pith. sign in

arxiv: 1906.11177 · v1 · pith:DHAAFQTBnew · submitted 2019-06-24 · ⚛️ physics.data-an · cs.LG· stat.ML

Data-driven prediction of vortex-induced vibration response of marine risers subjected to three-dimensional current

Signe Riemer-S{\o}rensen (2) , Jie Wu (2) Halvor Lie (2) Svein S{\ae}vik (3) , Sang-Woo Kim (3) ((1) SINTEF Digital , (2) SINTEF Ocean , (3) NTNU) This is my paper

Pith reviewed 2026-05-25 17:08 UTC · model grok-4.3

classification ⚛️ physics.data-an cs.LGstat.ML
keywords vortex-induced vibrationmarine risersthree-dimensional currentdata clusteringrandom forest regressionVIV predictionfatigue damage
0
0 comments X

The pith

Clustering 3D current experiment data groups riser VIV responses by flow direction and trains a random forest to predict them.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper takes results from one laboratory experiment in which a marine riser experienced vortex-induced vibrations under a three-dimensional current whose direction and speed change with depth. A clustering algorithm is applied to the measured flow profiles to extract parameters that appear to drive the observed responses. The responses themselves are then grouped by their statistical features, and these groups align with differences in current direction. A random forest regression model is fitted directly to the measured vibration data, and its output is compared with the predictions produced by the conventional semi-empirical tool VIVANA-FD. The motivation is that present industry tools rest on two-dimensional flow assumptions and must artificially flatten real three-dimensional profiles, which adds uncertainty to fatigue-life estimates.

Core claim

By clustering the experimental 3D flow data, measurable parameters that influence VIV responses can be identified; the riser responses group according to their statistical characteristics, which relate to the direction of the flow; and a random forest regression model fitted to the measured responses allows direct performance comparison with existing VIV prediction tools such as VIVANA-FD.

What carries the argument

Data clustering algorithm applied to experimental 3D flow profiles, followed by random forest regression trained on the resulting grouped VIV response statistics.

If this is right

  • Riser responses can be grouped by statistical characteristics that relate directly to current direction throughout the water column.
  • A random forest regression model can be trained on measured VIV data to produce predictions under three-dimensional flow conditions.
  • The data-driven predictions can be compared in performance against conventional tools that rely on two-dimensional flow assumptions.
  • The identified measurable parameters from clustering offer an alternative route to incorporating 3D current effects without first reducing profiles to two dimensions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If additional 3D-current experiments become available, the same clustering step could reveal whether the same direction-related parameters remain dominant.
  • The approach could be tested on other slender marine structures such as mooring lines or pipelines that experience similar 3D flow conditions.
  • Design workflows might eventually accept measured 3D current profiles directly rather than requiring manual conversion to equivalent 2D cases.

Load-bearing premise

The single available 3D-current experiment supplies a representative sample of the parameter space that governs VIV response under varying current directions.

What would settle it

Running the identical clustering-plus-random-forest procedure on data from a second, independent 3D-current riser experiment and finding that the model predictions no longer align with the measured responses or with VIVANA-FD outputs.

Figures

Figures reproduced from arXiv: 1906.11177 by (2) SINTEF Ocean, (3) NTNU), Jie Wu (2) Halvor Lie (2) Svein S{\ae}vik (3), Sang-Woo Kim (3) ((1) SINTEF Digital, Signe Riemer-S{\o}rensen (2).

