Soil Friction Along HDD Drilling

8/10/2019 Soil Friction Along HDD Drilling

1/46

Soil friction modeling in HDDs under operational conditions new insights in existing theories 1

PTC Conference, Hannover 2012

Soil friction modeling in HDD's under operational conditions - new

insights in existing theories

1 Problem definition

Since the development of the Horizontal Directional Drilling Technique a lot of research

has been performed with regard to the frictional effects during installation (pull back

operation) and more general on soil behavior (borehole stability), ref. [1], [2], [3] and [4].

Investigation and theoretical modeling on pipeline behavior in HDDs during operational

conditions has not been found. For pipelines under elevated temperature axial

displacements (expansion) in longer HDD configuration expansion increases quadratic-

ally and thus loads on upper bends of the HDD configuration also increase. In specificcases f.i. for longer smaller diameter HDD configurations this may lead to stresses

above the acceptable limits. To mitigate these stresses either the pipeline routing can

be changed by incorporating expansion loops or other expansion measures such as

placement of expansion cushions behind the bends.

The incorporation of expansion loops causes additional material and construction costs

and will often (especially in densely populated areas) lead to spatial problems. So it is

an engineersmission to design the HDD crossing as compact as possible.

2 Analysis approach

The soil friction along pipelines in general is depending on the type of surrounding soil

(and related soil properties). Typical for HDD configurations is that next to this the

curved shape of the pipeline section induces reaction forces which can be taken into

account. Another aspect which influences the reaction forces and thus the soil friction is

the deadweight of the pipe.

In this paper the models describing expansion and friction along HDDs are explained.

Based on these models a proposal is done how to deal with soil friction modeling along

HDD sections in the day to day engineering practice. After the definition of this

approach quantitative assessments are done on the schematized friction model itself

and its impact on expansion along the HDD section.

The schematized model is developed taking into account friction aspects acting on

subsoil pipelines as developed within the framework of the Dutch pipeline code since its

first issue in 1990 and its predecessors [5] and is a recommendation to amend the

current design rules.

7th Pipeline Technology Conference 2012


2/46


3/46


4/46



soil type. For pipelines in HDD crossings (long after installation) this is the case

because there is a relative constant shell of slurry-soil mix present around the pipe.

4 Soil friction model

The determination of the soil friction acting on the pipe in a HDD under elevatedtemperature loads is a crucial part of the assessment of expansion on the pipe bends.

As discussed earlier for the soil friction a linear elastic behavior is assumed (see figure

5) at which the maximum friction (Wmax) is reached at small relative displacements.

Figure 4: linear elastic behavior soil friction

Main aspects influencing the soil friction Wmaxare in general:

intergranular pressure around pipe;

adhesion between pipe and soil;

angle of friction between pipe and soil (dependent of soil friction angle and pipe

wall roughness).

To determine Wmax the following basic relationship is available for pipelines in open

trench:

| |(4)With:

soil friction along the pipe N/m outside pipe circumference m ratio horizontal/vertical intergranular pressure, in case of

neutral horizontal soil pressure equals -


5/46


6/46



Figure 5: soil friction acting on pipes in open trench compared to pipes in an HDD

4.1 Friction on pipelines installed by HDD technique

The friction on HDD installed pipelines is determined by:

the vertical soil load on the pipe (vertical intergranular pressure);

friction on interface plane pipe-bentonite;

friction on interface plane bentonite-soil;

additional soil friction due to the curvature of the pipe in the elastic bends.

4.1.1 Arching in soils and determination of intergranular pressure in HDDs

To determine the value of the vertical intergranular pressure acting on the HDD section

a parallel is assumed with the way the vertical soil load acts on a pipe installed in open

trench.

The maximal vertical soil load acting on a pipe is defined as:

(5)With:

maximal vertical soil load acting on a pipe N/mm effective soil load based on intergranular pressure N/mm outer diameter pipe mm

Soil Fricion

(Wmax)

Pipeline in

open trench

Intergranularsoil pressure

AdhesionSoil friction

angle and pipewall roughness

HDD-technique

Intergranularsoil pressure

Arching

Friction on pipe

- drilling fluidinterface

Friction on

drilling fluid -soil interface

Friction by soil

reaction incurved section


7/46



The bentonite shell round the pipe will stiffen up after installation but will act as a highly

compressible layer. As a result of the combination of this compressible layer and the

thickness of the soil column above the pipe, arching in the soil will occur and the vertical

soil load can be reduced (seefigure 6 ).

