OBJECTIVE:
Little quantitative information is available on ocean turbulence in storm and Loop eddy current environments. There is concern that forces due to ocean turbulence may contribute significantly to motions and fatigue loading on spar platforms as well as impact the operability of marine risers (top-tensioned and catenary). DeepStar has just completed a field measurement program on ocean turbulence and is currently processing the data. Among the numerous activities of the DeepStar 6402 CTR relative to experimental measurement of VIV is a recently completed model test program performed at the Institute for Marine Dynamics where one of the objectives was to investigate the sensitivity of VIV to various levels of upstream turbulence. Two separate OTRC projects are using CFD simulation techniques to examine ocean turbulence interactions with a spar platform and a marine riser, respectively. The project described herein seeks to tie all these efforts together, by supporting and coordinating transfer of information between OTRC and DeepStar projects and by developing an analytical modeling framework to incorporate the field measurement and CFD simulation results in a form that can be applied in design.
APPROACH:
As with the 2002-2003 (year 1) effort, three major thrusts are being pursued in parallel:
1. Ocean Turbulence Modeling - The framework being developed for modeling loading effects from ocean turbulence is being patterned after existing procedures used in wind engineering. This modeling approach was presented at the November 2002 DeepStar workshop to plan the field measurement program on ocean turbulence in storm and Loop current environments. Continued interactions with DeepStar to interpret the field measurements and to refine and calibrate the turbulence model are anticipated. In addition, numerical procedures based on these turbulence models are being developed to generate statistical realizations of 2D and 3D turbulent velocity fields for use in the OTRC CFD simulation projects on spar and riser VIV.
2. CFD Experiments – The ocean turbulence model referenced above will be applied to derive upstream velocity fields for use in CFD simulations. The design and analysis of CFD experiments investigating the effects of upstream turbulence on VIV will be conducted in collaboration with Dr. Kallinderis for spar platforms and with Dr. Chen for risers. Results and insights from these simulations will be used to propose guidelines and procedures for addressing VIV in design.
3. Laboratory Experiments - Access will be negotiated to select data sets from DeepStar’s recently completed high Reynolds number VIV experiments. The DeepStar 6402 committee is currently developing a plan for distributing test data to select CFD modelers for code development and benchmarking purposes. If the data is made available, it will be used as the basis for defining some test cases for CFD simulations that can be performed by Dr. Kallinderis in the OTRC project titled “CFD Simulation of Ocean Turbulence Interactions With Spar Platforms” and by Dr. Chen in the OTRC project titled “CFD Simulation of Riser VIV”. The experimental and numerical simulation data will then be analyzed and compared to seek insight on the nature of turbulence interactions with cylinder boundary layer and wake flows as it impacts VIV loads and responses (including external damping constraints).
DEPLOYMENT OF RESULTS:
Design guidance on structure forces and motions associated with turbulent current flows will be synthesized from the field and laboratory measurements (if and when available) and from the CFD simulations. Study results will be shared and validated through interactions with industry focal points as well as through conference and journal publications.
ANTICIPATED PROJECT DURATION: 3 years
PROJECT PLAN FOR YEAR 1 (2003-2004):
Scope of Work: Extend isotropic model of ocean turbulence developed in 2002-03 to incorporate non-isotropic structure. Seek access to DeepStar VIV experimental data from IMD experiments and field data from ocean turbulence measurement program. Interface with DeepStar and Dr. Kallinderis and Dr. Chen on the use of experimental data to compare with CFD simulations (if data is available). Collaborate with Dr. Kallinderis and Dr. Chen in the design, analysis and interpretation of CFD experiments of turbulence interactions with spar platforms and risers, respectively.
Anticipated Results: Insights on interaction of isotropic turbulence with bare and straked cylinders that may be fixed or elastically restrained (to allow VIV motions), derived from 2-dimensional CFD simulations.
