Improving prediction accuracy on flow-induced vibration analysis

Flow-induced vibration (FIV) from high-velocity multiphase flow can lead to fatigue failures and hydrocarbon leaks. A combination of computational fluid dynamics (CFD) and finite element (FE) modelling offers a potentially powerful tool for assessing and diagnosing multiphase FIV problems in hydrocarbon production piping systems. It can be extended to CO₂ and hydrogen systems, which are becoming ever more important, as we transition to alternative energy sources.

World Oil Editor-in-Chief Kurt Abraham discussed this issue with Xodus’ Paul Emmerson and Equinor ASA’s Steinar Orre, who have worked together on a variety of R&D projects and are spearheading changes to industry guidance, to reflect the increase of data and the impact of the energy transition.

World Oil (WO): Flow induced vibration (FIV) from high-velocity multiphase flow is a common source of vibration concern in process piping, but why is this phenomenon a concern for subsea systems in particular?

Steinar Orre (SO): Process piping in subsea systems often needs to accommodate movements from well growth (thermal growth), so it has to have some flexibility, Fig. 1. This is typically the case for piping downstream of subsea chokes from an individual well toward a manifold pipe. As these branch pipes or flex loops are typically more flexible than process piping in topside systems, they are more vulnerable to varying multiphase forces.

Fig. 1. Subsea piping, showing a production well jumper (grey) connecting into an in-line tee (ILT - yellow) subsea structure. The jumper has to be flexible, allowing expansion between the well and the ILT. The ILT consists of a complex combination of elbows and tees that are vulnerable to varying forces from multiphase flow. Both require careful design to mitigate any FIV issues. Image: Subsea 7 and Equinor.

Paul Emmerson (PE): As subsea piping (Fig. 2) is often quite inaccessible, it’s difficult to monitor. Visually, it’s not like topside systems, so FIV is something that should be taken very seriously. Subsea failures can be catastrophic, causing significant pollution damage to the environment and potentially leading to loss of life.

WO: While the consequences of inaccurate predictions of piping vibration vary, for a fatigue assessment, it could significantly affect life estimates and hence operational safety, or production rate limits, which may significantly impact operational profitability. What steps can be taken to avoid FIV?

PE: I would promote FIV screening on every piping system, both topside and subsea. It’s a simple check, a spreadsheet type of calculation, based on straightforward input data, such as fluid density, velocity, pipe diameter, wall thickness, and distance between supports. If it can be done as a matter of course, and as early as possible in the project lifecycle, it will help to identify or highlight any issues. Simple changes to the process conditions, such as reducing the flowrate, increasing pipe diameter or reducing the distance between supports can then be made. The other option is to proceed with a more detailed FIV analysis, Fig. 3.

SO: Quite often these subsea systems, like flex loops, are based on standard dimensions and the earlier you can start screening, the better to ascertain vibration potential. As you proceed through the project phases, the level of detail should increase. In my opinion, one should also work more towards instrumentation, which is typically not done with respect to vibration, with the use of online sensors like accelerometers to monitor vibration levels, Fig. 4. In an attempt to have an overview of the potential of vibration on each producing well, Equinor has started to carry out production monitoring, based on production parameters and has developed online FIV indicators on all subsea wells. Some of them are quite refined and advanced, while some are very simple.

WO: How does Equinor reduce the risk from FIV at different stages of the project lifecycle?

Fig. 2. Inspection of a typical pipe section on a subsea production system. This pipe connects a production well to a joint manifold pipe bridging several wells into a subsea flowline. These subsea pipes are vulnerable to varying forces from multiphase flow and can display low-frequency vibrations.

SO: Equinor’s technical requirements state that we have to do screening. We use the AVIFF approach from the UK’s Energy Institute (EI) and aim to start that as early as possible. We also work towards monitoring, either with proper instrumentation or just using production parameters. From research projects, we try to estimate the risk of vibration from the production parameters such as the oil rate, gas rate, water rate, pressure, and so on. Quite often, we see that using standard screening approaches is not sufficient, so we need to do a more detailed analysis. In some cases, we use expert companies like Xodus, for this.

WO: Little has been published on how well CFD performs as part of FIV modelling techniques for operational hydrocarbon production systems. Why is that, and how do these methods contribute to a better understanding of fatigue assessment and thus, improved production?

Fig. 3. Predicted liquid-gas flow (using CFD) and dynamic stresses with a snapshot of the exaggerated pipe displacement (using FEA) for a rigid jumper as illustrated in Fig. 1.

PE: Historically, much of the CFD validation has been performed against test data, where experiments have been conducted at atmospheric pressure, using just air and water. It’s much easier and cheaper to perform these types of tests, across a broader range of flow regimes, in laboratories than in the field. Getting good-quality field data from hydrocarbon production systems is actually quite difficult.

Much of the work, to date, has been done at low pressure, and that’s been good, up to a point. It has enabled better correlations for fluid forcing for these types of FIV assessments, as well as more accurate validation of CFD methods. But there’s still some way to go.

