The South Caspian AIOC Central/West Azeri Blowout – September 2008
In September 2008, nearly two years before the Deepwater Horizon explosion in the Gulf of Mexico, another BP rig had blown out in the Caspian Sea—which BP concealed from U.S. regulators and Congress. Had BP, Chevron, Exxon or the Bush State Department revealed the facts of the earlier blow-out, it is possible that events leading up to the Deepwater Horizon disaster could have been different.
Azeri-Chirag-Guneshli (ACG) Field Development
The Azeri offshore oil field lies in the South Caspian Sea, roughly 120 km off the coast of Azerbaijan and is a part of the larger BP led Azerbaijan International Operating Company [AIOC] Azeri-Chirag-Guneshli [ACG] field (Ref. 1). Formed in February 1995, AIOC is a consortium of 10 companies. The field contains 5.4 billion barrels of recoverable oil and was put into production in November 1997. During the first six months of 2012 it produced 124.5 million barrels and 11.5 million m3 of associated gas per day from 56 production wells. The AIOC participants include: BP as operator with 35.8% stake; Chevron 11.3%; SOCAR 11.6%; INPEX 11.0%; Statoil 8.6%; ExxonMobil 8.0%; TPAO 6.8%; Itochu 4.3%; and Hess with 2.7%. In September 2012, ONGC acquired Hess’s stake.
A BP presentation from 2006 entitled “Building A New Profit Centre: Azerbaijan” (Ref. 14), is an appropriately named summary of the BP story in Azerbaijan. This started with the infamous Mrs. Thatcher visit of 1992, delivering cheques to the Azeri government on behalf of BP, the initial AIOC “Early Oil” project in 1997 following the signing of the 1994 “Deal-of-the Century”(Ref. 15) between BP and the Azeri government, the development of ACG, the continually expanding Sangachal receiving terminal south of Baku, the Baku-Tbilisi-Ceyhan [BTC] and South Caucasus [SCP] oil and gas export pipelines and the new BP “flagship” Shah Deniz 1 and 2 projects currently under development.
Fig. 1. Map Showing Azeri-Chirag-Gunashli, Shah Deniz, Absheron & Other South Caspian Offshore Fields. Source: Greenfields-Petroleum.com
The Azeri field includes the Central, West and East Azeri production platforms and a compression and water injection platform (C&WP). Oil and gas is exported from the offshore fields through the Sangachal terminal across Azerbaijan, Georgia and Turkey via the 1,768 km (1,099 mile) long BTC and SCP pipelines (Ref. 4) connecting ACG to the south-eastern Mediterranean coast
Central Azeri is a production, drilling and quarters (PDQ) platform located in 128 m water depth in the central part of the Azeri field, constructed to produce approximately 420,000 bbl/d (67,000 m3/d). The facilities include a 48-slot PDQ platform 30-inch (760mm) oil pipeline and 28 inch (710mm) gas pipeline from the platform to the Sangachal terminal and operations started in February 2005. West Azeri is also a PDQ platform located in 120 m water depth, constructed to produce oil from the western section of the Azeri field. West Azeri adds 300,000 bbl/d (48,000 m3/d) to overall ACG production and facilities include a 48-slot PDQ platform and 30-inch (760 mm) oil pipeline from the platform to Sangachal. Operations started in December 2005 (Ref. 2). East Azeri is a PDQ platform in 150m water depth constructed to produce oil from the eastern section of the Azeri field and produces 260,000 bbl/d (41,000 m3/d), consisting of a 48-slot PDQ platform, which started operations in October 2006 (Ref. 3).
The C&WP supplies the Central, West and East Azeri platforms with water and gas injection services, manages gas export and provides electrical power using 10 Rolls Royce turbines and is bridge linked to Central Azeri platform. The gas injection capacity at the C&WP is 1 billion cubic feet per day (28 million m per day) with 5 gas injection wells. Water injection capacity is 1 million bbl/d (160,000 m3 per day) with 12 water injection wells. Gas export capacity stands at 250 million ft3/d (7.1 million m3). The Azeri C&WP has some of the largest water injection pumps and gas injection compressors among BP platforms worldwide. The topsides were constructed in the ATA (AMEC-Azfen-Tekfen) construction yard in Bibi-Heybat.
