Performance EvaluationPerformance EvaluationPerformance Evaluation Performance Evaluation of Buried Pipe Installationof Buried Pipe Installation

Preliminary ResultsPreliminary ResultsPreliminary ResultsPreliminary Results

Michele BarbatoMichele Barbato

Assistant ProfessorAssistant ProfessorDept. of Civil and Environmental EngineeringDept. of Civil and Environmental Engineeringp f g gp f g g

Louisiana State University, Baton RougeLouisiana State University, Baton Rouge

2009 Louisiana Transportation Conference2009 Louisiana Transportation ConferenceBaton Rouge, February 10, 2009Baton Rouge, February 10, 2009

OutlineOutline

• Problem Statement• Methodologygy• Finite Element Simplified Cross‐Section Model• Considered Modeling and PerformanceConsidered Modeling and Performance Parameters 

• Performance Sensitivity StudyPerformance Sensitivity Study• Validation of Installation Requirements for Realistic Implementation SituationsRealistic Implementation Situations

• Main results and preliminary recommendationsrecommendations

• Ongoing and Future Studies

Problem StatementProblem StatementMotivations: • Buried pipe performance and reliability is not

fully addressed by current specifications • New competing materials (PVC, HDP)

Objectives:Determine effects of geometric and mechanicalDetermine effects of geometric and mechanical parameters on soil-structure interaction developed in buried pipe installationin buried pipe installation

Scope:R i l b i f i i f i iRational basis for revision of current provisions on buried pipe installation in Louisiana and USA.

MethodologyMethodology• The soil-structure interaction between pipe

and soil is studied using the finite element method.

• Performance measures are defined fromPerformance measures are defined from finite element response quantities.

P i di f d diff• Parametric studies are performed on different geometries, materials, configurations, natural

il i isoil situations

• Analysis is focused on performance y psensitivity to design variables

Simplified CrossSimplified Cross--Section ModelSection ModelParameters considered

• Pipe material and geometry

Parameters considered in this study:

P

PavementHp

• Bedding properties• Fill properties and compaction• Surrounding soil properties

CH

Pavement BaseHb,pave

g p p• Importance of facility

CoverH

W The load P is obtained from AASHTO d i l d id i i t

BackfillD

t design load, considering an impact factor c = 1.5

Other data are obtained from AASHTO BeddingB material properties for pipe design and

are provided by LTRC in order to represent typical soil conditions for Louisiana

Bf

Louisiana.

CrossCross--Section Properties of PipesSection Properties of PipesCritical approximations:

D tCritical approximations:• Linear elastic materials• Linear geometry

Cross-section properties considered in this study:

• Simplified boundary conditions

• Pipe material: • Concrete: E = 2,900 ksi• Steel: E = 29,000 ksi• PVC: E = 400ksi• HDP: E = 140ksi

• Pipe diameter: D = 18”, 36”, 42”, 48”, 60”p , , , ,• Pipe diameter/thickness ratio, D/t:

• Concrete: D/t = 10, 20, 30 (referred to as RC10, RC20, RC30)• Steel: D/t = 50 100 200 (referred to as S50 S100 S200)Steel: D/t 50, 100, 200 (referred to as S50, S100, S200)• PVC: D/t = 10, 20 (referred to as PVC10, PVC20)• HDP: D/t = 10, 20 (referred to as HDP10, HDP20)

The design parameters considered are:Design Parameters for Sensitivity StudyDesign Parameters for Sensitivity StudyThe design parameters considered are:

1. Dip depth at the surface, , with design limit lim = 0.1in

2 Maximum stress at the pipe ring 2. Maximum stress at the pipe ring, max

A third design parameter (ring deflection) is not considered h i i it i t i t hi h t ti llherein, since it imposes constraints which are automatically satisfied when the requirements on are satisfied.

