Optimization of Frequency Domain Fatigue Analysis for Unbonded Flexible Risers

2021 ◽  
Author(s):  
Jiabei Yuan ◽  
Yucheng Hou ◽  
Zhimin Tan

Abstract Fatigue analysis of flexible risers is a demanding task in terms of time and computational resources. The traditional time domain approach may take weeks of time in global simulation, local modelling and post-processing of riser responses to get fatigue results. Baker Hughes developed a fast hybrid approach, which is based on a frequency domain technique. The new approach was first implemented at the end fitting region and then to all other regions of the riser. Studies showed that the hybrid approach achieved convenient and conservative results in a significant shorter period of time. To improve the accuracy and reduce conservatism of the method, Baker Hughes has further optimized the analysis procedure to seek better results approaching true solutions. Several methods were proposed and studied. The duration of representative cases and noncritical cases have been extended. The steps to predict stress spectrum based on transfer functions have also been updated. From previous studies, only one transfer function was built for fatigue load cases with similar response spectra. This assumption linearizes the system response and produces certain level of discrepancy against true time domain solution. In this study, multiple ways of spectrum prediction are evaluated and compared. The paper summarizes several techniques to further optimize the hybrid frequency domain approach. The updated fatigue results are found to be more accurate. The optimized approach therefore gives more flexibility to engineers to approach the true solutions, which were originally acquired from full 3-hr time domain simulations. The approach requires less analysis time and reduces iterations in pipe structure and riser configuration design, which leads to faster project execution and potential cost reduction.

Author(s):  
Catharine C. Marsden ◽  
Stuart J. Price

The results of an analytical investigation are presented for the aeroelastic response of a two-dimensional, structurally nonlinear airfoil subject to a forced excitation. The system is modeled as a two-dimensional, rigid airfoil section free to move in both the bending and pitching directions and possessing a rigid flap. The airfoil is mounted by torsional and translational springs attached at the elastic axis, and the flap motion is used to provide the forcing input to the system. The airfoil is immersed in an aerodynamic flow environment, modeled using incompressible thin airfoil theory for unsteady oscillatory motion. The equations of motion for the aeroelastic system are solved using a fourth-order Runge-Kutta numerical integration technique to provide time-history solutions of the response of the airfoil in the pitch and plunge directions. The time-histories are analysed using Fourier transform-based techniques to obtain frequency-domain response and transfer functions. Results show that the nonlinear response of the aeroelastic system contains frequencies other than the forcing frequency. When modal frequencies and damping values are calculated using standard Fourier-based techniques, it is shown that the super- and sub-harmonic frequency content in the nonlinear response can contribute to errors when results are compared to those obtained for the equivalent linear system. This paper describes an investigation of a method of analysis that, while based on the Fourier transform, has been modified to recognize and accommodate the nonlinear contribution to the system response. The method, developed by Bendat [1], uses a band-limited random input and separates the linear and nonlinear components of the response within the frequency domain. Results are given for the application of this method to the specific case of the structurally nonlinear aeroelastic system. It is shown that the method may be used to successfully recover the linear frequency response function using the input and output data for the nonlinear system.


Author(s):  
Krassimir Doynov ◽  
Christoffer Nilsen-Aas ◽  
Rune Haakonsen ◽  
Wan Kan ◽  
Robert Bjærum

Flexible risers are being deployed in more and more demanding applications in terms of water depth, remote locations, temperature, pressure and corrosive fluids. Focus has been put on long term riser integrity in general, and on fatigue performance in particular, as knowledge of pipe behavior and properties has been advanced over the last decade. In this context, accurate and consistent estimation of riser global and local response to external loading is essential. A methodology has been developed to efficiently calculate irregular wave stress time histories of tensile armour wires for flexible risers. The stress time histories are calculated directly from the global loads which are usually generated by using commercially available well proven global analysis tools. The methodology elevates the dynamic analysis of flexible risers from the conventional regular-wave approach to irregular-wave time-domain approach. This in turn allows a better assessment of the fatigue performance and provides a better fit-for-service assessment or an opportunity to reduce design conservatism. This methodology also allows for consistent stochastic fatigue evaluations to be performed in time domain simulations using the well established stochastic analysis approach. All flexible riser non-linear hysteretic effects are included and phase shift between tension and curvature is also fully accounted for. The key ingredient lies in the generation of transfer functions of all stress components using a validated local analysis (LA) tool based on finite element method. This is done because direct use of the LA tool for long time domain simulations is very computationally intensive and impractical. The stress transfer functions allow direct mapping of the tension and curvature readings to individual stress components, which are combined in a phase consistent manner to obtain the total stress-time histories. This methodology should also work well for other systems having complicated cross sections such as dynamic umbilicals and integrated production bundle, etc. Accuracy of the proposed methodology should be equivalent to that of using the LA tool directly provided that the stress transfer functions are constructed appropriately. In comparison with the traditional regular-wave methodology, this irregular wave approach has been shown to provide a significant fatigue-life improvement for the flexible riser tensile-wire in a deep water West Africa application.