Figure 6: arching around borehole

Based on the theory of Terzaghi it is considered that arching occurs when the thickness

of the soil mass extending above the pipe is larger than 4 times the width of the soil

column in shear (2B1). The width B1is defined as:

( ) (6)With:

is half the width of the soil column in shear m

is the outside diameter of pipeline m is the internal angle of friction is the radius of the borehole m

A reduction of the vertical soil load due to arching results in a lower vertical intergranular

pressure () on the pipe and thus reduces the overall friction.


8/46


9/46



Based onfigure 7 can be noted that for soil columns extending above the pipe smaller

than 8B1 the actual vertical intergranular pressure is taken into account. For larger soil

columns arching will occur and the vertical intergranular pressure then is assumed to be

zero.

Figure 8: schematization vertical intergranular pressure in multiple layer soil

For pipes in multiple layer soils (see figure 8) the approach for determination of the

vertical intergranular pressure is similar. It should be noted that according to the theory

arching will only occur in the sand layer at a depth of H8B1below the layer separation.

In the new approach arching is assumed to be effective as of the layer separation.

Immediately taking into account arching over this layer and thus neglecting the built up

of vertical intergranular pressure over the depth H


10/46



Figure 9: additional friction due to elastic curvature in bends

The distribution of the soil reaction forces is based on the theory for beams on elastic

foundations by Htenyi. The soil reaction forces will occur at the ends of each bend and

induce additional frictional effect acting on the pipe.

The additional normal force in the elastic bend as a result of this torque can be

calculated by:

(7)

With:

is the additional normal force for one bend in the borehole N is the maximum soil reaction near the end of the bend:

N/m

is the vertical modulus of sub grade reaction N/m

is the maximum displacement mis the pipe-soil stiffness characteristic

m-1

is the bending stiffness of the pipe Nm is the outside diameter m is the radius of the bend m

is the frictional coefficient between the pipe and the boreholewall

-


11/46



4.1.3 Proposal new friction model along HDDs

Parallel to the friction formula for pipelines in open trenches and based on the

elaborations above a slightly adjusted friction formula for HDD techniques can now beintroduced:

| |(8)Formula 8 represents the new approach to model friction along HDD sections.

Compared with basic friction formula (4) two new parameters are introduced WT3band

f2:

WT3b is introduced to take into account the friction as a result of the curvature of

the pipe in the elastic bends.

f2 replaces the adhesion component and takes into account the friction between

pipe and drilling fluid (bentonite).

The other parameters are identical with formula 4 but might have a different value

because of the application on HDD sections. In the paragraphs below the components

of formula (8) will be further explained. Friction will take place at two interfaces: the pipe

drilling fluid interface and the pipe-borehole wall interface (seefigure 10)


12/46



Figure 10: friction interfaces

4.1.3 Frictional effect due to pipe-drilling fluid interface

After installation of the pipe in the borehole the drilling fluid layer (bentonite) will stiffen.

The adhesion at the pipe-bentonite friction plane will as a result develop to a certain

level. Under open trench conditions (see formula 6) the adhesion contribution is 0,6a

and only applicable for cohesive soils (f.i. clay and peat); for granular soils there is no

adhesion.

Taking into account a pipe section installed in a borehole and surrounded with

bentonite, the adhesion is set equal to the dynamic friction of the bentonite. This value f2

is irrespective of the stiffness of the bentonite. Would a realistic adhesion for stiffened

bentonite (a=10 kN/m2 or 25 kN/m

2) been taken into account the calculated friction is

significantly higher.

The friction distribution at the pipe-bentonite friction plane (see figure 10) is mainly

determined by the shear properties of the bentonite. Under shear loading (friction) thebentonite shows snap back behavior (seefigure 11). At a certain level of displacement

the shear stress reaches a yield value (the static shear stress y,static). If the

displacement grows further the shear stress drops and a dynamic shear stress

(y,dynamic) value will act on the pipe. This value is significantly smaller and assumed

equal to the dynamic friction of bentonite: f2=50 N/m2.


13/46



Figure 11: schematized shear stress behavior of bentonite shell

Under operational conditions at elevated temperatures (especially for cyclic loads) along

the HDD section the movements at the pipe-bentonite friction plane are relatively big.

The static friction (adhesion) value will not able to withstand this displacements and the

pipe slips. Thus the dynamic friction will contribute and is to be used in formula (8).

4.1.4 Frictional effect due to pipe-borehole wall interface

Parallel to the pipe-drilling fluid interface a friction distribution at the pipe-borehole

interface must be taken into account (seefigure 12).


