Publisher’s version / Version de l'éditeur: Technical Report, 2011-02-01
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la
première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à [email protected].
Questions? Contact the NRC Publications Archive team at
[email protected]. If you wish to email the authors directly, please see the first page of the publication for their contact information.
NRC Publications Archive
Archives des publications du CNRC
For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.
https://doi.org/10.4224/17210689
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Numerical Work for Oceanide Project - A Progress Report
Derradji-Aouat, Ahmed
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
NRC Publications Record / Notice d'Archives des publications de CNRC:
https://nrc-publications.canada.ca/eng/view/object/?id=84bee1a3-ca30-4328-9bd3-29d27ce20799 https://publications-cnrc.canada.ca/fra/voir/objet/?id=84bee1a3-ca30-4328-9bd3-29d27ce20799
DOCUMENTATION PAGE
REPORT NUMBERTR-2010-22
NRC REPORT NUMBER DATE
November 2010
REPORT SECURITY CLASSIFICATION
Unclassified
DISTRIBUTION
Unlimited
TITLE
NUMERICAL WORK FOR OCEANIDE PROJECT – A PROGRESS REPORT
AUTHOR (S)
Ahmed Derradji-Aouat
CORPORATE AUTHOR (S)/PERFORMING AGENCY (S)
National Research Council, Institute for Ocean Technology, St. John’s, NL
PUBLICATION
SPONSORING AGENCY(S)
Oceanide, France (TOTAL E&P Research and Development, DORIS Engineering, Technip, Entropose Contracting, Saipem)
IOT PROJECT NUMBER
42_2406_26
NRC FILE NUMBER
KEY WORDS
Experimental, numerical, ice ridges, strain rates,
PAGES iii, 10, App. A FIGS. 6 TABLES SUMMARY
This is a progress summary for the numerical component of Oceanide project. The overall project contains both experimental and numerical components. The experimental work started last fiscal year while the numerical component started this year. Two main numerical tasks were planned: Task 1 is to develop and demonstrate the effectiveness of Finite Elements models for a First Year ice ridge interacting with a structure. Task 2 is to validate of the numerical results and final report next year. Task 1 and 2 were scheduled for this current fiscal year and next fiscal year, respectively.
The work is based entirely on the Finite Element Method (FEM); particularly explicit FEM. Explicit FEM was previously used to model ice behaviour, strength, fracture, ice blocks breaking, ice block interactions, as well as ice-structure interactions.
ADDRESS National Research Council
Institute for Ocean Technology Arctic Avenue, P. O. Box 12093 St. John's, NL A1B 3T5
National Research Council Conseil national de recherches Canada Canada Institute for Ocean Institut des technologies
Technology océaniques
NUMERICAL WORK FOR OCEANIDE PROJECT –
A PROGRESS REPORT
TR-2010-22
Ahmed Derradji-Aouat
November 2010
TABLE OF CONTENTS
List of Figures
...iii
1.0 INTRODUCTION ... 1
2.0 CURRENT
STATE
(AS OF OCT., 2010) ... 1
3.0 CURRENT
CHALLENGES... 3
4.0
NEXT STEP (FOR THE REMAINING PART OF YEAR 10-11) ... 3
5.0
PLANNED WORK (FOR NEXT YEAR 2011-12) ... 4
APPENDIX A: OCE - Numerical Presentation
LIST OF FIGURES
Figure 1: FY Ice ridge configuration ... 5
Figure 2: FY Ice ridge – Structure – Reaction Wall ... 6
Figure 3: FE mesh and solid models for the Spar 1/30 scale model (used in the
experiments)... 7
Figure 4: Structure - FY Ice ridge collision simulations... 8
(A to C at different times, D view from the back) ... 8
Figure 5: The Multi-physics approach - Water + Structure + FY Ridge + Reaction Wall 9
Figure 6: Buoyancy and Fluid-Structure Interactions (FSI) simulations... 10
should also be studied.
NUMERICAL WORK FOR OCEANIDE PROJECT – A PROGRESS REPORT
1.0
INTRODUCTION
This is a progress summary for the numerical component of Oceanide project. The overall project contains both experimental and numerical components. The experimental work started last fiscal year1 while the numerical component started this year. Two main numerical tasks were planned: Task 1 is to develop and demonstrate the effectiveness of Finite Elements (FE) models for a First Year (FY) ice ridge interacting with a structure. Task 2 is to validate of the numerical results and final report next year. Task 1 and 2 were scheduled for this current fiscal year and next fiscal year, respectively.