PROJECT PLAN FOR YEAR 2 (2004-2005):
Scope of Work: Refine and calibrate non-isotropic ocean turbulence model using DeepStar field data, if available. Interface with DeepStar on use of experimental data to compare with CFD simulations. Continue to collaborate with Dr. Kallinderis and Dr. Chen in the design, analysis and interpretation of CFD experiments of turbulence interactions with spar platforms and marine risers.
Anticipated Results: Preliminary insights on interaction of 3D non-isotropic turbulence with bare and straked cylinders that may be fixed or elastically restrained. Preliminary design guidance on structure forces and motions associated with turbulent current flows synthesized from lab measurements (if available) and CFD simulations.
PROJECT PLAN FOR YEAR 3 (2005-2006):
Scope of Work: Continue to collaborate with Dr. Kallinderis and Dr. Chen in the design, analysis and interpretation of CFD experiments of turbulence interactions with spar platforms and marine risers. Synthesize results into proposed design guidelines
Anticipated Results: Insights on interaction of 3D non-isotropic turbulence with bare and straked cylinders that may be fixed or elastically restrained. Design guidance on structure forces and motions associated with turbulent current flows will be synthesized from the field and laboratory measurements (if available) and from the CFD simulations. The guidance may include models for characterizing ocean turbulence, proposed ocean turbulence design criteria, force or motion allowances, recommended drag and lift force coefficients, force admittance functions, and procedures for setting up and interpreting model tests with combined wave and turbulent current environments (e.g. for testing spar platforms).
PRINCIPAL INVESTIGATOR (S) & OTHERS INVOLVED IN PROJECT:
PI: Dr. Richard Mercier
Collaborators: Dr. John Kallinderis, Dr. Hamn-Ching Chen
DATE: June, 2007Project Title: Ocean Turbulence Loads and Effects on Offshore Structures
MMS Project: 484 TO Number: 74486
Project PI: Richard Mercier
COTR: S. Buffington
Estimated Completion Date: 8/31/2007
Project Description:
This project includes several efforts related to improving the understanding of ocean turbulence and how to predict its effect on offshore structures, particularly spars:
1. Collaborate with DeepStar on planning of a field measurement program to gather data on ocean turbulence and on use of DeepStar experimental test data to validate a CFD model of turbulent flow interactions with a spar platform.
2. Collaborate with Dr. Kallinderis on design and analysis of CFD experiments to study the effect of ocean turbulence on Vortex-Induced Motions of spar platforms. The CFD model is being developed by Dr. John Kallinderis in a separate OTRC project.
3. Develop a model for generating turbulent velocity fields that can be incorporated as upstream boundary conditions in CFD simulations.Progress:
The December 2006 status report documented the successful validation of the numerical simulation procedure for isotropic turbulence modeled as a superposition of randomly oriented shear waves. The model was validated through comparison with benchmark theoretical results for the idealized case of single wavenumber isotropic turbulence. During this reporting period, the model was extended to simulate full spectrum isotropic turbulence and tested by conducting a large number of simulations over one-dimensional grids. In so doing a couple of important refinements were made to the model:
• A simple scaling law was derived and calibrated for assigning an energy level to each wavenumber component of the spectrum as a function of the Reynolds number, integral length scale, and number of Monte Carlo samples. The scaling is applied automatically in the numerical simulation so there is no need for adjustment of energy levels in a post-processing phase.
• The theoretical model and its numerical implementation were extended to include the transition in the turbulent velocity spectrum from the equilibrium subrange to the dissipation subrange. This results in better qualitative agreement between measured and simulated velocity spectra in the high wavenumber region. Quantitative comparisons will be made with available experimental data as part of the final validation effort.Simulations were performed on uniform 1D grids for Reynolds numbers ranging from 150 to 108 to investigate convergence of the simulated to the target velocity spectra. This analysis is not yet complete although clear trends are emerging from the results.
