You are on page 1of 9

Proceedings of the 2014 10th International Pipeline Conference

IPC2014
September 29 - October 3, 2014, Calgary, Alberta, Canada

IPC2014-33019

Measurements and Evaluation of Internal Wall Surface Roughness


of Small Diameter Pipes for High Pressure Natural Gas Systems

C. Hartloper, K.K. Botros, H. Golshan K. Jensen


J. Geerligs TransCanada Pipelines ATCO Pipelines
NOVA Research & Technology Limited. Edmonton, Alberta, Canada
Center Calgary, Alberta, Canada
Calgary, Alberta, Canada
colin.hartloper@novachem.com
kamal.botros@novachem.com

ABSTRACT NOMENCLATURE

The default roughness parameter values used in industry to Acronyms


determine the pressure loss through small diameter pipeline GERG Groupe Europeen de Recherches Gazieres
systems are much higher than the values employed in typical NIST National Institute of Standards and
large diameter gas transmission and lateral systems. It is Technology
uncertain whether these higher roughness values are due to REFPROP NIST Reference Fluid Thermodynamic and
higher topological roughness of the internal wall of the small Transport Properties Database
diameter pipes or if they are a result of other factors.
Measurements were taken on 17 small diameter pipe samples Symbols
in order to evaluate the pipe-wall roughness parameter. A A Cross-sectional flow through area of pipe
model to calculate the effective roughness parameter, which Ao Cross-sectional area of orifice
takes into account pressure losses due to the measured Ar Orifice area ratio
roughness as well as internal welds and scaling, has been Dh Hydraulic diameter of pipe
developed. The effective roughness parameter of these Di Inner diameter of pipe
samples is found to range from 20.4μm to 62.9μm, an f Darcy friction factor
increase of 11.0μm to 23.3μm over the measured pipe-wall feff Effective friction factor
roughness parameter. This range of effective roughness ff Fanning friction factor
parameters agrees well with the default range of 35μm to L Interweld spacing
65μm used in industry, as well as the literature quoted range P Pipeline model gas pressure
for clean pipe of 40μm to 100μm. The measured roughness ΔPf Pressure loss due to pipe-wall friction
parameter on average increases with increasing nominal pipe ΔPgw Pressure loss due to girth weld
size, a result that may be a characteristic of the extrusion or ΔPo Pressure loss due to orifice-like flow area
hot-rolling processes used to manufacture small diameter reduction at girth weld protrusion
pipes. Additionally, there is a large variation in the measured ΔPs Pressure loss due to scales
roughness parameters of pipe samples of the same nominal ΔPsw Pressure loss due to longitudinal seam weld
pipe size, indicating that surface roughness can vary ΔPtot Total pressure loss
depending on the manufacturing source of these pipes. pi Perimeter of pipe inner wall
Ra Arithmetic mean roughness parameter

1 Copyright © 2014 by ASME


Rz Peak-to-valley roughness parameter example, hot rolling and extruding result in higher surface
Re Reynolds number based on pipe hydraulic roughness than cold rolling. As such, the method used to
diameter manufacture pipe can influence the pipe’s internal surface
T Pipeline model gas temperature roughness and thus its frictional pressure loss. The Reynolds
v Pipeline model gas velocity number, a critical flow parameter for pipe applications, is
δgw Girth-weld protrusion proportionate to pipe diameter and thus would be lower in
δsw Seam-weld protrusion small diameter pipes. This could cause the flow regime in the
Δδgw Uncertainty in girth-weld protrusion pipe to be transitional, which would lead to a higher friction
Δδsw Uncertainty in seam-weld protrusion factor and thus higher pressure loss compared to a fully
ε Nikuradse equivalent sand grain roughness turbulent flow regime [5]. Pipe sections are joined together
parameter (referred to as the pipe-wall by means of girth welds, which would protrude further into
roughness parameter) the flow area of small diameter pipes relative to pipes of
εeff Effective roughness parameter larger diameter, acting as a periodic constriction to the flow
εs Scale roughness parameter and causing additional pressure losses [6].
Δε Uncertainty in pipe-wall roughness parameter
ζ Orifice loss factor
ζr Loss factor for rounded orifice
ζs Loss factor for sharp orifice
μ Pipeline model gas viscosity
ρ Pipeline model gas density
τ' Rounded orifice correction factor

