Read Anytime Anywhere Easy Ebook Downloads at ebookmeta.

com

Finite Element Programming in Non linear

Geomechanics and Transient Flow 1st Edition Nobuo
Morita

https://ebookmeta.com/product/finite-element-programming-in-
non-linear-geomechanics-and-transient-flow-1st-edition-
nobuo-morita/

OR CLICK HERE

DOWLOAD EBOOK

Visit and Get More Ebook Downloads Instantly at https://ebookmeta.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Application of Control Volume Based Finite Element Method

(CVFEM) for Nanofluid Flow and Heat Transfer 1st Edition
Mohsen Sheikholeslami
https://ebookmeta.com/product/application-of-control-volume-based-
finite-element-method-cvfem-for-nanofluid-flow-and-heat-transfer-1st-
edition-mohsen-sheikholeslami/
ebookmeta.com

Finite Element Method: Physics and Solution Methods 1st

Edition Sinan Muftu

https://ebookmeta.com/product/finite-element-method-physics-and-
solution-methods-1st-edition-sinan-muftu/

ebookmeta.com

Practical Finite Element Modeling in Earth Science using

Matlab 1st Edition Guy Simpson

https://ebookmeta.com/product/practical-finite-element-modeling-in-
earth-science-using-matlab-1st-edition-guy-simpson/

ebookmeta.com

Anti Catholicism in Victorian England Routledge Library

Editions The Victorian World Norman

https://ebookmeta.com/product/anti-catholicism-in-victorian-england-
routledge-library-editions-the-victorian-world-norman/

ebookmeta.com
Blue and Green Cities: The Role of Blue-Green
Infrastructure in Managing Urban Water Resources (2nd
Edition) Robert C. Brears
https://ebookmeta.com/product/blue-and-green-cities-the-role-of-blue-
green-infrastructure-in-managing-urban-water-resources-2nd-edition-
robert-c-brears/
ebookmeta.com

Risk Modeling: Practical Applications of Artificial

Intelligence, Machine Learning, and Deep Learning (Wiley
and SAS Business Series) 1st Edition Terisa Roberts
https://ebookmeta.com/product/risk-modeling-practical-applications-of-
artificial-intelligence-machine-learning-and-deep-learning-wiley-and-
sas-business-series-1st-edition-terisa-roberts/
ebookmeta.com

Into the Dark Dimension A Marvel Crisis Protocol Novel 1st

Edition Stuart Moore

https://ebookmeta.com/product/into-the-dark-dimension-a-marvel-crisis-
protocol-novel-1st-edition-stuart-moore/

ebookmeta.com

Classical Physics A Two Semester Coursebook 1st Edition

Bekir Karaoglu

https://ebookmeta.com/product/classical-physics-a-two-semester-
coursebook-1st-edition-bekir-karaoglu/

ebookmeta.com

Asteroid Impact Risk Impact Hazard from Asteroids and

Comets Impact Studies Trigo-Rodríguez

https://ebookmeta.com/product/asteroid-impact-risk-impact-hazard-from-
asteroids-and-comets-impact-studies-trigo-rodriguez/

ebookmeta.com
Waffles Mochi Get Cooking Learn to Cook Tomato Candy
Gratitouille and Other Tasty Recipes Yewande Komolafe

https://ebookmeta.com/product/waffles-mochi-get-cooking-learn-to-cook-
tomato-candy-gratitouille-and-other-tasty-recipes-yewande-komolafe/

ebookmeta.com
FINITE ELEMENT PROGRAMMING IN
NONLINEAR GEOMECHANICS AND
TRANSIENT FLOW
This page intentionally left blank
FINITE ELEMENT
PROGRAMMING
IN NONLINEAR
GEOMECHANICS
AND TRANSIENT
FLOW
NOBUO MORITA
Gulf Professional Publishing is an imprint of Elsevier
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, United Kingdom
Copyright © 2021 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system, without permission in writing from the
publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found
at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may
be noted herein).
Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any
information, methods, compounds, or experiments described herein. In using such information or methods they should be
mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any
injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions, or ideas contained in the material herein.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress

ISBN: 978-0-323-91112-2

For Information on all Gulf Professional Publishing publications

visit our website at https://www.elsevier.com/books-and-journals

Publisher: Charlotte Cockle

Senior Acquisitions Editor: Katie Hammon
Editorial Project Manager: Aleksandra Packowska
Production Project Manager: Anitha Sivaraj
Cover Designer: Victoria Pearson

Typeset by MPS Limited, Chennai, India

 Contents

Introduction ix 3. Finite element method with analytical

integration using simple elements 29
3.1 Discretization using 3D tetrahedral
I elements 29
Basics of the finite element 3.2 Analytical integrations 31
3.3 Assembling the elements 35
method 3.4 Nodal forces 37
3.5 Body forces 39
1. Fundamental equations of
poro-elasticity and fluid flow through 4. Finite element method with
porous media 3 isoparametric elements 41
1.1 Force, displacement, stress, strain, and
4.1 Isoparametric elements 41
displacement
strain relations 3
4.2 Brick elements 42
1.1.1 Three-dimensional stresses 3
4.2.1 Eight-node brick element (eight-node
1.1.2 For three-dimensional problems 5
3D solid element) 42
1.1.3 Displacement
strain relations 5
4.2.2 Twenty-node brick element 48
1.2 Equation of equilibrium and stress
strain
4.2.3 Elements compatible with the 8-node
relation 7
and 20-node brick elements 53
1.2.1 Equation of equilibrium 7
4.2.4 Ten-node parabolic tetrahedron
1.2.2 Stress
strain relations for isotropic
element 53
linearly elastic materials 7
4.3 Infinite element 55
1.2.3 Number of variables and
4.3.1 Twelve-node infinite element 59
equations 8
1.2.4 Stress
strain relations for porous
media 10 5. Numerical integration 63
1.3 Fluid flow through porous media 12
1.4 Matrix expression 13 5.1 Gaussian integration 63
5.2 Integration formula for triangle and tetrahedron
shape functions 65
2. Finite element methods 17
6. Solution of linear simultaneous
2.1 Discretization using the virtual work equations 69
principle 17
2.2 Discretization using the minimization of total 6.1 Matrix transformation for the boundary
potential energy 21 condition given by local coordinates 69
2.3 Discretization using the residual method 23 6.2 Solution of linear simultaneous equations 72
2.4 Discretization of the set of flow equations 6.2.1 Gaussian elimination 72
through porous media using the residual 6.2.2 Frontal method 75
method 25 6.2.3 Conjugate gradient method 84

v
vi Contents

7. Convergence and error analysis 87 9.2 Discretization using the virtual work
principle 130
7.1 Theoretical estimation of error 87 9.3 Discretization of transient flow equations
7.2 Numerical evaluation of error 91 through porous media 133
9.4 Coupling geomechanics and transient fluid
8. Application of the finite element flow 136
method to nonlinear geological 9.4.1 Full coupling 136
materials 93 9.4.2 Sequential coupling 137
9.5 Stability of the sequential methods 144
8.1 Standard triaxial rock test equipment and 9.5.1 One-dimensional compaction problem
typical test results 93 coupled with fluid flow and
8.1.1 Stress and strain invariants 97 geomechanics 144
8.1.2 Nonlinear elastic coefficients 98 9.5.2 Stability analysis of the coupled
8.2 Nonlinearity at a low-stress state 99 problem 145
8.3 Shear-type nonlinear strain 100 9.5.3 Stability of drained split method 148
8.3.1 Yield stress 100 9.5.4 Stability of undrained split
8.3.2 Maximum stress yield theory: Rankine method 150
theory 101 9.5.5 Stability of fixed strain method 152
8.3.3 Maximum strain yield theory (Saint- 9.5.6 Stability of fixed total stress
Venant theory) 102 method 156
8.3.4 Maximum shear stress yield theory 9.5.7 Numerical example of one-dimensional
(Tresca) 102 compaction problem using the fixed
8.3.5 Max octahedral shear stress yield theory total stress method 159
(von Mises yield criterion) 103 9.6 Sequential coupling with commercially
8.3.6 Mohr’s linear yield theory 104 available reservoir models 162
8.3.7 Mohr’s nonlinear yield theory 105 9.6.1 Sequential calculation of flow and
8.3.8 Drucker
Prager yield theory 107 geomechanics with uniaxial compaction
8.3.9 Lade yield theory 108 assumption 163
8.4 Yield envelope fitted to real polyaxial 9.6.2 Single or multiphase problems without
stress
strain empirical data 109 assuming uniaxial compaction 166
8.4.1 Mohr
Coulomb 109 9.6.3 One-step undrained method for short
8.4.2 Drucker
Prager 109 period production problems 173
8.4.3 Lade model 109 9.6.4 General multiphase problems 173
8.4.4 Solenhofen limestone 110 Further reading 176
8.4.5 Dunham dolomite 110
8.4.6 Fuji River sand 111
8.5 Incremental form of nonlinear stress strain for II
application of the finite element method 112
8.6 Application of the Newton
Raphson method Applications of Flow3D and
to nonlinear problems 115 Geo3D to real field problems
8.7 Calculation method of λ, Dep 116
8.8 Implementation 120 10. Pressure profile around
8.9 Construction of constitutive relations from perforations—field problems using
triaxial data 122
Flow3D 179
9. Coupling geomechanics and transient 10.1 Pressure profile around a single
fluid flow 127 perforation 179
10.1.1 Analytical solution 179
9.1 Fundamental equations for isotropic poro- 10.2 Numerical solution for pressure distribution
elasticity problems 127 around a single perforation 180
 Contents vii
10.2.1 Pressure distribution around a 12.2 Failure envelopes from empirical results 248
perforation without permeability 12.2.1 Castlegate sandstone 248
damage 181 12.2.2 Rozbark sandstone and sensitivity
10.2.2 Pressure distribution around a analysis 250
perforation with damaged 12.2.3 Dunham dolomite 253
permeability 182 12.2.4 Mizuho trachyte 255
10.2.3 Gas flow 182 12.2.5 Shirahama sandstone 258
10.3 Pressure distribution around a perforation for 12.2.6 Izumi sandstone 260
gravel packed well 183 12.2.7 Horonai sandstone 263
10.4 Quantitative analysis of the effect of 12.2.8 Yubari shale 265
perforation interaction on flow 12.2.9 Yamaguchi marble 267
efficiency 193 12.3 Stress state around an inclined well drilled
10.4.1 Evaluation of flow performance 195 through inclined formation 270
Nomenclature 202 12.3.1 Analytical solution to calculate stress
References 202 state around a borehole 270
12.3.2 Calculation of principal stresses 273
11. Evaluation of mechanical stability of 12.3.3 Zoback’s breakout angle
estimation 274
perforations using Geo3D 203 12.3.4 Breakout angle and depth for Mogi
11.1 Stability of perforations during oil and gas and Lade failure theories 276
production 203 12.3.5 Definition of UCS 279
11.1.1 Field applications 217 12.3.6 Breakout measurements 280
11.2 Field observation of sand-production 12.4 Comprehensive analysis of stress state around
problems 224 a borehole with temperature, swelling, and
11.2.1 How fluid flow affects sand production pore pressure change for layered and
from a three-dimensional perforation orthotropic formations 281
cavity 224 12.4.1 General statement 281
11.2.2 How drawdown necessary to produce 12.4.2 Parameter study with typical field
without sand changes during reservoir conditions 284
life 229 12.5 Failure theories to predict breakout angle
11.2.3 Why sand problems often occur after around a borehole 303
water cut 230 12.5.1 Rock failure criteria 303
11.2.4 Why sand flow is high when 12.5.2 Failure function to predict breakout
production is restarted 232 angle with various controllable
11.3 A quick method to forecast the possibility of parameters 308
sand problems: Perforation stability analysis 12.5.3 Breakout angle predicted by the four
using TWC or TPS test equipment 233 rock failure criteria of vertical well in
11.4 Concluding remarks 237 the base reservoir conditions 309
Nomenclature 238 12.5.4 Breakout angle predicted by the four
Further reading 239 rock failure criteria of horizontal well
in depleted reservoir condition 320
12.6 Effect of controllable parameters on safe mud
12. Numerical methods for the borehole window design 332
breakout problems using Geo3D 241 12.6.1 Introduction of safe mud weight
window 332
12.1 Rock failure and failure theories 241 12.6.2 Optimal safe mud weight window
12.1.1 Total failure, local failure, and internal design with various controllable
failure 241 parameters 333
12.1.2 Critical plastic failure theory 246 12.7 Conclusion 342
viii Contents

