On Bottom Stability

Pipeline Design Training

DNV 1981 - Lateral Stability Design (cont’d)

Appropriate hydrodynamic force coefficients are

used in the stability analysis of the pipeline.

The stability criterion is expressed by:

 W sub  FL 
Fs 
FD  FI 
DNV 1981 - Lateral Stability Design (cont’d)

Where:

Wsub = Submerged weight of pipeline per unit length (N/m)

FL = Hydrodynamic lift force per unit length (N/m)

FD = Hydrodynamic drag force per unit length (N/m)

FI = Hydrodynamic inertia force per unit length (N/m)

m = Coefficient of friction between pipe and seabed
(Varies between 0.3 for dense clay to 0.7 for sand,
normal friction coefficient =0.5)
Fs = Safety factor = 1.1
DNV 1981 - Lateral Stability Design (cont’d)

The drag force per unit length (FD) of the pipeline is

calculated as follows:

FD  1

2 W
C D Dt U d | U d |
DNV 1981 - Lateral Stability Design (cont’d)

The lift force per unit length (FL) is calculated as

follows:

2
FL  1
2
 W CL Dt Ud
DNV 1981 - Lateral Stability Design (cont’d)

The inertia force per unit length (FI) is calculated as

follows:

2
FI  1
4
  W Dt CI a sin 
DNV 1981 - Lateral Stability Design (cont’d)

Where:
w = Density of seawater (kg/m³)
CI = Inertia coefficient (3.29)
CD = Drag coefficient (Normally 1.2)
CL = Lift coefficient (Normally 0.9)

Dt = Total outer diameter of pipeline including coatings (m)

a = Horizontal water particle acceleration normal to

the pipe axis (m/s²)
Ud = Horizontal water particle velocity normal to
the pipe axis
= (Uc + Uw Cos ø) (m/s)
Uw = Wave induced horizontal water
particle velocity normal to pipe (m/s)
Uc = Horizontal steady current velocity normal to
pipe axis (m/s)
DNV 1981 - Lateral Stability Design (cont’d)

The hydrodynamic coefficients will be reduced to

care of trench effects if applicable as per Jacobsen et
al OTC paper “Fluid Loads on Pipeline: Sheltered or
Sliding” (what are the values?)
 DNV RP E305, 1988

DNV RP E305 is based on the PIPESTAB Joint

Industry Project conducted in the North Sea in the
mid-eighties. Three design methods are described in
this Code.
 Simplified Analysis

This analysis is based on quasi-static method with

results calibrated from the Generalised Stability
analysis.
 2. Generalised Analysis

This analysis is based on a set of non-dimensional

stability curves which have been derived from a
series of runs presenting pipe movement and strain
results with a dynamic response model. Net pipe
movement is permitted for pipe on sandy soil up to
40 pipe diameters. Pipe on clay, however, does not
allow net pipe movement.
 3. Dynamic Analysis

The analysis described involves a full dynamic

simulation of a pipeline resting on seabed with soil
resistance, hydrodynamic forces, boundary
conditions and dynamic response modelled for. It
forms the basis of reference for the Generalised
analysis.
 DNV RP E305, 1988 (cont’d)

The hydrodynamic coefficients allowing pipeline to

move a maximum of 20 m in sand and no
movement in clay.

Normally, a Pierson Moskovitz (PM) wave

spectrum is assumed in the analysis.
 W sub 
  FL    FD  FI
 Fw 
The stability criteria is expressed as:
Where:
Wsub = Submerged weight of pipeline (N/m)
FL = Hydrodynamic lift force per unit length (N/m)
FD = Hydrodynamic drag force per unit length (N/m)
FI = Hydrodynamic inertia force per unit length (N/m)
m = Coefficient of friction between pipe and seabed in
accordance with DNV RP E305.
(Friction coefficient varies between 0.15 to 1.3 for clayey soil
depending soil shear strength and K C Number. The frictional factor
for sand is 0.7 regards of flow parameters)

FW = Calibration factor depending on Keulegan Carpenter number

and velocity ratio. A safety factor of 1.1 is inherent in the calibration factor.
The drag force per unit length (FD) of the pipeline
is calculated as follows:
1
FD  2
 W C D D (U s Cos  U c ) | U s Cos   U c |