Figure 1
Figure 1. Figure 1: Left panel: Vortex Induced Vibrations (VIV) due to vortices shed in the wake of a slender marine riser (seen from above). The VIV manifest as movements e.g. in a figure-of-eight like pattern (dashed red line). Right panel: Examples of riser trajectories in the x/y plane perpendicular to the length of the riser. 1 Introduction Slender marine structures such as deep water marine risers are exposed to ocean c… view at source ↗
Figure 2
Figure 2. Figure 2: Upper panel: The geometrical configuration of the arms (blue) and the riser (red). In the 2D configuration (left), the angle between the horizontal arms was 0◦ leading to a 2D flow. In the mild 3D configuration (middle) the angle between the arms was 60◦ and in the strong 3D configuration (left) the angle was 120◦ , both leading to a 3D sheared flow. Lower panel: Normalised velocities along the riser in th… view at source ↗
Figure 3
Figure 3. Figure 3: An example of a time series interval for sensor 5 in the 2D geometry with rotational velocity of 1.629 ms−1 in the sample interval 5000 to 5500. The upper panels show the time series and the lower panels the real part of the fast Fourier transform. Time series: The response of the riser was measured with 10 pairs of bi-axial accelerometers mounted along the length of the riser. The accelerations were measu… view at source ↗
Figure 4
Figure 4. Figure 4: Upper panel: Distributions of the statistical parameters of the clusters. The bars show the normalised histograms while the thin lines are the kernel density esti￾mates. The cluster numbers correspond to the cluster members shown below. Lower panel: Examples of trajectories of cluster members. The colours and labels indicate the configuration. The last row shows examples of data points that are assigned to… view at source ↗
Figure 5
Figure 5. Figure 5: Top panel: Trajectories of the case study; the strong 3D configuration with a bottom flow speed of 1.105 ms−1 for sample range 2500-3000. Subsequent panels: The measurements (black crosses) and model predictions for the case study. The blue circles show the random forest prediction from fitting to data from all sensors with the sensor positions highlighted as larger circles. The orange squares show the pre… view at source ↗
read the original abstract

Slender marine structures such as deep-water marine risers are subjected to currents and will normally experience Vortex Induced Vibrations (VIV), which can cause fast accumulation of fatigue damage. The ocean current is often three-dimensional (3D), i.e., the direction and magnitude of the current vary throughout the water column. Today, semi-empirical tools are used by the industry to predict VIV induced fatigue on risers. The load model and hydrodynamic parameters in present VIV prediction tools are developed based on two-dimensional (2D) flow conditions, as it is challenging to consider the effect of 3D flow along the risers. Accordingly, the current profiles must be purposely made 2D during the design process, which leads to significant uncertainty in the prediction results. Further, due to the limitations in the laboratory, VIV model tests are mostly carried out under 2D flow conditions and thus little experimental data exist to document VIV response of riser subjected to varying directions of the current. However, a few experiments have been conducted with 3D current. We have used results from one of these experiments to investigate how well 1) traditional and 2) an alternative method based on a data driven prediction can describe VIV in 3D currents. Data driven modelling is particularly suited for complicated problems with many parameters and non-linear relationships. We have applied a data clustering algorithm to the experimental 3D flow data in order to identify measurable parameters that can influence responses. The riser responses are grouped based on their statistical characteristics, which relate to the direction of the flow. Furthermore we fit a random forest regression model to the measured VIV response and compare its performance with the predictions of existing VIV prediction tools (VIVANA-FD).

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a data-driven method for predicting vortex-induced vibration (VIV) responses of marine risers in three-dimensional (3D) currents. It uses experimental data from one 3D current test to apply clustering to identify influential measurable parameters, groups responses by statistical characteristics linked to flow direction, fits a random forest regression model to the VIV data, and compares its performance to the existing tool VIVANA-FD.

Significance. Should the approach prove robust, it would offer an alternative to semi-empirical VIV tools that are limited by their 2D flow assumptions, potentially improving fatigue predictions for risers in realistic ocean conditions. The integration of clustering and random forest regression demonstrates a practical application of machine learning to a complex hydrodynamic problem with nonlinear dependencies.

major comments (2)
  1. [Abstract] The abstract supplies no performance metrics, cross-validation details, data volume, or error bars; without these the claim that the data-driven method 'describes' 3D VIV cannot be evaluated against the stated comparison to VIVANA-FD.
  2. The single available 3D-current experiment supplies the entire dataset for both clustering and random-forest training, so extracted parameters and model performance may reflect test-specific artifacts rather than general VIV drivers. This undermines the claim that the identified parameters are general measurable drivers of VIV response.
minor comments (2)
  1. Clarify the specific clustering algorithm used and the criteria for grouping responses.
  2. Provide details on the input features and target variables for the random forest model.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on the abstract and data limitations. We address each major comment below and will make revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Abstract] The abstract supplies no performance metrics, cross-validation details, data volume, or error bars; without these the claim that the data-driven method 'describes' 3D VIV cannot be evaluated against the stated comparison to VIVANA-FD.