Nomenclature 343 14.4.3 Subsidence and compaction prediction

References 344 for undrained elastic moduli 412
Further reading 344 14.4.4 Subsidence and compaction prediction
for drained elastic moduli 416
13. Casing collapse for hydrostatic and 14.4.5 Field applications 418
geotechnical loads—Geo3D analysis 347 14.5 Parametric analysis of subsidence and
compaction 420
13.1 Casing collapse for hydrostatic load 347 14.5.1 Earth deformation during reservoir
13.1.1 Type A: the standard casing failure compaction 420
criterion applied to simple 14.5.2 Subsidence and horizontal movement
loading 347 maps 422
13.1.2 Type B: failure criterion under References 442
geotechnical load 352 Further reading 442
13.2 Concluding remarks 388
Nomenclature 391 Appendix A: Apparent elastic modulus
Further reading 391 with pore fluid 443
14. Three-dimensional reservoir
compaction problems with coupled Geo3D III
code 393
Programming of the finite element
14.1 Introduction of reservoir compaction methods
problems 393
14.2 Strain nuclei method 395
Appendix B: Computer program structure
14.3 Analytical solution at the center of a reservoir for transient fluid flow problems through
for a radial reservoir 405 porous media 459
14.4 Subsidence, pore pressure, and stress change in Appendix C: Program Geo3D 479
the overburden formation using the finite Appendix D: Program COEFI for
element method coupled with transient flow
determining the parameters of the
and geomechanics models 408
14.4.1 Methods to evaluate the subsidence nonlinear constitutive relations from a set
and compaction 408 of triaxial test results 505
14.4.2 Subsidence and compaction prediction
using the sequential coupling method Index 531
using the fixed total stress
method 409
 Introduction

Sedimentary rocks are formed by the Since the procedures of applying the
accumulation of solids and fluids. When oil finite element methods to field projects are
or gas is extracted from the sedimentary important, the latter half of the book
rock or when water is injected, stress describes the field projects that were per-
change and fluid flow occur. For some formed for oil industries. The field applica-
cases, fluid flows without significantly tions will enlighten the readers on how
deforming the rock. For other cases, stress fluid is flowing around a well, how sand is
change occurs without disturbing fluid produced from a production well, how
flow. Or, for some other cases, both fluid instability occurs in the drilling hole, how
flow and geomechanics are tightly coupled. casing is deformed, and how subsidence
This book is written to enhance the ability and compaction occur during oil and gas
of programming the finite element methods extraction.
in geomechanics for geophysicists and oil It takes several years to develop a 3D
and gas development engineers. finite clement program suitable for research
and field applications. The transient 3D
1. The finite element method for 3D
flow and Geo3D codes in this textbook
transient flow problems is developed
have been used for various field problems.
and the Flow3D program code is
Currently, many finite element programs
attached.
are available. However, good 3D transient
2. The finite element method for 3D
flow codes and the mechanical finite clem-
nonlinear geomechanics problems is
ent codes suitable for nonlinear geomecha-
developed and the Geo3D program code
nics are scarce. To enhance research and
is attached.
field analysis, the reader may create new
3. Several methods to couple the fluid flow
codes by adding several new subroutines
and geomechanics are developed.
to Flow3D and Geo3D codes attached in
Subroutines are attached to couple
this book.
Flow3D and Geo3D.
4. Necessary procedures are described to
couple Geo3D and commercial finite
difference flow code such as ECLIPSE.

ix
This page intentionally left blank
 P A R T I

Basics of the finite element

method
This page intentionally left blank
 C H A P T E R

1
Fundamental equations of poro-elasticity
and fluid flow through porous media
1.1 Force, displacement, stress, strain, and displacement
strain relations

The normal stress and the shear stress are the normal and shear forces (Fig. 1.1) per
unit area defined by
F
Stress: σ5
A
(1.1)
Normal stress σn 5 Fn =A
Shear stress τ 5 Fs =A

1.1.1 Three-dimensional stresses

Conventionally, the normal stresses are expressed by σ and the shear stresses are
expressed by τ. The planes perpendicular to x, y, and z coordinates are called x, y, and z
planes, respectively. The first subscript is the plane on which the stresses act and the sec-
ond subscript is the stress direction. For three-dimensional problems (Fig. 1.2), the stresses
are represented by the following nine stress components:
σxx τ xy τ xz σyy τ yz τ yx σzz τ zx τ zy

Strain
The deformation varies from point to point, and it is split into compressional and shear-
ing deformations as shown in Fig. 1.3. The normal strain component is defined as the
change in length per unit length and the shear strain is defined as the change in angle
between two original orthogonal directions. Conventionally, the normal strain is expressed
by ε and the shear strain or angular deformation is expressed by γ.
Normal strain: ε 5 ΔL=L
(1.2)
Shear strain: γ5θ

Finite Element Programming in Nonlinear Geomechanics and Transient Flow

DOI: https://doi.org/10.1016/B978-0-323-91112-2.00019-7 3 © 2021 Elsevier Inc. All rights reserved.
4 1. Fundamental equations of poro-elasticity and fluid flow through porous media

A: area

Fn
Fs

FIGURE 1.1 Force vector.

FIGURE 1.2 3D stresses.

V
'L
W
L
T

FIGURE 1.3 Visual illustration of strains.

I. Basics of the finite element method

 1.1 Force, displacement, stress, strain, and displacement
strain relations 5

1.1.2 For three-dimensional problems

The normal strains and the shear strains are expressed by ε and γ with subscripts for
the face applied and the direction of the deformation (Fig. 1.4).
εxx ; εyy ; εzz ; γ xy ; γ yz ; γ zx
h i h i
O0 A0 jx-component 2 OA O0 B0 jx-component 2 OB
εx 5 ; εy 5 (1.3)
OA OB
γ xy 5 ðθ1 1 θ2 Þ

1.1.3 Displacement
strain relations

Strains are expressed by
εx 5 ½uðx 1 ΔxÞ 2 uðxÞ=ðx 1 Δx 2 xÞ 5 ½u 1 ð@u=@xÞΔx 2 u=Δx 5 ð@u=@xÞ
θ1 5 ½vðx 1 ΔxÞ 2 vðxÞ=½x
 1 Δx 1 uðx 1 ΔxÞ 2 ðx 1 uðxÞÞ5 ½v 1 ð@v=@xÞΔx 2 v=Δx
θ2 5 ½uðy 1 ΔyÞ 2 u y =½y 1 Δy 1 uðy 1 ΔyÞ 2 y 1 v y  5 ½u 1 ð@u=@yÞΔy 2 u=Δy
γ xy 5 ðθ1 1 θ2 Þ 5 ð@v=@xÞ 1 ð@u=@yÞ
εx 5 ð@u=@xÞ
(1.4)
εy 5 ð@v=@yÞ
εz 5 ð@w=@zÞ
γxy 5 ð@u=@yÞ 1 ð@v=@xÞ
γyz 5 ð@v=@zÞ 1 ð@w=@yÞ
γ zx 5 ð@w=@xÞ 1 ð@u=@zÞ

FIGURE 1.4 Displacement and deformation of a square structure.

I. Basics of the finite element method

6 1. Fundamental equations of poro-elasticity and fluid flow through porous media

1.1.3.1 Number of variables

Stress: 9 components
σxx, σyy, σzz, τ xy, τ yz, τ zx, τ yx, τ zy, τ xz
Strain: 9 components
εxx, εyy, εzz, γ xy, γ yz, γ zx, γ yx, γ zy, γ xz

Displacement: 3 components
u, v, z
The total number of variables is 21.
The stress and strain components are reduced due to the symmetry for nonsingular
force condition. Then, the following equations are applied:
τ xy 5 τ yx ; τ zy 5 τ yz ; τ xz 5 τ zx
γ xy 5 γ yx ; γ zy 5 γ yz ; γ xz 5 γ zx
(Proof for static problems)
The moment around O (Fig. 1.5) is
ðτ xy wΔyÞΔx 5 ðτ yx wΔxÞΔy (1.5)
Hence, τ xy 5 τ yx holds. Similarly, other shear stress components are also symmetric.
If the shear stresses and strains are symmetric, then the number of the unknowns is
reduced to 15 as follows:
Stress: 6 components
σx, σy, σz, τ xy, τ yz, τ zx
Strain: 6 components
εx, εy, εz, γ xy, γ yz, γ zx

O A

w= width of the structure

FIGURE 1.5 Moment balance around the origin O for a static structure.

I. Basics of the finite element method

 1.2 Equation of equilibrium and stress
strain relation 7
Displacements: 3 components
u, v, w
Totally, the variables are reduced to 15.

1.2 Equation of equilibrium and stress
strain relation

1.2.1 Equation of equilibrium
The forces in the x direction for a cubic body are given by
   
Normal force: 2 σx ΔxΔz 1 σx 1 @σx =@x Δx ΔyΔz
   
Shear force: 2 τ yx ΔxΔz 1 τ yx 1 @τ yx =@y Δy ΔxΔz
    (1.6)
2 τ zx ΔxΔy 1 τ zx 1 @τ zx =@z Δz ΔxΔy
Body force: Fx ΔxΔyΔz
The sum of the forces is zero if they are in equilibrium. That is,
     
@σx =@x 1 @τ yx =@y 1 @τ zx =@z 1 Fx 5 0 (1.7)
Similarly, for y and z directions, the forces are in equilibrium. Hence,
     
@σy =@y 1 @τ xy =@x 1 @τ zy =@z 1 Fy 5 0 (1.8)
     
@σz =@z 1 @τ zy =@y 1 @τ zx =@x 1 Fz 5 0 (1.9)

1.2.2 Stress
strain relations for isotropic linearly elastic materials

Hook’s law for linearly elastic material assumes that if an axial stress is applied along
the x coordinate, the strain in the x coordinate becomes proportionally large (Fig. 1.6). The
proportional constant is called as Young’s modulus.
σx 5 Eεx or εx 5 σx =E

FIGURE 1.6 Concept of Poisson’s ratio.

I. Basics of the finite element method

8 1. Fundamental equations of poro-elasticity and fluid flow through porous media

The strains in y and z directions increase if the strain in the x direction is reduced. The
ratio of the strains in y and z directions and the strain in the x direction is called as
Poisson’s ratio denoted by ν. Or,
   
εy 5 2νεx 5 2 ν=E σx ; εz 5 2νεx 5 2 ν=E σx (1.10)
The above strains are generated when the stress is applied only in x direction. Similar
equations are applied to forces in y and z directions. Hence,
  
εx 5 1=E σx 2 νσy 2 νσz (1.11)
For y, z directions,
  
εy 5 1=E σy 2 νσz 2 νσx (1.12)
  
εz 5 1=E σz 2 νσx 2 νσy (1.13)
If a shear stress is applied as shown in Fig. 1.7, the shear strain increases proportionally; hence,
τ xy 5 Gγ xy
τ yz 5 Gγ yz (1.14)
τ zx 5 Gγ zx
where the proportional constant G is called as the shear modulus.

1.2.3 Number of variables and equations

So far, the variable definitions are described and the relations between variables are derived.
The followings are the summary of the number of the equations and the number of the variables.

15 variables used for fundamental elasticity

Stress: 6 components
σx, σy, σz, τ xy, τ yz, τ zx

xy

xy

FIGURE 1.7 Concept of shear modulus.

