INTRODUCTION
Drag reduction or flow improving in pipelines carrying petroleum products or
crude oils, was one of the main challenges especially in the last few decades
when the term power saving, raised up due to the rapid increase in global power
consumption. One of the major sectors that deal with huge amounts of power losses
is the liquids transportations through pipelines. It is known that, liquids
(water, crude oil and refinery products) are transported in pipelines in a turbulent
mode (Reynolds number higher than 2500) and that will lead to huge pumping power
losses along the pipelines (especially strategic pipelines). Drag Reducing Agents
(DRA’s) were used in the past few decades to improve the flow in pipelines and
to increase it without the need for any changes in the geometry of the pipeline
system. These agents can be classified into three major categories which are
(Virk et al., 1970): (i) polymers, (ii) surfactants
and (iii) suspended solids.
Surfactants with its major classifications (certain anionic, cationic,
nonionic and also zwitterionic) are proven to be a powerful drag reducers
in turbulent flow in pipes and can hence contribute to significant energy
saving. Their drag reduction ability at concentrations as low as part
per million is ascribed to the micelles present in the solution. These
micelles play a dominant role in the mechanism of turbulence suppression
and in the significant friction decrease, which can be even higher than
in some high polymer solution.
Siriluck and Sirivat (2007) investigated the influence
of ionic strength on the interaction between poly(ethylene oxide) (PEO) and
cationic surfactant, hexadecyltrimethylammonium chloride (HTAC) and the consequent
effect on turbulent drag reduction in aqueous PEO/HTAC solutions. For the measurement
of turbulent drag reduction in a couette cell, their data indicate that the
minimum wall shear stress in aqueous HTAC solutions occurs at an optimum HTAC
concentration, close to CMC and this optimum concentration value decreases with
increasing ionic strength. Finally, their findings provide evidences that the
turbulent wall shear stress does not always scale inversely with the hydrodynamic
volume of the polymersurfactant complex.
SungHwan et al. (2007) investigated the drag
reduction and heat transfer efficiency reduction of nonionic surfactant as
a function of fluid velocity, temperature and surfactant concentration. Their
results showed that existing alkyl ammonium surfactant had Drag Reduction of
60 to 80% at 10002000 ppm concentration with fluid temperature ranging between
50 and 60°C. They showed also that the percentage drag reduction decreases
when the fluid temperature was 7080°C. Finally, it was noticed that with
fluid temperature ranging between 70 and 80°C the Dr was 0.60.8 when the
concentration level was between 1000 and 2000 ppm.
Since the early nineties, many studies were published concerning the use of
certain additives as Drag reducing Agents in a media that is not compatible
with the physical behavior between the additive and the solvent (i.e., the solubility
of the additive in the solvent). This condition (the solubility) was one of
the basic conditions that any chemical must have to be classified as drag reducing
agent (Pereira and Pinho, 2002; Lim
et al., 2005; Usui et al., 1995; Mowla
and Naderi, 2006; Sher and Hetsroni, 2008; Inaba
et al., 1995; Wang et al., 1998).
Later, the classification of the suspended solids (insoluble in liquid media)
as Drag Reducing Agents opened the door wide for more research regarding the
availability of the solubility condition in the drag reduction phenomena (Mowla
and Nadari, 2006; Toorman, 2002; Zhu
and Peskin, 2007; Hideo et al., 2000). The
capability of insoluble material (solids) to act as a powerful Drag Reducing
Agent was the main motivation in the present investigation to prove the capability
of insoluble liquids to act as drag reducing agents in the transported hydrocarbon
media (kerosene).
MATERIALS AND METHODS
Pipe diameter, solution flow rate and additive concentrations were the
variables investigated in the present study. Each set of experimental
work deals with sodium stearate in kerosene flowing in one of three pipe
diameters within five addition concentrations and seven different solution
flow rates. This study was conducted in 2008 in Faculty of Chemical and
Natural Resources Engineering, University Malaysia Pahang.
Sodium Stearate (SS)
Sodium Stearate (SS) [CH_{3} (CH_{2})_{16}COONa]
is a white crystalline anionic surfactant with a molecular weight of 306
g/g mol and active substance of 87%.
SS can be found in most animal and vegetable oils and fats. It is used
in lubricating mixtures, water proofing materials and soap manufacturing.
SS can be made by reacting glyceryl stearate with water to form stearic
acid then the stearic acid reacts with coastic soda to form the sodium
stearate.
Flow System Description
The flow system apparatus constructed in the present investigation can be
shown from Fig. 1. The reservoir tank was supported with two
exit pipes connected to centrifugal pumps.