    Authors: We agree that the abstract should include quantitative details to support the claims. In the revised version, we will add specific performance metrics (such as R-squared values and mean absolute errors for the random forest predictions), cross-validation procedure (e.g., 5-fold cross-validation details), data volume (number of samples and features from the experiment), and error bars or uncertainty estimates. These additions will allow direct evaluation against VIVANA-FD results. revision: yes

  2. Referee: The single available 3D-current experiment supplies the entire dataset for both clustering and random-forest training, so extracted parameters and model performance may reflect test-specific artifacts rather than general VIV drivers. This undermines the claim that the identified parameters are general measurable drivers of VIV response.

    Authors: We acknowledge the validity of this concern. The analysis relies on data from a single 3D-current experiment because few such datasets exist. This introduces the risk that identified parameters and model performance are specific to the test conditions rather than broadly generalizable. We will revise the manuscript to explicitly discuss this limitation in a dedicated section, qualify the claims about the parameters being 'general measurable drivers,' and frame the work as a demonstration of the data-driven approach on available 3D VIV data, with a call for additional experiments to assess broader applicability. revision: partial

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external experimental data and independent benchmark.

full rationale

The paper describes applying a data clustering algorithm to 3D flow measurements from one experiment, grouping responses by statistical characteristics, fitting a random forest regression to the measured VIV responses, and comparing performance against the existing VIVANA-FD tool. No equations, self-citations, or steps reduce any claimed prediction or result to quantities defined by the model's own parameters or inputs by construction. The central result is a supervised fit evaluated on held-out experimental points against an external benchmark, satisfying the criteria for a self-contained derivation against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The work rests on the assumption that the single cited 3D experiment is informative and that standard supervised-learning assumptions hold; no new physical entities or ad-hoc constants are introduced.

axioms (1)
  • domain assumption The statistical properties of the measured 3D current profiles are sufficient to cluster and predict VIV response statistics.
    Invoked when the authors state that responses are grouped based on statistical characteristics relating to flow direction.

pith-pipeline@v0.9.0 · 5915 in / 1304 out tokens · 21157 ms · 2026-05-25T17:08:22.773337+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages · 1 internal anchor

  1. [1]

    Machine Learning 45(1), 5–32 (Oct 2001)

    Breiman, L.: Random forests. Machine Learning 45(1), 5–32 (Oct 2001). https://doi.org/10.1023/A:1010933404324

    work page doi:10.1023/a:1010933404324 2001

  2. [2]

    In: Pei, J., Tseng, V.S., Cao, L., Motoda, H., Xu, G

    Campello, R.J.G.B., Moulavi, D., Sander, J.: Density-based clustering based on hierarchical density estimates. In: Pei, J., Tseng, V.S., Cao, L., Motoda, H., Xu, G. (eds.) Advances in Knowledge Discovery and Data Mining. pp. 160–172. Springer Berlin Heidelberg, Berlin, Heidelberg (2013)

    work page 2013

  3. [3]

    Springer series in statistics, Springer (2009), http://www.worldcat.org/oclc/300478243

    Hastie, T., Tibshirani, R., Friedman, J.H.: The elements of statistical learning: data mining, inference, and prediction, 2nd Edition. Springer series in statistics, Springer (2009), http://www.worldcat.org/oclc/300478243

    work page arXiv 2009

  4. [4]