The work is based entirely on the Finite Element Method (FEM); particularly explicit FEM. Explicit FEM was previously used to model ice behaviour, strength, fracture, ice blocks breaking, ice block interactions, as well as ice-structure interactions.
2.0
CURRENT STATE (
as of Oct., 2010)
In this project, FE models were developed for the study of the interactions of ice ridges with offshore platforms. From the start, the modeling concept was based on three pillars, and they are: a) understand ice mechanical behaviour in cases of FY ice ridge-structure interactions, b) use a multi physics numerical solution, and c) validation of the numerical results for the target 95% confidence level. Other secondary pillars include realistic representation of the ice ridge geometry and dimensions, a self contact algorithm to model friction and collisions between ice blocks, a 2nd contact formulation to model the interactions between ice blocks and the platform structure. Further, if needed, the model should include the hydrodynamic effects on the results.
Traditionally, the mechanical behaviour of ice is categorized into low and high strain rates. At high strain rates (>10-3 m/s), linear elastic ice behaviour with a brittle mode of failure is used. However, at low strain rates (<10-3 /s), ice behaves as a time dependent elasto-plastic material with creep deformation. Ice ridges, however, are made of discrete ice blocks, the individual blocks can be small enough so that they will not bend and/or break into smaller pieces, and therefore, the behaviour of each block can be even viewed as rigid2. In this project, the calculation of ice strain rate, using Ralston equation, shows that for typical ice ridge drift velocities and the dimensions of the structure targeted in this program, ice should undergo linear elastic behaviour with a brittle mode of failure (the strain rates > 10-3/s). Two additional numerical challenges encountered when modeling ice ridges, and they are cohesion and frictional forces between ice blocks. The packing effects (how blocks can orient and pack themselves)
Figure 1 shows a FE model for a FY ice ridge. The keel, sail, and the parent ice sheet are modeled as a series of rectangular block (this is a very simple FE model). Real ice ridges, however, are made up of various ice blocks that are random in size, orientation, and shape. The rectangular blocks used at this time are easy to generate (one block is generated in a CAD system and then
1
Fiscal Year is April 1 to March 31.
2
Rigid block does not deform, but realistic material properties are needed so contact and friction between blocks is modeled properly.
copied in various directions). More importantly, at this stage, the main purpose of using this simple ice ridge model is to show how frictional and cohesive forces are modeled. Future model development should include ice ridges with random block sizes and shapes.
All numerical preprocessing work (all FE, models) presented in this study, were developed using ANSYS (www.ansys,.com) and the solution was obtained using LS-DYNA (www.lstc.com). Both ANSYS and LS-DYNA software were used because of their availability at the NRC-IOT. ANSYS has great flexibility and versatility to model any primitive volume blocks and surfaces, and also it has the capability to mesh complex forms and surfaces. LS-DYNA is based on explicit FE, which is mainly used for analysis of short-term impacts and collision problems (for example, it is well used for simulations of collisions of vehicles). In addition LS-DYNA provides an additional advantage by allowing users to model fluids (Water and Air), and Fluid-Structure Interactions (FSI).
een ice blocks. Naturally, ice ridges at any consolidated state varies between these two extrem .
sidered. This is simply achieved by using the appropriate friction coefficient between ice blocks.
spar and FPU into the final numerical model. For example, the spar FE model is shown in Fig 3. In the FE model for the ridge, the coincident nodes (within a prescribed tolerance) are welded together via spotweld beam elements (Figure 1c). Spotwelds are very short beam elements (their lengths can be zero or very close to it). The beam elements have stiffness, and if desired they can be programmed to fail (and/break) at prescribed stress an/or deformation conditions and/or limits. Therefore, it follows that the strength of cohesion between ice blocks is modeled by varying the stiffness of the spotweld beams. Ice blocks in the consolidated layer can be spotwelded by beams with high stiffness values. Very low (or zero) stiffness values are used for the spotweld beams in non-consolidated regions of the ice ridges. In the first case (moderate to high stiffness values), cohesion is accounted for, while for the 2nd case (very low or zero stiffness), there are no cohesion forces betw
es
Also, the stiffness for spotweld beams can be varied from one region of a ridge to another. For example, cohesion forces can be varied with keel depth. However, regardless of how the ridge is spotwelded, the frictional forces are also need to be con
At this stage of model development, the structure of the platform is represented by a simple inclined plate (Figure 2) since this current model is for the numerical concept development and demonstration. However, it is our objective, in the future, to integrate the actual models for both
To simulate the ice ridge–structure interactions, the plate is moved forwards (in the + X direction) at a given velocity3. First, the structural plate approaches the ice, contacts it, and then it pushes through the ice blocks. The ice blocks resist initially, and then they break away (see Figure 4 and the corresponding video Animations). For clarity, it is important to note the fundamental difference in the vocabulary between two basic numerical expressions: “element contact” and “self-contact”. 1) The blocks of the ice ridge contact the structural plate, and that is known as element contact, and 2) the ice blocks contact and push each other during the simulations, and this is known as self-contact.