I. Basics of the finite element method

 1.2 Equation of equilibrium and stress
strain relation 9

Strain: 6 components
εx, εy, εz, γ xy, γ yz, γ zx
Displacement: 3 components
u, v, w

And, 15 equations are derived:

15 equations for linear elasticity problems

Displacement
strain relations
 
εx 5 @u=@x
 
εy 5 @v=@y 
εz 5 @w=@z
    (1.15)
γ xy 5 @u=@y 1 @v=@x
   
γ yz 5 @v=@z 1 @w=@y
   
γ zx 5 @w=@x 1 @u=@z

Equations of equilibrium are given by

     
@σ =@x 1 @τ yx =@y 1 @τ zx =@z 1 Fx 5 0
 x     
@σ =@y 1 @τ xy =@x 1 @τ zy =@z 1 Fy 5 0 (1.16)
 y     
@σz =@z 1 @τ zy =@y 1 @τ zx =@x 1 Fz 5 0

Stress
strain relations are given by

  
εx 5 1=E σx 2 νσy 2 νσz
  
εy 5 1=E σy 2 νσz 2 νσx
  
εz 5 1=E σz 2 νσx 2 νσy (1.17)
τ xy 5 Gγ xy
τ yz 5 Gγ yz
τ zx 5 Gγ zx

Note that for the tensor calculations, the shear strains are halved since the following
definition is more convenient for calculations:
   
γ xy 5 0:5 @u=@y 1 @v=@x
   
γ yz 5 0:5 @v=@z 1 @w=@y
   
γ zx 5 0:5 @w=@x 1 @u=@z
τ xy 5 2Gγ xy
τ yz 5 2Gγ yz
τ zx 5 2Gγ zx

I. Basics of the finite element method

10 1. Fundamental equations of poro-elasticity and fluid flow through porous media

Hence, the variables can be determined by solving the set of the above equations if the
boundary conditions are specified.

1.2.4 Stress
strain relations for porous media

Some equations used for nonporous media are applied to porous media. Both the
equations of equilibrium and strain
displacement relations have the same form as
the nonporous media. However, the stress
strain relations for porous media must
include the deformation induced by pore pressure. For porous media, uniform and
hydrostatic pressure is applied to each grain if the pore pressure is increased without
increasing the effective stress. Suppose the pore pressure p is applied on each grain.
The pore pressure is applied on all the solid surfaces uniformly so that the stress
induced within the solid is also uniform if the solid material is uniform. The solid
reduces the size proportionally; hence the porosity remains the same as shown in
Fig. 1.8.
Suppose the hydrostatic stress is applied everywhere within the solid, that is, uniform
stress σX ~ σ ~ σ ~ ーｐ is applied everywhere within the solid, then the strain induced
y Z

by the hydrostatic stress is given by

   
εx 0 5 1=Em ½1 2 2ν m  2p
   
εy 0 5 1=Em ½1 2 2ν m  2p (1.18)
   
εz 0 5 1=Em ½1 2 2ν m  2p
where E m and ν m are matrix Young’s modulus and Poisson’s ratio, respectively. The
hydrostatic stress applied throughout the grains is called neutral stress since the
grains shrink proportionally with pressure p; hence, the shape of each grain does
not change. Since the shape of each grain does not change, the porosity remains
unchanged.

FIGURE 1.8 Pore pressure applied on the grain surfaces.

I. Basics of the finite element method

 1.2 Equation of equilibrium and stress
strain relation 11
Now, let apply external stress without changing the pore pressure. Then, the additional
stresses are given by
     
σex 5 σx 1 p ; σey 5 σy 1 p ; σez 5 σz 1 p (1.19)

The above stresses are called effective stresses. Note that the compressive stress is nega-
tive. If the compressive stress is positive, the sign in front of p becomes minus. With the
effective stress, the grain shape changes where the external stresses are transmitted
through the grain
grain contact points. The bulk stress
strain relation is given by the fol-
lowing equations:
 h i
εx 0 0 5 1=E σex 2 νσey 2 νσez
 h i
εy 0 0 5 1=E σey 2 νσez 2 νσex (1.20)
  h i
εz 0 0 5 1=E σez 2 νσex 2 νσey

where Young’s modulus and Poisson’s ratio are the bulk Young’s modulus and the bulk
Poisson’s ratio. The strain induced by the neutral stress and the effective stress is superimposed.
Stress
strain relation for porous media:
εx 5 εx 0 1 εhx 0 0 i 
  
5 1=E σex 2 νσey 2 νσez 2 1=Em ½1 2 2ν m p
εy 5 εy 0 1 εhy 0 0 i 
   (1.21)
5 1=E σey 2 νσez 2 νσex 2 1=Em ½1 2 2ν m p
εz 5 εz 0 1 εz 0 0h i 
  
5 1=E σez 2 νσex 2 νσey 2 1=Em ½1 2 2ν m p

The shear stress is not affected by the neutral stress so that the effective stresses are
written by the following tensor form:
σeij 5 σij 1 pδij (1.22)

A more detailed discussion follows since the effect of the neutral stress is actually very com-
plex and the normal rocks contain various minerals. The rock grains deform proportionally
only if all the grains consist of the same minerals and all the pores are connected so that the
pore pressure is the same throughout the pores. Then, all the grains are proportionally shrunk
due to the neutral stress; hence, the porosity remains the same. If the porosity remains the
same, the rock properties are not significantly affected by the neutral stress. The permeability
slightly changes even with the same porosity since the pore capillaries become slightly smaller.
However, such changes are normally small enough so that the changes are negligible.
The proportional change in grain shape is actually considered to be ideal since the actual
rock consists of heterogeneous grains and disconnected voids. Actually, each grain consists of
different minerals so that the deformation of each grain is not uniform even with the same neu-
tral stress. Some voids are also isolated so that the pore pressure in the isolated voids may be
different from the pore pressure at the surrounding pores. Rocks have laminations during the
sedimentation process, inducing vertical changes in grain properties. All these cause distortion

I. Basics of the finite element method

12 1. Fundamental equations of poro-elasticity and fluid flow through porous media

of formation when the neutral stress is applied. In fact, it is not difficult to construct an artificial
rock that does not deform uniformly with the neutral stress. However, passt tests show that
99% of pores are connected for the standard sedimentary rock; hence, the uniform grain defor-
mation due to the neutral stress is a good approximation.
The displacement
strain relations and the equation of equilibrium are the same regardless
of the porous or nonporous media if the total stress concept is used. Hence, if the pore pressure
is specified, the number of variables is 15 and the number of equations is also 15 for porous
media. To find the solution, the effective stresses or the total stresses may be selected for the
stress method. However, selecting the total stress as the primary variables may reduce the con-
fusion. The effective stress can be later calculated once the total stress is found.

1.3 Fluid flow through porous media

According to Darcy’s law, the volumetric rate ui of flow per unit cross-sectional area is
proportional to the pressure gradient and inversely proportional to the fluid viscosity. The
equation is given by
K
u52 δrP (1.23)
μ
where K is the proportional coefficient called as permeability and δ is the non-Darcy coeffi-
cient (Fig. 1.9).

FIGURE 1.9 Flow in and out from a small hexahedron.

I. Basics of the finite element method

 1.4 Matrix expression 13
The continuity equation is derived based on the material balance which states that the
volumetric rate of flow into the small volume element ΔxΔyΔz minus the flow out from
the element is the accumulated liquid volume in the element. Or,
 
2 ½ρux 1 Δðρux Þ 2 ρux 1 ½ρuy 1 Δðρuy Þ 2 ρuy 1 ½ρuz 1 Δðρuz Þ 2 ρuz (1.24)

Δt 5 ρφΔxΔyΔzt1Δt 2 ρφΔxΔyΔzjt

or
     
@ðρux Þ @ðρuy Þ @ðρuz Þ
2 ρux 1 Δx 2 ρux 1 ρuy 1 Δy 2 ρuy 1 ρuz 1 Δz 2 ρuz
@x  @y @z
Δt 5 ρφΔxΔyΔzt1Δt 2 ρφΔxΔyΔzjt

or

@ðρux Þ @ðρuy Þ @ðρuz Þ 

2 Δx 1 Δy 1 Δz Δt 5 ρφΔxΔyΔzt1Δt 2 ρφΔxΔyΔzjt (1.25)
@x @y @z

Simplifying the above equation gives fluid flow equations.

Fluid flow equations


Continuity equation: r ½ρu 5 2 @ðρφÞ=@t 2 q PCΩ (1.26)

Assuming the liquid compressibility is small, we have

1 dV
Phase behavior c 5 2 or ρ 5 ρs ecðp2ps Þ (1.27)
V dp

The boundary conditions are given by

P 5 Pb PCΓ2
P 5 Pi t50 (1.28)
q 5 qb qCΓ1

In the above equations, the variables are ux ; uy ; uz , ρ, and p, and the number of equations
is 5; hence, if the boundary conditions are given, the system of the equations may be
solved.

1.4 Matrix expression

A matrix expression is used for developing equations for the finite element method.

I. Basics of the finite element method

14 1. Fundamental equations of poro-elasticity and fluid flow through porous media

The displacement
strain relation is given by

2 @ 3
0 0
6 @x 7
6 7
6 @ 7
6 7
6 0 0 7
6 @y 7
6 7
6 @ 7
6 72 3
6 0 0 7
6 @z 7 u
6 74 5
ε56 @ @ 7 v (1.29)
6 7
6 0 7 w
6 @y @x 7
6 7
6 @ @ 7
6 0 7
6 7
6 @z @y 7
6 7
6 @ @ 7
4 0 5
@z @x

Equation of equilibrium is given by: rσ 1 F 5 0

where
2 3
@ @ @
6 @x 0 0 @y
0
@z 7
6 7
6 7
6 @ @ @ 7
6 0 7
r 5 6 0 @y 0 @x @z 7 (1.30)
6 7
6 7
6 @ 7
4 0 0 @ 0
@ 5
@z @y @x
 T  
σ 5 σx σy σz τ xy τ yz τ zx ; F 5 Fx Fy Fz

The stress
strain relation is given by

σe 5 Dε 2 ð1 2 c^m ÞIpo
 T
σe 5 σex σey σez τ exy τ eyz τ ezx
 T (1.31)
ε 5 εx εy εz γ xy γ yz γ zx
 T
I5 1 1 1 0 0 0

I. Basics of the finite element method

 1.4 Matrix expression 15
2 ν ν 3
1 0 0 0
6 12ν 12ν 7
6 7
6 ν 7
6 1 0 0 0 7
6 12ν 7
6 7
6 7
6 1 0 0 0 7
Eð1 2 νÞ 6 1 2 2ν 7
6 7
D5 6 0 0 7
ð1 1 νÞð1 2 2νÞ 6 2ð1 2 νÞ 7
6 7
6 1 2 2ν 7
6 Symmetry 0 7
6 2ð1 2 νÞ 7
6 7
6 7
6 1 2 2ν 7
4 5
2ð1 2 νÞ

c^m 5
1 2 2ν m
Em
= 1 2E2ν

For fluid flow through porous media, the matrix expression is given by

Darcy’s flow
K
u52 δrP
μ
0 1
@
B @x 0 0C
0 1 0 1 B C
B C
ux Kx Kxy Kxz B @ C
u 5 @ uy A; K 5 @ Kyx Ky Kyz A; r 5 B
B 0 @y 0C C (1.32)
B C
uz Kzx Kzy Kz B C
B @C
@ 0 0 A
@z

where μ, δ, and p are scalar.

Continuity equation


r ½ρu 5 2 @ðρφÞ=@t 2 q PCΩ (1.33)
Or
@ðρux Þ @ðρuy Þ @ðρuz Þ
1 1 5 2 @ðρφÞ=@t 2 q (1.34)
@x @y @z
Phase behavior: assuming the liquid compressibility is small, we have

ρ 5 ρs ecðp2ps Þ (1.35)

I. Basics of the finite element method

This page intentionally left blank
 C H A P T E R

2
Finite element methods
2.1 Discretization using the virtual work principle

Virtual work principle: The virtual work principle uses two forms of general principle, those
of virtual displacement and virtual forces. The virtual displacement form states that for a body
in equilibrium with body forces and applied boundary forces, the sum of the energy of the
applied loads and the strain energy stored during the virtual displacement is equal to zero. That
is, the following equation holds.
δW 1 δU 5 0 (2.1)
where δW 5 work done by the boundary stresses and δU 5 internal work done by stresses and
body forces.