14/46



Figure 12: friction distribution at drilling fluid interfaces

The generated friction at this plane depends in case of no arching on the vertical

intergranular pressure (vertical soil load Qn) and is independent of arching on the

buoyancy effects acting on the pipe in the borehole. The friction value at this plane is

based on these vertical loads and the friction angle between the pipe and the borehole

wall. Results on direct shear tests show that at low levels of bentonite mixed with sand

the internal frictional angle (and thus friction value) drops significantly. I.e. for 100%

sand mixture =22 and for a 10% sand-bentonite mixture =12. Since no empirical

values of other types of mixture are available and the internal friction angle seems to be

relatively independent of the soil surrounding the bentonite shell this value is assumed

to be generally applicable for pipes in HDD sections.

For the soil friction this results in a component of (tan=tan 12=) 0,2 which can be

taken into account.

4.2 Assessment of new approach

Based on formula (8) and the approach and assumptions above for determining the

overall friction a break down for each component contributing in the friction along 3

typical HDD configuration is assessed in this paragraph. By this the effect of each

friction component can be evaluated per typical.


15/46


16/46



h=1,5 m is calculated based on the basic formula 4 (taking into account the full

intergranular pressure).

For the calculation of the frictional effect of buoyancy a volumetric weight of 1100 kg/m3

is used for the bentonite slurry. The deadweight of the transported medium is neglected

since a gas pipeline is considered.

To calculate the friction component as a result of the soil reaction in the elastic bends

the additional normal force for one bend (T3b) is calculated and divided by the arc length

of the considered bend.

To evaluate the separate friction components along the complete HDD configuration the

friction components acting on each part (section) of the HDD are summated and

presented in the paragraphs below. The four typical borehole sections to be taken into

account are:

A. upper bends (based on unconsolidated vertical soil load with pipe installed in

open trench);

B. tipping point for the occurrence of arching at reference level H=8B1;

C. elastic bends;

D. floor pipe (straight pipe).


17/46



4.2.1 Friction assessment DN100 typical

The calculated friction contribution along the total HDD section for the DN100 typical is

graphically presented infigure 13.

Figure 13: graphical presentation friction along DN100 HDD

A breakdown of the friction per borehole section of the HDD is presented in

figure 14.The contribution of each friction component is also indicated as percentage of

the total calculated friction.

0

200

400

600

800

1000

1200

1400

1600

1800

0 50 100 150 200

Friction

in[N/m]

Friction in borehole

Friction

-6

-5

-4

-3

-2

-1

0

0 50 100 150 200

Depth[m-surface

level]

Length borehole path [m]

HDD DN100


18/46



Figure 14: DN100, soil friction component per borehole section

Based on the friction assessment for the DN100 typical the following is noted:

Arching occurs along the entire HDD except at the upper bend as a result of theconstruction split field pipeHDD.

In the elastic bend s of the HDD arching occurs and the friction for this section ismainly (app. 88%) determined by the friction induced by soil reaction supportingthe bend. The absolute amount of friction at section C however is significantlylower than at sections A and B.

At the floor pipe section (section D) the total friction is determined by theadhesion between the pipe and the drilling fluid and the buoyancy effects of thepipe.

The friction in the elastic bends is approximately 8 times higher than the friction

at the floor pipe (seefigure 13).

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

A: Upper bend B: Tipping point

arching

C. Elastic curved

bends

D: Floor pipe

Distribution of friction component per section in the HDD section

Top soil load component Tipping point arching Drilling fluid component

Buyoancy effect Soil reaction curved bends


19/46



4.2.2 Friction assessment DN600 typical

The calculated friction contribution along the total HDD section for the DN600 typical is

graphically presented infigure 15.

Figure 15: graphical presentation friction along DN600 HDD

A breakdown of the friction per borehole section of the HDD is presented in figure 16.

The contribution of each friction component is also indicated as percentage of the total

calculated friction.


0

5000

10000

15000

20000

25000

0 100 200 300 400 500

Friction[N/m]

Friction in borehole

Friciton

-16

-14

-12

-10

-8

-6

-4

-2

0

0 100 200 300 400 500

Depth[m-minussurfacelevel]

Length borehole path in [m]

HDD DN600


20/46



Based on the friction assessment for the DN600 typical the following can be noted:

When arching doesnt occur (sections A and B) the total friction is mainly

determined by the vertical intergranular pressure (k) in non-cohesive soils and

by adhesion in cohesive soils.