Fig. 2. ACG Gunashli, Chirag and Azeri Platforms. Source: APC Engineer.com
Central Azeri Platform Gas Blowout
On 17th September 2008, a gas leak was discovered in the seabed immediately beneath the Central Azeri platform after gas bubbles were observed in the water, with a drilling mud blowout occurring to the drilldeck reportedly via a gas injection well (Refs 5, 6, 7 and 19). The platform was in production at the time, so was shut down and evacuated, the biggest emergency evacuation in BP’s history. It was extremely fortunate that all 212 workers were able to removed safely without ignition and explosion. The incident resulted in two fields being shut and output being reportedly cut from 900 k to 300 k barrels a day, representing a loss of 40 to 50 million USD per day to the Azeri government, with production disrupted for months (Refs 5 and 6). West Azeri was powered by a cable from Central Azeri, so this was also shut down (Ref. 8). Production at West Azeri resumed on 9th October 2008 and at Central Azeri in December 2008 (Refs 10 and 11).
Gas or water injection is a technique commonly used to raise pressures and improve well extraction and enhance waning pressure within formations. Gas injection wells effectively “sweep” the formation for oil, thus boosting production. At Azeri, BP explicitly stated the blowout was due to a “bad cement job” (Ref. 9) but no further details were provided and public statements indicate there was considerable uncertainty as to the real cause. It is reported that nitrogen foam cement (Refs.20 and 21) as used at the Gulf of Mexico Macondo well two years later was adopted at Azeri. A “number of wells” were shut off, which were presumably subsequently investigated in order to log the cement integrity throughout. These may have been repaired at a later date, but again further details were not provided.
Fig. 3. AIOC-ACG Seabed Rendition Showing Extent of Mud Volcanoes, Slope Failures and Faulted Zones. Source: Azerbaijan International Operating Company
Apsheron Ridge Geohazards
The ACG fields situated along the subsea Apsheron Ridge contain several extreme geological hazards (Ref. 12). The seabed is notoriously faulted, unstable and weak and susceptible to medium strength earthquakes. The ACG ridge is dominated by a continuous 5,000 year old submarine slope failure escarpment. In order to generate “first oil” in the late 1990’s to fund future development BP always planned to use the half completed Soviet Chirag jacket, the only structure present at the time, which lay very close to the failed slope edge. This was achieved by designing and constructing a new topside for the jacket, including a seismic isolation module to guard against the prevailing levels of seismic shaking.
It may be reasonably assumed that it will have determined at what depth the gas problems initially occurred, based upon well logs and previous geophysical surveys, and along exactly which stratigraphic horizon(s) flows occurred. This will have indicated the likely source of the gas, possibly high permeability ancient Palaeo-breccia deposits of varying ages and depths associated with mud volcano flows. The ACG mud volcanoes and associated seabed and subsurface features were mapped during the Soviet period and were well understood, based upon extensive Russian and other geological literature. The seabed was then geophysically surveyed and mapped in considerable detail, including initial regional BP surveys in the late 1990’s and subsequent detailed well site, facility and pipeline surveys. It has always been strongly suspected that gas hydrates may be present across the area at relatively shallow depth (Ref. 24).
The articles listed in a Wikipedia link (Ref. 13) summarise the situation that arose in September 2008: A 1998 article describing planned geohazard studies at ACG in relation to facilities siting and drilling risks summarises the geohazard situation and views of AIOC at the time (Ref. 12).
Previous Well Drilling Problems
There had already been well drilling problems from 2003 onwards at West Azeri, causing severe well casing buckling problems. These have been described in detail and were not gas related, but caused by the uniquely weak and variable clay formation mineralogy present across ACG (Ref 18). This well buckling was directly caused by formation weaknesses at certain stratigraphic levels as a result of the upper ~200 m of clay strata consisting of a unique geochemical make-up and particle structure, including highly saline pore fluids all resulting from the presence of vast quantities of upwelling gas/mud/saline deep waters associated with the mud volcano on a geological time scale. It is for this reason that the a “Riserless Mud Recovery” drilling technique [RMR] was specifically developed at great expense and subsequently used successfully at West Azeri. Some of this information is in the public domain (Refs. 16 and 17), although only success stories tend to be publicised, rarely the initial problems. The (in)famous “Lessons Learned” in the oil industry are usually for internal consumption and are rarely made fully available.
Fig. 4: Locations of Pipelines, Platforms and “Traffic Light” Risk Zones at Azeri-Chirag-Gunashli. Source: Azerbaijan International Operating Company.