The figures below illustrate the comparison of the simulated and theoretical spectra for the in-line and transverse velocity fluctuations for a case with Reynolds number Rl = 104 and integral length scale l = 1 m. The simulation was performed by discretizing the target velocity spectrum into 1023 equally spaced wavenumber bands and representing the contribution of each wavenumber band by the superposition of 104 Monte Carlo samples of randomly-oriented shear waves. The simulation was performed on a uniform one-dimensional grid with spacing ? = 1 cm and dimension L = 211 ? = 20.48 m. Wavenumber averaging over 11 bands was applied to smooth the spectra of the simulated velocity fields. The solid black line is the theoretical spectrum and the dashed lines are the numerical modeling results.
Work remaining on this project involves performing simulations of isotropic turbulence at increasing Reynolds numbers on 2D and 3D grids and verifying that the spatial velocity correlation structure is in agreement with theoretical results. If time permits, an attempt will be made to extend the model to simulate non-isotropic turbulence that is characteristic of stratified flow conditions.
DATE: December, 2006Project Title: Ocean Turbulence Loads and Effects on Offshore Structures
MMS Project: 484 TO Number: 74486
Project PI: Richard Mercier
COTR: S. Buffington
Estimated Completion Date: 10/31/2007
Project Description:
This project includes several efforts related to improving the understanding of ocean turbulence and how to predict its effect on offshore structures, particularly spars:
1. Collaborate with DeepStar on planning of a field measurement program to gather data on ocean turbulence and on use of DeepStar experimental test data to validate a CFD model of turbulent flow interactions with a spar platform.
2. Collaborate with Dr. Kallinderis on design and analysis of CFD experiments to study the effect of ocean turbulence on Vortex-Induced Motions of spar platforms. The CFD model is being developed by Dr. John Kallinderis in a separate OTRC project.
3. Develop a model for generating turbulent velocity fields that can be incorporated as upstream boundary conditions in CFD simulations.Progress:
The remaining work on this project is focused on development of a model for generating turbulent velocity fields for use as input boundary conditions in CFD simulations of fluid-structure interaction. A preliminary model was developed at the outset of the project and used to generate 2D velocity fields for Reynolds number 150 and 1,000 for Dr. Kallinderis. These early CFD simulations showed that turbulence in an otherwise steady incident velocity field could have a substantial influence on the drag and lift forces exerted on a bare cylinder. Unfortunately the development effort became stalled in attempting to generate a third 2D velocity field for Dr. Kallinderis at Reynolds number 10,000 because of excessive computational time requirements. Initial efforts to streamline the computational algorithms were marginally successful.
The modeling approach assumes that turbulence can be represented as a superposition of randomly-oriented shear waves in an infinite medium with wavenumber and energy structure configured to match the target velocity spectrum, as initially suggested by Corrsin (1959). A useful theoretical result for the one-dimensional velocity spectra for a random 3D field of single wavenumber shear waves is provided in Tennekes and Lumley (1972). A previous comparison of numerical and theoretical results for this benchmark case indicated that the numerical model was able to reproduce the theoretical wavenumber spectrum of transverse velocity fluctuations but not the spectrum of in-line (longitudinal) velocity fluctuations.
After substantial diagnostic effort, the source of the modeling error was traced to a subtle statistical sampling bias in the Monte Carlo algorithm for generating randomly-oriented shear waves. Several alternative algorithms were devised and tested, but they also displayed different forms of statistical bias. With the insight gained from these unsuccessful efforts, a simple new algorithm was developed which appears to be free of bias errors and provides a modest improvement in computational efficiency. Figure 1 illustrates the quality of the agreement between the numerical and theoretical results for the benchmark single wavenumber shear wave case.
Figure 1 – In-line and transverse velocity spectra of isotropic shear waves of wavenumber ?* = 3000, simulated with 106 randomly-oriented wave components on a uniform rectangular grid with spacing ? = 0.001 and dimension L = 214. Wavenumber averaging over 26 bands was applied to smooth the spectra of the simulated velocity fields. The solid black line is the theoretical spectrum and the dashed lines are the numerical modeling results.Now that we are confident in the numerical implementation of the underlying theory and have been able to streamline the computational algorithms, we will repeat the low Reynolds number simulations of 2D velocity fields and perform test simulations at increasing Reynolds numbers to demonstrate feasibility. Both 2D and 3D velocity fields will be generated. The simulations will be analyzed to determine the spatial velocity correlation structure and, where possible, comparisons will be made with fundamental theoretical and experimental results.