INTRODUCTION
Figure 1: Schematic representation of the cold rolling
In this paper, a small diameter pipe system is defined as that process, adapted from Degarmo et al. [2].
which comprises of pipes with nominal pipe size ranging
from 1¼” – 4”. For these small diameter pipelines, roughness
parameter values between 35μm and 65μm are typically used
in industry in order to simulate the pressure loss in pipe
systems. These values are much higher than the roughness
parameter value of 15μm – 20μm typically used to simulate
the pressure loss in uncoated large diameter pipes. It is
uncertain whether these high roughness parameter values are
in fact due to higher topological pipe-wall roughness or if
they are due to other sources of pressure loss in the small
diameter pipe systems. Since this larger than expected
pressure loss impacts the flow capacity and/or leads to higher
operating cost, further investigation is warranted.
Figure 2: Schematic representation of a hot rolling and
There are many possible causes for the relatively higher than piercing process, adapted from Brensing et al. [3].
expected pressure loss for small diameter pipe. Pipes can be
manufactured either welded or seamless [1]. For welded pipe,
the pipe is generally rolled from plate as shown in Figure 1 The objective of the present paper is to identify the root
[2]. The subsequent welding process depends on the pipe cause(s) of the apparent high pressure loss observed in small
diameter. For smaller diameters, a hot pressure welding diameter pipe systems. With this purpose in mind, the
process is often applied. For mid-range diameters, an electric following sections are organized as follows. First, a model
resistance technique is used, and for large diameters electric used to quantify the pressure loss in small diameter pipes,
arc welding is commonly employed [3]. Seamless pipe is which was developed based on measurements on 17 pipe
either extruded or manufactured by a combination of piercing samples, is explained in detail. Next, the pipe sample
and hot rolling as shown in Figure 2 [3]. These processes are measurements as well as the total pressure loss estimated by
not economical for large diameter pipes, therefore seamless the pressure-loss model are presented. Here, an effective
pipe usually do not exceed 660mm in diameter [3]. All of the roughness parameter is calculated for each pipe sample and
above-mentioned manufacturing techniques result in varying compared to both industry standard and literature reported
internal surface roughness of the pipe [4]. Figure 3 shows the values. Finally, the results from the samples are compared to
average surface roughness resulting from a variety of one another and conclusions are drawn on the root cause of
manufacturing techniques, where it can be seen that, for the high pressure loss in small diameter pipes.

2 Copyright © 2014 by ASME


Figure 5 shows a schematic of a typical pipe section with
inner diameter Di, pipe-wall roughness parameter (ε), scale
roughness parameter (εs), girth welds with associated
protrusion (δgw) and spacing (L), and seam welds with
associated protrusion (δsw). Based on this simplified diagram
a model can be constructed from which the pressure loss
from each of the above-mentioned sources can be calculated
on a per-length basis. From this model, the total pressure loss
can be used to calculate an effective roughness parameter
(εeff).

Figure 4: Images of 1½” pipe samples illustrating the


seam weld (left) and girth weld with associated scales
(right).
Figure 3: Average surface roughness as a result of various
manufacturing processes, adapted from Oberg et al. [4].

