone can cost £3,000 each.

The workstation can be installed under yourdesk, or you can access it remotely, with theworkstation based in a data centre. You can haveit dedicated (so only one remote person uses theworkstation) or ‘virtualised’ (so the same work-station is shared between multiple users).

Lenovo has also engineered its workstation sys-tem for the needs of finance users. They also doa lot of Monte Carlo type analysis, which re-quires fast computation, which can be done bythe GPUs using their own memory.

Watch Mr Leach’s talk on video and downloadslides athttp://www.findingpetroleum.com/video/1324.aspx

Configured specifically for the oil and gas industry – Lenovo workstations

Lenovo's exhibition stand



Attendee list – Transforming Subsurface Interpretation, London, Apr 13 2015

What did you enjoy most about the event?

Christian Bukovics, Partner, Adamant VenturesPaul Murphy, Key Account Manager, Oil andGas Division, Airbus Defence and SpaceRoberto Ruiz, Geophysicist, ARK CLS Ltd.Christian Richards, VP Sales EAME, ARKeXClaire Husband, Senoir Geophysicist, ARKeXAnne-Marie Liszczyk, Geophysicist, ARKeXLimitedSteve Callan, VP Sales & Marketing, BGPMarineJonathan Watson, Geophysicist, BridgeporthltdBryn Austin, Director & Geological/Geophysical Consultant, Brynterpretation LtdRobert FE Jones, Regional Exploration Manager, Cairn Energy plcRoger Taylor, Technical Marketing Manager,CGGMarianne Parsons, Geophysicist, CGGSean Waddingham, Data Library, CGGVeritasWill Thornton, Geologist, CGLAnne Benfedda, Marketing Manager,Chemostrat LtdRoger Doery, Consultant, ConsultantAndrew McCarthy, Exploration Manager /New Ventures Manager / Geophysicist, Con-sultantDominic Davey, Information Systems, ConsultantDan Kunkle, Director, Count GeophysicsTagbo Ndefo, Data Manager, DegeconekNigeria LimitedJean Martinie, Representative, DGIHelena Zapata Suarez, Interpretation Geoscientist, Dolphin GeophysicalNicolas Hand, Geoscientist, Dolphin Geophysical Ltd

Brian Donnelly, ConsultantTimothy Culwick, Solutions Architect,DrillinginfoJane Wheelwright, Technical Application Specialist, Dynamic Graphics LtdToyin Solanke, Consultant, Eknalos GeoTekMartin Riddle, Technical Manager, EnvoiSerje Heyer, Director, Feather TechSalar Golestanian, Managing Director, Finity AssetAvinga Pallangyo, Finance Administration,Future Energy PublishingKen Armitage, MD, GeoDirk LtdAllan Induni, GeoscientistAlexandra Kenna, Managing Director,GEOSERVE LimitedDavid Bannister, Marketing Manager, GeotraceChris Boot, Business Development, GetechNorman Hempstead, Director, HempsteadGeophysical SvcsSuzanna Bailey, Hewlett-PackardPaul Muscat, director for energy and utilitiesindustries in UK, Hewlett-PackardJohn Brennan, HPMartin Hodge, Data Processing & Time Imaging Geophysicist, IExSTRavi Chandran, Director, Kalki ConsultantsLimitedMike Leach, Workstation Technologist / Business Development Manager, LenovoRebecca Clare, Business Development Manager, LenovoNick Grealy, Director, No Hot AirMark Enfield, Managing Director, P.D.F. Limited

Robert Parker, Consultant, ParkerAdebola Akin-Odidi, Principal Geophysics,PetrofacVincent Sheppard, Chief Geophysicist, PetrofacJohn Hother, MD, PronetaJames Page, BDM, PVE ConsultingJosh King, Analyst, RAB CapitalSimon Fleckner, Graduate, Kes Heffer, Director, Reservoir Dynamics LtdDavid Sendra, SubSurface Manager/Petrophysicist, RSI Geophysical Ltd.Keith Holdaway, Principal Solutions Architectand O&G Domain Expert, SAS InstituteDominic McCann, Director, SAS InstituteDavid Webber, Seismic Operations Supervisor, Sceptre Oil and Gas LtdPaul Day, Director / Consultant, SCGIS Ltd.John Moses, Seabed GeosolutionsGlyn Roberts, Director, Spec Partners LtdRichard Lee, Sub-Surface Data Technologist,Sub-Surface Data TechnologistFrank Eisenhower, Operations Manager, TGSSean Akinwale, Business Development Man-ager, TGS Geophysical CompanyNigel Quinton, Head of Exploration, TowerResources plcJay Sahota, Lead Geoscientist, Tullow OilAlec Robinson, President & CEO, Valient EnergyPeter Lancaster, Director, Valioso LtdRana Wallwork, Technical Analyst, White Rose Energy Ventures LLPBen Couzens, www.CV.Couzens.bizAndrew Zolnai, Consultant, zolnai.ca