Two different safety coefficients are introduced and used toTwo different safety coefficients are introduced and used to evaluate design conditions:

1. SC = lim/1. SC lim/

2. SC = lim/ max

IT IS FOUND THAT SC IS USUALLY SMALLER THAN SCDeformations control the pipe design and installation

Surrounding soil stiffness (E )

Sensitivity Study (1)Sensitivity Study (1)Surrounding soil stiffness (Esoil)

1.2

1

1.1

RC10RC20RC30HDP10The natural soil stiffness is crucial

0.8

0.9SC HDP20

PVC10PVC20S50S100S200

Different minimum requirements are needed between stiff and yielding soil

0.6

0.7

0 1 2 3 4 5 6 7 8 9 10 11

S200stiff and yielding soil

Stiff and yielding soils can be specified in terms of initial stiffness choosing a al e in the inter al 5 10 ksi0 1 2 3 4 5 6 7 8 9 10 11

Esoil (ksi)

Model properties: D = 60”; H = 12”; W = 18”; B = 6”

initial stiffness, choosing a value in the interval 5-10 ksi

Calibration of this value requires more advanced Efill = 30 ksi; Eb_pave = 45 ksi

Pavement type: Asphalt, h = 4”, hbase = 10”Sensitivity variable: soil stiffness Esoil = 1-5-10 ksi

nonlinear FE analysis studies, which will allow also to study the effects of soil strength

Trench width (W)

Sensitivity Study (2)Sensitivity Study (2)Trench width (W)

1.1

1

1.05

RC10RC20RC30HDP10

Efill/Esoil = 6

0.9

0.95SC HDP20

PVC10PVC20S50S100

0.8

0.85

S200

12 18 24 30 36 42 48 54 60 66

W (in)

Model properties: D = 60”; H = 12”; B = 6”Efill = 30 ksi; Eb = 45 ksi; E il = 5 ksi WEfill 30 ksi; Eb_pave 45 ksi; Esoil 5 ksi

Pavement type: Asphalt, h = 4 ”, hbase = 10”Sensitivity variable: trench width W = 18-30-60 in

D

Trench width (W)Sensitivity Study (3)Sensitivity Study (3)

0.9

0.95

Trench width (W)

0.8

0.85RC10

RC20

RC30

HDP10

Efill/Esoil = 30

0 7

0.75

HDP20

PVC10

PVC20

S50

S100

S200

SC

The trench width has a moderate effect0.65

0.7 The trench width has a moderate effect on the displacement safety coefficient SC

0.612 18 24 30 36 42 48 54 60 66

Model properties: D = 60”; H = 12”; B = 6”W (in)

W

The effect depends on the ratio Efill/Esoil and becomes significant for Efill/Esoil ≥ 10

Efill = 30 ksi; Eb_pave = 45 ksi; Esoil = 1 ksiPavement type: Asphalt, h = 4 ”, hbase = 10”Sensitivity variable: trench width W = 18-30-60 in

DFor W > D/2 (i.e., total width of the trench > 2D), this effect is always very small

Co er height (H)

Sensitivity Study (4)Sensitivity Study (4)

1 05

1.1

Cover height (H)

1

1.05

SCd

RC10RC20RC30HDP10

The cover height (for H ≤ D) has a significant positive effect

0.9

0.95

S

HDP20PVC10PVC20S50S100S200

a significant positive effect on the displacement safety coefficient SC

This effect increases for

0.8

0.85

This effect increases for increasing flexibility of the pipe

6 12 18 24 30 36 42H (in)

Model properties: D = 60”; W = 18”; B = 6” H

Further study is required to evaluate the cover height effects for H ≥ D

p p ; ;Efill = 30 ksi; Eb_pave = 45 ksi; Esoil = 5 ksi

Pavement type: Asphalt, h = 4”, hbase = 10”Sensitivity variable: Cover height H = 12-24-36 in

DNonlinear FE analysis is required to model soil arch effects

Bedding thickness (B)B ddi thi k (B)

Sensitivity Study (5)Sensitivity Study (5)g ( )

1.05

1.1

Ebedding/Esoil = 1: uniform material

Bedding thickness (B)

1

C

RC10RC20RC30HDP10HDP20

0.9

0.95

SC

HDP20PVC10PVC20S50S100S200

0.8

0.85

3 6 9 12 15 18 21 24 27

B (in)

Model properties: D = 60”; H = 12”; W = 18”;Efill = 30 ksi; Eb_pave = 45 ksi; E 5 ksi E 5 ksi DEsoil = 5 ksi; Ebedding = 5 ksi

