Overview of wellhead fatigue monitoring

Paul Forman, Dan Walker, John Henderson, Jim Maher

October 14, 2013

Wellhead fatigue is acknowledged as a technical challenge in the oil and gas industry. In an effort to advance industry capabilities and better inform operational decisions, BP is developing a system for instrumentation and data interpretation to provide enhanced insight regarding wellhead loading on a real time basis. The following activities were taken to define the system concept:

These activities led to development of simpler, more robust system concept that provides better data, faster. Prototypes are currently in development. The system is not dependent on specific mechanical or control details of the BOP, and can therefore be retrofitted to existing BOPs.

History

The original and current riser instrumentation systems are “data loggers.” These self-contained systems were designed to obtain better understanding of vortex induced vibration (VIV) on deepwater risers. Their use has been extended to measure BOP accelerations, which can be used to estimate wellhead fatigue.

However, there is significant delay between measurement and availability of data for interrogation. Therefore operational decisions are often made based on extrapolation of pre-analysis results. Real time systems have been proposed but rely on the use of cables to carry the high-bandwidth data to the surface for processing. Cables can impact riser running and are susceptible to the major hydrodynamic loads in the splash zone.

Historical data comparison

BP’s riser monitoring database dates back to the early 2000s and includes data from the Deepwater Gulf of Mexico and the North Sea, as well as more recent data from the Mediterranean, Caspian Sea, and other areas. This database was analysed to better understand how well standard industry analytical methods predict riser and BOP motions. The conclusion is that although the analytical predictions are conservative, they are adequate for riser dynamics. However, the predictions are generally more conservative for BOP dynamics and typically lead to higher fatigue loading predictions than would be calculated from observed BOP motions.

Fig. 1 illustrates a common data disparity, in LMRP RMS accelerations vs. wave height. Original predictions are in blue and summary statistics for the observed BOP motions are in red. Although some scatter is expected, in an accurate prediction the observed data points would spread equally above and below the blue analytical prediction line. Instead, the data is bounded by the analytical results and most data is well below the line.

The location of the majority of the data points below the blue line indicates substantially less loading was accumulated than originally predicted. The highlighted band for Seastate 3, for example, shows less than 10% of the originally predicted loading.

Similar comparisons were performed on other data from a variety of regions with similar results. However, there were cases in which more vibration was observed than was predicted. Practical mitigation options exist to reduce the accumulation rate for those rare circumstances.

Based on this analysis, a real time system concept was developed to:

Maximize safety and operational efficiency by more accurately calculating the fatigue loading on the wellhead.

Provide rapid feedback on events resulting in greater than anticipated loading to inform timely operational decisions.

Correlation of displacement to stress and simplification of calculations

Most riser instrumentation systems were designed to detect stresses along the riser above the flexjoint (i.e. above the LMRP/BOP). These complex systems require resolution of riser dynamic modes and allocation of various amplitudes to the participating modes. Additionally, a large number of sensors (8-10) and full time series data is required at the point the computations are performed. Wellhead loading is traditionally calculated by imposing the time series data on a finite element analysis (FEA) model. This process is computationally intensive and typically performed when a large data set is available for processing. The complexity and need for large amounts of data in this method negates the use of a real time system.

Due to a near-linear relationship between wellhead and BOP stresses, the number of sensors and amount of data can be reduced. A portion of the processing can also be performed on the sensor prior to transmission to the surface. Therefore, data transmission needs are reduced and the need for a cable is eliminated.

Points of interest for wellhead fatigue are typically in the vicinity of the Low Pressure Housing (LPH) or at the first connector (~30ft below mudline). The stress-displacement relation at the first connector is almost perfectly linear. The relation at the LPH is close to linear and can be approximated by a linearity assumption. Fig. 2 shows the full displacement time series plotted against the full stress time series, thereby eliminating the time dimension and illuminating the displacement-stress relationship.

A linear slope can be determined from Fig. 2 and used to recreate the stress time series as long as the displacement time series is known. In Fig. 3, this slope is used as a transfer coefficient and the two signals look the same, demonstrating that the displacement data can be used to approximate the stress signal.

Fatigue calculations are performed using either time domain or frequency domain. If using time domain, a linear transfer coefficient can be used in conjunction with the sensor displacements and the resulting signal can be processed to obtain cycle amplitudes. The histogram is transmitted to the surface for further fatigue processing. If using frequency domain, transfer functions can be derived based on the results of the pre-analysis. Suitable wideband frequency-domain fatigue methods are available to process the signals. Transmitting the spectrum to the surface is advantageous as it can be used for diagnostic purposes. The amount of data transmitted is reduced by at least a factor of 10 and perhaps as much as 100, depending on the method used. This is an advantage since acoustic transmission at low bandwidth is more battery efficient and more reliable.

