Production riser design: Integrated approach to flow, mechanical issues
Environmental conditions at a number of deepwater sites that BP participates in to determine influences on risers. [34,084 bytes] Temperature profiles for a 100-mile flowline and 10,000-ft riser system indicate that there can be significant expansion cooling effects in the riser, leading to potential flow restrictions or blockage. [10,361 bytes]
Predicting riser-to-riser interaction
A. Moros, P. Fairhurst
- Environmental conditions at a number of deepwater sites that BP participates in to determine influences on risers. [34,084 bytes]
- Temperature profiles for a 100-mile flowline and 10,000-ft riser system indicate that there can be significant expansion cooling effects in the riser, leading to potential flow restrictions or blockage. [10,361 bytes]
- Current data from Nyk High (Norwegian deepwater) and Schiehallion (UK deepwater). While Schiehallion currents exhibit typical shear profiles, Nyk High currents show strong currents at the surface decaying at about 200 meters below the surface and increasing again before decaying with a monotonic shear profile. [9,082 bytes]
- Vortex induced vibration measurements obtained from accelerometers on drilling riser assemblies at Nyk High and Schiehallion show there is pronounced change in current direction with water depth at Nyk High. [7,890 bytes]
- Coplanar impacts betewwn risers [24,404 bytes]
- Design steps and consideration. [8,307 bytes]
For deepwater developments with floating production systems, there are four main risers systems which are considered:
- Top tension vertical risers
- Steel catenary, shaped catenary or wave catenaries
- Hybrid, a combination of vertical bundled pipes supported by buoyancy at a certain water depth below the surface
- Flexible pipes connecting the top of the vertical bundle to the floating host facility. The latter is the system which will be used in the Girassol development.
Water currents will cause large deflections and vortex induced vibrations (VIV) with a tendency of reducing the life of the riser system. These effects can be mitigated by the addition of vortex suppression devices which eliminate the VIV, but not necessarily the drag loads in the system. Waves can cause large forces that need to be taken into consideration in a design.
A variety of internal flow conditions also exist depending on the field. The fluids transported can range from viscous oils to volatile gas condensates and the characteristics can also change significantly with depletion of the reservoir. Most notably, the water cut can rise to levels of 90% or more.
Most deepwater flowlines and risers are required to transport unprocessed multiphase flows in which the flow regime is more likely than not in slug flow. This gives rise to variations in the gas and liquid density and velocity at any point in the transportation system. These variations produce fluctuations in the loads at bends and also affect the mass of fluid in the riser. Even if normal slug flow is not present, flow rate ramp-ups, shutdowns, restarts, and pigging can lead to slugging, and this must be allowed for in the mechanical design.
In addition to these hydraulic factors, there are also flow assurance issues caused by the low temperatures that result from long distance transportation of the unprocessed fluids. Temperature profiles for a 100-mile flowline and 10,000-ft riser system show in an accompany graphic indicates that there can be significant expansion cooling effects in the riser. This can lead to potential flow restrictions or blockage as a result of wax or hydrate formation. Insulating risers may not resolve this problem and some form of adding heat to the fluids may be required.
In deepwater developments, the risers can be exposed to severe currents resulting in large deflections and vortex induced vibrations as well as severe internal flows associated with slugging. The riser design needs to take account of all these complex and interactive phenomena.
All these affect the mechanical design of the risers and the mechanical features also feed back into how the fluids are transported efficiently. For example, the shape of the riser can have a large impact on the internal slugging characteristics and this can lead to large variations in the gas and liquid flows that have to be dealt with by the topsides processing facilities.
In the remainder of this paper we will briefly describe our thoughts and the work we, together with partners in the industry, have been doing to address some of the areas which affect the design of the risers mainly:
- Flow-induced vibrations (VIV)
- Riser array behavior for top tension risers
- Effects of internal flows on the dynamic behavior
Flow induced vibrationsWhen a cylindrical structure or a riser are exposed to a current, vortices are shed alternatively from side to side, resulting on an oscillating force on the cylinder. If the frequency of oscillation is close to the natural frequency of the riser, the resulting vibrations can be substantial with an amplitude equal to the riser diameter. Such high oscillations introduce severe stresses and can cause early fatigue failure of drilling, work over and production risers and wellheads.
The industry relies on semi-empirical models to predict the vortex induced vibration (VIV) related fatigue. Such code that we use in BP is "Shear7." The designers need to know the accuracy of the predictive tools to establish the necessary conservatism and determine with confidence the requirement for VIV suppression devices. The accuracy of these models depends on good validation against appropriate data. Calibration at present, is limited to data obtained from scaled models, and until now, there was no appropriate full scale data to calibrate the models.