PIPE SAMPLES AND PRESSURE LOSS MODEL

A total of 17 small diameter pipe samples were analyzed. The


samples were 600mm in length and had nominal pipe sizes
between 1¼” and 4”. Five of the samples (two 1¼”, one 1½”,
one 2” and one 3”) were centered on a girth weld, the rest of
Figure 5: Schematic depicting the model used to calculate
the samples were taken away from a girth weld. Six of the
the total pressure loss in small diameter pipes.
samples (one 1¼”, three 1½”, one 3” and one 4”) had a
longitudinal seam weld, while all of the other samples were
seamless. Thirteen of the samples (three 1¼”, three 1½”, In the following subsections, the estimation of the pressure
three 2”, three 3” and one 4”) were Schedule 40, two of the loss due to the pipe-wall roughness, the girth weld and the
samples (one 1¼” and one 1½”) were Schedule 80 and the seam weld is described in detail. Note that, since the scales
other two samples had a non-standard wall thickness. are produced by the weld splatter from the girth weld, the
pressure loss of the girth weld is taken to be the sum of the
Images in Figure 4 show the seam weld on a 1½” pipe pressure loss due to the orifice-like cross-sectional area
sample on the left and the girth weld on a 1½” pipe sample reduction of the weld protrusion and the effect of the scales
on the right. Scales that were produced by the girth weld due to the girth weld splatter. After each of the mechanisms
splatter can be seen in the vicinity of the girth weld in Figure contributing to the pressure loss is described, the calculation
4. Clearly, the seam weld, girth weld and scales will cause of the effective roughness parameter is presented.
pressure losses in the pipe system in addition to the pressure
loss due to the pipe-wall roughness. The seam weld reduces 1) Pressure Loss Due to Pipe-Wall Roughness
the hydraulic diameter of the pipe, while the girth weld
protrusion acts similar to an orifice to reduce the cross- The pressure loss due to the pipe-wall roughness (ΔPf, in Pa)
sectional area of the pipe along the width of the weld. The is [5]:
scales serve to increase the pipe-wall roughness. However, it
is not immediately apparent how the magnitudes of these
additional pressure losses compare to the pressure loss due to (1)
the pipe-wall roughness.

3 Copyright © 2014 by ASME


where f is the Darcy friction factor, ρ is the fluid density (in The loss-factor for a round orifice protrusion is:
kg/m3) and v is the fluid velocity (in m/s). For turbulent pipe
flow f is a function of Reynolds number (Re) and the relative
( ) √ ( ) ( ) (8)
roughness (ε/Di). The original correlation of the Darcy
friction factor was given by Colebrook [7] as:
where τ’ is defined as:
( ) (2)
( ) (9)
Equation (2) was subsequently plotted as a function of Re
and ε as the Moody diagram. At moderate Re the flow regime Figure 6 shows a comparison of the pressure drop between
is transitional and equation (2) is implicit. However, at high hypothetical sharp and round orifices with identical area
Re the flow is considered fully turbulent. In the fully ratio. Here it can be seen that, for all area ratios, ΔPo is higher
turbulent flow regime, Reynolds number effects can be for a sharp orifice than for a round orifice. Consequently, the
neglected and f is simply: girth weld was chosen to be modeled as a sharp orifice since
a sharp orifice give a more conservative pressure-loss
( ) estimation and the girth weld protrusion (see Figure 4) was
(3) shaped more like a sharp orifice than a round orifice.

2) Pressure Loss Due to Girth Weld – Orifice and Scales

The pressure loss due to the girth weld is subdivided into the
pressure losses due to both the orifice-like cross-sectional
area reduction due to the girth weld protrusion (ΔP o, in Pa)
and the scales resulting from the girth weld splatter (ΔP s, in
Pa):

(4)

From Idelchik [6], the general formula for pressure loss due
to an orifice is: Figure 6: Modeling a hypothetical girth weld as a sharp
orifice versus a round orifice.

(5)
3) Pressure Loss due to Scales

The scales caused by the weld splatter on the girth weld serve
where ζ is the orifice loss factor and depends on the orifice to increase the pipe-wall roughness parameter in the
protrusion shape. Two orifice shapes could potentially model proximity of the girth weld. Idelchik [6] suggests a
the cross-sectional area reduction due to the girth-weld roughness parameter value of approximately 400μm for
protrusion – round or sharp. The loss-factor for a sharp stainless steel pipe with small depositions of scale. As
orifice protrusion is: indicated in Figure 5, this higher roughness parameter is
applied to only the region with scale depositions, observed to
be 6” on both sides of the girth weld in the samples
[ ( ) ( )] (6) considered. Then, the pressure drop due to the scales can be
calculated as:
where Ar is the area ratio between the reduced cross-sectional
area (Ao, in m2) at the orifice and the pipe cross-sectional ( ) ( ) (10)
area (A, in m2):

where ΔPf(εs) and ΔPf(ε) are calculated from equation (1)


using the scale roughness parameter and measured pipe-wall
( ) (7)
roughness parameter, respectively.