“

”

“

”

“

”

“

”

“

”

The variety ofbreadthand depth of topics being presented, as well as the time given to Q&APeter Lancaster,Valioso Ltd

Opportunity tomeet a cross section of various industrystake holdersSean Waddingham,CGG

Variety of topics and exposure to information outside of normalday-to-day workDominic Davey

Some mix of disciplines albeitheavily geophysicalKes Heffer

It was interesting tohear about thingsfrom the data analysis side ofthings, ie from those 'outside' the industry and to seesome applicationsof the technologies


11Digital Energy Journal - Special report, Apr 13 2015 Finding Petroleum forum "Transforming Subsurface Interpretation"”

Geodirk – using geology to understand seismic

Ken Armitage, CEO of UK company Geo-dirk, has developed a methodology andcomputerised system for understandingseismic data taking geology into account,which he believes can make a big impact inavoiding dry wells.

Using geological understanding in seismicinterpretation can “at least double the infor-mation available from post stack seismic,”he said.

He presented the method and system atFinding Petroleum’s April 13 2015 Londonconference “Transforming Subsurface Inter-pretation”.

The method uses computer tools to analysefor geological ‘shapes’ (such as faults andthrusts) in the seismic data.

A geologist can look at this geological pic-ture and see if it makes geological sense.The picture with “geological sense” meansthat she can see how the subsurface couldhave ended up with these features, based onwhat might have happened over geologicaltime.

In geology, if a geological story looksroughly correct, it probably is correct, hesaid.

The geologist can work out how the rockmight have been deposited, and how itmight have changed since deposition (dia-genesis).

“It is necessary to build something which asedimentologist and structural geologist canagree is a sensible picture,” he said.

Once you have a sensible geological story,you can process your seismic data in theusual way, taking into account the rockproperties you expect from the geology.

You can put the model in reservoir model-ling software such as Landmark and Petrel.Then people from all disciplines can have alook at it and check it makes sense to them.

The end result is that you should have amodel you feel surer of, which can be usedto better make drilling decisions, he said.Instead of having a fat bell curve with lots

of possibilities of what the truth might be,you have a narrow one.

Usually the geological interpretation is onlydone after the seismic interpretation hasbeen finished, he said.

“Working out geology at the same time asdoing seismic interpretation can be quitedaunting to think about, which is probablywhy nobody has in the past, or made muchprogress with it,” he said.

Reducing dry wells

The oil and gas industry really needs to im-prove its exploration success rate, he said,as it is reported to have dropped from 25 percent to 15 per cent over the past few years.

“Exploration is a small part of E&P cost,but there's still a significant amount ofwaste,” he said.

“Investors and oil company owners aretelling us we've got to solve this problem.Stop getting money to spend wastefully ondry holes.”

Finding more reservoirs

With methods such as this one, it should bepossible to accurately target much smallerreservoirs, he said, including fields whichhave many layers of reservoirs and seals.

“We are moving to a situation where we'vegot to start looking at interbedded sediment,not just massive sand and massive shale.”

“All the big easy structural traps seem tohave been found,” he said. “We've got tofind the missing ones and do it with lessrisk.”

With better subsurface understanding, youcan also find ways to reduce water produc-tion, and extend the production life of thefacilities, he said.

Permeability and porosity

Mr Armitage did an analysis of dry wells,and found that one of the biggest causes is

poor predictions of porosity and permeabil-ity (“poroperms”), he said.

A reservoir needs to have high porosity andpermeability, and be sealed by a rock withlow porosity and permeability, he said.

In order to make a good porosity / perme-ability estimation, you need to know aboutthe rock types – which means a geologicalunderstanding, he said.

There are 30 or 40 different geological rea-sons why porosity and permeability mightchange moving across horizontally, he said.

Some areas of subsurface have a mix ofrock fragment sizes, from very fine clasticsto much coarser clastics. This all means adifference in rock properties, which willmess up your seismic processing if youdon’t know about it.

About 60 per cent of rock does actuallyhave predictable properties, he said. For ex-ample 60 per cent of all shale sediment willcompact in a similar way due to the weightof rock above it.

So rock in many areas of the world has in-terchangeable properties, including the GulfCoast, Norway and some of the North Sea,he said.

The way the rock compacts will drive itsdensity, velocity of seismic waves throughit, porosity and permeability.

But there’s also a 40 per cent probabilitythat it will have different properties, whichwill mean that the calculations will providethe wrong result, he said.

For example if the rock has had non verticalstress on it, or pushed up from below bysalt. “That will stop this process working.”

But some rock will compact at faster rates– including chalk, and material made ofcoarser grains. This will lead to a higherthan expected loss in porosity.