 The lift force per unit length (FL) is calculated as

follow:

FL  1
2
 W C L D (U s Cos   U c ) 2
The inertia force per unit length (FI) is calculated as
follow:

1 2
FI   w D C M a sin 
4
Where:  w = The mass density of seawater (kg/m³)
CM = Inertia coefficient
= 3.29
CD = Drag coefficient
= 1.2 for Re < 3 x 105 and M  0.8
= 0.7 ??
CL = Lift coefficient
= 0.9
D = Total outer diameter of pipeline including external coatings (m)
a = Horizontal water particle acceleration normal to the pipe axis
(m/s²)
Us = Horizontal water particle velocity normal to the pipe axis due to wave
(m/s)
Uc = Horizontal steady current velocity normal to the pipe axis due to wave
(m/s)
 = Wave phase angle (deg.)
M = Current velocity to wave velocity ratio (Uc/Us)
 AGA (PRCI) Method

 The PRCI/AGA Stability software was developed based on analytical research

and large-scale test model sponsored by Pipeline Research Council
International, Inc. (PRCI).

 The software represents the state-of-the-art design in pipeline stability and

models the complex behaviour of pipe/soil interaction which includes:
1. Hydrodynamic forces which account for the effect of wake (generated by flow over
pipe) washing back and forth over the pipe in oscillatory flow;
2. Embedment (digging) into clay or sand which occurs as the pipe resting on the
seabed is exposed to oscillatory loading and small oscillatory deflections.

 Three levels of analysis are provided by PRCI (AGA) Stability software, namely
Levels 1, 2 and 3. The general characteristics of each level of analysis is
summarised as follows:
 PRCI (AGA) STABILITY SOFTWARE
Software Analysis Type Description

Performs a simplified analysis using traditional

Level 1 Simplified Static
methods.

Performs a static analysis based on:

 Realistic hydrodynamic forces
Level 2 Simplified Quasi-Static  Realistic pipe embedment calculated by
quasi-static simulation of wave induced
pipe oscillations.

Consists of 3-program suite, WinWave,

WinForce and WinDynamics.
 WinWave generates wave kinematics for
Dynamic Time Domain with 3-D random seas.
Level 3 Wave Kinematics for 3-D  WinForce generates wave forces based
Random Seas on time history of wave kinematics
 WinDynamics analyses pipe dynamics
with external forces and a history
dependent soil model.
 AGA (PRCI) Method (cont’d)

Level 1 analysis:
This approach is based on traditional stability
analysis methods where the Morrison type
hydrodynamic forces and frictional soil resistance
are considered. Its design methodology corresponds
to that described in DNV 1976 and DNV 1981 Codes.
A ‘no movement’ pipeline stability design criteria is
assumed.
 AGA (PRCI) Method (cont’d)

Level 2 analysis:
With the similar ‘no movement’ stability criteria, it is
based on quasi-static analysis where it simulates
pipeline embedment process as in the Level 3
analysis.
 AGA (PRCI) Method (cont’d)

Level 3 analysis:
This level of analysis is most detailed where pipeline
is simulated in a finite element time domain
software. Detailed information on pipeline
movement and stresses obtained are basis of pipeline
safety assessment.
 AGA (PRCI) Method (cont’d)

Normally, a Level 2 analysis is adequate. Level 3

analysis is carried out only if further optimisation of
concrete weight coating thickness is required.
 Analytical Methodology (Level 2)

According PRCI (AGA) a pipeline exposed to wave

flow will experience a hydrodynamic force, which is
expressed by two components: the in-line drag force
and the lift force.

