Top Banner
Numerical Analysis of Effect of Mass Diffusion of gas on gas pipeline Dr. Yongming Liu Vishal Chandrasekhar Varun Desai Arizona State University 1
19
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: vishal.ppt

Numerical Analysis of Effect of Mass Diffusion of gas on gas pipeline

Dr. Yongming LiuVishal Chandrasekhar

Varun Desai

Arizona State University

1

Page 2: vishal.ppt

Previous work

Extended finite element method (XFEM) and cohesive zone modeling (CZM) for the crack initiation and propagation simulation.

Parametric study and sensitivity analysis for the optimized prognosis algorithms.

2

Page 3: vishal.ppt

OUTLINE

Introduction and Objectives

Steps for diffusion + XFEM

Modeling & Simulation

Results, Conclusion and Future work

3

Page 4: vishal.ppt

INTRODUCTION

The research objectives of this project are:Evaluate the mass diffusion using numerical simulation on the PE and PA11 pipeline sections and compare the variations in the amount of gas diffusedObtain an insight into the degradation of the physical properties of the materials of pipeline and their influence on the life of materialsAddress the reduction in physical properties of the pipe due to mass diffusion and its influence using numerical simulation.

4

Page 5: vishal.ppt

INTRODUCTION

Gas diffusion occurs when The gas near the impingement may dissolve with

nearby solvent like water In areas of gross contamination when liquid

hydrocarbon condensates form in gas pipelines. when soil surrounding the pipe is heavily

contaminated with liquid hydrocarbons (diesel, gasoline, etc.)

Page 6: vishal.ppt

INTRODUCTION

6

In the first part, a simulation of the mass diffusion of methane gas has been simulated for particular conditions. The simulations gives us a clear picture on the variation of the gas concentration along the radial direction of the pipeline.

Page 7: vishal.ppt

Problems in performing mass diffusion + XFEM The gas diffusion analysis done in Abaqus cannot be

preceded with another type of analysis i.e. multistep analysis is not possible

The gas diffusion simulation and XFEM analysis uses two different element types.

Direct transfer of results from one simulation to another is not possible

7

INTRODUCTION

Page 8: vishal.ppt

Steps to infuse diffusion in XFEM

These are the steps to be followed to infuse the gas diffusion results in XFEM model Obtain the model with the element type required

for XFEM along with mesh Obtain the gas diffusion results for the same

geometry and mesh Perform a node to node mapping of the diffusion

into the model prepared for XFEM.

8

Page 9: vishal.ppt

Creating the model for XFEM analysis

The model is first created for the following dimensions and then meshed to obtain required number of layers.

Dimensions Inner radius: 100 mm Outer radius: 120 mm Thickness: 20 mm Cavity radius: 5 mm Length of pipe: 180 mm (90mm in sym. model) Mesh: 2 layers radially

9

Page 10: vishal.ppt

GEOMETRY & MESH

10

Modeling and simulations were carried out in ABAQUS.

Major parameters Geometry & mesh BC’s Material Model Convergence criterion

The final geometry is created as a 3D model and meshed to get 2 layers and the element type is 3D stress element

Page 11: vishal.ppt

Gas diffusion results mapping for XFEM analysis

11

The gas diffusion simulation is done in a different model and the process of simulation are explained

The main steps in the gas diffusion process are The model is defined with same geometry as

required for XFEM analysis with same mesh. The material model is defined as per the material

and the gas chosen for analysis The concentration of gas is given as load and the

boundary conditions are defined

Page 12: vishal.ppt

These are standard values for diffusion constants which are obtained through time lag method for measurement* which has been used throughout the project

12

Material Temperatur

e

(°C)

Pressure

(MPa)

Diffusion

coefficient (D)

(cm2/s)

Solubility coefficient

(S)

( cm3 (STP)/cm3·MPa)

Polyethylene +

Methane

40 7 1.6 0.57

Polyethylene +

Carbon dioxide

40 4 4 1.4

Polyamide 11 +

Methane

70 10 0.62 0.44

Polyamide 11 +

Carbon dioxide

70 4 1.0 2.3

*Flaconneche, B., J. Martin, and M. H. Klopffer. "Permeability, diffusion and solubility of gases in polyethylene, polyamide 11 and poly vinylidene

fluoride.“ Oil & Gas Science and Technology 56.3 (2001): 261-278.

MATERIAL MODEL

Page 13: vishal.ppt

Diffusion result

13

The gas diffusion simulation is shown in the picture

The concentration values in the diffusion can be defined in the new model as temperature using thermal-diffusion analogy.

Page 14: vishal.ppt

Importing meshed model for XFEM analysis

14

A new model database is created in abaqus and the meshed model with the 3D stress element is imported

This step is required for a diffusion + XFEM analysis because of two reasons The elements for diffusion model databases and

XFEM model database are different. It is possible to give properties layer wises as we

import an already meshed model The gas diffusion occurred radially and hence we are

giving different properties for different layers depending on concentration

Page 15: vishal.ppt

Mapping of diffusion result into new model

15

After importing the meshed model into new model database then next step is to import the diffusion values

This step is done in two ways If there is physical deformation, then node to node mapping can

be done. If there is no physical deformation, the whole diffusion result is

mapped as a field variable and defined in the new model

Page 16: vishal.ppt

Layer wise material definition

16

Once the gas diffusion values are defined as field variable the degradation f physical properties due to diffusion has to be addressed.

The reason for importing an meshed model can be well understood now with the properties being defined layer wise

The picture shows the selection of inner layer for which a reduction of about 6% in physical property was given.

Page 17: vishal.ppt

17

Layer wise material definition

Once the outer layer is defined the inner layer is also given a different property which is about 3% less than the actual value.

This enables us to give properties even element-wise which makes it a great method to be adopted while we do an XFEM analysis due to impingement.

Page 18: vishal.ppt

Results and conclusion

18

The work so far has been to study the effect of gas diffusion and its impact on the pipeline.

We have found a methodology by which concentration data can be incorporated in a XFEM analysis

We have been successfully in developing a method by which we can define material properties for any number of elements or region for the pipe structure.

Page 19: vishal.ppt

Future work

The work are being done for small mesh size but still it takes a lot of time for computation.

Having successfully incorporated the diffusion into the new model, we are currently working on post diffusion XFEM analysis.

The progress over the next meet should be XFEM analysis of model with diffusion results infused in the model for coarse meshed model

To use the diffusion mapping technique to give diffusion concentration near the impingement and defining a lower physical property around impingement area

19