Lessons for The Future For Drilling in Gas Hydrate Areas
It is fairly certain that BP will have learned from the problems encountered at ACG and the lessons learned should have been transferred across to the adjacent Shah Deniz project. However, and this is a key question, how much of that information was taken onboard by BP Houston is another question. In most people’s experience, oil companies in different centres operate quite independently at project levels, with separate teams and different ways of working and reporting and even different procedures and processes. The “top-hole” shallow section drilling hazard reports produced by a group operating from one location may be quite different to those generated by another, even though ostensibly containing a similar make-up of technical expertise and experience. Despite there ostensibly being some cross-pollination, any detailed consultation between such groups is often somewhat superficial, with hazard reports from each often not even being viewed, much less reviewed, by the other teams.
Certain questions should be asked with respect to how much the technical people on the ground in BP Houston involved in planning and designing the Macondo MC 252 well knew about what had happened years previously offshore ACG South Caspian at West and Central Azeri. It should be determined how much those experiences and the (theoretically anyway) “lessons learned” were transferred from Baku, through Sunbury out to Houston and how much, if at all, these affected the Macondo initial well design, shallow hazard assessments and the defined drilling and well completions procedures.
Locating of Facilities in Relation to Surface and Drilling Geohazards
At Central and West Azeri, the closer the pair of platforms were placed to the mud volcano cores, the higher the risk to both surface facilities and drilling operations and these were positioned very close indeed. It was estimated that very 100 m of shift away from the target Balakhany X reservoir axis (the synclinal hydrocarbon “trap”, see risk map above) cost in excess of a million dollars in extra well length costs alone for a 48 well slot directionally drilled platform. There will have been great pressure from management to position the platforms as close as possible to the mud volcano cores and slope edges, since the “axis” is more or less coincident with the line of the mud volcano cones and the distinct seabed slope escarpment failure seen along the ACG field is just downslope of that. To minimise these costs, the platforms were placed very close, possibly intersecting permeable gas charged breccias flow deposits tied back into the mud volcano “gas core” which plunges more or less vertically many km below seabed. .The breccia layers are old flow deposits, many 100s of thousands or even millions of years old, which were very well mapped geophysically. If these were drilled too close without having stepped the facility back sufficiently in mitigation, there would have been some degree of risk.
Methane Gas Hydrates in South Caspian and Gulf of Mexico
It is likely that frozen gas hydrates are present in the seabed at some depths/locations at ACG. Indeed conditions are such that there are possibly widespread features of the deep water of the South Caspian, characterised by
- Depth-restricted, lenticular bodies well beneath the seafloor,
- The apparent accumulation of free gas within the underlying sediment, and
- Evidence of associated recent slope failures in the overlying strata.
For the SOCAR/Total field, Absheron, adjacent to ACG. hydrates were assessed as likely to be present from predicted thermobaric modeling, perhaps developing in water depths as shallow as ~150 m, and could form layers as thick as 1350 m It is believed that an assessment of the likely presence and depth of hydrates at Shah Deniz adjacent to ACG (see Figure 1) has been carried out. This may be comparable to the assessment carried out for Macondo in the Gulf of Mexico by BP in Houston, where there is brief mention in the pre-drilling shallow hazards report (Ref. 22). However the Macondo calculations are incorrect since they assume pure methane throughout, using an outdated relationship between temperature, pressure, salinity and methane content. The presence of substantial quantities of ethane and propane (~13%) known to be present at MC252 (and in the South Caspian) was ignored, such that the predicted depth/thickness of hydrates below seabed was too shallow at 1,279 to 1,905 feet (390 to 580 m), for temperature gradients of 35 to 25 deg./km (see Section 7.2.4, p.9, entitled “Hydrates”).
The offshore industry may never find out exactly how and why shallow gas blew through which specific wells at Central Azeri unless further details of well numbers, locations, azimuths, stratigraphic depths, dates, times and geology are provided. Investigative reports on how many “bad cement jobs” were discovered and how these were assessed and determined would need to be provided, with information as to whether or not these wells were repaired/remediated and eventually put back into commission or not.
Ultimately, “Dual Gradient” RMR riserless drilling was developed and adopted at West Azeri, although at Macondo, BP do not appear to have done anything different to that done previously in the GoM, based upon their corporate experiences in Azerbaijan. Indeed based upon what is known in the public domain, the company appears to have simply moved ahead with the development and design of the Macondo well from 2009 onwards as a deepwater exploration well to be converted to a producer as if nothing had occurred in the South Caspian which was out of the ordinary.