As a final task, an attempt will be made to extend the model to simulate non-isotropic turbulence that is characteristic of stratified flow conditions in the ocean, particularly near the thermocline. This will require developing a suitable procedure for distorting the vertical wavenumber components in the turbulent field. This model should be of interest in 3D CFD modeling of riser VIV due to sheared currents.
Date: December, 2004Project Title: Ocean Turbulence Loads and Effects of Offshore Structures
MMS Project: 484 TO Number: 74486
PI: Richard Mercier
COTR: A. KonczvaldEstimated Completion Date: October 2006
Project Description:
This project includes several efforts related to improving the understanding of ocean turbulence and how to predict its effect on offshore structures, particularly spars:
1. Collaborate with DeepStar on planning of a field measurement program to gather data on ocean turbulence and on use of DeepStar experimental test data to validate a CFD model of turbulent flow interactions with a spar platform.
2. Collaborate with Dr. Kallinderis on design and analysis of CFD experiments to study the effect of ocean turbulence on Vortex-Induced Motions of spar platforms. The CFD model is being developed by Dr. John Kallinderis in a separate OTRC project.
3. Develop a model for generating turbulent velocity fields that can be incorporated as upstream boundary conditions in CFD simulations.Progress:
Work on the numerical model for generating turbulent velocity fields continued. The modeling approach assumes that turbulence can be represented as a superposition of randomly-oriented shear waves in an infinite medium with wavenumber and energy structure configured to match the target velocity spectrum. Comparison of numerical and theoretical results for a random field of single wavenumber shear waves indicated some discrepancies, most of which have been resolved. The model is able to reproduce the theoretical wavenumber spectrum of transverse velocity fluctuations but there are still discrepancies in the in-line velocity spectrum to be worked out. Some progress has been made in streamlining the algorithms to improve computational speed. Once all the necessary benchmark tests have been completed and we are confident that the underlying theory has been correctly implemented we will re-focus our attention on generating higher Reynolds number simulations.
Reports & Publications:
Date: June 2004
Project Name: Ocean Turbulence Loads and Effects on Offshore Structures
Project Number: 484 Task Order: 74486
Principal Investigators: Richard S. Mercier
Estimated Completion Date: October 2006
Project Description:
This project includes several efforts related to improving the understanding of ocean turbulence and how to predict its effect on offshore structures, particularly spars:
1. Collaborate with DeepStar on planning of a field measurement program to gather data on ocean turbulence and on use of DeepStar experimental test data to validate a CFD model of turbulent flow interactions with a spar platform.
2. Collaborate with Dr. Kallinderis on design and analysis of CFD experiments to study the effect of ocean turbulence on Vortex-Induced Motions of spar platforms. The CFD model is being developed by Dr. John Kallinderis in a separate OTRC project.
3. Develop a model for generating turbulent velocity fields that can be incorporated as upstream boundary conditions in CFD simulations.Progress:
Negotiations with DeepStar on use of their experimental data to validate a CFD model of turbulent flow interactions with a spar platform are underway. We are reviewing test documentation provided by DeepStar with the objective of selecting relevant data sets to request from DeepStar.
An attempt to generate a turbulent velocity field at Reynolds number of 10,000 for Dr. Kallinderis was not successful due to excessive computational time required by the preliminary numerical algorithm described in previous status reports. Efforts to improve the efficiency and robustness of the algorithm have been marginally successful to date, however there are a number of avenues remaining to be explored.