To replicate real life, we want to be able to get the methods working more accurately at high pressure for hydrocarbon systems. In 2018, Equinor’s team carried out tests at Porsgrunn, where they looked at both low- and high-pressure flow tests¹ on a hydrocarbon fluid system, on a simple flexible loop. While only vibration was measured, they were able to back-calculate the fluid forcing from the vibration measurements, using an FEA model of the loop. The results showed that the vibration actually reduced as the pressure increased. Some work was also done with CFD by Xodus (Fig. 5) and other collaborators² to replicate the test conditions, where we partially captured this effect, getting the trend right, but tending to overestimate the forces.

Fig. 4. Inspection of vibration levels with a camera from an ROV (remotely operated vehicle). A ruler placed in front of subsea equipment can be used for a first estimate of low- frequency vibration displacement.

It’s important, particularly for high-pressure systems, that we continue to improve our prediction accuracy. We’ve come a long way in the last ten years, but if we can use CFD, we can get more accurate solutions, which are a bit less conservative.

SO: One of the reasons that little has been published on this is that when we do have an operational vibration issue, we do not necessarily have precise production conditions in the pipe. Although modern platforms are heavily instrumented, flow measurements between oil separators are generally not available. This is one of the reasons why it’s difficult to take a CFD model and run cases and compare with measured vibrations.

A good example of an operating topside production system is the Tyrihans inlet pipe toward the Kristin platform, where Xodus performed some modelling work, and Equinor compared the measured vibrations. This was a useful exercise, where the modelled vibration response in the pipework compared favorably to the measurements. It was published in 2020 at the OMAE conference, winning best paper in its category (CFD and FSI Symposium).³

WO: Together, Equinor and Xodus have been involved in several R&D projects over the past decade, looking at FIV. What current R&D projects are ongoing?

Fig. 5. Predicted liquid-gas flow, using CFD through a bend of the Porsgrunn test loop, investigating the effect of fluid forcies at low (10 bar) and high (80 bar) operating pressures.

SO: We have perfomed—I think for the first time—an investigative experimental campaign on two-phase CO₂ flow in a relatively flexible pipe. This was a repetition of the high-pressure experiments in Porsgrunn¹ three years ago. There may be some CFD attempts on that data in the future.

We are now looking into a new campaign to study the effect of vertical piping in the Porsgrunn loop. We have some indication, based on data from operations and some CFD attempts from TechnipFMC. We think there is an effect in vertical piping, which is not yet fully understood. Together with Xodus, TNO and two subsea contractors—TechnipFMC and Aker Solutions—we have a small project underway to set up a proper “recipe” on how to do detailed FIV analysis.

PE: We have learnt so much from that project as we started looking at different approaches to the modelling for the first time. So, for example, comparing how structural modelling is performed, whether it’s done in the frequency domain, as a harmonic model, as opposed to doing it in a time domain, as a transient model, and comparing the results directly.

We are also starting to look at multiphase fluid damping. This is quite a complex situation in terms of modelling, and computationally, it is time-consuming. But as Steinar said, we are trying to incororate these types of approaches into the “recipe,” to make sure that the necessary assessment and analysis steps are taken correctly.

WO: What research projects would Equinor and Xodus like to see in the future, to improve FIV prediction capabilities?

Fig. 6. Predicted slug flow, using CFD through the JIP test loop.

SO: We have come a long way since 2014, when we initiated the joint industry project (JIP) with Xodus and TNO,⁴ Fig. 6. That was a huge step forward. We have since had a number of smaller projects after that JIP, and I believe that trend will continue.

One aspect that hasn’t been studied in great detail is single-phase liquid. Using standard screening methods, you often end up with a very conservative design, such as for subsea water injectors. Recent investigations on an injector indicated insignifcant vibration levels with water injection rates far beyond those being permitted by standard screening tools. We would like an improved methodology for how to assess single-phase liquid potential for vibration. More operational data subsea would be very useful for ongoing knowledge of response to real-life systems.

The other is the impact of vibration on renewables and CO₂, both single-phase or two-phase, depending on the pressure-temperature diagram. So far, we have done one experimental campaign, and there will be more. We also need prediction tools for CO₂, which is different than hydrocarbons, and also flow with hydrogen is quickly gaining prominence. Hydrogen is much lighter than natural gas, it’s a much smaller density, so to produce volumes, you typically end up with extremely high velocites, which can have potential for vibration. These are very important topics for future research.

PE: My wish list would include having more data at high pressure. There were a few gaps in the Porsgrunn flow regimes that were not possible to test, due to operational constraints. For example, annular flow conditions were not possible, and that would have been great to have, to add to the matrix of cases. Though expensive, where possible, we should perform tests using a larger piping system and measure the fluid forcing as well. Pursuing further improvements in the CFD modelling approach for high-pressure operating conditions and investigations into multiphase fluid damping would also be on the list. These are areas where we can do more work and could improve overall prediction methods. We could be spending more time evaluating and extracting information from the field data we already have access to.

WO: Care and a degree of balanced conservatism are key aspects to any kind of research. Are there any hurdles to overcome with this type of investigation and analysis, such as level of investment and where that funding comes from?

