Session: 04-01-05 Flexible Pipes V

Submission Number: 182298

Machine Learning With Motion Spectrum for Prediction of Flexible Riser Fatigue Life 

Fatigue life of tensile wire in offshore flexible riser is an essential factor both in design stage and condition monitoring. Traditional fatigue prediction relies on deterministic method which evaluates various sea-states and operational conditions. However, this approach is computationally intensive and usually performed before pipe manufacture and installation. Frequency domain technique is available to compute the transfer function and predict the total stress spectrum of tensile wires with given wave scatter diagram. This approach significantly expedites the fatigue analysis in the design stage.

 

As more sensors are installed in the field, field data like vessel motions, pipe operating pressure and temperature, wave and surface current are available for evaluation of flexible pipe fatigue accumulation in the field.  However, time domain simulation with live information is computationally challenging and might not be feasible in the field. On the other hand, frequency domain approach itself with transfer function between wave and stress spectrum might no longer be feasible for field monitoring, as more non-linearity factors are introduced into the system.

 

This paper presents a new method of using vessel motion spectrum to predict the total stress spectrum of tensile wires. Artificial Neural Network (ANN) is used in combination with frequency domain technique to learn the governing spectral feature such as heave, pitch and roll, the built model can capture the underlying fatigue-driving behavior without the need for extensive, repetitive FEA simulation. Case study is presented to compare the difference from standard time domain simulation and ANN prediction.  It is shown that the data driven method offers an alternative solution for field fatigue life assessment for flexible pipe risers.

 

Presenting Author: Jiabei Yuan Baker Hughes

Presenting Author Biography: Principal engineer, 14 yrs for working in flexible pipe industry
PhD from the University of Texas at Austin in Civil Engineering

Authors:

Jiabei Yuan Baker Hughes
Linfa Zhu Baker Hughes
Yucheng Hou Baker Hughes
Greg Baker Baker Hughes
Zhimin Tan Baker Hughes