Along the total length of the elastic bend arching will occur. The upper ends of

the elastic bends are positioned near the arching tipping point (point B) and thus

no arching occurs. The vertical intergranular pressure (k) is governing at this

point and the influence of the soil support reaction as a result of the curvature of

the bends at this section is limited: approximately 3% of the total friction (see

figure 16,point B).

Along the other part of the elastic bend (where arching occurs) the friction is for

40% determined by the friction between pipe and drilling fluid (adhesion) and the

buoyancy effects (see figure 16, point C). The other 60% of the friction is

determined by the influence of the soil support reaction as a result of the

curvature. At the floor pipe section the total friction is dominated by the buoyancy effects of

the pipe: 80%, see graph 4 point D.

The absolute friction along the elastic bends is approximately 2 times higher than

the friction at the floor pipe section (seefigure 15).

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

A: Upper bend B: Tipping point

arching

C. Elastic curved

bends

D: Floor pipe





21/46


22/46




Based on the friction assessment for the DN1200 typical the following can be noted:

When arching doesnt occur (sections A and B) the total friction is mainly

determined by the vertical intergranular pressure (k) in non-cohesive soils andby adhesion in cohesive soils.

Along the total length of the elastic bend arching will occur. The upper ends ofthe elastic bends are positioned near the arching tipping point (point B) and thusno arching occurs. The vertical intergranular pressure (k) is governing at this

point and the influence of the soil support reaction as a result of the curvature ofthe bends at this section is limited: approximately 1% of the total friction (see,figure 18,point B).

Along the other part of the elastic bend (where arching is assumed) the friction isfor 65% determined by the friction between pipe and drilling fluid (adhesion) andthe buoyancy effects (see figure 18, point C). The other 35% of the friction isdetermined by the influence of the soil support reaction as a result of thecurvature.

At the floor pipe section the total friction is dominated by the buoyancy effects ofthe pipe: 90% (seefigure 18,point D).

The absolute friction along the elastic bends is approximately 1,5 times higherthan the friction at the floor pipe section (seefigure 17).

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

A: Upper bend B: Tipping pointarching

C. Elastic curvedbends

D: Floor pipe





23/46



4.2.4 Overall conclusions friction along the borehole

To assess the differences in friction contribution for each typical diameter in a plot is

made of the friction along the borehole per section for each typical pipe diameter.

Figure 19: distribution of friction along the borehole per typical diameter

Based on the results infigure 19 for the DN600 typical the following can be concluded:

For large pipe diameters installed by HDD-technique in general the largest

friction occurs between section A and C, from the upper bend up till the end of

the elastic bend.

For the larger diameters the biggest contribution in the friction along HDDs is

determined by the top soil load component.

For smaller pipe diameters the soil reaction influence at the elastic bends isrelatively high in relation to larger diameters.

For large pipe diameters the friction along the floor pipe is dominantly determinedby the buoyancy effect. In general arching will occur at this section and the

contribution of the adhesion between the pipe and drilling fluid is relativelylimited. For smaller diameters the contribution of the drilling fluid adhesionhowever is considerable.

0

10000

20000

30000

40000

50000

60000

70000

80000

90000

DN100

DN600

DN1200


24/46


25/46


26/46



the system stiffness of the total HDD crossing: pipe diameter, wall thickness, bendconfiguration, material.

The soil parameters are schematized and calculated according to the formulas andspring models of the soil as defined in the Dutch design code NEN 3650-1: 2003 except

for friction. Based on moderate sand (for soil properties see table 6). In table 7 anexample set of the input soil parameters are given for the DN100 typical.

Table 6: soil properties moderate sand

Moderate sand

Volumic weight dry 18 kN/m3

Volumic weight wet 20 kN/m3

Angle of internal friction 32,5

Youngs modulus sand 10 N/mm2

Table 7: example PLE soil input parameters for DN100 typical

Description parameter Value PLE Parameter

Upward vertical spring acting on pipe 0,286 N/mm /mm KLT

Maximum upward soil reaction on pipe 0,370 N/mm RVT

Downward vertical spring acting on pipe 0,089 N/mm2/mm KLS

Maximum downward soil reaction on pipe 2,04 N/mm RVS

Horizontal spring acting on pipe 0,0391 N/mm /mm KLH

Maximum horizontal reaction on pipe 1,21 N/mm2 RH

The friction applied in the PLE models is according to the values and graphs presentedinfigure 13,figure 15 andfigure 17.