4 Copyright © 2014 by ASME


4) Pressure Loss due to Seam Weld to provide a conservative estimate for pressure loss, it was
assumed that all samples had girth and seam welds. The
The seam weld serves to decrease the hydraulic diameter (D h, girth- and seam-weld protrusions were measured on all pipe
in m) of the pipe. From visual inspection, the seam weld has samples that had girth and seam welds, and the mean of these
a roughly square shape. The seam weld protrusion serves to measured protrusions was used for all samples. A
both reduce the cross-sectional area of the pipe as well as conventional natural gas composition was used in the present
increase the inner-wall perimeter (pi, in m). Thus, the analysis (Table 1). The mean flow velocity (v), temperature
hydraulic diameter is reduced from Di to: (T) and pressure (P) of the gas were assumed to be 15m/s,
10°C, and 5MPa-a, respectively. With these conditions, the
( ) GERG-2008 Equation of State [8], as implemented in
(11) REFPROP [9], was used to calculate the gas density (ρ) and
viscosity (μ) to be 43.1kg/m3 and 11.6x10-6 Pa-s,
respectively.
Thus the pressure drop due to the seam weld is:
Table 1: Model natural gas composition
( ) ( ) (12) Component % by Mole
Methane 91.502
Ethane 4.679
where ΔPf(Dh) and ΔPf(Di) are calculated from equation (1) Propane 1.329
using the hydraulic diameter and inner diameter, respectively. n-Butane 0.232
i-Butane 0.174
5) Total Pressure Loss and Effective Roughness Parameter i-Pentane 0.055
n-Pentane 0.040
Once all of the individual pressure loss mechanisms have Hexane 0.039
been accounted for, the total pressure loss can be calculated Nitrogen 0.762
by simple addition: Carbon
1.186
Dioxide
(13)
Roughness Measurements

Next, an effective Darcy friction factor can be calculated by The results of the measurements performed on all 17 pipe
rearranging equation (1) and substituting ΔP tot in for ΔPf: samples are presented in Table 2, where each of the samples
is assigned an arbitrary sample number. The Nikuradse
equivalent sand grain roughness (ε), also known as the pipe-
(14) wall roughness parameter, was calculated as:

Finally, the effective roughness parameter can be found (16)


either iteratively from equation (2) if the flow is in the
transitional/turbulent regime or can be calculated by
where the peak-to-valley roughness parameter (Rz) was
rearranging equation (3) if the flow is fully turbulent:
measured in 24 locations on each pipe sample according to
the procedure outlined in Botros et al. [10]. These 24
( )
(15) measurements were then averaged to obtain the Rz parameter
for each sample. The girth and seam weld protrusions were
measured with callipers in two locations on each sample. Δε
represents the 95% confidence interval for the pipe-wall
roughness parameter measurements. Samples with a * do not
RESULTS have a girth weld and samples with a ** do not have a seam
weld.
The model described above was used to estimate the total
pressure loss and effective roughness parameter. A weld Table 2: Measured pipe dimensions.
spacing of L = 10m, which is on the lower end of the typical
manufactured pipe length, was assumed. A total of 17 NPS # Di ε Δε δgw δsw Sch
samples, with nominal pipe sizes ranging from 1¼” to 4”, (mm) (μm) (μm) (mm) (mm)
were investigated. Recall that not all pipe samples contained 1.25" 1 35.4 56.0 22.2 2.1 ** 40
girth or seam welds, which according to the pressure loss 2 35.6 9.4 5.5 * ** 40
model, will contribute to additional pressure losses. In order 3 35.4 25.9 8.3 2.1 ** 40

