Engineering Practice

Pipe Insulation: Finding the Optimal Thickness

Avoid tedious numerical calculation by using this quick, shortcut method
Alejandro Anaya Durand, Allan L tions that need to be solved numeri-
Osornio Maldonado, Alexander cally. After this, a shortcut method is
Álvarez Holly, Itzayana Montesinos presented, with examples to com-
Aguilar, Adalberto Fuentes pare the results of using the numeri-
Martínez and María Daniela Ávalos R1 cal versus the shortcut method.
Galván R2 Heat transfer. The general equa-
National Autonomous University of tion for the rate of heat transmission
Mexico, Faculty of Chemistry through a pipe having a homoge-
Pipe

I
neous insulation material can be de-
nsulated pipes are important in Insulation rived by starting with Fourier’s law,
many applications in the chemical represented by Equation (1):
Figure 1. The problem is to determine the opti-
process industries (CPI). When it mal insulation thickness, R2 – R1
comes to the insulation itself, the
two main questions are: which ma- pipelines, the cost of insulating ma-
terial should be used; and how much terials and more. (1)
of it (thickness)? Three variables are In order to find a system’s minimal
considered in order to find the most cost and therefore the optimum insu- Where:
economical thickness of an insulated lation thickness, numerical methods qi = the heat-transfer rate in the i di-
pipe: the inner radius of pipe, the are required. This process involves rection, Btu/h
cost of energy and the cost of the finding an equation describing the A = the area normal to the direction i
insulation material itself. cost as a function of the variables, (direction of the heat flow), ft2
Determining the most economical then setting the derivative of that dt/dxi = the temperature gradient in
thickness of insulation for a pipe- equation equal to zero. Although it the i direction, °F/ft
line is a problem of finding the right is possible to solve such a problem K = the thermal conductivity, Btu/h.
balance between using the least numerically, using software such as ft.°F
amount of insulating material with the Solver for Excel, it is tedious and For an insulated pipe of length L,
least amount of energy loss. When time consuming. (Figure 1) the heat transfer is in the
increasing the thickness of the insu- This article first outlines the numer- radial direction, R, and the area is
lating material, the energy loss de- ical approach, and then presents a 2πRL
creases, but the cost of the insulating simple shortcut method that can be
material rises. When decreasing the used as a quick, practical tool to de-
thickness of the insulating material, termine the most economical thick- (2)
the material cost decreases, but the ness of different insulating materials.
energy lost rises — that translates An example calculation using both Equation (2) is integrated from R1
into monetary losses. methods is then given, which shows (inside radius of pipe insulation) to
The equation that solves this prob- the same results are achieved. R2 (outside radius of pipe insula-
lem is complicated. It is a function of tion), and from T1 to T2, where T1
many variables, such as inside and Fundamentals is the temperature in the inside ra-
outside temperatures, cost of en- To demonstrate the complexity of the dius of insulation. For practical pur-
ergy, conductivities, nominal sizes of problem, we first develop the equa- poses, T1 is considered the same

Nomenclature
Ca Least squares fit of mineral wool costs (Table 2) mp Cost of heat loss, $/ft.yr
d0 Outer diameter of insulation = 2r2, in. np Cost of insulation per year, $/ft.yr
FT Adjustment factor for temperature qR Heat transfer rate in radial direction, Btu/h
FN Adjustment factor for pipe size R1, R2 Insulation inner and outer radius, ft
FM Adjustment factor for temperature r1, r2 Insulation inner and outer radius, in.
FD Adjustment factor for pipe size T1, T2 Temperature at R1 and R2, °F
h2 hc + hr Ta Ambient temperature, °F
hc Surface heat-transfer coefficient, radiation U2 Heat flowrate from R2 to outside air, Btu/h
hr Surface heat-transfer coefficient, convection U Heat flowrate (per foot) from R2 to outside air, Btu/h.ft
K Thermal conductivity, Btu/h.ft.°F Y Number of hours operation, h/yr
L Length of pipe insulation, ft yp Total cost (insulation plus energy), $/ft.yr
M Cost of heat, $/million Btu ξ Emissivity of insulation surface

Chemical Engineering www.chemengonline.com october 2015 61

 as the operating temperature inside the pipe. This gives
Equaton (3)

(8)