(Proof):
The strain energy induced by virtual displacement δui is given by
ð ð
δU 5 σij δεij dΩ 2 Fi δui dΩ (2.2)

The following equation is used to modify Eq. (2.2)

ð ð ð
@ðσδuÞ @σ @ðδuÞ
dΩ 5 δu dΩ 1 σ dΩ (2.3)
@x @x @x
Then,
ð ð ð ð
@ðδui Þ
δU 5 σij δεij dΩ 2 Fi δui dΩ 5 σij dΩ 2 Fi δui dΩ
ð ð ð@xj
@ðσij δui Þ @σij
5 dΩ 2 δui dΩ 2 Fi δui dΩ (2.4)
@xj @xj

Finite Element Programming in Nonlinear Geomechanics and Transient Flow

DOI: https://doi.org/10.1016/B978-0-323-91112-2.00004-5 17 © 2021 Elsevier Inc. All rights reserved.
18 2. Finite element methods
Ð Ð
Using the Gauss divergence theory fi;i dΩ 5 fi ni dΓ, Eq. (2.4) becomes
ð ð ð ð ð
@σij @σij
δU 5 σij δui nj dΓ 2 δui dΩ 2 Fi δui dΩ 5 Ti δui dΓ 2 δui 1 Fi dΩ (2.5)
Γ Ω @xj Γ Ω @xj
The left term of the last equation is the work done by the surface stress and is written by
ð
δW 5 2 Ti δui dΓ (2.6)
Γ
Hence,
ð
@σij
δU 1 δW 5 2 δui 1 Fi dΩ (2.7)
Ω @xj
If a body is in equilibrium with applied boundary forces, then,
@σij
1 Fi 5 0; hence; (2.8)
@xj
δU 1 δW 5 0
We now derive a set of discretized equations based on the virtual displacement princi-
ple. Suppose now a body is in equilibrium state with body forces and applied boundary
forces. Or, the following equation holds.
δU 1 δW 5 0
Suppose u is the displacement in the body. The approximate form of u is obtained using
the displacements ue at element nodes with a proper interpolation function.
u 5 Nue (2.9)
The displacement strain relation becomes
ε 5 Bue (2.10)
where

2 3
@
6 @x 0 0 7
6 7
6 7
6 @ 7
6 0 0 7
6 @y 7
6 7
6 7
6 @ 7
6 0 0 7
6 @z 7
6 7
B56 @ @ 7½N1 N2 . . . . . . . . . : Nn  (2.11)
6 7
6 0 7
6 @y @x 7
6 7
6 @ 7
6 0 @ 7
6 @z @y 7
6 7
6 7
6 @ @ 7
4 0 5
@z @x

I. Basics of the finite element method

 2.1 Discretization using the virtual work principle 19

3 2
u
u54 v 5 (2.12)
w

 T
ε 5 εx εy εz γ xy γ yz γ zx (2.13)

Now, consider virtual displacement δue at the nodes. Then, the displacement and strain
are given by (Fig. 2.1)

δu 5 Nδue (2.14)

δε 5 Bδue (2.15)

The stress strain relation is given by

σ 5 Dε (2.16)

 T
σ 5 σx σy σz τ xy τ yz τ zx (2.17)

FIGURE 2.1 Virtual displacements at nodes.

I. Basics of the finite element method

20 2. Finite element methods

2 ν ν 3
1 0 0 0
6 12ν 12ν 7
6 7
6 ν 7
6 1 0 0 0 7
6 12ν 7
6 7
6 7
6 1 0 0 0 7
Eð1 2 νÞ 6 7
6 1 2 2ν 0 0 7
D5 6 7
ð1 1 νÞð1 2 2νÞ 6 2ð1 2 νÞ 7
6 7
6 1 2 2ν 7
6 7
6 Symmetry 0 7
6 2ð1 2 νÞ 7
6 7
6 1 2 2ν 7
4 5
2ð1 2 νÞ

The work done from external stress Ti is given by

ð
δW 5 ðδuÞT TdΓ (2.18)
Se

The strain energy stored in the body is given by

ð
δU 5 ðδεT σ 2 δuT FÞdΩ (2.19)
Ve

According to the virtual work principle, if the body is in equilibrium, the following
equation holds.
δU 1 δW 5 0
Hence,
ð ð
ðδuÞT TdΓ 1 ðδεT σ 2 δuT FÞdΩ 5 0 (2.20)
Se Ve
ð ð
ðδue ÞT N T TdΓ 1 ðδue ÞT ððBÞT σ 2 N T FÞdΩ 5 0 (2.21)
Se Ve

Since Eq. (2.21) holds for arbitrary virtual displacement, the following equation is obtained.
ð ð ð ð
N T TdΓ 1 ððBÞT σ 2 N T FÞdΩ 5 0 N T TdΓ 1 ððBÞT DBue 2 N T FÞdΩ 5 0 (2.22)
Se Ve Se Ve

Hence,

The finite element method

Ke ue 5 Fe 1 T e (2.23)
where
ð ð ð
Ke 5 ððBÞT DBÞdΩ; Fe 5 N T FdΩ; T e 5 2 N T TdΓ
Ve Ve Se

I. Basics of the finite element method

 2.2 Discretization using the minimization of total potential energy 21
The above equation is the discretized form of the fundamental equations of elasticity. It
applies to one element. If it is assembled for all the elements, a set of linear equations are
obtained with respect to the nodal displacements.

2.2 Discretization using the minimization of total potential energy

Minimization of potential energy: If a body is in equilibrium with body forces and applied
boundary forces, the change in potential energy with virtual displacement is zero. And, if the
potential energy is minimum, the equilibrium is stationary. Suppose the external work done by
the boundary stresses is W and the internally stored potential done by stresses and body forces
is U, then, the following equation holds if the body is in the equilibrium state.
δUp 5 δðW 1 U Þ 5 0 (2.24)

(Example):
Consider a coiled spring (spring constant 5 K) with a weight F. Suppose D is the
extension due to the weight (Fig. 2.2).
Then, the potential energy stored is given as
1
Up 5 KD2 2 FD (2.25)
2
Since the potential energy becomes minimum at the equilibrium, the following equation
holds
@Up
5 KD 2 F 5 0 (2.26)
@D

FIGURE 2.2 Extension of a coiled spring with force F.

I. Basics of the finite element method

22 2. Finite element methods

Hence, the stretched spring length becomes

D 5 F=K (2.27)
Now, we use the minimum potential energy principle to discretize the set of the funda-
mental elasticity equations. Let u be the displacement in a body in equilibrium with body
forces and applied boundary forces. The displacement u is approximated with the nodal
displacements ue given by
u 5 Nue (2.28)
The relation between displacement and strain is given by
ε 5 Bue (2.29)
The stress strain relation is given by
σ 5 Dε (2.30)
Let Ti be the boundary stress and the energy applied by the external stress is given by
ð
W 5 ðuÞT TdΓ (2.31)
Se
The strain energy stored in the body becomes
ð
1 T
U5 ε σ 2 uT F dΩ (2.32)
Ve 2

Hence, the potential energy becomes

ð ð
1 T
Up 5 U 1 W 5 ε σ 2 uT F dΩ 1 ðuÞT TdΓ (2.33)
Ve 2 Se
Substituting Eq. (2.28) into Eq. (2.33) gives
ð ð  ð
e T 1 eT T
Up 5 ðu Þ N TdΓ 1 ðu Þ
T
ððBÞ DBu ÞdΩ u 2 N T FdΩ
e e
(2.34)
Se 2 Ve Ve
If the body is in the stationary equilibrium, the displacement minimizing the potential
is found by
@Up
50 (2.35)
@ue
Hence,
Ke ue 5 Fe 1 Te (2.36)
where
ð
K 5
e
ððBÞT DBÞdΩ
ð Ve
Fe 5 N T FdΩ
Veð

T 5 2 N T TdΓ
e
Se
The above equation is the discretized form of the fundamental equations of elasticity.

I. Basics of the finite element method

 2.3 Discretization using the residual method 23

2.3 Discretization using the residual method

Let us consider the following forms of differential equations (Fig. 2.3).

Residual method
AðuÞ 5 0 uCΩ
(2.37)
BðuÞ 5 0 uCΓ
We now consider the integral form of equation with weight wT .
ð ð
wT AdΩ 1 wT BdΓ 5 0 (2.38)

where w is a set of arbitrary weight functions.

We now prove if Eq. (2.37) holds, then Eq. (2.38) holds. And, if Eq. (2.38) holds for arbitrary
weight function wT , Eq. (2.37) holds.

(Proof by contradiction):
Suppose there exists a point x 5 x where AuðxÞ 6¼ 0 or BuðxÞ 6¼ 0. Since these functions are
continuous, AuðxÞ 6¼ 0 or BuðxÞ 6¼ 0 within the neighborhood Ωε of x. Since wT is an arbi-
trary set of functions, select wT 5 1 within xAΩε and wT 5 0 outside of Ωε . Then,
ð ð ð ð
w AdΩ 1
T
w BdΓ 5
T
w AdΩ 5
T
AdΩ
Ω Γ Ωε Ωε
Ð Ð
~
Since AuðxÞ 6¼ 0 within Ωε , there exists xAΩ e such that ~
Ωε AdΩ 5 AðuðxÞÞ Ωε dΩ 6¼ 0.

FIGURE 2.3 Neighborhood of x.

I. Basics of the finite element method

24 2. Finite element methods
Ð Ð
Hence, Ω wT AdΩ 1 Γ wT BdΓ 6¼ 0 within the neighborhood of AuðxÞ 6¼ 0 or BuðxÞ 6¼ 0.
This
Ð T is contrary
Ð T toÐ Eq. (2.38). Similarly,
Ð for the neighborhood of BuðxÞ 6¼ 0,
Ω w AdΩ 1 Γ w BdΓ 5 Γε BudΓ 5 Bðuð ~
xÞÞ Γε dΩ ¼
6 0. This is also contrary to Eq. (2.38).
Therefore, for arbitrary weight function wT , Eq. (2.38) must hold.
Ð T Ð T
Conversely, if AðuÞ 5 0 uCΩ and BðuÞ 5 0 uCΓ, obviously Ð Tw AdΩ 1Ð Tw BdΓ 5 0 holds.
Hence, the condition AðuÞ 5 0 uCΩ and BðuÞ 5 0 uCΓ, and w AdΩ 1 w BdΓ 5 0 are nec-
essary and sufficient conditions.
Using the residual method, the system of the fundamental elasticity equations is
discretized.
Ð Ð
Using wT AdΩ 1 wT BdΓ 5 0, the following equation should hold:
ð ð ð
@σij 2 2
T
w 1 Fi dΩ 1 w ðTk 2 T k ÞdΓ 1
T
wT ðuk 2 uk ÞdΓ 5 0 (2.39)
@xj Γ1 Γ2

The finite element methods with the nodal displacements

Ð as the unknown variables
2
use the boundary condition uk 5 uk at Γ2 . Hence Γ2 wT ðuk 2 uk ÞdΓ 5 0 is satisfied. Hence,

ð ð ð ð ð
@σij
wT
@xj
1 Fi dΩ 1
Γ1
2
wT ðTk 2 Tk ÞdΓ 5 2 ðrwÞT σdΩ 1
Γ1 1Γ2

wT σ ndΓ 1 wT FdΩ

ð
2
1 wT ðTk 2 T k ÞdΓ
Γ1

ð ð ð ð (2.40)
2
52 ðrwÞT σdΩ 1 wT FdΩ 2 wT T k dΓ 2 wT Tk dΓ 5 0
Γ1 Γ2

On Γ2 boundary, uÐk 5 uk and regardless of any Tk , the results should be the same.
Hence, setting wT 5 0, Γ2 wT Tk dΓ 5 0.
Choosing the shape function as the weight, we have now ｗ ＝ Ｎ and B 5 rw. Hence,
ð ð ð
2
2 ðrwÞT σdΩ 1 wT FdΩ 2 wT T k dΓ
Γ1
ð ð ð
2 (2.41)
5 2 B DBdΩ u 1 N FdΩ 2 wT T k dΓ 5 0
T e T
Γ1

Hence,

Ke ue 5 Fe 1 Te (2.42)

where
ð ð ð
Ke 5 ððBÞT DBÞdΩ; Fe 5 N T FdΩ; T e 5 2 N T TdΓ
Ve Ve Se

I. Basics of the finite element method

 2.4 Discretization of the set of flow equations through porous media using the residual method 25

2.4 Discretization of the set of flow equations through porous media

using the residual method

The continuity equation, Darcy’s law and density with small compressibility are:


r ½ρu 5 2 @ðρφÞ=@t 2 q PCΩ
Kk
u52 δ r P (2.43)
μ
ρ 5 ρs ecðp2ps Þ

The boundary conditions are given by (Fig. 2.4)

P 5 Pb PCΓ2
P 5 Pi t50 (2.44)
q 5 qb qCΓ1

Discretization using the weighted residual method uses the following equation.
ð ð ð
T

w ðr ½ρu 1 @ðρφÞ=@t 1 qÞdΩ 1
Γ1
T 2
w ð2 q 1 q ÞdΓ 1
Γ2
2
wT ðp 2 p ÞdΓ 5 0 (2.45)

Choosing pressure as the unknown variable, the following equation on Γ2 boundary is

automatically satisfied.
ð
2
wT ðp 2 p ÞdΓ 5 0 (2.46)
Γ2
Hence,
ð ð
T

w ðr ½ρu 1 @ðρφÞ=@t 1 qÞdΩ 1
Γ1
2
wT ð2q1 q ÞdΓ 5 0 (2.47)

FIGURE 2.4 Boundary conditions for fluid flow.

