Raises design criteria questions
Michael G. Dwyer
Kerry J. Kessler
J. Ray McDermott
As the oil and gas industry continues to recover from Hurricane Ivan, an opportunity arises to consider Ivan’s impact on future designs, maintenance, and assessment of platforms and pipelines.
Industry experts are studying and evaluating Ivan and its consequences, and an industry response should emerge. This response may take several forms, including modifications to operating procedures, maintenance requirements, and design practices.
Ivan, a Category 4 storm with winds exceeding 140 mph, passed on a north-northeast path striking the Florida-Alabama coast on Sept. 15, 2004. Of the 4,000 structures and 33,000 mi of pipelines in the Gulf of Mexico, 150 platforms and 10,000 mi of pipelines were in the direct path of Hurricane Ivan, according to Minerals Management Service estimates.
At the height of the storm, data obtained from National Data Buoy Center Buoy No. 2,040 recorded a significant wave height of 52.5 ft. The buoy is in deepwater off the Alabama coast in the vicinity of offshore operations. Given the duration of the storm, a significant wave height of 52.5 ft results in a maximum wave height in excess of 90 ft. The buoy subsequently broke free from its mooring.
Observed platform damage is another tool for estimating the severity of Hurricane Ivan. In one instance, industry experts noted obvious wave inundation damage in a platform’s deck at an elevation of 60 ft. This could have been a single rogue wave, not typical for a storm with the intensity of Ivan. Based on this observation, the estimated height of the wave would be approximately 90 ft. These estimates of maximum wave height indicate that the intensity of Ivan exceeds the intensity of the 100-year return period storm upon which API RP 2A design criteria is based.
The industry has attributed a substantial amount of the deferred production after the storm to damage that occurred to pipelines rather than actual structural damage to the producing platforms.
The industry has reported to the MMS that Ivan destroyed seven structures. Two were braced caissons in relatively shallow water. Four were typical jacket type structures in 250 ft of water. The seventh was a typical jacket type structure in 479 ft of water.
At least six additional platforms sustained major damage. Examples of major damage include bent structural supports, collapsed rig derricks, severely damaged production vessels and piping, overturned helidecks, and collapsed living quarters. The integrity of the structure does not have to be compromised for the MMS to consider the damage major.
It is encouraging to note that the majority of the damage reported is primarily associated with production equipment and drilling rigs on the deck, rather than damage that would compromise the integrity of the structure itself. Nevertheless, several jacket type structures in relatively shallow water suffered damage to primary braces and jacket legs. Five drilling rigs suffered major damage.
Ivan damaged 12 large-diameter pipelines (10 in. or larger) in federal waters. In addition, the mooring system failed for several drilling rigs. As the drilling rigs drifted, their mooring systems damaged several pipelines.
Structural failures
Two braced caissons in relatively shallow water, installed between 1985 and 1988, failed. The water depth at one location was less than 80 ft. The water depth at the other was less than 120 ft.
The latest edition of American Petroleum Institute RP 2A, which was released in December 2000, is the 21st edition. It includes a provision specifically defining a braced caisson as a minimum structure subject to special recommendations regarding its design. Special recommendations include a limit of 0.85 on stress interaction ratios and a safety factor of 1.5 on the foundation overload case. Primary API wave load criteria for L-1 structures are the 100-year return period storm. However, if the water depth is less than 100 ft, API may class a braced caisson as an L-3 structure (low consequence), as opposed to category L-1 (high consequence) or L-2 (medium consequence).
The L-3 classification has a reduced design wave height requirement. Thus, for braced caissons in greater than 100 ft the latest API criteria is more stringent than when these braced caissons were designed. For braced caissons in less than 100 ft, the net effect of API RP 2A revisions cannot be readily determined, but could possibly be less stringent than earlier editions of API RP 2A.
Four of the platforms destroyed, installed between 1969 and 1972, were in water depths between 232 and 255 ft with deck heights between 40 ft and 46 ft. All followed design requirements based on earlier editions of API RP 2A. API released the fourth edition of its RP 2A in 1972.
As a consequence of structure failures believed caused by waves inundating decks, API has increased the minimum deck height requirement. Current API RP 2A criteria require the minimum elevation of the underside of the deck to be 49 ft above Mean Lower Low Water (MLLW) for these structures. Given their water depth, location relative to the continental shelf, and the magnitude of the waves, perhaps a shoaling effect occurred, that is the elevation of the wave crest above the still water level increased as the water depth decreased. This increases the probability of the wave inundating the deck.
We believe the failure of the eight-pile fixed platform installed in 1984 in 479 ft of water was due to mudslide movement in conjunction with the direct effects of Ivan. The intensity of the soil movement during Ivan exceeded expectations.
Damage
One could categorize major platform damage into two groups: failure of primary structural components such as main braces, jacket legs, deck legs, and piles, and displacement of deck equipment such as quarters modules and drilling rigs.