PE: When we worked on the JIP back in 2014, it was at a time when the oil price was reasonably high, and there was a good level of funding from quite a few operators and EPCs. With the oil price crash, funding became, for obvious reasons, restricted. More recently, the pandemic has not helped. I do think it would be great if we could get more collaborative projects, because that would help to increase the knowledge base on our understanding of FIV to improve assessment methods.

SO: In Equinor, we’ve always viewed FIV as an integrity risk for our operations. We try to be open and share experiences and, in quite a few cases, share data and operational process conditions. However, I think the industry would benefit from working more closely on this. This could mean having more operating companies involved. We could have larger, more-expensive experimental campaigns for these issues and see this is being done during the ongoing revision of the EI AVIFF guidelines. It’s very helpful when different operators share case studies and experiences, as it helps provide a safer environment for our people and safer designs for our systems.

WO: Does there need to be further regulation or guidance for FIV within the upstream oil and gas industry, or should there be any changes to the existing guidance?

SO: We had tried to develop an alternative screening method, but we ended up going back to the standard AVIFF screening for the first screening in projects. We think the AVIFF guideline is the place to start, but the section on how to proceed when you get red flags from the AVIFF screening could be more detailed and perhaps revised.

PE: I don’t think it is intended to be a step-by-step guide, but it should at least give some advice on things to look out for, and potential pitfalls. Hopefully, it will evolve, and there will be a section on this in the next issue. There are also plans to revise the subsea guideline document in 2022, with updates to the topside document as well as other industry learnings.

SO: One of the improvements in the current AVIFF guideline for the topside version is the less-conservative scaling for wet gas or gas condensate production, which typically means multiphase flow with a relatively small amount of liquid. In Equinor, we recommend this for subsea systems, although it’s not part of the official AVIFF guideline.

PE: The good news is that this will be brought into the subsea guideline, and it’s a perfectly acceptable approach. When we update the guidelines for the subsea version, we will have a lot more learnings over the last four or five years from the subsea sector, as there’s been more focus on vibration on subsea infrastructure than previous years across the industry. For single-phase liquids and high liquid volume fractions, Xodus is looking at previous multiphase test data, and we see it being at a similar correction factor for gas systems. We hope to have this included in the forthcoming revision of the guidelines; if not, then it will be at a later revision.

WO: Are there any additional items that you would like to mention that we haven’t already touched on?

SO: We support the development of FIV, as it aids our interest to develop digital tools to corroborate predictions and mitigate vibration problems. We also need more detailed plans of how to instrument subsea systems to actually monitor vibration levels. While there have been various attempts, there is no strategy on how to actually build subsea systems and how to monitor them to avoid dangerous vibrations. While some analysis can be done by just looking into the production data, I think there should also be a strategy for physical measurements of vibration levels, like having online accelerometers streaming data towards the topside.

PE: I think it would be good to bridge the gap between simulation and real-life responses of our systems. Though I appreciate it’s much easier for us to do that on topsides than subsea, it would be an improvement for the industry.

It's important to stress that we are getting faster all the time at doing these types of FIV assessments with improving computational power through cloud computing. Five or ten years ago, an analysis might take a week or two weeks. Now, we’re able to turn this around in about one or two days. So, there’s been a lot of progress made on that front and hopefully that will continue.

REFERENCES

1. Belfroid, S, N.Gonzalez-Diez, K. Lunde and S. Orre, “Multiphase flow induced vibrations at high pressure,” paper PVP2020-21139, Proceedings of the ASME 2020 Pressure Vessels & Piping Conference, Minneapolis,
   Minn. July 19-24, 2020.
2. Emmerson, P. R. et al, “Multiphase flow induced vibrations at high pressure: CFD analysis of multiphase forces,” paper OMAE2021-62873, OMAE Conference Online, June 21-30, 2021.
3. Emmerson, P. R., M. J. Lewis and N. A. Barton, Xodus Group Limited, London, UK and S. Orre and K. Lunde, Equinor, Stavanger, Norway, “Flow induced vibration analysis of topside piping at high pressure,” paper OMAE2020-18760, Fort Lauderdale, Fla., June 28-July 3, 2020.
4. Belfroid, S. P.C., E. Nennie, A. van Wijhe, H. Pereboom and M. Lewis, “Multiphase forces on bend structures – overview large scale 6-in. experiments”, 11th International Conference on Flow Induced Vibration, The Hague, The Netherlands, July 4-6, 2016.

About the Authors

Paul Emmerson

Xodus

Paul Emmerson is principal consultant at Xodus, and has 28 years of experience applying analytical and CFD modelling skills to solve complex fluid and thermal engineering problems. He is an expert in CFD, flow modelling and piping vibration and has published numerous papers at conferences and in technical journals.

Steinar Orre

Equinor ASA

Steinar Orre is the principal researcher for flow technology with Equinor ASA. He has worked in the oil and gas industry for more than a decade and specializes in fluid flow modelling, analysis and piping vibration. He has worked with Equinor since 2012.