I. Basics of the finite element method

26 2. Finite element methods

Choosing the shape function as the weight function, Eq. (2.47) becomes
ð ð ð ð ð

ð
 ð Γ1
2

ð ð ð Γ1
2
N T ðr ½ρuÞdΩ 1 N T ð2 q 1 q ÞdΓ 5 N T ðρuT nÞdΓ 2 ðrNÞT ðρuÞdΩ 1 N T ð2q 1 q ÞdΓ
2 2
5 N T qdΓ 1 ðrNÞT TðrNÞdΩPe 1 N T ð2q 1 q ÞdΓ 5 N T q dΓ 1 ðrNÞT TðrNÞdΩPe
Γ1 Γ1 Γ1
(2.48)
where
T 5 ρkδK=μ
The following matrix form of equations are used in the above equations

q 5 qx nx 1 qy ny 1 qz nz (2.49)

@ðρux Þ @ðρuy Þ @ðρuz Þ


r ρu 5
@x
1
@y
1
@z
(2.50)
0 1
nx
ρuT n 5 ρðux uy uz Þ@ ny A (2.51)
nz
2 3
@N1 @N1 @N1
6 @x @y @z 7
6 7
6 @N @N @N 72 3
6 2 2 27
6 7 ρux
6 @z 74 ρu 5
ðrNÞT 5 6 @x @y
7
6
6   7
7 ρuz
y
 (2.52)
6 
6 @N @N @N 7  7 
4 n n n5
@x @y @z

Eq. (2.47) becomes

ð ð ð
T

N ðr ½ρu 1 @ðρφÞ=@t 1 qÞdΩ 1
Γ1
T 2
N ð2 q 1 q ÞdΓ 1
Γ2
2
N T ðp 2 p ÞdΓ

ð ð ð ð (2.53)
2
5 N T q dΓ 1 ðrNÞT TðrNÞdΩPe 1 N T qdΩ 1 N T ð@ðρφÞ=@tÞdΩ
Γ1

Substituting ρ 5 ρ0 ecðP2P0 Þ ,
ð ð
N T ð@ðρφÞ=@tÞdΩ 5 N T ðρφc @P=@tÞdΩ
ð (2.54)
5 N T ðρφcNð@P=@tÞe ÞdΩ

I. Basics of the finite element method

 2.4 Discretization of the set of flow equations through porous media using the residual method 27
Evaluating ρφc at the element center and expressing the pressure with respect to the
nodal pressure, the time term becomes

ð@P=@tÞe 5 P
e (2.55)

and
ð ð ð
N ð@ðρφÞ=@tÞdΩ 5 ðρφcÞcenter N
T T 
N dΩðP Þ 1 ðρÞcenter N T N dΩðφ_ 1 1 φ_ 2 Þe
e
(2.56)

φ1 5 φecr ðp2po Þ and φ2 5 φeco ðp2po Þ where cr is the rock compressibility and co is the miscella-
neous rock compressibility, which cannot be handled by the standard rock compressibility.
The following discretized form of equations is derived.
ð ð ð ð
T 2
N q dΓ 1 ðrNÞ TðrNÞdΩP 1 N qdΩ 1 ρφðcf 1 cr 1 co Þ N T NdΩðP Þe 5 0
T e T
(2.57)

Γ1

The volume integration of the source and sink terms of Eq. (2.57) are actually the pro-
duction and injection of fluids from wells. Hence, it may be written as a line integration as
ð ð
N qdΩ 5 N T ðx; y; zÞqðx; y; zÞdl
T
(2.58)

The transmissibility T is evaluated at the upstream point to enhance the stability of the
solution. Therefore, Eq. (2.53) becomes
ð ð
2 ρkδ
N T q dΓ 1 ðrNÞT KðrNÞdΩðPe Þn11
Γ1 μ up
ð ð (2.59)
 
1 N T qdΩ 1 ρφðcf 1 cr 1 co Þ N T NdΩ ðPe Þn11 2 ðPe Þn =Δt 5 0

The final discretized form of fluid flow

Ke ðpe Þn11 5 Fe 1 T e (2.60)

where
ð h ið
ρkδ
Ke 5 ðrNÞT KðrNÞdΩ 1 ρφðcf 1cr 1co Þ center =Δt N T NdΩFe 5
μ up
ð h ið ð (2.61)
2
2 N T qdl 1 ρφðcf 1cr 1co Þ center =Δt N T NdΩðPe Þn T e 5 2 N T q dΓ
Γ1

Note that φ_ 2 5 Δφ2 =Δt, which is assigned at each cell during only time step, which has
induced the porosity change.

I. Basics of the finite element method

This page intentionally left blank
 C H A P T E R

3
Finite element method with analytical
integration using simple elements
3.1 Discretization using 3D tetrahedral elements

Consider now a tetrahedral element given by Fig. 3.1. We determine the shape function Ni .
2
Now, let choose Ni in the following form:
2
N i ðxÞ 5 ðai 1 bi x 1 ci y 1 di zÞ=ð6VÞ (3.1)
The coordinates of the four nodes are given by
2 3
xi
xi 5 4 yi 5for i 5 1; 2; 3; 4
2
(3.2)
zi
2
Since Ni 5 1 at i-node and 0 at the other three nodes, the shape function is given by
   
1 x y z   1 xi yi zi 

2  1 xj yj z j   1 xj yj z j 
Ni ðxÞ 5   =
 

 (3.3)
 1 xm ym z m   1 xm ym z m 
 1 xp yp z p   1 xp yp z p 

This function satisfies

2 2
Ni ðxα Þ 5 δiα (3.4)
Or, the shape function is calculated in the following form.
2
N i ðxÞ 5 ðai 1 bi x 1 ci y 1 di zÞ=ð6VÞ
 
 1 x1 y 1 z 1 
 
 1 x2 y2 z 2 
6V 5  
 (3.5)
 1 x3 y3 z 3 
 1 x4 y4 z 4 

Finite Element Programming in Nonlinear Geomechanics and Transient Flow

DOI: https://doi.org/10.1016/B978-0-323-91112-2.00009-4 29 © 2021 Elsevier Inc. All rights reserved.
Discovering Diverse Content Through
Random Scribd Documents
 II. Size relatively large, body-length 6 inches or more, teeth distinctly
anisodont, skull with nearly even posterior table, limbs very long, ventral
armature highly developed Cephalerpeton

Genus AMPHIBAMUS Cope, 1865.

Cope, Proc. Phila. Acad. Nat. Sci., 1865, pp. 134-137. Geol. Surv. Ills., 11,
pp. 135-141, pl. xxxii, 1 text-fig.
Hay, Proc. Am. Phil. Soc., XXXIX, p. 120, 1900.
Moodie, Jour. Geol., XVII, p. 81, fig. 24, 1909.

Type: Amphibamus grandiceps Cope.

The publication of the type species of this genus began the
researches of Professor Cope on the extinct Amphibia of North
America, which he continued for so many years with such excellent
results (105-177). The description was based on a single specimen
(plate 3, fig. 7) belonging to Mr. Joseph Evans, of Morris, Illinois,
who loaned it to Dr. Worthen for the Illinois Geological State Survey
(107), in order that it might be described. The type has been
destroyed by fire; so I am informed by Mr. L. E. Daniels, of Rolling
Prairie, Indiana. There are two other known specimens of the
species. One is in the collection of Mr. Daniels and the other No. 794,
of Yale University Museum.
This genus may be clearly separated from all the other
microsaurians by characters which are peculiar to the form. Among
these may be mentioned the possession of sclerotic plates in the
eyes; the large size of the orbits in comparison with the dimensions
of the skull; the short, broad form of the body; the very short tail;
the possession of a calcified cartilaginous pubis; clawed phalanges;
presacrals 22. The character which places the genus distinctly in the
Microsauria is the possession of long, slender, curved ribs, first
detected on Mr. Daniels's specimen (plate 14, figs. 1, 2), by Dr. Hay
(316). Its stegocephalian characters are evident in every particular
of its anatomy the roofed skull, the arrangement of the cranial
elements, the presence of a well-developed ventral armature, and
the digital formula (4 for the hand and 5 for the foot).
The genus Amphibamus was regarded by Cope as a
representative of a new order of vertebrates which he called (105)
Xenorachia. He later (123) abandoned this, however. Fritsch (251),
Zittel (642), and others regarded Amphibamus as a branchiosaurian.
The exact position of the form was uncertain until 1900, when Dr.
Hay (316) described the long, curved ribs and suggested its place
among the Microsauria. He, however (Cat. Foss. Vert., p. 410), made
the mistake of including the branchiosaurian family Protritonidæ,
under Microsauria, thus confusing the subject further. The genus
(462) has not the slightest relationship with the Branchiosauria.

Amphibamus grandiceps Cope.

 Cope, Proc. Phila. Acad. Nat. Sci., pp. 134-137, 1865; Geol. Surv. Ills., 11,
pp. 135-141, pl. xxxii, and 1 woodcut, 1866.
Hay, Proc. Am. Phil. Soc., XXXIX, p. 120, 1900.
Moodie, Jour. Geol., XVII, No. 1, p. 82, fig. 24, 1909.
Moodie, Kan. Univ. Sci. Bull., VI, No. 2, pp. 343-349, pl. 1, figs. 1 and 2;
pl. 5, fig. 3; pl. 7, fig. 1; pl. 11, 12, 13, 1912.