These two forces are calculated based on the physics

of the water-pipeline interaction.
 Analytical Methodology (Level 2) (cont’d)

The AGA hydrodynamic force model is expressed as:

1

FD   D U s2  C oD 
2
C iD cos i ( wt   iD )  (1)

FL 
1
2

 DU s2  C oL  C iL 
cos i ( wt   iL ) (2)

2
FI   DU s2 C M  a(t ) (3)
4
 FL = Hydrodynamic lift force per unit length (N/m)
FD = Hydrodynamic drag force per unit length (N/m)
FI = Hydrodynamic inertia force per unit length (N/m)
 = Density of sea water (kg/m3)
D = Pipe outer diameter (m)
Us = Near seabed water velocity (m/s)
C iD , C iL = Fourier coefficient
iD iL = Fourier phases
CM = Inertia coefficient
a (t ) = Water acceleration (m/s2)
 Analytical Methodology (Level 2) (cont’d)

The Fourier coefficient and phases are determined

from extensive model test programme, which
includes the effects of steady current, waves, pipe
roughness, and seabed roughness. These values have
been stored as a database, which is implemented in
the AGA Level 2 stability analysis.
 Analytical Procedures

 The AGA Level 2 is a quasi-static analysis program,

which is designed to take advantage of the results from
the AGA’s hydrodynamic and pipe/soil interaction tests.

 The procedure of the program analysis is as follows:

1. Based on user inputs, the program calculates the

significant bottom velocity, Us
2. Maximum and minimum in-line hydrodynamic forces
for the Largest 200 waves contained in an assumed 4-
hour long build-up sea state are calculated.
3. Maximum and minimum inline forces for the largest 50 waves
during a subsequent 3-hour long design sea state are
calculated.

4. Based on the forces calculated, a conservative estimate of pipe

embedment at the end of the 4-hr storm build-up period is
calculated.

5. Based on the forces calculated and the pipe embedment

calculated, the amount of additional pipe embedment that can
be produced by the 50 largest waves in the design sea state is
calculated in a similar fashion similar. This embedment and the
associated soil resistance force are saved for future processing.
6. Hydrodynamic forces for a complete wave cycle are
calculated for four statistically meaningful wave induced
bottom velocities which are expected in a 3-hr long
design sea state.

7. The four bottom velocities, and, the most likely number

of wave induced velocities exceeding each are:

U1/3 = 1.00 Us (135 exceedances)

U1/10 = 1.27 Us (40 exceedances)
U1/100 = 1.66 Us (4 exceedances)
U1/1000 = 1.86 Us (0 exceedances)
8. Using the soil resistance values obtained and the
hydrodynamic forces calculated, the minimum
factor of safety against lateral sliding is calculated
for the pipe embedment at the end of the 4-hr long
build up period, and at the end of the 3-hr long
design sea state.
9. The factor of safety is calculated at one-degree
intervals of wave passage for a complete 360-degree
from:
 (Ws  FL (t ))  Fh
Factor of Safety (FOS) 
FD (t )  FI (t )
10. The recommended Level 2 stability criteria should
satisfy the following aspects. At the end of the 4-
hour storm build up, the pipeline should be stable
in the U1/100 condition, i.e. FOS ≥ 1.0 At the end
of the additional 3-hour storm period, the pipeline
should be stable in the U1/1000 condition, i.e. FOS
≥ 1.0
COMPARISON BETWEEN AGA AND DNV RP E305

In general, DNV RP E305 designs are more

conservative than the AGA designs. This is true for
most designs where pipeline is laid on clay and all
designs where the soil is sand. No net pipe movement
criterion is assumed.

For cases in DNV RP E 305 design where movement

is allowed for pipe on sand, the concrete
requirements is significantly less, and often similar to
the AGA designs which movement is not permitted.
COMPARISON BETWEEN AGA AND DNV RP E305
(cont’d)

 Compared to the traditional method (DNV 1976 / 1981), both AGA

Level 2 analysis and DNV RP E305 designs result in concrete
weight coatings that are more sensitive to soil strength / density.

 With no pipe movement criteria, this sensitivity is similar in both

AGA Level 2 analysis and DNV RP E305. However, concrete weight
coating is less sensitive to soil density when pipe net movement is
allowed for in DNV RP E305 design.

 AGA designs produce concrete weight coating less sensitive to water

depth than DNV RP E305 designs. This is due to the reduction in
pipe embedment in the AGA Level 2 analysis at deeper waters.
 PIPELINE STABILITY IN OPEN TRENCH

Both AGA and DNV RP E305 are not applicable for

pipeline resting in an open trench. Thus, to analyse
pipeline stability in an open trench, an alternative
method would have to be used.