Loadings caused by wave inundation of the deck appear to be the prevalent cause of damage to the integrity of the structure. Inundation of the deck dramatically increases the horizontal load and overturning moment on the structure, resulting in the potential failure of primary structural members and collapse. Wave inundation of a cellar deck can increase the wave load by 50% and the corresponding overturning moment by 75%.
Damage of primary structural members due to wave inundation occurred mostly on shallow water fixed base jacket type structures designed when deck elevation requirements were less stringent.
A significant portion of reported structural damage is associated with the displacement of production equipment, quarters, and drilling rigs on the deck. For some structures, damage can be attributed to inundation of the deck by the wind as well as the wave. Perhaps deck accelerations contributed to the displacement of equipment and drilling rigs, particularly for floating and compliant structures.
Most of the jacket type structures that experienced major damage did not have drilling rigs in place at the time of the storm. Thus failures occurred even though these structures were not subject to the largest gravity loads considered in the design. An inadequate deck height for this storm was probably the greatest contributing factor.
The exposure to Ivan and subsequent performance of the deepwater fixed (compliant composite leg platforms and guyed towers) and floating platforms (spar or TPL) validate these design concepts. The major damage reported on these platforms is not integral to the integrity of these structures, but rather associated with displacement of items on the deck. It should be noted that Ivan passed quite close to several of these deepwater platforms.
The MMS has stressed that all production wells lost to Ivan (approximately 80) will require complete abandonment. Given that some of these wells were not producing, the operator could have plugged and abandoned the wells before Ivan at a considerable cost savings.
Pipeline damage falls into two categories: a physical impact and a lack of stability to withstand the effects of storm-induced loading. Examples of physical impact include the failure of the host facility or an anchor drag. Examples of lack of stability include bottom current loading or foundation failure. In particular, pipelines in mudslide sensitive areas near the mouth of the Mississippi River were damaged. Locating and repairing damage to these pipelines will require some effort. As a result of Ivan, as much as 20-30 ft of mud covers some pipelines.
The risk of pipeline damage is reduced when the risk of losing the host structure is reduced. Burial of pipelines in deepwater to reduce loss due to anchor drag is not cost effective. However, burial may be cost effective for key trunk lines when the loss of revenue is factored in.
Whenever possible, pipelines should avoid areas subjected to major slope failures and subsequent mud flows. Burial may protect a pipeline at the toe of a slope failure, but soil flow could carry off a pipeline in the heel. Pipelines with sufficient wall thickness to allow them to plastically deform may allow the pipeline to survive displacements when located in area subjected to minor slope instability.
Design criteria
In addition to federal regulations, the primary source of design criteria for structures in the GoM is API publications such as API RP 2A. As noted previously, primary API wave load criteria for L-1 structures is the 100-year return period storm.
Oceanographers within the industry are gathering and analyzing Ivan’s data. It is unlikely that the data from a single event, when incorporated into a database that includes a large population of storms at multiple locations across the Gulf, would significantly alter the distribution of the database. Thus, one would not expect Ivan to result in significant change in the definition of a 100-year storm event. Conceivably, API could modify its design event without changing the definition of a 100-year storm. For example, the design event could be the 200-year event rather than the 100-year event, or the design event could remain the 100-year event, but minimum deck elevations could be increased.
It is interesting to compare estimated maximum wave heights associated with Ivan to current API design wave height recommendations. API RP 2A prescribes a design wave height versus water depth for structures in each exposure category L-1, L-2, or L-3. For a typical L-2 exposure category structure in water depths between 250 ft and 400 ft, the design wave height is approximately 63 ft. For a structure in the L-1 exposure category, the design wave height for water depths between 250 ft and 1,000 ft is approximately 70 ft. Recall that estimated actual maximum wave height for Ivan is approximately 90 ft, which exceeds the L-1 exposure category recommendation.
A number of changes in recent editions of API RP 2A directly impact the survivability of offshore platforms. The 20th edition of API RP 2A, released in 1993, included an updated wave load recipe. This included consideration of wave directionality and increased minimum deck height criteria. The 21st edition includes a provision specifically defining a braced caisson as a minimum structure subject to special recommendations regarding their design. This 21st edition of API RP 2A also included Section 17, The Assessment of Existing Structures. This does not impact new designs, but does address the safe use of older existing platforms.
The theme of API RP 2A Chapter 17 is managing risk. In API RP 2A, this is expressed in terms of an Exposure Category based on Life Safety and Consequence of Failure. Passing the current API RP 2A assessment criteria does not guarantee a structure will survive a Category 3 (winds between 111 and 130 mph) or greater storm, but rather that the MMS will accept the risk of losing a structure where there is no threat to life or the environment. The owner may be willing to accept this risk on less important structures, but monetary considerations may dictate a structure with increased capacity in other cases. An L-3 structure passing the assessment criteria has a 50% chance of being lost if subjected to a Category 3 storm. For L-2 structures, the risk of failure is reduced. For L-1 structures, the risk of failure is small.