5 Copyright © 2014 by ASME


4 32.4 11.5 3.2 * 1.0 80
1.5" 5 37.9 15.6 11.2 * ** 80
6 41.7 11.1 4.8 * 1.1 40
7 41.6 20.3 19.4 2.7 1.1 40
8 41.3 11.2 3.4 * 0.8 40
2" 9 52.7 22.9 32.3 * ** 40
10 53.1 32.1 12.2 * ** 40
11 52.4 15.4 21.3 1.0 ** 40
3" 12 78.0 37.4 14.6 * ** 40
13 78.5 31.8 18.2 * ** 40
14 78.2 27.3 18.5 1.3 ** 40
15 80.9 86.4 40.9 * 1.0 NS
4" 16 102.6 39.7 18.6 * ** 40
17 104.5 39.4 16.8 * 1.0 NS
Figure 8: Measured roughness parameter, and associated
95% confidence interval, for the 17 pipe samples.
From the calliper measurements, the girth weld protruded
between 1 to 2.7mm, while the seam weld protruded between
0.8 to 1.1mm. The measured roughness parameter from all of First, in general, the roughness parameter increases with
the samples is between 9.4 and 86.4μm. Two samples, nominal pipe size. In pipe manufacturing, almost all small
numbers 1 and 15, were identified as outliers due to their diameter pipe are manufactured by either extrusion or hot
high roughness. Upon inspection, these samples were rusted rolling [3]. The trend of increasing pipe-wall roughness
excessively on their internal surface. Figure 7 compares the parameter with nominal pipe size may therefore be a
measured roughness profile from a typical, non-rusted characteristic of the extrusion and/or hot rolling
sample (#3) to a sample that is excessively rusted (#1). Here manufacturing techniques.
it can be seen that the roughness profile of the rusted sample
is much less uniform, displaying large peaks and valleys that Second, the roughness parameter of the majority of samples
contrast the consistent profile of the non-rusted sample. is greater than 15 – 20μm, which is a typical roughness
parameter for large diameter pipes. Therefore, the topological
internal surface roughness of small diameter pipes is larger
than that of large diameter pipes. Recall that the hot rolling
and extrusion techniques, that are more commonly used to
manufacture small diameter pipe, result in higher surface
roughness than the cold-rolling technique, which is more
commonly used to manufacture large diameter pipe. Thus, it
is likely that the difference in the methods used to
manufacture small and large diameter pipe is a root cause for
the higher than expected pressure losses through small
diameter piping systems.
Figure 7: Comparison between the internal-surface
profile of a rusted (#1, ε=56.0μm) and non-rusted (#3, Third, in Figure 8 there is a large deviation observed between
ε=25.9μm) sample. the measured pipe-wall roughness parameter of pipe samples
of similar nominal pipe size. Recall from Figure 3 that the
average roughness of material produced from the same
Figure 8 visually presents the roughness parameter manufacturing method can vary by up to an order of
measurements, and associated 95% confidence intervals, for magnitude, depending on the specific implementation of the
all 17 samples. Several distinguishing factors are identified method. The deviation in roughness parameter of similar
for each sample, such as: the nominal pipe size, the pipe sized pipe samples supports this, and suggests that the
schedule, whether the sample was seamed or seamless, and similarly sized samples were manufactured from different
whether the sample was rusted or non-rusted. Three mills. Consequently, the mill at which pipe is manufactured
observations can be seen in Figure 8. should be taken into account when designing small diameter
pipe systems.

Based on the measurements and typical pipeline operating


conditions, the Reynolds number based on the internal
diameter of the pipe samples ranges from 2.0x10 6 to 5.8x106.
Figure 9 shows the Moody chart with the flow-regime range

6 Copyright © 2014 by ASME


of the current investigation highlighted in red. Here it can be
seen that the flow regime for all of the pipe samples ranges
between the transitional/turbulent and fully turbulent zones.
Since the flow is almost fully turbulent, transitional effects
will not significantly increase the friction factor or the
frictional pressure loss. Therefore, the higher apparent
pressure drop in small diameter pipes is likely not due to the
lower Reynolds number for flow through these pipes.

Figure 10: Pressure loss per meter due to seam weld (top),
girth weld (middle) and pipe-wall friction (bottom) for
weld spacing of L=10m. Error bars indicate the 95%
confidence interval.