Type: Specimen has been destroyed. There is an excellent

specimen (plate 4, figs. 5. 6), No. 794 (1234), in Yale University
Museum, and another nearly as good in the possession of Mr. L. E.
Daniels, of Rolling Prairie, Indiana.
Horizon and locality: Mazon Creek shales, near Morris, Illinois.
The form of the skull of Amphibamus grandiceps Cope is not
unlike that of Tuditanus minimus Moodie (462) from the Linton,
Ohio, beds, but it is less acuminate than in that form. The large size
of the orbits is especially striking. The shape of the skull is triangular,
with concavities in the posterior table which correspond to the ear-
slits so characteristic of Metoposaurus (242) from the Keuper of
Germany. The narrowed posterior table of the skull is truncate, as in
several other genera of Microsauria, notably Tuditanus and
Saurerpeton. In structure the skull differs but little from many of the
other Carboniferous forms, but the arrangement of the elements of
the skull is more regular than in other genera.
The premaxillaries are very small elements in the anterior tip of
the skull. They border the nares. The skull is rather peculiar among
the Microsauria in the possession of a distinct lacrimal. I have
detected this element in the cranium of Stegops divaricata Cope. As
here defined the lacrimal is triangular, with its posterior border
formed exclusively by the prefrontal. Its other relations are the
normal ones. The nasal is elongate, with the usual relations of that
element. The frontal is slightly longer and broader than the nasal. It
apparently forms a portion of the inner border of the orbit. The
parietal foramen lies in the anterior fourth of the parietal, a rather
unusual position for this structure. The parietals, as in so many of
the Microsauria, together form the largest element of the skull and
are roughly a triangular area in the postero-median portion of the
skull. The postparietal and the tabulare are clearly distinguishable
and they have the usual relations for those elements. The maxillary,
jugal, and quadratojugal together form the greater part of the
maxillary border. The postero-lateral angle of the skull is, as visual,
formed by the squamosal. The orbit is bounded posteriorly by the
postorbital and the postfrontal, which include in the angle between
them the quadrangular squamosal. The orbit is especially remarkable
for its size as compared with the dimensions of the skull, being
without a parallel among other known Microsauria. Around the
border of the orbit in the specimen Cope studied (105) there were
found 14 quadrangular plates which he called "superciliary plates."
Hay (316) was inclined to regard them as sclerotic plates. In the Yale
Museum specimen (plate 4, figs. 5, 6) there are 20 of these plates,
and there seems to be no doubt that they are sclerotic elements. In
the restoration (fig. 26) 29 sclerotic plates are given, but there is no
assurance that this number is the exact one. They may also have
been slightly larger, but not as large as in Branchiosaurus.
The vertebral column is preserved nearly entire in the Daniels
specimen and quite entire (478) in the Yale specimen. Cope, in his
study of the type (105, 107), thought there could be no more than
13 presacrals, but the specimen was poorly preserved and indecisive
on this point. Dr. Hay (316) was inclined to the opinion that there
were less than 20. The Yale specimen shows 22 centra, which are
elongate, hour-glass-shaped bodies, with the neural spine a long,
low crest running the entire length of the centrum, with a median
elevation, so that in lateral view the spine would be triangular in
form. The body of the centrum is expanded laterally into a
diapophysis which extends anteriorly. The posterior vertebræ, at
least, had the notochord largely persistent. The osseous part of the
vertebra seems to have been but a thin shell, and the structure of
the zygapophyses can not be determined. That they were dorsal in
position is, however, evident from several vertebræ. The points of
these structures project laterally.
The tail is short and the caudal
vertebræ weakly developed.
There are distinct
impressions of at least 12 pairs
of ribs in the Daniels specimen.
They are long, slender, and
curved, and there is no definite
assurance that there were as
many ribs as are indicated (fig.
26) in the restoration (462). The
ribs are intercentral (469) and
probably occupied the full length
of the vertebral column. There
may have been as many as
indicated in the restoration.
One of the most interesting
features of the Yale specimen is
the preservation of a small patch
of skin, evidently from the back,
lying to one side near the head,
measuring 5 mm. in length by 3 Fig. 26. Restoration of body outline and
skeleton of Amphibamus grandiceps
mm. in width. The fragment
Cope, from Mazon Creek, Illinois, shales.
shows the skin to be of Restoration is based on complete
tuberculated scales, 4 of which specimens of the species and on Cope's
occupy the length of 1 mm. The drawing. Form of body is indicated in
scales are somewhat hexagonal, one specimen, that in possession of Mr.
Daniels. × 1.5.
almost rounded, and were
Skull: pmx, premaxilla; n, nasal; fr, frontal;
relatively quite thick. They lie in par, parietal; la, lacrimal; pf, prefrontal; pof,
a close mosaic (fig. 27). postfrontal; po, postorbital; pp, postparietal;
spt, supratemporal; mx, maxilla; j, jugal; qj,
quadratojugal; sq, squamosal; tab, tabulare.
The Yale specimen has, very
Skeleton: ic, interclavicle; cl, clavicle; sc,
well preserved, a portion of the scapula; h, humerus; r-u, radius, ulna; r,
ventral scutellæ, of the throat, carpus; pu, pubis; il, ilium; f, femur; t, tibia; fb,
fibula; ts, tarsus; x, ischium.
chest, and belly. The
arrangement of the plates on
the throat and chest is almost
exactly the reverse of what
Credner has described (190) for
Branchiosaurus amblystomus
Cred. On the throat, in the
present form, the chevron
points anteriorly, and it is the
anterior prolongation of the
belly scutes with the postero-
lateral projection of the gular
scutes which form the chest
and arm scutellation. The belly
chevrons point anteriorly, as in
Branchiosaurus, the rods
formed by the scutes being Fig. 27. Skeleton of Amphibamus
grandiceps Cope. × 1.4.
straight and not curved as in
Branchiosaurus. The entire c, carpus; cl, clavicle; cr, caudal rib; cv,
ventral armature preserved is caudal vertebra; f, femur; h, humerus: il,
displaced to the left of the ilium; s, skin: or, orbit: r, radius; ul, ulna;
animal and only the anterior sc, scapula; sp, sclerotic plates; t, tibia
and fibula; ts, tarsus; vs, ventral
portion is preserved. scutellæ. Specimen No. 794, Yale
University Museum.
The pectoral girdle is only
partially known. The scapula is
crescent-shaped. The other
elements are indicated only by fragments and nothing is known of
their form.
The arm elements are nearly all known. The humerus is slender
and expanded at the ends, with its articular surfaces well developed.
The separate radius and ulna are of approximately the same size
and length. The carpus is unossified. The complete phalangeal
formula for the hand of Amphibamus is unknown. The third digit
seems to have 4 elements. The formula 2-2-3-2 has been suggested
(462).
 The pelvis is very satisfactorily known. The ilium is a long,
slender, straight rod, with expanded ends. The ischium is shown on
both sides of the vertebral column in the Yale specimen. Its form is
almost identical with that of Paleohatteria longicaudata Credner,
from the Rothliegenden of Saxony. The ischia are apparently
approximate in the median line, though this character is somewhat
obscured by the impression of the caudal vertebræ. Their relation
with the ilium, other than that they were posterior to it, is uncertain.
The pubis is, apparently, calcified cartilage. It is a squarish plate,
somewhat corrugated, lying anterior to the ilium in the Daniels
specimen. The elements of the pelvis were undoubtedly hung loosely
in the flesh, as in modern salamanders, since there is no indication
of articular surfaces.
The hind limb is well known, the type having a nearly complete
leg with the foot. The Daniels and the Yale specimens supplement
and substantiate the type. The femur is longer than the humerus,
but more slender, with its articular surfaces about as well developed
as in the humerus. The element is a simple rod of bone without
muscular crests of any kind. The tibia and fibula are, likewise,
slender separate rods of bone. The tarsus is unossified. The
phalangeal formula is 2-2-3-4-3, and is fairly definite.

Fig. 28. Restoration of probable appearance of

Amphibamus grandiceps Cope on the basis of the
material described herewith. × 1.5.

In the type specimen the matrix in the orbit was blackened as if

by the pigmentum nigrum of the choroid. The same has been
noticed in other specimens. Professor Cope thought this indicated
that the animal was nocturnal.
There are many characters in Amphibamus which seem to
approximate the reptilian type of structure. Among these may be
mentioned the character of the articular surfaces of the limb bones,
the intercentral position of the ribs, the incipient double-headedness
and the curvature of the ribs, the presence of a cartilaginous
calcified pubis, the length of the limbs, and the clawed character of
the phalanges.
Amphibamus was a low, flat, short, and undoubtedly a creeping,
crawling animal, possibly spending a portion of its time in the water;
but it could not have been a swimmer. It was one of nature's first
attempts at constructing a land vertebrate.

Measurements of Amphibamus grandiceps Cope.

Collection of Mr. L. E. Daniels, of Rolling Prairie, Indiana:
mm.
Entire length of specimen 62
Posterior width of head 15
Length of head 15
Posterior height of skull 3
Length of orbit 5
Width of orbit 3.5
Interorbital width 4
Width of skull in front of orbits 11
Width of skull just back of orbits 16
Length of presacral region of the vertebral column 30
Length of tail 13
Length of fore limb 13.5
Length of humerus 4
Length of radius and ulna 3
Length of right hand as preserved 3.5
Length of rib along curve 5.5
 Length of hind limb 17
Length of ilium 4
Length of vertebral centrum 1.75
Length of portion of scapula (?) preserved 4.5
Length of foot 6.5
Width of impression of body midway 16

No. 794 (1234), Yale University Museum:

mm.
Length of skeleton 67
Length of skull 15
Posterior width of skull 15
Depth of tympanic notch 4
Width of tympanic notch 6
Long diameter of the orbit 7
Transverse diameter of the orbit 5.5
Interorbital width 4.5
Diameter of pineal foramen .75
Length of cervical series of vertebræ 9
Length of dorsal series 35
Length of caudal series 13
Length of a centrum of the dorsal series 1.5
Length of dorsal rib 3.5
Length of arm 20
Length of humerus 7
Length of radius and ulna 4
Width of carpal space 3
Length of third digit 5
Length of leg 25
Length of ilium 3
Length of femur 9
Length of tibia and fibula 5
Length of carpal space 4
Length of 1st digit 3
Length of 2d digit 4.5
 Length of 4th digit 7
3 ventral scutellæ in 1 mm.

Amphibamus thoracatus Moodie.

Moodie, Proc. U. S. Nat. Mus., 40, pp. 431-433, fig. 2, 1911.
Moodie, Kans. Univ. Sci. Bull., VI, No. 2, pp. 347-349, pl. 5, fig. 2, 1912.

Type: Specimen No. 4306, U. S. National Museum.

Horizon and locality: Mazon Creek shales, near Morris, Illinois.
The type is a part of the collection of Mr. R. D. Lacoe, in the U. S.
National Museum. The fossil is very poorly preserved, but the
remains are to be seen on both halves of the nodule, so that
considerable can be made out as to its structure.
The chief diagnostic characters which will at once distinguish the
species are the elongate arm, large interclavicle, shape of the
vertebra, and triangular skull.
The portions of the animal which are preserved are the
impression of the skull with one orbit, the right humerus and radius
with portions of others, and traces of ventral scutellæ. These
remains are so intermingled with the remains of plants that it has
been quite difficult to distinguish bone impression from plants. This,
however, has been done by whitening the fossils with ammonium
chloride, when the texture of the fossils serves to distinguish the one
from the other. Parts of the plants have been converted into galena
and kaolin, as have also parts of the bones, so the task has been
rendered doubly difficult. There can be no doubt, however, that the
observations recorded below are correct. The position of the arm in
relation to the pectoral girdle and the position of the girdle in
relation to the skull impression first called attention to the possible
presence of a fossil amphibian.
There is little to be said of the skull. It is merely an impression in
the nodule. It is triangular in form, with the snout an acute angle.
The angle is, however, exaggerated by the compression to which the
fossil has been subjected. The right side of the skull lies over a
portion of some plant. The animal is preserved on its back, so that
this gives a good opportunity for the study of the pectoral girdle,
which is partially preserved. The interclavicle is very large and from
it the species has been given its specific name (thoracatus—armed
with a breast plate). It is an exaggerated T, with the stem very short
with its anterior margin curved, and ending in a rather sharp,
elongate point. The interclavicle recalls, in a measure, the same
element of the Branchiosauria, although it is much more expanded
anteriorly and has a shorter spine. In these respects it resembles
more nearly a reptilian interclavicle (fig. 14 B) .
The clavicle is of the simple triangular shape so characteristic of
the Microsauria. It is somewhat displaced backward and its inner
margin is slightly obscured. The humerus is elongate, apparently
cylindrical, and with expanded ends, resembling very closely the
humerus of Amphibamus grandiceps, although its proportions are
much greater than in that species. Its length is almost equal to the
length of the skull, while in A. grandiceps the length of the humerus
is only half that of the skull. The radius (ulna?) resembles in its
general proportions those of the humerus. It is a more elongate,
slender, lighter bone. The impression of the other bone of the
forearm is obscured.
A portion of a single vertebral centrum from the posterior part of
the dorsal series is preserved. It is apparently amphicœlous; its
width is nearly half greater than its length.
Measurements of the Type of Amphibamus thoracatus Moodie.
(No. 4306, U. S. National Museum.)

mm.
Length of entire specimen, as preserved 60
Length of skull impression 18
Greatest width of same 15.5
Long diameter of right orbit 4
 Transverse diameter of same 3
Transverse width of interclavicle 14
Long diameter of same 7(?)
Long diameter of clavicle 9
Greatest transverse diameter 3
Length of humerus 10
Greatest diameter of same 4
Least diameter of same 1.5
Length of radius (ulna?) 11
Length of vertebral centrum 2
Width of same 3

Genus CEPHALERPETON Moodie.

Moodie, Kans. Univ. Sci. Bull., VI, No. 2, p. 340, 1912.

Type: Cephalerpeton ventriarmatum Moodie.