this project is to simulate interactions of offshore structures with non consolidated ice ridges.
During the simulations, the spotweld beams stretch and break, and therefore ice blocks pile up on top of each other (see Figure 4). The stiffness of the spotweld beams are assumed to be very small because the objective of
3
Note that the above ice ridge model was built on the basis of the assumption that buoyancy forces resist body and gravity vertical loads. In the simulations (Fig 4), buoyancy is simply represented by upwards-vertical small nodal force (the some of the buoyancy nodal forces is enough to balance the downwards gravity loads).
y using a more accurate numerical model to take into account the fluid domain and its effects.
fixed elements). This is opposed to the Lagrangian model, where elements can deform in space.
ater and ice are calculated via an ALE (Arbitrary Lagrangian ulerian) algorithm in LS-DYNA
he above summary is for work produced this year (mainly, April – September 2010).
.0
CURRENT CHALLENGES
section parameters need to be reviewed and considered more carefully.
ed, finding resources (time) for the numerical work on this project is becoming a hallenge.
.0
NEXT STEP (FOR THE REMAINING PART OF YEAR 10-11)
separate models (structure and ice model and fluid and ice). In the former, no fluids and no FSI
Although, buoyancy can be represented by upwards-forces (this is especially valid for large ice sheets interacting with structures), the validity of using springs or upwards forces for discrete ice masses (such as a ice ridge, an iceberg, …etc.) is to be demonstrated, and therefore, the actual hydrodynamic forces and motions should be represented b
The fluid domain (water and air) is represented by two Eulerian computational spaces (Figure 5, one for air and one for water). Together, these two spaces allow modeling the free surface4, hydrodynamic motion of the ice ridge, and buoyancy forces. The fluid domains are made up of fixed computational spaces called Eulerian Spaces (in Eulerian FE method the mesh is fixed and the calculations are done with respect to the mass/velocity of fluid going in and out of the
In this current model, the structure and ice are modeled as Lagrangian, while the water and air are modeled as Eulerian. By combining Lagrangian structural and ice models with the Eulerian water and air spaces, hydrodynamic forces and fluid structure interaction (FSI) problems are taken into account. Figure 6 shows an example for how buoyancy is modeled in this simple ridge model. The example shows the case if the ice ridge is assumed to be rigid (a rigid large block) and it is lifted way above the surface of the water, and then let go - free fall. The animation associated with the figure shows how the ridge bobs up and down in the water for a time until it become stable. The interactions between w
E T
3
Lately, it was noticed that in some cases, computations crash half way through the simulations time. Efforts are being made to investigate and fix the problem. It is believed that, at this time, the input of the spotweld beam
It appears that the initial NRC-IOT budget (time and money) allocated for this project was underestimat
c
4
It is important to point out that the computer runs shown above are made by considering two
4
calculations. In the later, the ice ridge considered as a rigid block, no spotwelds, no self-contact, and no contact with structure.
The objective for next step (remaining of this year) is to integrate the two numerical models mentioned above (the Lagrangian model and the Eulerian space). Together, one integrated model will be obtained, and it includes spotweld beams, blocks self contact, interactions between ice and structures, hydrodynamic effects, the proper buoyancy model, and fluid structure interaction forces.
5.0
PLANNED WORK (FOR NEXT YEAR 2011-12)
After the integrated is completed and demonstrated, the following tasks are needed
1. Generate an ice ridge with random blocks with random sizes, to reflect typical ice ridge in real life
2. Import the mesh for the actual model, spar and the ship shaped FPU
3. Simulate case scenarios conducted in the ice tank and predict the actual test results 4. Compare numerical predictions with the test data and provide a validation assessment 5. Discuss the performance of the numerical model and provide conclusions and
recommendation.
6. Others, input and direction from clients/partners are very welcome.