So if you don’t do a geological analysis,your work will only be accurate 60 per centof the time – when the rock is deposited ina standard way, he said.

12

By integrating geological understanding with seismic processing, you can get a much better understandingof the subsurface, which can lead to fewer dry wells, said Geodirk’s Ken Armitage



If you have a reservoir model built up usingcells, you can make sure that the geologicalmodel you evolve from the seismic is sensi-ble, for every single cell, he said.

“It could mean going through the laborioustasks of figuring out all the things that couldmake geology change - then filtering the datafor its presence of absence.”

“Geology is giving us surprises, and thereforeporoperm surprises.”

Dong case study

In one example for oil company Dong, a litho-logical understanding was built into the seis-mic interpretation, including water,anhydrides and dolomites.

There were also volcanic rock in the subsur-face volume being studied, which is hard tospot.

The target reservoir was actually beneath thesource rock, with everything over pressurised,he said.

Computer system

With Mr Armitage’s method, the seismic in-terpreter could access a standard database of

rock properties, which can be used to calcu-late seismic velocity, porosity and permeabil-ity taking geological factors intoconsideration, such as rock type and com-paction rate.

Then you work out how this would change ifvarious seismic shapes or geological attributeswere present, such as faults.

You could also add in factors such as temper-ature and pressure, and how these might im-pact porosity and permeability. None of thisinformation is available directly from seismic.

You can develop databases of rock propertieswhich have undergone similar compactionand transformation (diagenesis).

“We basically bundle all the lithologies intodifferent families of behaviour,” he said.

“This is where you need processing power,because you've got to filter the seismic dataand shapes,” he said.

“You've got to make this as objective as youcan and minimise subjectivity.”

“It only takes one project to enable an assetteam to be expert in how this works,” he said.“It is 1000 different applications fitted to-gether in one system.”

Some of the computer processing tasks mighttake 4-5 days to run on an older PC, but withtoday’s workstations it could be done instan-taneously, with the results checked in realtime, he said.

Background

Geodirk’s processes were first developed inthe early 1990s by oil companies who foundthat they did not have a good idea of the ve-locity of seismic data, with rock imaged by2D seismic. “We tried to make a geologicalway of working out seismic sequences,” hesaid.

Mr Armitage was involved in early work todevelop algorithms to try to spot various geo-logical ‘shapes’ in the data, for example look-ing for signs of non-vertical stress, inversion(something pushing up from below) andfaults.

When 3D seismic was developed, much betterpredictions of velocity could be made. But thevelocity still isn’t made accurately enough topredict porosity and permeability, he said.

View Ken Armitage’s talk on video and download slides atwww.findingpetroleum.com/video/1272.aspx


Delegates at the coffee break


14

Using well fibre optic data together with the reservoir model



Integration and visualization of well fibreoptic dataDynamic Graphics is developing computer tools which can help display acoustic and temperature data,gathered from fibre optics in wells, together with other reservoir data, such as seismic and production data

Dynamic Graphics Inc (DGI), a company specialising in data visualisation and analysis,has developed a computer tool to displayacoustic and temperature data, generated fromfibre optic cables in wells, together with all ofyour other reservoir data and models.

This should make it easier for oil and gascompanies to work with well fibre optic DASand DTS data, to understand what the data isshowing, and how it relates to what else ishappening in the surrounding reservoir.

“My experience is that a lot of data is beingcaptured but not used to its full capacity,” saidJane Wheelwright, Technical Application Spe-cialist at DGL, speaking at the Finding Petro-leum forum in London on Apr 13,“Transforming Subsurface Interpretation”.

Also, “a key thing is not looking at a well inisolation but getting an understanding of whatelse is happening in the reservoir,” she said.“The results are more powerful when they canbe integrated with other data. It gives a far bet-ter understanding of what is happening with awell.”

Displaying it

Dynamic Graphics provides two ways of dis-playing well fibre optic data. Firstly the tem-poral downhole view, where the temperatureor acoustic data for the well can be visualisedtogether with any other well and reservoirdata, at any specific point in time.

This enables the data to be viewed at differentpoints in time with other time varying data –so, for example, you could see a 4D seismicsurvey from a year ago, together with theacoustic data from the well at that time.

Another way is the ‘waterfall’ display whereyou see how the data recorded from the wellis changing over time.

“The capability to see both these displays isvery useful, they complement each other,” shesaid.

“You want to have capability to integrate withexisting data, quantitively and qualitatively,and be able to back interpret and cross plot theinformation.

CoViz 4D

Ms Wheelwright showed a view of a matureoffshore oilfield using Dynamic Graphics’CoViz 4D software, with many different datasets available.

By looking at the field at different points intime, you can see how the status of each wellchanged, and the history of the field. You cansee the various fluid flows, with oil, gas andwater in different colours.