Pavement type: Asphalt, h = 4”, hbase = 10”Sensitivity variable: Bedding depth B = 6-12-24 in

DB

Bedding thickness (B)Sensitivity Study (6)Sensitivity Study (6)

1.1

1.15

Ebedding/Esoil = 2 (6in)Ebedding/Esoil = 6 (below 6in)

Bedding thickness (B)

1

1.05RC10

RC20

RC30

HDP10

HDP20

PVC10

0.9

0.95 PVC20

S50

S100

S200SC

For uniform distribution of stresses, a layer of 6” f l b ddi t i l i i d ( d

0.8

0.85

3 6 9 12 15 18 21 24 27

B (i )

6” of loose bedding material is required (and sufficient) immediately below the pipe

Model properties: D = 60”; H = 12”; W = 18”;Efill = 30 ksi; Eb_pave = 45 ksi; E 5 ksi E 10 ksi

B (in)In presence of yielding soil, the bedding thickness can be increased and the additional

material (at depth larger than 6” below the pipe)Esoil = 5 ksi; Ebedding = 10 ksi

Pavement type: Asphalt, h = 4”, hbase = 10”Sensitivity variable: Bedding depth B = 6-12-24 in

DB

material (at depth larger than 6 below the pipe) should be compacted at the same level of the

backfill material

Sensitivity Study (7)Sensitivity Study (7)Backfill soil stiffness (Efill)

1 2

1.3

1

1.1

1.2

RC10RC20RC30

0.7

0.8

0.9

SC

HDP10HDP20PVC10PVC20S50S100

0.4

0.5

0.6

S100S200

The stiffness of the backfill material is 5 10 15 20 25 30 35 40 45 50

Efill (ksi)

Model properties: D = 60”; H = 12”; W = 18”; B = 6”E = 45 ksi; E = 5 ksi

very important

Adequate compaction and material quality mustEb_pave = 45 ksi; Esoil = 5 ksiPavement type: Asphalt, h = 4”, hbase = 10”Sensitivity variable: Backfill soil stiffness Efill = 10-30-45 ksi

Adequate compaction and material quality must be ensured everywhere and required in the

specifications

Road pavement typeRoad Pavement Type

Sensitivity Study (8)Sensitivity Study (8)

2

2.2

2.4Road Pavement Type

1.6

1.8

2

SC

RC10RC20RC30HDP10HDP20

1.2

1.4

S HDP20PVC10PVC20S50S100S200

0.8

1

Asphalt 4in Concrete 8in Asphalt 10inThe road pavement type is a crucial factor in the sp a t Co c ete 8Asphalt 10in

Model properties: D = 60”; H = 12”; W = 18”; B = 6”Efill = 30 ksi; Eb_pave = 45 ksi; Esoil = 5 ksi

Sensitivity variable: Road pavement type

p ypperformance of buried pipes

In presence of rigid pavement (concrete)y p ypAsphalt, h = 4”, hbase = 10”Asphalt, h = 10”, hbase = 10”Concrete, h = 8”, hbase = 10”

In presence of rigid pavement (concrete), any pipe configuration performs satisfactorily

if the pavement is not damaged

Diameter (D)Diameter (D)

Sensitivity Study (9)Sensitivity Study (9)

1.5

1.6

1.7

Diameter (D)

1.3

1.4

1.5

SC

RC10RC20RC30HDP10HDP20PVC10

1

1.1

1.2

S PVC10PVC20S50S100S200

0.8

0.9

18 24 30 36 42 48 54 60

D (i )The displacement at the road surface increases

D (in)

Model properties: H = 12”; W = 18”; B = 6”E = 30 ksi; E = 45 ksi; E = 5 ksi

palmost linearly for increasing pipe diameters

Minimum requirements for satisfactoryEfill = 30 ksi; Eb_pave = 45 ksi; Esoil = 5 ksiPavement type: Asphalt, h = 4”, hbase = 10”Sensitivity variable: Pipe diameter D = 18-36-42-48-60 in