It is clear that at ACG. BP played a dangerous game in locating major new offshore facilities as close as possible to extremely hazardous geological features. The price of taking those risks was paid very shortly afterwards and that price could have been far higher. In situations such as that, one should never arrogantly play dice with Mother Nature and hope to win for the sake of a few dollars saved. But this was the all-new slimmed down cost-cutting rather secretive BP, more concerned with reputation, cost control, their bottom line and keeping information from partners and the industry at large. Certainly more concerned than informing those partners and the global oil industry of the background and detailed reasons behind what could easily have been a major fatal blowout disaster, rather than just another big profit loss.
- “Azeri-Chirag-Guneshli; The Largest Oil Field under Development in the Azerbaijan Sector of the Caspian Basin”.
- “BP Begins Production at West Azeri Field in Caspian Sea”, Rigzone, 5th January 2006
- “Production Begins at East Azeri in the Caspian Sea” (BP Press Release), 23rd October 2006.
- “Golden Weld Ceremony Links Azerbaijan and Georgia”, Azerbaijan International, Sec. 12.4, pp. 84 – 87, Winter 2004.
- “BP Shuts Down Two Azeri Oil Platforms After Gas Leak”, Bloomberg, 17th September 2008.
- “WikiLeaks: BP’s ‘Other’ Offshore Drilling Disaster”, Time World, 18th December 2010.
- US Embassy in Azerbaijan; 8th October 2008; Original Title: “Azerbaijan Seeks to Develop ACG Deep Gas, can Supply Georgia with Winter Gas; US Embassy Cables: BP May Never Know Cause of Gas Leak, US Told”, Guardian, 15th December, 2010.
- US Embassy in Azerbaijan; 15th January 2010; Original Title: “Azerbaijan: BP Downbeat on 2009 Shah Deniz Phase Two Progress; US Embassy Cables: BP Blames Gas Leak on Bad Cement Job, 15th January 2009.
- “BP Partially Resumes Production at Azeri Platform”, Reuters Edition UK,.23rd December 2008.
- “Mud Volcanoes Top Hazards for Future Azeri Operations”, Offshore, Pennwell, 3rd January 1998.
- Azeri Oil Field Wikipedia.
- “BP in Azerbaijan, September 2006”, Presentation, p. 87.
- “Pipe Dreams, Part 1; Azerbaijan’s Riches Alter the Chessboard”, 4th October 1998.
- Alford & Asko (M-I Swaco), Stave, R. (AGR Subsea), “Riserless Mud Recovery System and High Performance Inhibitive Fluid Successfully Stabilize West Azeri Surface Formation”, 2005 Offshore Mediterranean Conference, Ravenna, March 2005, Paper No. OMC 038.
- Stave, R. (2007), “Riserless Mud return Technology Solves Shallow Wellbore Instability Problem: A Case History”, Presentation to American Assoc. Drilling Engineers, 23rd May 2007, p. 30.
- Allen, J.D., Hampson, K., Vermeijden, C. And Clausen, C.J.F. (2005), “Well Deformations at West Azeri, Caspian Sea”, Proceedings of the International Symposium. on Frontiers in Offshore Geotechnics (IS-FOG 2005), 19th – 21st September 2005, Perth, WA, Australia
- US Embassy Cables: “WikiLeaks Cables: BP Suffered Blowout on Azerbaijan Gas Platform”, 16th Secember 2010.
20. Earth Island Journal, “Part 2: BP Covered Up Blow-out Prior to Deepwater Horizon”, April 24th 2012.
- Environment News Service, “Halliburton May Have Pumped Unstable Cement Down BP Oil Well”, 29th October, 2010.
- BP “Shallow Hazards Assessment; Proposed Macondo Exploration Well MC 252 #1; Surface Location in Block 252 (OCS-G-32306); Mississippi Canyon Area, Gulf of Mexico. 8th June 2009, p. 46.
- Diaconescu, C.C. and Knapp, J.H. (2002), “Gas Hydrates of the South Caspian Sea, Azerbaijan: Drilling Hazards and Sea Floor Destabilizers”, Proc. Offshore Tech. Conf, Houston, May 2002, Paper No. OTC 14036.
- Bagirov, E. and Lerche, I. (1997), “Hydrates Represent Gas Source, Drilling Hazard”, Oil and Gas Journal, 12th January, 1997.