The seven structures that failed met the API RP 2A 21st edition, Chapter 17 assessment criteria for existing structures, which has been adopted by MMS. With regard to braced caissons, the 21st edition of API RP 2A designates braced caissons as minimum structures subject to more stringent criteria. Regarding the four- and eight-pile platforms, where deck height is believed to be a contributing factor, the 20th edition of API RP 2A (1993) increased the minimum deck height design requirement. With regard to the eight-pile structure where a mudslide appears to be a contributing factor, API RP 2A does state that earth movement is a potential environmental load to be considered. However, API RP 2A makes no specific recommendations for the consideration of mudslides in combination with the design event. API does address the unique nature of mudslides by requiring a Certified Verification Agent (CVA) review regardless of water depth.
It is difficult to determine if the platforms Ivan destroyed would have survived had they met current API RP 2A design criteria. Perhaps the criteria used in the design of the failed platforms exceeded the requirements at the time of their design. It can, however, be said that deck height is the primary factor in determining the survivability of a structure for an event such as Ivan.
Future design criteria
The industry is reviewing data obtained from Ivan, and the MMS has requested proposals for studies of the consequences of Ivan. Information obtained from Ivan may be incorporated into API and MMS activities via existing committees and joint industry projects.
For example, API has a committee whose charge is to reorganize RP 2A. This task began long before Ivan. New platforms will continue to be addressed in API RP 2A. Those sections of the current edition of RP 2A associated with the assessment of existing platforms will form the basis of a new API publication RP SIM (Structural Integrity Management).
In addition, API will remove some sections of RP 2A associated with specific design requirements, such as fire and blast, creating a third API document. This reorganization will provide a risk management perspective, which should be a format more conducive to the inclusion of lessons learned from Ivan.
From a risk perspective, some damage is acceptable. Damage is not acceptable if it results in the loss of life or environmental damage. From an economic perspective, for a given probability of occurrence, the investment required to avoid damage must exceed some fraction of the cost required to repair the damage. Thus, given no lost lives or environmental damage, for a low probability of occurrence event such as Ivan, one would expect minimum impact on design.
Full implementation of API RP 2A, Section 17 recommendations may, in some cases, prompt more rigorous maintenance. This will reduce the number of members compromised by corrosion.
Given the dislocation of deck equipment, quarters, and drilling rigs in the absence of wave inundation, tie-down requirements of these items necessitate a close review. Floating structures may require additional consideration where the major damage was limited to dislocation of deck equipment, quarters, and drilling rigs, but was not necessarily associated with deck inundation. Review of the requirements for securing equipment, quarters, and drilling rigs on the deck, giving particular attention to the inertial loads created from the accelerations associated with floating structures is necessary.
Mudslides played a major role in pipeline damage and contributed to the failure of one structure. Perhaps future editions of API RP 2A will provide more specific recommendations regarding the intensity of mudslides considered in the design event.
A review of deck height is needed to ascertain if the risks associated with the current deck height are acceptable. Recall the four structures that failed in approximately 250 ft of water with deck elevations between 40 ft and 46 ft.
Based on current API RP 2A criteria, the minimum deck height would be 49 ft. The estimated maximum wave height associated with Ivan is approximately 90 ft. This would, including storm tide, result in a crest elevation of 58 ft above the water surface. Comparing this estimated crest elevation to the actual deck height of these four structures it can be concluded that Ivan’s waves probably inundated the decks of these structures, resulting in their failure. The current API minimum deck height requirement of 49 ft, which includes an air gap of 5 ft, is less than this estimate of Ivan wave crest elevations. Given this minimum deck height of 49 ft, a wave with a height of 74 ft would contact the deck.
The API RP 2A design criteria provide some latitude in the analysis requirements. For example, a simplified fatigue design is acceptable for structures meeting certain criteria such as water depth less than 400 ft for a period less than 3 sec. Given the advances in computers, designers are performing more analyses and employing more sophisticated design tools for conventional designs in shallow water. Included in these more sophisticated design tools are dynamic and non-linear analysis and ultimate strength analysis. Ivan may stimulate this trend. Additional data obtained from more sophisticated analyses can aid in the design. For example, analysts can use acceleration from motion studies in the design of tie-downs for the equipment, quarters, and drilling rig on the deck.
There are many lessons to be learned from Ivan. Investigations continue, followed by evaluations, reports, and technical papers. Perhaps API will change some design criteria. In many respects, Ivan is a validation of current offshore design methodology. There were no failures of structures designed in accordance with the latest criteria. The probable causes of failure in older structures have been addressed in the latest criteria.