It should be noted that for the assessment in normal engineering practice modeluncertainty values and load factors have to be taken into account on these parameters.In the framework of this study these are not taken into account in order to establish acomparable set of data.

The external loads taken into account in the calculation models are summarized intable

8.Per calculation model two load combinations have been taken into account: p d+ T

(load case 4) and T (load case 3).

Table 8: external loads

DN100 DN600 DN1200Internal pressure (pd) 40 barg 80 barg 80 barg

Temperature load (T) 35 C 45 C 45 C

In the calculation of the expansion of the HDD sections in paragraph 4.2.5 the influenceof the upper bends has not been taken into account. The stiffness of this bend is ofimportance in the way the new friction model will act and distribute its expansiondisplacements and forces along the HDD. To take into account this influence, the upper


27/46



bends and a section of adjacent straight field pipe are connected to the HDDconfiguration in PLE. Based on normal engineering practice for gas pipelines in theNetherlands typical upper bends with a radius of 40Doare chosen.

5.2 Expansion results

In the next paragraphs the expansion results for each typical is presented based on thePLE calculation. To assess the influence of the modeled friction the axial displacementsand expansion force over the HDD are plotted in graphs and presented. Also a tablewith numerical results at some characteristic points in the HDD is presented. The fivecharacteristic points along the borehole sections of the HDD for evaluation and analysisof results are presented infigure 20.

Figure 20: overview points along borehole sections to be evaluated


28/46



5.2.1 Presentation and analysis of results DN100 Typical

Figure 21: schematized HDD configuration

Figure 22: schematized ultimate soil friction along HDD

-6

-5

-4

-3

-2

-1

0

1

2

0 100 200 300 400 500 600 700 800

Level[mNAP]

Length [m]

Pipeline Configuration DN100

Ground level Z-coordinate pipe axis (Ground)water level

0

500

1000

1500

2000

2500

3000

0 100 200 300 400 500 600 700 800

Friction[N/m](partialfactorexcluded)

Length [m]

Ultimate soil friction DN100

Ultimate soil friction in HDD = 0 N/m Ultimate soil friction in HDD calculated


29/46



Figure 23: calculated axial displacements (LC4: pd+ T; LC3: T)

-0,60

-0,50

-0,40

-0,30

-0,20

-0,10

0,00

0,10

0,20

0,30

0,40

0,50

0,60

0 100 200 300 400 500 600 700 800

Axialdisplacement[mm]

Length [m]

Axial displacements DN100 (LC4)

Axial displacement F = calculated Axial displacement F = 0 N/m

-0,60

-0,50

-0,40

-0,30

-0,20

-0,10

0,00

0,10

0,20

0,30

0,40

0,500,60

0 100 200 300 400 500 600 700 800


Length [m]




30/46



Figure 24: calculated axial forces (LC4: pd+ T; LC3: T)

149,00149,25

149,50

149,75

150,00

150,25

150,50

150,75

151,00

151,25

151,50

151,75

152,00

152,25

0 100 200 300 400 500 600 700 800

Axial(compressive)forcekN

Length [m]

Axial force DN100 (LC4)

axial force F = 0 N/m axial force F = calculated

171,50

171,75

172,00

172,25

172,50

172,75

173,00

173,25

173,50

173,75

174,00

174,25

0 100 200 300 400 500 600 700 800


Length [m]




31/46



Table 9: numeric results per typical borehole section

Load case 4:

pd+ T F F=0 F vs F=0 F F=0 F vs F=0

point node F-ax [kN] F-ax [kN] relativedeviation

u-ax [mm] u-ax [mm] relativedeviation

A1 225 149,46 149,65 0,19% -0,42 -0,44 4,8%B1 225 149,36 149,65 0,19% -0,42 -0,44 4,8%

C1s 255 149,45 149,69 0,16% -0,41 -0,44 5,6%

C1e 443 149,86 149,85 -0,01% -0,30 -0,33 10,0%

D 1123 150,13 149,85 -0,19% 0 0 0

Load case 3:

T F F=0 F vs F=0 F F=0 F vs F=0



A1 225 171,70 171,96 0,15% -0,38 -0,40 4,8%

B1 225 171,70 171,96 0,15% -0,38 -0,40 4,8%

C1s 255 171,78 172,01 0,13% -0,37 -0,39 5,4%

C1e 443 172,17 172,17 0,00% -0,28 -0,30 7,3%

D 1123 172,42 172,17 -0,14% 0 0 0

Based on the displacements presented in Fehler! Verweisquelle konnte nichtgefunden werden.andtable 9 the following can be noted:

The axial displacements do not tend to differ a lot when friction is applied alongthe HDD (see Fehler! Verweisquelle konnte nicht gefunden werden.).