Figure 9: Moody chart, adapted from White [5], with


Effective Roughness Parameter
flow-regime parameter space highlighted in red.
Equation (15) was used to calculate the effective roughness
Total Pressure Loss parameter based on the total pressure loss for each pipe
sample. Figure 11 presents εeff alongside ε for all pipe
The pressure losses due to the pipe-wall roughness, girth samples. Here it can be seen that, although the pressure
weld and seam weld were calculated for all 17 pipe samples losses due to the girth and seam weld were found to be small
shown in Table 2. The pressure losses are presented in relative to the frictional pressure loss, these pressure losses
Figure 10. Note that equation (2) was used to calculate the increase the effective roughness parameter significantly. For
pressure loss due to pipe-wall roughness. Also, note that the all pipe samples, εeff is between 11 – 23μm higher than ε.
95% confidence intervals in the girth- and seam-weld Excluding the rusted samples, the maximum εeff is 62.9μm,
protrusions, Δδgw and Δδsw, were calculated based on the for the seamless 4” sample. The overall highest maximum
uncertainty in the girth- and seam-weld protrusion εeff is 119.6μm, for the rusted, seam-welded 3” sample. While
measurements. it is unlikely for an in-service pipe to be this rusted, this
number represents a good worst-case scenario for pipe-wall
As shown in Figure 10, the girth and seam welds cause a roughness parameter. Although this worst-case roughness
small pressure loss relative to the pipe-wall friction, with the parameter is very high relative to the 20μm used as a default
girth and seam welds together representing a maximum of for large diameter pipes in industry, it is only marginally
about 10% of the total pressure loss. As expected from the larger than 100μm, which is the upper range of roughness
form of equation (1), the frictional pressure loss decreases parameter values for welded steel tubes in good condition
with an increase in nominal pipe size (i.e., increase in Di). cited in the literature [6].
The total pressure loss ranged between 0.8kPa/m and
3.2kPa/m, with the minimum and maximum pressure loss for
the 4” and 1¼” pipe samples, respectively.

7 Copyright © 2014 by ASME


thus between 17.9μm and 67.5μm, which agrees well with
the default range used in industry of 35μm to 65μm. Since
the pressure loss due to the girth- and seam-weld protrusions
increases as the protrusion gets larger, control of the girth-
and seam-weld protrusions during pipe manufacture and
pipeline assembly is critical, as higher than necessary
pressure losses have negative economic implications.