Reports & Publications:
Date: December 2003
Project Name: Ocean Turbulence Loads and Effects on Offshore Structures
TEES Project Number: 32558-47600 MMS Task Order: 74486 MMS Project Number: 484
Principal Investigators: Richard S. Mercier
Estimated Completion Date: October, 2006
Project Description:
This project includes several efforts related to improving the understanding of ocean turbulence and how to predict its effect on offshore structures, particularly spars:
Progress:
- Collaborate with DeepStar on planning of a field measurement program to gather data on ocean turbulence and on use of DeepStar experimental test data to validate a CFD model of turbulent flow interactions with a spar platform.
- Collaborate with Dr. Kallinderis on design and analysis of CFD experiments to study the effect of ocean turbulence on Vortex-Induced Motions of spar platforms. The CFD model is being developed by Dr. John Kallinderis in a separate OTRC project.
- Develop a model for generating turbulent velocity fields that can be incorporated as upstream boundary conditions in CFD simulations.
The previously developed preliminary model for generating turbulent velocity fields was exercised to generate 2D flow fields for two test cases of interest to Dr. Kallinderis, the first at Reynolds number Re=150 and the second at Re=1,000. Dr. Kallinderis used these velocity fields as upstream boundary conditions in CFD simulations of the flow around a bare cylinder that is fixed in space. For the flow conditions assumed, the CFD simulations show that the effect of the turbulence on the lift coefficient is modest while the effect on the peak drag coefficient is quite strong, driven primarily by the large-scale components of the turbulent field. These preliminary results underscore the importance of better understanding the large-scale structure of turbulence in the ocean. A turbulent velocity field is currently being generated for a third test case at Re=10,000.
While the turbulence model can easily be extended to generate 3D turbulent velocity fields, the preliminary numerical algorithm employed is computationally slow and inefficient, even for 2D simulations. Consequently effort is being focused on improving the efficiency and robustness of the algorithm. These improvements are needed for future 3D CFD simulations envisaged by Dr. Kallinderis.
Reports & Publications: None
Date: June 2003
Project Name: Ocean Turbulence Loads and Effects on Offshore Structures
Project Number:
Principal Investigators: Richard S. Mercier
Estimated Completion Date:
Project Description:
This project includes several efforts related to improving the understanding of ocean turbulence and how to predict its effect on offshore structures, particularly spars:
1. Collaborate with DeepStar on planning of a field measurement program to gather data on ocean turbulence and on use of DeepStar experimental test data to validate a CFD model of turbulent flow interactions with a spar platform.
2. Collaborate with Dr. Kallinderis on design and analysis of CFD experiments to study the effect of ocean turbulence on Vortex-Induced Motions of spar platforms. The CFD model is being developed by Dr. John Kallinderis in a separate OTRC project.
3. Develop a model for generating turbulent velocity fields that can be incorporated as upstream boundary conditions in CFD simulations.Progress:
PI participated in the DeepStar Ocean Turbulence Workshop held Nov. 6-7, 2002 to determine the requirements for a field measurement program to characterize the large scale (up to 100 m) structure of ocean turbulence in the Gulf of Mexico. The Workshop report is available from DeepStar (CTR 6801). PI participated in subsequent meetings of the DeepStar Metocean Committee to review various proposed turbulence measurement systems. A system has been selected and the field program is planned for 3Q 2003.
A simple spectral model of the turbulent velocity structure in the upper mixed layer of the ocean has been adapted from isotropic turbulence theory. As field data becomes available characterizing the effect of thermal stratification on the large scale structure of turbulence in the upper 200 m of the Gulf of Mexico, the spectral model will be modified to better reproduce the field observations.
A preliminary algorithm has been developed to generate statistical realizations of 3-dimensional turbulent velocity fields for specified integral turbulence scales (e.g. rms velocity fluctuation and integral length scale) using the spectral model mentioned above. Further testing and development of the algorithm will be pursued as a Master’s thesis project.
The algorithm will now be used to provide Dr. Kallinderis with 2D turbulent velocity fields for use as upstream boundary conditions in his CFD simulations.