The difference in axial displacements at the upper bends is for the situation withfriction equal to the situation without friction.

The relative increase of hindrance of displacement along the HDD seems to besignificant: 5% to 10%. However taking into account the absolute axialdisplacement figures the difference is small.

For both load cases the axial displacements show more or less the same relativeincrease of hindrance of displacement.

Based on the axial forces presented in

figure 24 andtable 9 the following can be noted:

When applying friction along the HDD a shift of force (redistribution) is noted from

the floor pipe towards the upper bends (see graph 11). In case of friction the compressive axial force in the elastic bends and floor pipe

tends to increase in favor of the axial force in the upper bend which is reduced.

The expansion forces along the HDD do not differ significantly for each loadcombination (with or without internal pressure) when friction is applied comparedwith when friction is neglected.


32/46



Result analysis DN100 typical:

The difference between applying friction along HDD or neglecting it (F=0) issmall.

As a result of applying friction redistribution of forces along the HDD occurs infavor of the forces at the upper bend.

Comparison between the calculated expansion without including the upper bend

and connected field line (see table 3) with the analyzed results in softwarepackage PLE (see table 9) shows that modeling of the total pipeline systemreduces the occurring free displacement by an order 100 (approximately 40 mmvs 0,40 mm).


33/46






-16

-14

-12

-10

-8

-6

-4

-2

0

2

0 200 400 600 800 1000

Level[mNAP]

Length [m]



0

5000

10000

15000

20000

25000

0 200 400 600 800 1000Friction[N/m](partialfactorexcluded)

Length [m]



A

B

D

C


34/46


35/46




050

100

150200250300350400450500550600650700750800850900950

1000

0 200 400 600 800 1000


Length [m]



15001550

16001650170017501800185019001950200020502100215022002250

23002350

0 200 400 600 800 1000


Length [m]




36/46




Load case 4:




A1 524 285,33 603,28 111,4% -10,30 -20,16 95,7%C1s 541 329,72 602,36 82,7% -9,45 -19,49 106,3%

B1 580 469,92 599,24 27,5% -7,94 -18,46 132,4%

C1e 969 813,41 586,52 -27,9% -3,02 -9,260 207,0%

D 1163 830,10 586,52 -29,3% 0 0 0

Load case 3:




A1 524 1862,43 2108,56 13,2% -6,31 -11,07 75,6%

B1 541 1893,85 2107,54 11,3% -5,82 -10,80 85,8%

C1s 580 1990,69 2104,80 5,7% -4,86 -10,29 111,5%

C1e 969 2212,83 2092,07 -5,5% -1,84 -5,19 182,0%

D 1163 2223,10 2092,07 -5,9% 0 0 0

Based on the displacements presented in

figure 27 andtable 10 the following can be noted:The axial displacements differ significantly when friction is applied along the HDD(see

figure 27). The influence of the friction in riser and the elastic bends is shown clearly in

figure 27.Between point B and C the axial displacement graph is curved with asteeper gradient than the linear part of the graph towards the midpoint of theHDD. The difference in shape can be explained by the increase of frictionbetween point B and C. On this section the friction due to elastic bending incombination with unarched vertical soil load is dominating. Towards the midpointof the HDD (between point C and D) arching occurs and the friction decreaseswhich results in a less steep linear relation for the axial displacement towards themidpoint of the HDD.

The displacement curve under load case 4 shows a significantly higher maximum

axial displacement than for load case 3. This is caused by the internal pressureacting on the bend and results in an additional axial displacement.

The difference in axial displacements at the upper bends is for the situation withfriction significantly less than for the situation without friction. For load case 4 theaxial displacement is reduced with 96%. For load case 3 the reduction is about76%.

In the case of friction a significant shift of displacement is noted along theseparate sections of the HDD.


37/46


38/46



table 4)with the analyzed results in software package PLE (seetable 10)showsthat modeling of the total pipeline system reduces the occurring freedisplacement about 20 times (approximately 129 mm vs 6 mm).