This genus is founded on remains of a nearly entire individual of
a relatively large microsaurian from the Mazon Creek shales. The
genus is most immediately related to the Amphibamidæ, of which
two species are already known, Amphibamus grandiceps Cope and
A. thoracatus Moodie. The present genus differs from these species
in many respects, notably in size. The skull in Cephalerpeton is
nearly as long as half the entire body of Amphibamus grandiceps
Cope, inclusive of the tail. Other structural differences are the
anisodont teeth, the large size and the more median position of the
orbits, and the absence of the posterior tympanic notch in
Cephalerpeton. The form of the skull recalls that of Melanerpeton
and Pelosaurus (190) of Europe, but those genera are
branchiosaurian, while the present form, from the structure of the
vertebræ and the long, curved ribs, is an undoubted microsaurian.
Nothing like it occurs in any of the amphibian faunas thus far made
known. It is most nearly approached by a member of the genus
Erpetosaurus, but from this genus the present form is readily
distinguished by the smooth skull bones, the absence of a posterior
table to the skull, and the presence of a highly developed ventral
armature. The interorbital width is less than the transverse diameter
of the orbit.

MOODIE PLATE 18

1. Type specimen of Erpelosaurus sculptilis Moodie, from the

Cannelton Shales of Pennsylvania. Original in the University of
Chicago, Walker Museum.
 2. Skeletal elements of Eryops sp. indet., from the Pittsburgh Red
Shale at Pitcairn, Pennsylvania. a=nearly complete vertebra; b
and c=ribs; d=pleurocentrum; f=neural arch and spine. Originals
in the Carnegie Museum at Pittsburgh. After Case.
3. Photograph of amphibian footprints, Dromopus aduncus Branson,
from the Mississippian shales of Giles County, Virginia. × 1/3.
Courtesy of Dr. Branson. Original in the Museum at Oberlin
College.
4. Photograph of type of Thinopus antiquus Marsh, the amphibian
footprint from the Devonian of Pennsylvania. × 1/4. Courtesy of
Dr. Lull. Original No. 784, Vale University Museum.

Cephalerpeton ventriarmatum Moodie.

Moodie, Kans. Univ. Sci. Bull., VII, No. 2, pp, 350-352, pl. 1, fig. 4; pl. 7,
fig. 2, 1912.

Type: Specimen No. 796, of Yale University Museum.

Horizon and locality: Collected at Mazon Creek in 1871, near
Morris, Illinois.
The remains on which the present species is based consist of an
almost entire skull, 26 consecutive vertebræ, both fore limbs, 20 ribs
preserved on the right side of the body, and a portion of the ventral
armature (plate 4, fig. 4).
The skull is very broad posteriorly, its width being one-third
greater than its length, with due allowance for crushing. A pineal
foramen is not preserved. The sutures bounding the premaxillaries,
the maxillæ, the nasals, the prefrontals, the frontals, a portion of the
parietals, the squamosal, the supratemporal, the quadratojugal, and
the quadrate (?) are fairly well preserved. The arrangement of these
elements can be discerned by reference to figure 29. The prefrontals
are unusually large and are triangular in shape. The supratemporal is
also quite large. The surface of the skull bones is smooth and there
is nowhere an indication of sculpture.
Portions of 4 sclerotic plates are
preserved in the right orbit. These
measure 0.5 by 0.75 mm. The orbits are
large and the interorbital space is less
than the transverse diameter of the orbit.
Thirteen teeth, apparently pleurodont,
are preserved on the left maxilla. They
are short, sharply pointed, smooth, and
unequal. The first 2 left maxillary teeth
from the anterior end are short; then
follows a tooth which is one-third longer
than these two; the fourth tooth is
somewhat shorter than the third; the
fifth and sixth are still shorter and are
practically equal in size, though
somewhat larger than the first two.
Fig. 29. Skeleton of
The right mandible is preserved Cephalerpeton ventriarmatum
Moodie. × 1.
almost entire, though so badly eroded
that little can be said of its structure. pf, prefrontal; cl, clavicle; m,
Impressions of 12 teeth are present on mandible; h, humerus; j,
the mandible and all are, apparently, jugal; mx, maxilla; or, orbit;
equal. The cotylus seems to have been ph, phalanges of hand; par;
parietal; po, postorbital; r,
far posterior and an angle of the radius; sp, sclerotic plates; u,
mandible projected slightly back of the ulna; vs, ventral scutellæ.
skull.
There remain only a few indefinite
impressions of the cervical vertebræ. The union of the skull with the
vertebral column is obscured and lost. Impressions of the dorsal
vertebræ are well preserved, and wax molds made from these show
the structure of the dorsal vertebræ surprisingly well. They are long
and cylindrical, with the median portions slightly constricted by a
deep pit on each side of the low neural ridge, which takes the form
observed in Thyrsidium, Molgophis, Phlegethontia, Dolichosoma (fig.
8) and other genera. The vertebræ are strongly amphicœlous and
the notochord was probably persistent. The sides of the vertebræ
are smooth.
The ribs are all intercentral in position; the anterior ones very
broad near the base, recalling the broadly expanded ribs described
by Schwarz (540) for Scincosaurus, Ptyonius, Thyrsidium, and other
genera. Posteriorly the ribs become slender and cylindrical. They are
all rather long and distinctly curved, with probably a cartilaginous
tip.
There is preserved a single element of the right side of the
pectoral girdle. This is, I think, the coracoid, an element which has
hitherto escaped observation among the American Microsauria. It is
long and spatulate at both ends, with the median portion apparently
almost cylindrical, not unlike that described by Credner (181) for the
coracoid of Branchiosaurus, save that the lower end of the
branchiosaurian coracoid is acuminate. In the present form it is
spatulate. Its relations with the other elements of the pectoral girdle
have never been satisfactorily determined.
The fore limbs are both partially preserved. The humerus of the
right side is complete. It is greatly elongated for a microsaurian. The
form of the element is not unlike that of a lizard, with the lower end
of the bone spatulate and endochondrium well developed. Very little
difference can be seen between the form of the arm bones, which
represent the radius and ulna. They are both elongated, with
constricted median portion and expanded truncate ends. The carpus
is unossified and the cartilage has left no trace of the elements.
The right hand has two metacarpals preserved, which are fully
half as long as the radius and ulna. They are separated some little
distance from the ends of these elements, though this may be due
to post-mortem shifting. The carpus may, however, have been broad.
On the left side are preserved portions of the humerus, radius, ulna,
and 3 metacarpals, lying close to the vertebral column. The carpal
space is not so large on the left as on the right. The ventral
armature is well preserved in a narrow patch about an inch in
length. The chevron-shaped rods are quite large, there being 2 of
them in 1 mm.
Measurements.

mm.
Entire length of fossil 98
Length of skull 22
Width across base of skull 28
Long diameter of eye 10.5
Transverse diameter of eye 8
Interorbital space 4
Length of mandible 26
Depth of mandible at coronoidal region 3.5
Depth of dentary 2
Length of long tooth 2
Diameter of long tooth at base .5
Length of preserved portion of vertebral column 64
Length of a centrum 3
Median width of a centrum 1.5
Length of rib 6.5
Width of rib at base .33
Length of coracoid 5
Width of coracoid at anterior end 2.5
Length of carpal space 5
Length of humerus 18
Width of shaft 1
Distal width of humerus 4
Length of radius and ulna 10.5
Length of metacarpal 6
Length of ventral armature preserved 24
Number of rods in length of 5 mm 10
MOODIE PLATE 19

Type specimen of Ctenerpeton alveolatum Cope, from

the Coal Measures of Ohio. × 1.33. Original in U. S.
National Museum.
MOODIE PLATE 20

1. Skull of Erpetosaurus minutus Moodie, from the

Cannelton slates of Pennsylvania. Original in U. S.
National Museum. Enlarged × 3.3.
2. Skull and anterior part of body of Ptyonius pectinatus
Cope, from the Coal Measures of Linton, Ohio.
Original in U. S. National Museum. × 1.
3. Skeleton of Eosauravus copei Williston, from the Coal
Measures of Linton. Ohio. "The oldest known reptile
from North America" and closely related structurally
 to the Microsauria. Original in U. S. National
Museum. × 1.
4. Part of the ventral scutellation and ribs of
Sauropleura digitata Cope, from the Coal Measures
of Linton. Ohio. Original in American Museum of
Natural History. × 1.
 CHAPTER XVIII.

THE MICROSAURIAN FAMILY

NYRANIIDÆ, FROM THE COAL
MEASURES OF OHIO.
Family NYRANIIDÆ Lydekker, 1890.
Lydekker, Cat. Fossil Reptilia and Amphibia, p. 166, 1890.
Skull with the palatines situated near the middle line, internally to the
vomers and pterygoids, and the palatine vacuities small and placed far back.
Vertebræ (Ichthyerpeton) discoidal. Teeth less complex than in the
Anthracosauridæ. A ventral armor present and the entire body covered with
small cycloid imbriated scales.

The type genus of this family was placed by Fritsch (251) with
the Archegosauridæ, although its resemblance to Anthracosaurus
was pointed out; it was subsequently made the type of a family by
Lydekker (393) in 1890, and placed next the Archegosauridæ.
Known from the Coal Measures of Bohemia, Ireland, and Ohio.
Two genera from North America, Ichthyerpeton and
Cercariomorphis, are assigned tentatively to this family, both known
from the Coal Measures (462) of Linton, Ohio, and both with the
body completely scaled. The distinguishing characters are found
chiefly in the shape and arrangement of the scales, the structure,
form, and size of the body, all of which are given full treatment in
the discussion below.

Genus ICHTHYERPETON Huxley, 1866.

 Huxley, Trans. Roy. Irish Acad., XXIV, p. 195, pl. xxiii, fig. 1; Scientific
Memoirs, III, p. 195, pl. 23, fig. 1, 1866.

The genus was founded by Huxley (334) for the reception of the
species Ichthyerpeton bradleyæ from the Kilkenny Coal Measures of
Ireland. The remains of the type specimen represent "the hinder
moiety of the trunk, with the greater part of the tail, of an animal
whose scaly integument and laterally compressed, fin-like tail might
easily lead one to take it for a fish, were not its true position among
higher vertebrata settled at once by the digitate hind limb; while its
alliance with the labyrinthodonts is indicated by the delicate spicular
ossicles, which form a rudimentary dermal shield along the belly."
(Huxley.)

Ichthyerpeton squamosum Moodie.

Moodie, Jour. Geol., XVII, No. 1, p. 69, 1909.
Moodie, Proc. U. S. Nat. Mus., 37, p. 24, 1909.

Type: Specimens Nos. 4476 and 4459, U. S. National Museum.

Locality and horizon: Linton, Ohio, Coal Measures.
The present species is based on well-preserved remains from the
Linton, Ohio, beds. There are two specimens of the species
preserved on blocks of coal and together they represent the greater
part of the length of the animal. The species is located in the genus
Ichthyerpeton, which was founded by Huxley (334, p. 351) on
remains from the Coal Measures of Ireland, on account of the
character of the dermal covering, which consists of small scales such
as Huxley described in the form from Ireland. The specific characters
of this form are the small size of the rounded scales, the attenuated
tail, the apparent absence of limbs, the character of the ventral
scutellation, and the slightly curved condition of the ribs.
It is estimated, from the portions preserved, that the animal
attained a length of not less than 3 feet and its body was long and
slender. It may have had an appearance similar to the modern
caudate genus Siren, though there were doubtless 4 limbs present
instead of 2. The slenderness of the body is at variance with the
condition found in the type species Ichthyerpeton bradleyæ Huxley,
in which the trunk was rather stoutly built. The character of the
anterior portion of the body in the present species can not be
determined and the skull is wanting. There are no evidences of
anterior limbs, although the ventral scutellation preserved would
seem to include the pectoral region. No pectoral shields are
preserved, nor are there any traces of pelvic girdle or limbs.
The preserved portions on one block include nearly the entire tail
and the posterior region of the body, and on the other block the
dorsal region of the body and the anterior portion of the tail, so that
the two specimens supplement each other in an interesting manner.
There are impressions of several vertebræ preserved. They are much
the same in character as Huxley has described for the type species
(I. bradleyæ). They are short and thick and were probably
amphicœlous. There are likewise preserved the remains of rather
slender recurved ribs mingled in with the remains of the ventral
scutellation and distinguished from the elements of the abdominal
shield by their size and curvature. They are, apparently, single-
headed, but the character of their articulation can not be
determined. The ventral scutellation consists of fine continuous rods
arranged in the regular chevron pattern. They do not seem to be
divided into oat-shaped scutes, as is the case with the form
described by Huxley. The ventral rods are closely packed for a
distance of more than 6 inches, but as they are scattered their exact
arrangement can not be determined. They seem to have extended
to the cloacal region, but there are no evidences of the specialized
clasping organs such as Fritsch (251) has described in the ventral
scutellæ of Ophiderpeton. The scales, which are well preserved on
the tail, may have covered the entire body, since there are many
scattered scales in the dorsal region of one of the specimens. They
are slightly oval, tuberculate, and measure scarcely 1 mm. in their
longest diameter. They show but slight evidences of having been
imbricated, though it is likewise possible that they were simply
inclosed within the integument, and somewhat separated from one
another. The most posterior part of the tail preserved seems to
indicate that the tip was attenuated. It was probably flattened from
side to side. We may thus regard Ichthyerpeton squamosum as an
elongate aquatic animal with a long, flattened tail, and since there
were possibly no limbs or very small ones, it would be an animal
highly adapted for life in the water. The present species is of interest
because it represents an additional discovery of the scaled Amphibia
in North America. The species previously known from the Linton,
Ohio, deposits is Cercariomorphus parvisquamis Cope. Dermal scales
have also been observed in specimens of Amphibamus grandiceps
Cope and Micrerpeton caudatum Moodie (462, 478) from the Mazon
Creek, Illinois, beds, and Sir William Dawson (208) described scales
accompanying several forms from the Joggins deposits of western
Nova Scotia.
MOODIE PLATE 21