CONCLUDING REMARKS

1. The measured pipe-wall roughness parameter of the 15


non-rusted small diameter pipe samples ranged from
9.4μm to 39.7μm. The pipe-wall roughness parameters
Figure 11: Effective roughness parameter (open symbols) of the two rusted small diameter pipe samples were
for all pipe samples is plotted with reference to the 56.0μm and 86.4μm.
measured roughness parameter that was originally shown 2. For the non-rusted samples, taking into account the
in Figure 8. effects of the girth and seam welds, the effective
roughness parameter of the small diameter pipe samples
increased by 11.0μm to 23.3μm over the measured pipe-
wall roughness parameter. The effective roughness
DISCUSSION parameters ranged from 17.9μm to 67.5μm, a range that
agrees well with the default roughness parameter range
The results of the study suggest that, for small diameter of 35μm to 65μm used in industry.
pipes, the pipe’s roughness parameter increases with nominal 3. The extrusion and hot-rolling techniques that are used to
pipe size. Since this relationship between roughness manufacture small diameter pipe result in higher pipe-
parameter and nominal pipe size is not present on large wall roughness compared to cold-rolling techniques that
diameter pipes, which most often manufactured by cold- are used to manufacture large diameter pipe. This may
rolling methods, this relationship seems to be a characteristic explain the higher pipe-wall roughness parameter on
of the extrusion and/or hot rolling methods that are often small diameter pipes compared to large diameter pipes,
used to manufacture small diameter pipes. It is unknown up which typically have roughness parameters between
to which nominal pipe size this relationship would continue, 15μm – 20μm.
since the largest pipe in the current study was 4”. 4. The measured roughness parameter of small diameter
Nevertheless, an interesting trend is noted from comparing pipes, with nominal pipe size ranging from 1¼” to 4”,
the sample’s roughness parameter (ε) to their relative increases as a function of inner diameter. This result may
roughness (ε/Di). Including all of the 15 non-rusted samples, be a characteristic of the manufacturing techniques used
the average ε is 23.4μm with a standard deviation of 10.5μm for small diameter pipes.
(45% of the average), while the average ε/Di is 4.1x10-4 with 5. Pipe produced from different mills can show significant
a standard deviation of 1.3x10-4 (30% of the average). Since variance in pipe-wall roughness parameter. This
the standard deviation of the sample’s ε/Di is relatively conclusion is supported by the wide range in pipe-wall
smaller than the standard deviation of the sample’s ε, if one roughness parameters measured on seamless pipe
wanted to characterize the roughness of all small diameter samples.
pipes with a single value, it would be more representative to 6. Transitional effects do not contribute significantly to the
use a single relative roughness rather than a single roughness pressure loss through small diameter pipes as the flow
parameter. For example, for this study, it is more precise to regime is almost fully turbulent even through 1¼”
say that all samples have a relative roughness of 4.1x10 -4 diameter pipes.
than it is to say that all samples have a roughness parameter
of 23.4μm. This result has implications in the design of small
diameter pipe systems, as it may be more appropriate to
assume a single relative roughness throughout the system ACKNOWLEDGEMENTS
when estimating pressure losses over the system.
The author wishes to acknowledge the discussion and
For small diameter pipes, it is shown that the girth- and technical dialogue with Gene Poissant and Landen Stein on
seam-weld protrusions cause the effective roughness the operation of both small and large diameter pipe systems.
parameter to be 11.0μm – 23.3μm higher than the pipes The support of TCPL Technology Management; Thomas
roughness parameter, which was measured to range between Robinson, Anthony Tse and Tracy Cairns during the course
9.4μm and 39.7μm. The effective roughness parameter is of this work is greatly appreciated. This paper is part of a

8 Copyright © 2014 by ASME


research program sponsored by TransCanada Pipelines
Limited., and permission to publish it is gratefully
acknowledged.

REFERENCES
1. M. L. Nayyar, Piping Handbook, 7th ed., McGraw-Hill, 2000.
2. E. P. Degarmo, J. T. Black and R. A. Kohser, Materials and
Processes in Manufacturing, 9th ed., John Wiley & Sons, 2002.
3. K. Brensing and B. Sommer, "Steel Tube and Pipe
Manufacturing Processes," Salzgitter Mannersmann
Rohrenwerke, [Online]. Available:
http://www.smrw.de/files/steel_tube_and_pipe.pdf. [Accessed
26 August 2013].
4. E. Oberg, F. D. Jones, H. Horton and H. H. Ryffel, Machinery's
Handbook (29th Edition), 29th ed., Industrial Press, 2012.
5. F. White, Fluid Mechanics, 6th ed., McGraw-Hill, 2008.
6. I. E. Idelchik, Handbook of Hydraulic Resistance, 3rd ed.,
CRC Press, 1994.
7. C. F. Colebrook, "Turbulent flow in pipes, with particular
reference to the transition region between smooth and rough
pipe laws," Journal of the Institution of Civil Engineers
(London), 1939.
8. O. Kunz, R. Klimeck, W. Wagner and M. Jaeschke, "The
GERG-2004 Wide-Range Equation of State for Natural Gases
and Other Mixtures," Groupe Europeen de Recherches
Gazieres (GERG), Technical Monograph, GERG TM15, 2007.
9. "NIST Standard Reference Database 23, NIST Reference Fluid
Thermodynamic and Transport Properties Database
(REFPROP): Version 9.1," [Online]. Available:
www.nist.gov/ts/msd/srd/nist23.cfm. [Accessed 30 July 2013].
10. K. K. Botros, M. Piazza and D. Abayarathna, "Tools and
Methods for Internal Pipe Coating Evaluation for Gas
Transmission Systems," in Proceedings of IPC 2012, Calgary,
Canada, September 24-28, 2012.

9 Copyright © 2014 by ASME

You might also like