Geophysics ◽  
1969 ◽  
Vol 34 (5) ◽  
pp. 739-752 ◽  
Author(s):  
A. Becker

The response of a time‐domain electromagnetic system over a thin conducting sheet may be simulated by purely electronic means and without recourse to scale model experiments. The simulation is based on the similarity between the frequency domain response function for a thin sheet and the transfer function of certain RC active networks. Since this type of experiment employs actual field equipment, the proposed technique also constitutes a valid means of data quality control. It is difficult to carry out an analog simulation for conductors which do not resemble a thin sheet. If, however, the frequency domain response function for the situation in question is known, the simulation may be carried out on a digital computer. The digital simulation process involves a numerical Fourier decomposition of the primary field waveform (as seen by the receiver), the calculation of the effect of the ground on each harmonic component, and the recombination of the secondary field harmonics to form the observed transient. The technique is illustrated with some calculations of theoretical responses for an EM system over a homogeneous ground and over a thin horizontal conducting sheet. The digital simulation technique is more useful than the analog.


1988 ◽  
Vol 110 (1) ◽  
pp. 24-29 ◽  
Author(s):  
J.-N. Juang ◽  
H. Suzuki

This paper demonstrates the close conceptual relationships between time domain and frequency domain approaches to identification of modal parameters for linear systems. A frequency domain eigensystem realization algorithm, via transfer functions, is developed using a known procedure formulated for a time domain eigensystem realization algorithm, via free decay measurement data. An important feature is the capability of windowing to concentrate analysis on the frequency range of interest. The procedure of overlap averaging is used to produce smoother spectra to reduce the effect of noise on identified modal parameters. Examples from simulation and experiments are given to illustrate the validity of formulations derived in the paper.


Author(s):  
Claudio Marcio Silva Dantas ◽  
Marcos Queija de Siqueira ◽  
Gilberto Bruno Ellwanger ◽  
Ana Lu´cia F. Lima Torres ◽  
Marcio Martins Mourelle

The steel catenary riser was adopted by Petrobras as a cost-effective alternative for oil and gas export and for water injection lines on deepwater fields, where large diameter flexible risers present technical and economic limitations. The installation of the P-18 SCR was a pioneer project of a free-hanging steel catenary riser linked to a semi-submersible [1] and demonstrated the technical feasibility of the concept. Fatigue damage verification is an important issue in SCR design, demanding a high number of loading cases to be analyzed. The random time domain nonlinear analysis is considered an attractive and reliable tool for fatigue analysis as nonlinearities are properly modeled and the random behaviour of environmental loadings is considered. As time domain analysis is high computer time consuming, the frequency domain analysis has been considered as an alternative tool for the initial phases of riser design to be used mainly for fatigue damage verification. This paper presents a methodology developed to perform a linearized frequency domain analysis aiming at fatigue damage verification. Two drilling risers were analyzed with the frequency domain procedure developed. The model of a steel lazy-wave riser was analyzed both in frequency and time domain in order to compare fatigue damage results. The analyses were performed using the Petrobras’s in-house computer codes ANFLEX, ALFREQ and POSFAL developed and implemented as part of projects from CENPES/PETROBRAS with “COPPE/UFRJ -The Engineering Post-Graduating Coordination of the Federal University of Rio de Janeiro”.


Author(s):  
Wenbin Dong ◽  
Ingar Scherf ◽  
Gudfinnur Sigurdsson

A bridge between platforms needs to operate safely and continuously over its lifecycle. This paper focuses on the fatigue assessment of the bridge pin connection due to relative movements between platforms. A nonlinear time domain stochastic fatigue analysis of the pin connection in a bridge in the North Sea using a combined model of the jacket platforms and the interconnecting bridge is presented. The fatigue life is compared to the fatigue life from a linear frequency domain stochastic analysis. The facility has been in operation for more than 40 years and the operator requested an update of the inspection plans for the bridge. An RBI analysis has been done according to [1] based on fatigue results from wind gusts and relative movements. Regarding the fatigue assessment due to relative movements there are uncertainties related to selection of the friction coefficient. It was assessed that a friction coefficient of 0.4 is slightly conservative in this case. The fatigue life of the pin was calculated based on a linear frequency domain stochastic analysis, assuming that the bridge was fixed at both ends and this was considered reasonable conservative for fatigue estimation. Efforts have been made in the study presented here to assess the conservatism through a nonlinear time domain stochastic fatigue analysis. The sliding connections of the bridge are simulated by nonlinear springs. The effects of assuming different friction coefficients and different nonlinear spring models for a certain friction coefficient on the fatigue damage of the pin are investigated by a sensitivity study. The fatigue lives of the pin thus computed for a series of short-term sea states for the different assumptions for the friction coefficient and the nonlinear spring model are then compared to the result from a corresponding frequency domain approach.


Sign in / Sign up

Export Citation Format

Share Document