Minimum requirements for satisfactory performance of buried pipes need to be

functions of the pipe diameter

The installation requirements are verified for all combinations of

Validation of Installation RequirementsValidation of Installation RequirementsThe installation requirements are verified for all combinations of the following conditions: • Road pavement type: p yp

• Asphalt pavement with h = 4in;• Asphalt pavement with h = 10in;

Concrete pa ement ith h 8in• Concrete pavement with h = 8in.• Surrounding soil properties:

• Moderately stiff soil (E = 5ksi);Moderately stiff soil (E 5ksi); • Yielding soil (E = 1ksi);

• Pipe diameter:p• D = 18in• D = 60in

Pi t i l C t St l PVC HDP• Pipe material: Concrete, Steel, PVC, HDP• Importance of facility: Low ADT, Moderate ADT, High ADT

Results for roads with Low ADT

Current Performance (1)Current Performance (1)Results for roads with Low ADT

1.5

1.6Asphalt pavement with h = 4in

1.2

1.3

1.4

P f i t ti f t !

0.9

1

1.1

SC

D = 18'' (MSS)D = 18'' (YS)D = 60'' (MSS)D = 60'' (YS)

Performance is not satisfactory!

Moderately stiff soil (MSS):Esoil = 5 ksi

Yielding soil (YS):E il = 1 ksi

0.6

0.7

0.8

ete

eel

VC DP

Modeling assumptions:The road pavement type is asphalt with h = 4” and hbase = 10”.

Esoil 1 ksi

Con

cre

Ste

PV HD

p yp p baseThe pavement base material has stiffness Eb_pave = 30 ksi.The backfill material has stiffness Efill = 30 ksi. The bedding material has stiffness Ebedding = 10 ksi.

Results for roads with High ADT

Current Performance (2)Current Performance (2)Results for roads with High ADT

4.4

Concrete pavement with h = 8in

3.6

4

V hi h f t f t !

2.8

3.2SC

D = 18'' (MSS)D = 18'' (YS)D = 60'' (MSS)D = 60'' (YS)

Very high safety factors!

2

2.4

ete

eel

VC DP

Moderately stiff soil (MSS):Esoil = 5 ksi

Yielding soil (YS):E = 1 ksi

Con

cre

Ste

PV HD

Modeling assumptions:The road pavement type is concrete with h = 8” and hbase = 10”.

Esoil = 1 ksi

p yp baseThe pavement base material has stiffness Eb_pave = 45 ksi.The backfill material has stiffness Efill = 45 ksi. The bedding material has stiffness Ebedding = 10 ksi.

Possible Improvements (1)Possible Improvements (1)

1.3

Asphalt pavement with h = 4in

15”

1.1

1.2

As isH =

D/4

= 1

2 =

30”

= 30

”

60”

0.9

1Increasedcover

D = 60”

H

H =

D/2

H =

D/2

=

H =

D =

6

H =

12”

H =

24”

H =

24”

8”

0.7

0.8

rete

teel

PVC

HD

P

D = 60Yielding soilFlexible pavementLow ADT

H H H

H =

48

Recommendations:1. Increase minimum cover to

• Concrete: H ≥ max(D/4 12”)

Con

cr S P H

Modeling assumptions:The road pavement type is asphalt with h = 4” and hbase = 10”.

• Concrete: H ≥ max(D/4, 12 )• Steel and PVC: H ≥ max(D/2, 24”)• HDP: H ≥ max(D, 48”)

p yp p baseThe pavement base material has stiffness Eb_pave = 45 ksi.The backfill material has stiffness Efill = 45 ksi. The bedding material has stiffness Ebedding = 10 ksi.