As part of our continuous commitment to improving and validating the models used for design, BP conducted measurements on a 375-meter drilling riser in West of Shetlands over the entire drilling period of three months. In parallel, we are working with our partners in the Norwegian Deep Water Programme (NDP), and now have collected an extra three data sets from drilling campaigns in Nyk High (BP lead) at 1,300 meters water depth, Vema (Statoil lead) at 1,290 meters water depth, and Helland Hansesn (Shell lead) at about 800 meters water depth. The riser response was measured with accelerometers at four locations along the riser.
The current profile was monitored simultaneously with acoustic doppler techniques. Water column current studies for Schiehallion (UK deepwater and Nyk High (Norwegian deepwater) show that although the Schiehallion currents exhibit typical shear profiles, the Nyk High currents show strong currents at the surface decaying at about 200 meters below the surface and increasing again before decaying with a monotonic shear profile.
Again, the current direction for West of Shetlands (Schiehallion) is fairly uniform, apart from near the seabed. In Northern Norway, on the other hand, there is more pronounced change in current direction with water depth. The available VIV models cannot handle such complexities and can only have a monotonic current profile as input.
The riser response was measured with accelerometers (the Robit DACOS system), positioned at different locations on the risers. The units were battery powered, collecting data over the entire drilling campaign. Data generation showed the variation of Reynolds number with reduced velocity for different modes and a variation of the measured shear fraction, versus the potentially excited mode, based on the Schiehallion data. There did not appear to be a clear boundary between multi and single mode response at a mode response of 3. In the region defined as single mode, it appears that there is a multi mode.
Our work on the validation of Shear7 continues and we are now in the process of examining the best set of code parameters that fit the experimental data.
Riser array behaviorWhen top tension risers are used, one needs to determine the spacing between the risers such that the forces generated by the currents do not result in the risers interfering with each other. In most current designs, the approach followed is to avoid clashing. Various softwares and models exist which are used to estimate the required spacing, indicate whether clashing occurs and where.
As the industry moves toward deeper waters and harsh environments, there will be scenarios where clashing may have to be accepted in order to have an economically feasible design. To ensure that clashing between the risers is acceptable, the levels of impact need to be estimated and its consequences assessed. The design sequence can be summarized in an accompanying flow chart.
To achieve this, we have put together three main tasks which need to be addressed in order to have an integrated approach to the design. We believe that one needs to understand:
- What are the impact forces or energy levels during clashing?
- What is the damage caused to risers by such impacts and is it acceptable?
- What should be put in place to mitigate any impacts?
For this scenario, the equivalent mass before the impact worked out to be the same as that of 21-meters length. The effect of both single and multiple impacts has been considered. In this analysis, only coplanar impact was considered.
Impact simulations indicate that an impact event occurs typically over a 6.5 msec period. Drilling risers (both steel and titanium) experience elastic deformations with no residual strains after the impact. In general, because of the larger wall thickness, the drilling risers are not affected by the impact for the energy levels studied. The production risers showed high levels of residual stresses and wall ovalization following the impact. The impact is taken as circumferential stress. Successive impacts result in increased plastic strain of 50% of the first impact.
With a knowledge of the generated stresses, one can then design protecting barriers to withstand the impacts.
Effects of internal flowsDuring the production of hydrocarbons in deepwater, slugging is expected in the production risers. Changes in the internal fluid density of deepwater risers could have a significant effect on the dynamic riser configuration and can give rise to large displacements.
Previously, in relatively shallow water scenarios, it was proved that the slugging will not affect the dynamics of a riser. In water depths greater than 800 meters, the slugging is more severe and the impact could be significant.
Our objective in BP has been to quantify these effects, and set guidelines when such loads due to internal flow should be included in the riser design. The original BP work has recently been extended under the NDP, where the slugging characteristics and their effect on riser dynamics have been examined for two riser configurations. Typical slugging characteristics at different locations along a wave catenary riser in 1,200 meters were studies (see accompanying figure).
The impact of slugging can result in large excursions at the hog bend, up to 180 meters, and large bending moments at the touchdown point. Fatigue due to slugging alone could be an issue. The benefits one has with the wave catenary riser over a simple steel catenary riser (SCR) in particular at TOP, might be superseded when severe slugging is present.
Our work has indicated that this effect was very pronounced for wave catenary risers, but less for SCRs. At present, the effects of the internal flow behavior are not included in the mechanical design of the risers. It is our view, that although the effect of slugging loads is not show stopper in the feasibility of a riser system, such loads should be considered in detailed design on a case by case basis.
ConclusionsWe are continuing the calibration work for the VIV predictive tools such as Shear7 and VIVA in collaboration with the MIT and NDP consortia. It is expected that by the end of next year, both codes will be validated against the full scale data.
A large number of experiments have been planned for this year to quantify the impact loads during clashing and the loads that will cause damage to the risers as part of NDP. We are also working on issuing guidelines on the effect of slugging loads for different riser systems and when such loads should be considered during detailed design.
Editor's Note: This article is a shortened version of a paper presented at the Deep Offshore Technology Conference (DOT '98), held in New Orleans in November, 1998.
Copyright 1999 Oil & Gas Journal. All Rights Reserved.