1. Mandible of Micrerpeton deani Moodie, from

the Linton. Ohio, Coal Measures. Original in
American Museum of Natural History, No.
2934. × 0.6.
2. Portion of the skull of Micrerpeton deani
Moodie, possibly of the same individual as
the mandible. From the Linton, Ohio, Coal
Measures. Original in American Museum of
Natural History, No. 3535 G. × 0.4.
3. Type of Cercariomorphus parvisquamis Cope,
from the Linton, Ohio, Coal Measures.
 Original in American Museum of Natural
History. × 1.
4. An additional specimen of Cercariomorphus
parvisquamis Cope, from the Linton, Ohio,
Coal Measures. Original in American Museum
of Natural History. × 1.
5. Skull of Sauropleura scutellata Newberry. From
the Coal Measures of Ohio. × 1.
6. Tooth of Mastodonsaurus sp. indet. of the
Carboniferous of Kansas. Original in
University of Kansas Museum. × 1.
7. Tooth of Mastodonsaurus giganteus Jaeger,
from the Triassic of Germany. Introduced for
comparison with the tooth from the Kansas
Carboniferous. × 1.

Measurements of the Types of Ichthyerpeton squamosum Moodie.

mm.
Length of animal as estimated from two impressions 3 ft.
Length of longest impression 21 in.
Length of specimen containing tail impression 9 in.
Width of tail impression: Maximum 50 mm.
Minimum 6 mm.
Width of a single scale 1 mm.
Distance from base of tail to tip 125 mm.
Length of specimen as preserved 225 mm.
Width of chevron rod space 30 mm.
Length of rib 25 mm.
8 chevrons in a distance of 3 mm.

Genus CERCARIOMORPHUS Cope.

Cope, Proc. Amer. Phil. Soc., 1885, p. 405.

Type: Cercariomorphus parvisquamis Cope.

 The type specimen of this genus is supplemented by a portion of
the body of another specimen which adds a little to our knowledge
of the animal's form, but nothing as to structure. Cope's original
description is as follows:
"Represented by a fusiform body which terminates in a long, slender,
cylindrical tail, and which is covered with small subquadrate scales
quincuncially arranged. No fins or limbs are preserved, and the form of the
head can not be made out. Probably a portion of the skull is preserved.
There are some scattered bodies in the body portion, which look like deeply
concave vertebræ with the zygapophyses of batrachians. There are some
linear impressions at one point, which resemble the bristle-like rods on many
Stegocephali. They are so few as to be of little importance. The scales are
like those of fishes. There are traces of segmentation in the axis of the long
tail.
"The position of this curious form is quite uncertain. It is quite different
from anything observed hitherto in the American Coal Measures."

Cercariomorphus parvisquamis Cope.

Cope, Proc. Amer. Phil. Soc., 1885, p. 405.
Moodie, Science, n.s., XLI, No. 1056, p. 463, 1915.

Type: Specimen No. 2560, Newberry Collection, American

Museum of Natural History.
Horizon and locality: Discovered by Samuel Huston at the Linton,
Ohio, Coal Mines. (Plate 21, figs. 3, 4; 24, figs. 2, 3.)
The scales (plate 24, fig. 2) in their present condition are entirely
smooth. At a distance of 20 mm. from the base of the tail they are in
20 longitudinal series. At that point the transverse diameter of the
body is 140 mm. The outline contracts rather abruptly to the tail, of
which 66 mm. are preserved. The surface of the tail is obscured by a
thin layer of carbonaceous matter not sufficiently thick to obscure
scales, which are evident at distances of 16 mm., 43 mm., and 52
mm. from the tip. The scales on the tail are smaller than those on
the body and are without markings of any kind. The anterior half of
the body is depressed and distorted, but the remainder is well
preserved and shows a fairly good outline of an apparently limbless
body.
An additional specimen (No. 8683 G, of the Newberry Collection,
American Museum of Natural History) reveals no new facts as to
structure, but serves to show that the body of the animal was long
and slender (plate 21, fig. 4). The portion studied comes
undoubtedly from the middle of the body. No limb elements are
preserved. The scales are somewhat larger, especially toward the
sides of the body, than in the type. The fragment measures 70 mm.
in length by 18 mm. and 26 mm. in width. One of the largest scales
measures 1 mm. in diameter.
Measurements of the Type.

mm.
Length of entire remains 180
Greatest width 22
Greatest width of undisturbed portion 15
Length of an individual scale .75
 CHAPTER XIX.

THE AISTOPODOUS
MICROSAURIAN FAMILY
PTYONIIDÆ, FROM THE COAL
MEASURES OF OHIO.
Family PTYONIIDÆ Cope, 1875.
Cope, Geol. Surv. Ohio, II, pt. II, p. 357, 1875.
Elongate, slender, weak-limbed, aquatic microsaurians. Neural and hæmal
spines of vertebræ elongated, expanded and sculptured. Ventral armature
weakly developed or absent. Skull lanceolate, with long, slender teeth.

Three genera are assigned to this family: Ptyonius,

Œstocephalus, and Thyrsidium. The forms are very closely related,
and when additional material is secured the three genera may be
found to be identical. The species included in this family are:
Ptyonius pectinatus Cope, P. vinchellianus Cope, P. marshii Cope, P.
nummifer Cope, P. serrula Cope, Œstocephalus remex Cope, O.
rectidens Cope, Thyrsidium fasciculare Cope. The species are all
exclusively from the Linton, Ohio, Coal Measures, and most of them
are known from abundant material.

Genus PTYONIUS Cope, 1875.

Cope, Geol. Surv. Ohio, II, pt. II, p. 373, 1875.

Cope designated no species as the type, but we may regard

Ptyonius pectinatus as typical.
 Form elongate, with long tail and lanceolate cranium. Limbs
weak, a posterior pair only discovered. Three clavicular elements;
abdomen protected by packed osseous rods, which are arranged en
chevron, the angle directed forward. Neural and hæmal spines of
caudal vertebræ expanded and fan-like. Ribs well developed. The
various species vary in length from 3 to 10 inches. They are the
most abundant amphibian in the Linton beds. The present genus
resembles Lepterpeton Huxley (334), of the Kilkenny, Ireland, Coal
Measures. But that genus possesses divided abdominal rods, or "oat-
shaped scales," and the form of the cranium and proportions of the
body are different.
The genus is closely related to, possibly identical with,
Œstocephalus, but additional material will be required to settle this
point.
Cope (123) gives the following key for the separation of the 5
species:
x. Abdominal rods coarser, not more than 10 in
5 mm.
Median pectoral shield discoid, radiate-
ridged; muzzle short P. nummifer
Median pectoral shield oval, pitted and
ridged P. marshii
xx. Abdominal rods hair-like, 15 or more in 5
mm.
Median pectoral shield with radii from the
center, the principal forming a cross; form
wider P. vinchellianus
Middle pectoral with pits at the center
P. pectinatus
and few or no radii; form narrow
Middle pectoral shield narrow, closely
reticulate medially, and radiate towards the
circumference; size half that of last P. serrula

Ptyonius pectinatus Cope.

 Cope, Proc. Acad. Nat. Sci., 1868, p. 216.
Cope, Trans. Am. Phil. Soc., XIV, p. 20, 1869.
Cope, Trans. Am. Phil. Soc., XV, p. 266, 1874.
Cope, Geol. Surv. Ohio, II, pt. II, p. 377, pl. xxvii, fig. 7; xxviii, figs. 2, 3,
6; pl. xxix, fig. 2; pl. xxx, fig. 2; pl. xxxv, figs. 1-3; pl. xli, fig. 1, 1875.
Moodie, Proc. U. S. Nat. Mus., 37, p. 24, pl. 8, fig. 3, 1909.
Schwarz, Beiträge zur Paleontologie und Geologie Osterreich-Ungarns und
des Orients, Bd. XXI, p. 83, figs. 23, 24, 26, 1908.

Type: It is impossible to determine which one of the specimens is

the type. There are numerous representatives of the species, as
follows: Nos. 140, 1096 G, 8345 G, 8555 G, 1089 G, 2, 132, 133, no
number, 1094 G, 8545 G, 8677 G, 1159 G, 105, no number, 1091 G,
7a, 1092 G, 1093 G, 1095 G, 153, and others unnumbered in the
American Museum of Natural History; in the U. S. National Museum
are the following: Nos. 4458, 4463, 4464, 4514. (Plate 20, fig. 2.)
Horizon and locality: Linton, Ohio, Coal Measures.
The most abundant species of the Linton Coal Measures. There
are over three dozen specimens preserved in the Newberry
collection. The species is a clearly marked one, as a rule, though
there is great variation in the size of the body and the form of the
vertebræ. Though there are several apparently complete skulls
preserved in the collection, it is impossible to make out the
morphology of the elements on account of the amount of crushing
to which the skulls have been subjected.
The head is lancet-shaped, and the muzzle very elongate,
slender, and acute at the extremity. The head is in fact a miniature
of an ichthyosaur cranium. (Plate 20, fig. 2.) The orbits are large and
posterior to the median line. The anterior portion of the skull is
narrow, posteriorly truncate, and the mandibular angle is projecting.
The posterior portion of the mandible is sculptured. Possibly the
entire cranium was also, and this has been lost; in fact, this
sculpturing is indicated in one or two specimens. The teeth are
conical and sharp, longitudinally
striate, and anisodont. There
seems to be evidence of
palatine or pterygoid teeth,
though this needs confirmation.
The pectoral plates are well
preserved, with the interclavicle
a narrow oval, with anterior and
posterior prolongations. In one
specimen it is sculptured. The Fig. 30. Restoration of Ptyonius. × 1.
clavicles are narrow and slightly
sculptured. The abdominal
scutellæ are bristle-like.
The vertebræ are short, with expanded neural and hæmal
spines. The expanded condition of the neural spines begins over the
thoracic region, where they are low. They become well developed in
the posterior dorsal region. The caudal fan-shaped spines are larger.
The dilated portions form equilateral triangles which stand on
moderately short pedicels. They are weakly ridged, and each ridge is
prolonged into a narrow acute tooth beyond the margin, 11 of which
may be counted on one of the best preserved spines. The
longitudinal striæ are terminated near the pedicel by two others
which cross obliquely from each side, and, meeting, present the
appearance of the margin of a cup sculptured in relief, from which
the striæ arise. Pedicels smooth. The spines are in contact at their
angles, thus forming a continuous line. In a typical specimen there
are 6 in half an inch, in another 7, and in a third 8. The ribs are well-
developed and slender.
No traces of fore limbs have been detected in the numerous
specimens, but elements of hind limbs are preserved. In one of
these the femur is a small bone, contracted at the middle. The form
of the body is snake-like.
There were probably from 75 to 100 vertebræ in a single animal.
The form may be well compared to the modern Amphiuma so far as