Fig. 1: 
Experimental rig 
Table 1: 
Relative roughness for pipes used 

Table 2: 
Minimum entrance length for the pies used 

The first exit pipe with 0.0508 m ID was connected to the main centrifugal
pump which delivers the fluid to the testing sections. The other exit is of
0.0127 m ID was connected to the other centrifugal pump for the draining of
the solution after each run.
Three seamless carbon steel pipes of various inside diameters (0.0508, 0.0254
and 0.01905 m) with relative roughness shown in Table 1 were
used in constructing the flow system. The three sections A, B and C represent
the three testing sections with the three pipe diameters investigated.
A complete closed loop piping system was build. Piping starts from the reservoir
tank through the pump, reaching a connection that splits the pipe into two sections.
The first section returns back to the tank (bypass) and the other splits into
three sections with different pipe diameters (testing section). For each pipe
diameter, the minimum entrance length required for a fully developed velocity
profile in turbulent flow was calculated from the relationship suggested by
Desissler (1950).
Therefore, the minimum entrance length for the present study according
of the pipe diameter is shown in Table 2.
Calculations
The average velocity (V) and Reynolds number (Re) were calculated
using the solution volumetric flow rate readings (Q), density (ρ),
viscosity (μ) and pipe diameter (D), for each run as follows:
Pressure drop readings through testing sections before and after drag reducer
addition, were needed to calculate the percentage drag reduction Dr (%) as follows
(Virk et al., 1970):
Where:
Δp_{a} 
= 
Pressure drop after drag reducer addition 
Δp_{b} 
= 
Pressure drop before drag reducer addition 
Fanning friction factor was calculated using the following equation (Yunus
and Cimbala, 2006):
RESULTS AND DISCUSSION
The results of the experimental work showed that the percentage Drag reduction
increases by increasing the transported fluid flow rate presented by Reynolds
number (Re) as shown in Fig. 2 and 3. This
behavior is due to the increase of the degree of turbulence that provides a
suitable media for the drag reducing agent to act efficiently in the media by
suppressing the turbulence structures formed.

Fig. 2: 
Effect of Reynolds number on percentage drag reduction for
SS within different concentrations dissolved in kerosene flowing through
0.0254 m ID pipe 

Fig. 3: 
Effect of Reynolds number on percentage drag reduction for
SS within different concentrations dissolved in kerosene flowing through
0.01905 m ID pipe 
Increasing the degree of turbulence means increasing the number of eddies that
absorb the energy from the main flow to complete its shape. The penetrated (diluted)
molecules of the surfactant added (sodium stearate) will be part of these eddies
which will create a new media that these eddies works with and it will be more
difficult for these eddies to form and complete its shape with the new addition
to its structure done by these additives.
From the results it is also can be noticed that the percentage drag reduction
increases by increasing the additive concentration from 50 to 300 and
that is due to the increase of the number of the surfactants molecules
involved in the drag reduction operation. This behavior also can be seen
from Fig. 4 and 5.
The results of this study showed that within certain surfactant type and concentration,
Dr (%) Increases by decreasing the pipe diameter, which means that the surfactant
will have a better media to work in within smaller pipes (Fig.
6). Decreasing the pipe diameter means increasing the velocity inside the
pipe which will increase the turbulence. Although, the flow inside the three
pipes is turbulent but the degree of turbulence differs. For smaller pipe, the
energy absorbed by the turbulence (eddies) from the main flow will be higher
than that for larger pipes. That, whenever the degree of turbulence becomes
higher, the number of collisions between eddies will be higher, which will produce
smaller eddies. These collisions provide extra number of eddies absorbing energy
from the main flow to complete their shape. Overcoming smaller eddies is easier
by surfactants than larger once, because of the amount of energy absorbed by
smaller eddies is lower.