39/46






-23

-18

-13

-8

-3

2

0 200 400 600 800 1000 1200 1400 1600

Level[mNAP]

Length [m]



0

10000

20000

30000

40000

50000

60000

70000

80000

90000

0 200 400 600 800 1000 1200 1400 1600Frictio

n[N/m](partialfactorexcluded)

Length [m]



A

C

B

D


40/46



Figure 31: calculated axial displacements (LC4: pd+ T; LC3: T)

-30

-25

-20

-15

-10

-5

0

5

10

15

20

25

30

0 200 400 600 800 1000 1200 1400 1600


Length [m]



-20

-15

-10

-5

0

5

10

15

20

0 200 400 600 800 1000 1200 1400 1600


Length [m]




41/46





0

200400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

0 200 400 600 800 1000 1200 1400 1600


Length [m]



5000

5250

5500

5750

6000

6250

6500

6750

7000

7250

7500

7750

8000

8250

0 200 400 600 800 1000 1200 1400 1600


Length [m]




42/46



Load case 4:




A1 127 139,37 1841,01 1221,0% -10,58 -24,30 129,6%

C1s 165 845,86 1815,50 114,6% -5,90 -22,28 277,8%B1 225 1717,06 1791,19 4,3% -3,07 -20,69 573,3%

C1e 553 2396,31 1732,21 -27,7% -0,94 -11,36 1109,7%

D 748 2417,69 1732,21 -28,4% 0 0 0

Load case 3:




A1 127 6285,52 7543,00 20,0% -7,01 -14,59 108,2%

B1 165 6799,17 7518,48 10,6% -3,95 -13,63 245,2%

C1s 225 7394,55 7494,17 1,4% -2,06 -12,70 516,5%

C1e 553 7840,10 7435,20 -5,2% -0,65 -7,019 972,5%

D 748 7855,01 7435,20 -5,3% 0 0 0

Based on the displacements presented in

figure 31 andtable 11: numeric results per typical borehole section the following can benoted:

Compared to the DN600 typical the linear gradient of the axial displacement near

the midpoint section is less steep. This is an indication that the DN1200 typical ismore hindered along the HDD compared to the DN600 typical.

The influence of friction in riser and the elastic bends for the DN1200 typical islarger than for the DN600 typical as can be read of the steeper curved sectionbetween point B and C.

The displacement curve under load case 4 shows a significantly higher maximumaxial displacement than for load case 3. This is caused by the internal pressureacting on the bend and results in an additional axial displacement.

The difference in axial displacements at the upper bends is for the situation withfriction significantly less than for the situation without friction. For load case 4 theaxial displacement is reduced with 130%. For load case 3 the reduction is about

108%. Based on the low absolute values of axial displacement near the midpoint of the

HDD can be concluded that the pipe at this point is fully hindered: the relativedisplacement between pipe and soil friction springs nears zero.

The relative increase of hindrance of displacement along the HDD is significant:for load case 4 from 130% at the upper bends up to over 1100% near midpoint ofthe HDD. For load case 3 the value varies from 108% at the upper bend up to970% near midpoint.


43/46



Based on the axial forces presented in

figure 32 andtable 11 the following can be noted:

When applying friction along the HDD a shift of force (redistribution) occurs from

the floor pipe towards the upper bends (see

figure 32). Compared with the DN600 typical the redistribution of axial forces ismore pronounced.

In case of friction the compressive axial force in the elastic bends and floor pipetends to increase in favor of the axial force in the upper bends which is reduced.For load case 4 this range of this shift is between -30% (increase) for the axialforce in the floor pipe up to 1221% (decrease) for the upper bends.

The expansion forces along the HDD differ significantly per load case. In case offriction 4 a reduction of 1221% for load case 4 is reached and for load case 3 thisreduction is 20%.

For the case with friction the axial force at the upper bends for load case 4 isabout 98% less than for load case 3.

Result analysis DN1200 typical: The difference between applying friction compared to applying no friction is

substantial both for axial displacements as for axial forces.

The flat linear distribution of axial displacement near midpoint of the HDD incombination with the constant axial force of table 11 in points C1e and D showthat the pipe is fully hindered.

The axial forces at the midsection of the HDD will increase in favor of the axialforces at the upper bends as a result of the increased hindrance.

The difference in axial displacement and axial forces between load case 4 andload case 3 is significant and caused by the effect of internal pressure at thebends (the so called convex bottom effect).

Comparison between the calculated expansion without including the upper bendand connected field line (see table 5) with the analyzed results in softwarepackage PLE (see table 11) shows that modeling of the total pipeline systemreduces the occurring free displacement about 36 times (approximately 254 mmvs 7 mm).


44/46



6 Technical Conclusion

Based on the results presented quantitatively and qualitatively the following technical

conclusions relating the frictional schematization can be drawn:

Based on the typical calculations can be concluded that modeling friction along

the HDD leads to a more refined distribution of axial displacement and axial

forces along the HDD and in absolute sense is valuable to mitigate and reduce

expansion at the upper bends in the HDD design.