Instrumentation capabilities

A survey of current industry capabilities was conducted in order to assess how these findings could be combined with equipment to create a more optimal system to achieve the operational and knowledge-generation goals. The intent of the survey was to collect information on advances in a broad range of industry niches, including subsea positional survey, dynamic positioning/beacons, drilling riser angle measurement systems, etc. The concept is to combine technological advances with the improved algorithms to provide an improved system.

Two of the most relevant equipment advances are:

1. Acoustic systems – Given the data simplifications, an acoustic modem can be used to transmit data. This has several advantages including near real time data without the need for a cable and minimal interfaces with the BOP/riser mechanical systems. The latter is important because this system can be retrofitted onto existing facilities.

2. Low power sensors – MEMS (microelectro- mechanical systems) accelerometers reduce the power requirements by 50-100 times. This allows systems to be deployed for 1-2 years at a time, operating continuously without needing a change-out operation. However, there is a reduction in resolution.

Although the above listed two improvements are enabling, the following improvements provide substantial increases in potential functionality of such a system:

1. Optical modems – High bandwidth downloads through short distances of water have progressed recently. Although the system described results in little need for full time series data, there is an advantage in the ability to periodically download full data for diagnostic purposes. An optical modem and associated ROV download device can enable this without introducing significant complexity.

2. Processing capabilities – The miniaturization of processing and storage advances the use of on-board algorithms.

3. Improved error rejection algorithms – Many algorithms from digital signal processing and inertial navigation can assist with identifying errors at an early stage. This is important for on-sensor processing because errors are less visible. Moving beyond the traditional “person in the loop” processing will require the use of additional algorithms.

Synthesis, system definition Based on the improvements and simplifications listed above, it is possible to assemble a system with the following features:

Conclusions

A real time system to measure wellhead loading is in development. This system will provide data that more accurately reflects the loads on the wellhead. Additionally, there is a reduction in the amount of data and a simplification of the data processing. This allows for more rapid data transmission to the surface without the need for a cable. This system will help inform operational decisions in a more accurate and timely manner.OE

Paul ForemanPaul Forman is Vice President for Wells Engineering in BP Exploration Operating Co. Ltd. He is responsible for strategic direction and business delivery of Wells Engineering. He leads a team of engineers in the organization’s drive for consistency, rigor and standardization in how BP conducts engineering design for well construction across the full lifecycle of the well. Forman has 24 years’ industry experience and earned a BS in Engineering at RGIT (now Robert Gordon University), Aberdeen.
Dan WalkerDan Walker is Global Wellhead Integrity Program Manager for BP America Inc. He graduated from The University of Oxford in Engineering Science, completing a M.Eng in 1999 at St. Catherine’s College and D.Phil at Magdalen College in 2003. He was recruited by BP as a technical specialist in the area of offshore hydrodynamics. Walker has worked in Azerbaijan, Russia, Alaska, Norway and Trinidad, and has covered technical areas in BP ranging from deepwater oil and gas exploration and production to offshore wind projects. Walker has also worked for BP in the Gulf of Mexico and Angola and continues his role as relationship manager for BP’s engagement with the University of Oxford.
John HendersonJohn Henderson is a senior drilling engineer working in BP’s Global Wells organization and is currently based in the UK. He has worked in various operational engineering roles in the North Sea, Middle East, Africa, and South America. Most recently he has been involved in the subsea wells development planning for Shah Deniz Stage 2 project in the Caspian Sea. Henderson has 28 years of industry experience.
James MaherJames Maher is a consultant focused on innovative solutions for floating systems, risers, and equipment for the floating drilling and production industry. He has founded several deepwater technology companies and has been involved in the development and commercialization of many technologies. He was involved in the development of several generations of spar technology, in his role as the Spar Product Manager for Technip and other engineering management roles. Maher has a BS in Mechanical Engineering and a BA in Government Studies from the University of Notre Dame and an MSE in Civil Engineering from Purdue University.
Michelle EdwardsEdited by:
Michelle Edwards
is a technical writer and editor for Barrios Technology. Dr. Edwards has worked in biotechnology where she assisted in the management of a small start-up company. She then moved on to NASA’s Johnson Space Center where she led a team in the development of an architecture to identify and analyze risks to humans from long duration space flight. Dr. Edwards has a B.A. in biology and a Ph.D. in neuroscience.