2. Require high quality and high compaction of fill material(Efill = 45 ksi)

Possible Improvements (2)Possible Improvements (2)

Concrete pavement with h = 8inResults for roads with Low ADT

2.5

2.2

2.3

2.4

In presence of rigid road pavement (concrete)

1.9

2

2.1

D = 60”

In presence of rigid road pavement (concrete), the safety coefficient are very high also for the

worst case scenario

1.6

1.7

1.8 Yielding soilFlexible pavementLow ADT

The minimum dimensional requirements can be relaxed

A more accurate study based on nonlinear FE1.5

Con

cret

e

Stee

l

PVC

HD

P

A more accurate study based on nonlinear FE analysis is required to calibrate the relaxed

requirements

Based on the current results, it is safe to keep the same requirements used for low ADT

also for moderate and high ADT

1 P f i t f i d fl ti (di d th)

MainMain ResultsResults1. Performance in terms of maximum deflection (dip depth)

at the surface is often the most demanding requirement for buried pipe installation

2. The performance of the soil-structure system is crucially dependent on the road pavement type. For rigid pavement (concrete) pipe deformation is generally very small(concrete), pipe deformation is generally very small

3. Natural soil stiffness, backfill stiffness, cover height are very important

4. Bedding thickness and trench width have smaller effects on deformation. These effects depend on the ratio between the bedding and fill material stiffness and the natural soilthe bedding and fill material stiffness and the natural soil stiffness.

5. Pipe diameter, D, significantly affects surface p g ydisplacement. For D ≥ 48in and flexible pavement (asphalt), cover height should be function of D

Preliminary RecommendationsPreliminary Recommendations

1. For rigid pavement (concrete), minimum requirements can be relaxed. Continuity of the pavement is crucial

2. For D ≥ 48in and flexible pavement, the minimum requirement on minimum cover in presence of yieldingrequirement on minimum cover in presence of yielding soil should be function of the pipe diameter, e.g.:• Concrete: H ≥ max(D/4, 12”)• Steel and PVC: H ≥ max(D/2, 24”)• HDP: H ≥ max(D, 48”)

3. For uniform distribution of stresses, a layer of 6” of loose bedding material is required immediately below the pipe. Additional bedding thickness should be compacted as the backfill material

Ongoing StudiesOngoing StudiesDifferent linear and non linear finite element models and programs are currently being

d t t dused to study:1. effects of direct modeling of natural soil

surrounding the pipe trench

2. finite element model mesh sensitivity

3. three-dimensional effects

4 effects of different modeling options for the4. effects of different modeling options for the applied loads

Mesh Expansion for D = 18" (HDP)

Direct Modeling of Natural Soil Direct Modeling of Natural Soil

-0.02

Mesh Expansion for D 18 (HDP)

-0.05

-0.04

-0.03

(in)

-0.08

-0.07

-0.06

form

atio

n (

-0.11

-0.1

-0.09Def

02

4

0

5

10

-0.12

46

8

10

15

Horizontal (ft)Vertical (ft)

Stress in D18 RC Pipe for Varying Model Thickness

ThreeThree--Dimensional EffectsDimensional Effects

14

16

Stress in D18 RC Pipe for Varying Model Thickness

0 Vertical 0 Horizontal

0 Vertical 11.25 Horizontal

8 Vertical 0 Horizontal

10

12

)

8 Vertical 0 Horizontal

8 Vertical 11.25 Horizontal

8

10

omen

t (k-

in)

4

6Mo

0

2

0 1 2 3 4 5 6 7 8 9

Mesh Thickness (ft)

Suggested Model for Nonlinear Finite Suggested Model for Nonlinear Finite Element AnalysisElement Analysis

Nonlinear staged Finite Element analysis will be

Current and Future ResearchCurrent and Future ResearchNonlinear staged Finite Element analysis will be used for:

1 C lib ti f th il tiff l t di ti i h1. Calibration of the soil stiffness value to distinguish between stiff and yielding soil

2. Study of effects of soil strength

3. Study of effects of material (cracking, fracture, plasticization) and geometric (buckling) nonlinearities

4. Study of effects of installation procedures

5. Calibration of other important parameters for p pcases in which the performance estimated from linear analysis is not satisfactory

AcknowledgementsAcknowledgementsLA D t t f T t ti d D l t• LA Department of Transportation and Development

• Louisiana Transportation Research Center p(project 08-6GT)

• LA DOTD PIPE Committee• LA DOTD PIPE Committee

• Dr. Zhongjie “Doc” Zhang, Pavement and GeotechR h Ad i i t tResearch Administrator

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect those of thematerial are those of the authors and do not necessarily reflect those of the sponsors.

Thank you!Thank you!Thank you!Thank you!Questions?Questions?