The influence of the upper bends and the adjacent field sections leads to a much

stiffer overall system by which application of friction increases the favorable

effect on the resulting axial displacements and axial forces acting on these

bends. By the convex bottom effect acting at these bends under operational

conditions (in situations with internal pressure) the effect of hindrance along the

HDD is increased compared to the situation without this effect.

Within the framework of the frictional schematization there seems to be some

diametrical dependent effects noticeable. The following diameter dependent

conclusions can be drawn:

Increase of diameter leads to an increase of hindrance of the total HDD system.

In case friction is modeled the hindrances effect grows progressively. For the

DN100 typical this effect is hardly noticeable but for the DN600 typical and

DN1200 typical this effect is substantial.

Based on the calculated axial displacements for the various typicals the increase

of hindrance seems not to be developing linearly with the diameter. This is aresult of the influence of axial pipe stiffness at the upper bends acting on the rest

of the HDD section. Apparently the effect of larger diameters (DN600 and

DN1200) on the total stiffness is much larger than for smaller diameters.

In case of friction along the HDD an increase of diameter leads to more

concentrated axial forces near the upper bends in combination with a decrease

of the absolute value. Near the midpoint of the HDD the axial force will increase

and shows a more constant gradient. This is in line with the first conclusion.

Based on the conclusions above can be said that under application of friction an

increase of diameter will lead to a fully hindered HDD pipeline section in which therelative displacements near the midpoint are small and the axial force is relatively

constant. This acts favorable for the axial forces on the upper bends.


45/46



9 General conclusion

Applying friction along a HDD by the presented approach is valuable and practical for

engineering purposes. The presented approach is developed on the existing soil

mechanical theories for subsoil pipelines. Based on the assessment of the approach

and the calculated results, neglect of friction along the HDD is not realistic. The effect ofthe proposed schematization of friction is large on axial displacements and axial forces

occurring in the HDD section under operational conditions; especially for large

diameters. Application of friction along the HDD leads to substantial a reduction of axial

forces on the upper bends by which costly mitigating expansion measures (f.i. cushions

loops) can be prevented.

8 Recommendations

To further evaluate the proposed schematization of friction along the HDD andthe influence of the upper bends onto the total HDD section, it can berecommended to set up a field monitoring program on axial displacements oninstalled HDDs under operational conditions.

To determine the diametrical effect of the proposed approach in relation to thedevelopment of the axial forces acting on the upper bends it is recommended tofurther investigate more specifically on this subject; f.i. in the form of a masterthesis.

Based on the noted expansion effects for a small diameter gas pipeline it can bevaluable to determine the expansion effects for pre-insulated pipe sectionsinstalled in HDDs under the proposed frictional schematization. Taking into

account that as a result of increased surface area (on which the friction will beacting) and increase of buoyancy effect it is assumed that the hindrance effect inthe HDD will increase and the influence of applying friction becomes morenoticeable (lower axial displacements and thus lower expansion forces at theupper bends). Especially at smaller diameters this influence is expected to besignificant and might lead to a compacter HDD design (without expansionmeasures) of high temperature gas flow lines and district heating water lines. Theassumed influence of the proposed frictional schematization for these types ofpipeline materials under even more challenging conditions seems to be morediscriminative in relation to expansion problems and is worthwhile to be further

investigated.


46/46

References

1 Stability of borehole during Horizontal Directional Drilling, Thomas Viehfer et al.

2. Soil deformations due to Horizontal Directional Drilling Pipeline Installation, G.M.

Duyvestyn and M.A. Knight, Proceedings NORTH AMERICAN NO-DIG 9-12 April 2000.

3 Experimental Investigation of Borehole and Surface Friction Coefficients During HDD

Installations, G. El-Chazli et al., Proceedings NASTT NO-DIG 24-27 April 2005.

4 Modeling the soil pipeline interaction during the pull back operation of horizontal

directional drilling, J.P. Pruiksma and H.M.G. Kruse

5 Dutch Standard NEN3650-1 + A1 Requirements for pipeline systems part 1 General

Quire 1 to 6, distributed by NNI, Delft, August 2006.

6 Tebodin report 42380.01-1931205 Berekenen van de wrijving voor leidingen gelegd in

open ontgraving d.m.v. HDD techniek, Hengelo, November 2011

Soil Friction Along HDD Drilling

Documents