    Lie, H., Braaten, H., Jhingran, V., Sequeiros, O.E., Vandiver, K.: Comprehen- sive riser viv model tests in uniform and sheared flow. vol. 5. ASME 2012 31st International Conference on Ocean, Offshore and Arctic Engineering (2012). https://doi.org/10.1115/OMAE2012-84055

    work page doi:10.1115/omae2012-84055 2012

  5. [5]

    Offshore Technology Conference, OTC-8700-MS, Houston, Texas (1998)

    Lie, H., Mo, K., Vandiver, J.: Viv model test of a bare- and a staggered buoyancy riser in a rotating rig. Offshore Technology Conference, OTC-8700-MS, Houston, Texas (1998). https://doi.org/10.4043/8700-MS

    work page doi:10.4043/8700-ms 1998

  6. [6]

    The Journal of Open Source Software 2(11) (mar 2017)

    McInnes, L., Healy, J., Astels, S.: hdbscan: Hierarchical density based clustering. The Journal of Open Source Software 2(11) (mar 2017). https://doi.org/10.21105/joss.00205

    work page doi:10.21105/joss.00205 2017

  7. [7]

    Passano, E., Larsen, C., Lie, H., , Wu, J.: VIVANA - Theory Manual Version 4.4 (2014)

    work page 2014

  8. [8]

    Journal of Offshore Mechanics and Arctic Engineering 131 (2009)

    Srivilairit, T., Manuel, L.: Vortex-induced vibration and coincident current velocity profiles for a deepwater drilling riser. Journal of Offshore Mechanics and Arctic Engineering 131 (2009). https://doi.org/10.1115/1.3058684

    work page doi:10.1115/1.3058684 2009

  9. [10]

    Marine Structures 51, 134 – 151 (2017)

    Thorsen, M., Sævik, S., Larsen, C.: Non-linear time domain analysis of cross-flow vortex-induced vibrations. Marine Structures 51, 134 – 151 (2017). https://doi.org/10.1016/j.marstruc.2016.10.007

    work page doi:10.1016/j.marstruc.2016.10.007 2017

  10. [11]

    Offshore Technology Conference, OTC-10931-MS, Houston, Texas (1999)

    Triantafyllou, M., Triantafyllou, G., Tein, Y.D., Ambrose, B.D.: Pragmatic riser viv analysis. Offshore Technology Conference, OTC-10931-MS, Houston, Texas (1999). https://doi.org/10.4043/10931-MS

    work page doi:10.4043/10931-ms 1999

  11. [12]

    Journal of Fluids and Structures 21(3), 335 – 361 (2005)

    Trim, A., Braaten, H., Lie, H., Tognarelli, M.: Experimental investigation of vortex- induced vibration of long marine risers. Journal of Fluids and Structures 21(3), 335 – 361 (2005). https://doi.org/10.1016/j.jfluidstructs.2005.07.014, marine and Aeronautical Fluid-Structure Interactions

    work page doi:10.1016/j.j 2005

  12. [13]

    Flow-Induced Vibration Conference, Lucerne, Switzerland (2000)

    Vandiver, J.: Predicting lock-in on drilling risers in sheared flows. Flow-Induced Vibration Conference, Lucerne, Switzerland (2000)

    work page 2000

  13. [14]

    Vandiver, J., Li, L.: Shear7 v4.5 Program Theoretical Manual (2007)

    work page 2007

  14. [15]

    Vandiver, J.K., Swithenbank, S.B., Jaiswal, V., Jhingran, V.: Fatigue damage from high mode number vortex-induced vibration. vol. 4. ASME 2006 25th International Conference on Ocean, Offshore and Arctic Engineering (2006). https://doi.org/10.1115/OMAE2006-92409

    work page doi:10.1115/omae2006-92409 2006

  15. [16]

    Westfall, P.H.: Kurtosis as peakedness, 1905 - 2014. r.i.p. American Statistician 68(3), 191–195 (aug 2014). https://doi.org/10.1080/00031305.2014.917055

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1080/00031305.2014.917055 1905