Fig. 4: 
Effect of concentration on percentage drag reduction for SS
dissolved in kerosene flowing through 0.0254 m ID pipe 

Fig. 5: 
Effect of concentration on percentage drag reduction for SS dissolved in kerosene flowing through 0.0508 m ID pipe 

Fig. 6: 
Effect of pipe diameter on percentage drag reduction at different
volumetric flow rates, with 200 ppm concentration of SS dissolved in kerosene 

Fig. 7: 
Friction factor versus Reynolds number at different concentrations
of SS surfactant dissolved in kerosene flowing through 0.0254 m ID pipe 
This indication was supported by large number of the experimental results of
the present study. That generally, Dr (%) values for pipes of 0.0254 and 0.01905
m ID are close to each other and far from those of 0.0508 m ID.
Another representation to the effect of all variables used in this investigation
can be seen using friction factor, which was calculated from Eq.
4.
Figure 7 shows the friction factor for various Re pipe diameters,
surfactant concentrations. It can be noticed that, when the surfactant concentrations
is zero (pure solvent), most of the experimental data points are located at
or close Blasuis asymptote, which give an indication that the starting points
of the operation are close to that of the standard operation conditions suggested
in the literatures. When the surfactant is presented in the flow, the experimental
data points are positioned in the direction of lowering friction towards Virk
asymptote that represent maximum limits of drag reduction, which will give the
idea that, to reach such an asymptote, higher additive concentration and Re
are needed for each pipe diameter. But, it must be considered that higher concentrations
should not affect solvent properties, also by considering the economical costs
of raw material of drag reducing agents, therefore it was difficult to reach
virk asymptote without affecting the investigated solvent properties.
From the results, it can be noticed that the SS can be classified as
drag reducing agent although it is not soluble in the transported media.
This phenomenon may be due to the idea of DRA molecules penetration instead
of solubility which will open the door for anew explanations depending
on those regarding the drag reduction caused by suspended solids. Which
will lead to the idea of having these penetrated molecules as part of
the turbulence inside the transport pipe which will prevent eddies from
completing its shape and that will prevent the absorption of more.

Fig. 8: 
Predicted versus observed values of friction factor for SS
dissolved in kerosene 
In the present investigation, the dimensional analysis could be used
for grouping the significant quantities into a dimensionless group to
reduce the number of variables appearing and to make the results so compact
and applicable to all similar situations.
The choice of the appropriate variables that influent the friction factor
(f) in the case of drag reduction is a great task. Since it is influenced
by solvent physical and flow properties. Starting with the following relation:
ΔP = f (D, μ, ρ, V, C, L) 
Therefore by applying the dimensional analysis using Buckingham π
theorem, the following nondimensional relation was proposed:
Or:
The essential problem now is to find the values of the constants a, b
and k, that give the best fitting of the experimental data.
The resulting equation is as follows:
f = 82.3 (Re)^{0.9}. (C)^{0.12} 
(7) 
Figure 8 shows the relation between the observed values
of friction factor taken form experimental data and the predicted values
from mathematical correlation. It can be noticed that most points lie
at or close to the straight line, which means a good agreement between
theoretical and experimental data.
CONCLUSIONS
It was proven in the present investigation that the condition of additive
solubility is not always stands for all types of drag reducing agents.
The experimental results showed that the sodium stearate is working well
as drag reducing agent in nonaqueous media (kerosene) and the mechanism
of drag reduction for suspended solids in turbulent flow can be adopted
to explain this phenomenon.
It was proven that the percentage Drag Reduction (Dr%), increases by increasing
the flow rate inside the pipe (the degree of turbulence) and the addition concentration
and reduced by increasing the pipe diameter and all that due to the changes
in the turbulence media the drag reducer works with.
Comparing the experimental results with those obtained before ( by comparing
the friction factor) shoed good agreement.