“As water is injected and oil produced,changes can be seen,” she said.

On the same view, you can see a 4D seismicdata set, a geological model, and a reservoirsimulation, drawn from other software pack-ages (in this case Seisworks, LandmarkNexus, and RMS).

The CoViz 4D software has a Global timeslider which amalgamates all the time stepsfrom each file, so you can see the correct sta-tus of all the data loaded at different points in


Frontier Exploration?

Screen your acreage with broadband airborne gravity for informed prospect identifi cation.

time,(for example, hourly production data andyearly updates to the 4D seismic).

You can look in detail at any part of the field.For example, you might want to look at datafrom a producer and injector well pair.

With temperature data, you could see a ther-mal slug moving up the well in real time.

You can see well temperature data togetherwith the well completion data, because ele-ments of the completion could be causing achange in temperature.

If you have production data together, youcould see (for example) that the temperaturerose after the well was shut in.

You can see where valves were open andclosed.

“There are wide ramifications in using thistechnology,” she said. “It gives quite uniquedata and information for the reservoir man-agement team, and it allows for understandingwhat is happening in a reservoir and optimis-ing decision making.”

Well fibre optics

Fibre optic cables can measure acoustics andtemperature by understanding how theychange the flow of light through the cable.One cable can provide many different sorts ofdata.

DAS (“Distributed Acoustic Sensing”) is away to use fibre optic cables in wells to makean acoustic (sound) recording of what is hap-pening in the well.

This can be used to spot problems (for exam-ple damage to a component), and also torecord seismic data from inside the well.

DTS (“Distributed Temperature Sensing”) isusing the fibre optic cable to measure temper-ature at all points in the well.

Fibre optics are already used to replace pro-duction logging tools, as a means of under-standing which zones in a well are providingmost of the oil, and to spot flow leaks insidethe well.

The systems are also used to monitor for re-strictions in the well bore (which can causethe fluid flowing past them to make a sound).It provides information about completions. Itcan be used in fracking, to monitor how wellthe fracking is going.

Oil companies need the data to work out thebest way to frack a certain well, and the bestway to optimise hydrocarbon recovery.

The data generated by DTS (temperature) ismuch smaller volume than the DAS (acoustic)data. Most DAS files are larger than 1 TB,which makes them very hard to work with. “Akey issue is to reduce the size of data files,”she said.

The DAS data is currently usually provided inHDF5 format, which can be imported intoCoViz 4D.

View Jane’s talk on video at www.findingpetroleum.com/video/1289.aspx

HEADQUARTERS

Dynamic Graphics, Inc. 1015 Atlantic Avenue | Alameda, CA 94501-1154

Phone 510.522.0700 | Fax 510.522.5670

www.dgi.com | [email protected]

OFFICES & REPRESENTATIVES

Aberdeen Scotland, [email protected]

Aerospace, Defense, and [email protected]

Bakersfield, California USA

P. R. [email protected]

Houston, Texas [email protected]

London, [email protected]

Paris, [email protected]

Washington, DC [email protected]

LEGAL NOTICE

© 2015 Dynamic Graphics, Inc. All Rights Reserved.

No part of this publication may be reproduced,translated, or

transmitted in any form or by any means, electronic or mechanical,

including photocopying, recording, use or capture in any information

storage or retrieval system, or otherwise, without the express prior

written permission of Dynamic Graphics, Inc.

The information contained in this document is subject to change

without notice and should not be taken as a commitment,

representation, or warranty on the part of Dynamic Graphics, Inc.

Further, Dynamic Graphics, Inc. assumes no responsibility for errors

that may appear in this document.

Dynamic Graphics, CoViz, and the logos shown below are

trademarks of Dynamic Graphics, Inc. that are registered trademarks

or the subject of pending applications in various countries.

All other trademarks belong to their respective owners.

[email protected]

INTEGRATION

INTEGRATE / VISUALIZE / ANALYZE DATAIN CONTEXT WITH COVIZ 4D

of multi-disciplinary data, including reservoir simulation, 4D seismic, seismic attribute extractions, structure model, and production data

Data for all images used with permission of owner.

monitored usingmicroseismic data. Event locations, errorbars, and focal mech-anisms are visualizedin 4D in relation to the surrounding geologicmodel, well locations,and surface infra-structure. Treatment curves are also inte-grated into the display.

HYDRAULICFRACTURING

SEISMIC

amplitude values aresampled into both a reservoir simulation grid and well log data, enabling statistical analysis and comparison of diverse datasets

Documents

Special report83a7383a5e33475eed0e-e819cda5edf0a946af164bb0b2f2ae3c.r0.cf1.rackcdn.com/...Special report Transforming Subsurface ... Mr Holdaway was followed by a talk on seabed seismic