(3) Finally, an expression for Newton’s law of cooling is

given as Equation (9):
Because we are concerned with the heat flow between
the inside surface of the pipe insulation and the outside
air (ambient temperature = Ta), Equation (3) is divided by
the outer surface area of the insulation, and the right side (9)
expressed in terms of thermal resistances (the inverse
of the heat-transfer coefficients), to which is added the To summarize thus far, Equation (8) expresses the
thermal resistance due to surface radiation and surface heat flow (per foot) from r1 (the inner surface of the in-
convection. This gives Equation (4). sulation, in inches) to the environment, and this should
be equal to Equation (9), the heat flow (per foot) from
r2 (the outer surface of the insulation, in inches) to the
environment.
Costs. If Y is the number of hours of operation per year
and M is the cost of heat (in dollars per million Btu) then
the cost of heat lost for the insulated pipe (mp, $/ft.yr)
(4) can be expressed as Equation (10):

Where: (10)
Rs = 1/h2
h2 = hr + hc The next step is to find a function that describes the
cost of insulation per foot per year. This function will be
The heat flowrate per linear foot (U, Btu/h.ft) is found named np. For this, it is necessary to find the price of
by rearranging Equation (4) to give: different insulation materials. For each insulation mate-
rial, it was found that there are tables of prices for a
specific thickness and nominal pipe size. Table 1 shows
such information for mineral wool, for example. These
data were plotted for each pipe size, as shown in Figure
2. The y-axis is the cost per foot, and the x-axis is the
thickness of the insulation. In the context of the previous
(5) discussion, x = r2 – r1.
For the case of mineral wool, the plots of Figure 2 can
Expressions for the surface radiation, hr, and surface be fit to linear equations, as shown in Table 2. These
convection, hc, are given by Equations (6) and (7): linear relationships can be generalized as the function
Ca. Multiplying this function by the the cost of installation
(assume 4% per year) plus the annuity, which accounts
for the annual interest, i, for amortization over a period of
n years, gives the general equation for insulation costs
as follows:
(6) Mineral wool
35

30 Pipe size, in.

—1
(7) 25 —2
—3
Ca = Cost, $/ft

20 —4
Where: —5
d0 = surface diameter, in., = 2r2 15 —6
ξ = surface emissivity (≈1)
10

Equation (5) can be rearranged to give Equation (8), 5

which includes a conversion from inches to feet, since
0
pipe sizes and insulation thicknesses are normally ex- 0 0.5 1 1.5 2 2.5 3 3.5
pressed in inches. Insulation thickness, in.
Figure 2. The cost of mineral wool depends on the pipe size. The data can be
approximated by the function Ca (Table 2)
62 Chemical Engineering www.chemengonline.com october 2015
 Table 1. The cost of mineral wool insulation, $/ft
Iron pipe 1 2 3 4 5 6
size, in.
Insulation
(11) thickness, in.

For example, imagine we have a pipe with a nominal 1.0 5.133 6.727 7.383 8.167 9.256 9.867
size of 3 in. insulated with mineral wool. In this case, we 1.5 8.500 9.613 10.927 11.813 12.620 13.000
would use the equation y = 8.0272x – 0.5719 for Ca. If 2.0 11.607 13.607 16.440 17.247 19.143 19.900
we add 5 years as the time factor for depreciation and 3.0 20.340 21.600 23.167 25.660 28.450 31.283
amortization, with an annual interest for amortization of 6.0 42.800 44.020 47.591 52.394 58.038 66.793
15% and adding cost of installation per year of 4% (20%
between 5 years). Then np is: Table 2. Ca function for mineral wool*
Pipe size, in. A B Correlation coefficient, r2
1 7.6133 2.88 0.9933
2 7.5474 1.2648 0.9961
3 8.0272 0.5719 0.9906
To summarize thus far, the costs for energy and insula- 4 8.8903 0.9476 0.9968
tion material, yp, is given by Equation (12). 5 9.8596 1.1194 0.9922
6 11.05 2.2071 0.9888
(12)
*Least squares fit to data in Table 1 (Figure 2): Ca = Ax – B
Finding the economic thickness. For a particular pipe
with nominal diameter of r1, the value for the most eco-
nomical insulation thickness is determined by taking
the derivative of yp with respect to r2, and setting this
equal to zero:

(13)

For the previous example of the 3-in. pipe with mineral

wool insulation, Equation (13) becomes the following:

To solve Equation (13a), it is necessary to use a nu-

merical method. This is done by varying r2 until the ex-
pression on the left equals zero. But at the same time,
one must also vary T2 so that Equation (8) is equal to
Equation (9). When all of these conditions are met, the
value of r2 found is the optimum value, and thus gives
the optimal thickness (r2 – r1).
Obviously, solving Equation (13') is not an easy task.
And this expression is very specific to mineral wool in-
sulation for a 3-in. pipe. For other insulation materials,
it is necessary to find equations that give the price of
the material with thickness, as was done with mineral
wool (Table 2). To avoid this burdensome task, we have
developed a shortcut method to estimate the calculated
result. This shortcut is described below.
(13a)
Overview of the shortcut
Where the following constants (a–d) and the expression To develop the shortcut method, we start with a base-
for σ were defined to simplify mp in order to more easily case scenario in which all variables are known. Then, sev-
perform the differentiation: eral graphs are generated that show how a given variable
changes with respect to the others as they change from the
base case. Changing only one variable at a time enables us
to observe how the economic thickness behaves towards a
certain variable. With this information we can generate sev-
eral adjustment factors, which can be applied to the base
case in order to get the values for the scenarios of interest.

Chemical Engineering www.chemengonline.com october 2015 63

 3.5
Table 3. Values of FM
Temperature 3 — 2 million Btu
— 3 million Btu
Material — 4 million Btu
200°F 250°F 300°F 350°F 400°F 450°F 500°F — 6 million Btu
2.5
Mineral wool 1 1 1 1 1 1 1 — 8 million Btu

FT
Calcium silicate 1 1.02 1.02 1.02 1.01 1.01 1 2
Fiberglass 1 1.02 1.04 1.07 1.08 1.1 1.11
Cellular glass 1 1.08 1.13 1.14 1.15 1.16 1.15 1.5
Perlite 1 1.02 1.03 1.04 1.03 1.03 1.02
1
200 250 300 350 400 450 500
Table 4. Values of FD T1, °F
Nominal Figure 4. The adjustment factor for mineral wool, FT, for temperatures other
diameter than the base case (200°F)
Material
1 in. 2 in. 3 in. 4 in. 5 in. 6 in. 2.4
Mineral wool 1 1 1 1 1 1
2.2
Calcium silicate 1 0.98 0.88 0.77 0.8 0.81
2
Fiberglass 1 0.9 0.88 0.84 0.82 0.88
Cellular glass 1 0.91 0.87 0.9 0.81 0.85 1.8

FN
Perlite 1 0.97 0.79 0.87 0.77 0.80 — 2 million Btu
1.6 — 3 million Btu
— 4 million Btu
1.4 — 6 million Btu
The shortcut method — 8 million Btu
The shortcut method consists of using three graphs and 1.2
two tables to find the optimal thickness. We consider 1
five different insulation materials, but the the graphs and 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
tables can be further developed to handle any type of Pipe nominal dia., in.
insulation material. Figure 5. The adjustment factor for mineral wool, FD, for pipe sizes other than
The factors that we need to know for this method are the base case (1 in.)
the following:
• Cost of energy, $/million Btu
• Temperature at R1, °F temperature of 200ºF.
• Pipe size, nominal dia., in.
• Type of insulation material
As a starting point, we choose a base case to general-
ize many instances graphically. We select 200ºF as the (14)
temperature at radius r1, a 1-in. pipe and an ambient
temperature of 77ºF. The graphs in Figure 3 show the A plot of FT versus T1 is shown in Figure 4 for different
behavior of the optimal thickness of insulation for a pipe- energy costs (mineral wool is the reference case here).
line versus energy costs for the five different insulation In a similar fashion, a correction factor, FN, is determined
materials for this base case. for when the nominal pipe size is different from 1 in., in
For temperatures other than 200ºF at radius r1, it is order to find the optimal insulation thickness.
necessary to adjust the optimal thickness using a fac-
tor called FT. This factor is determined from the char-
acteristics of mineral wool and is basically a relationship
between thickness at one temperature (T1) and the base (15)
0.7
— Mineral wool
Figure 5 shows the behavior of FN with nominal pipe
0.6 — Fiberglass diameter for various energy costs (for mineral wool).
0.5
— Calcium silicate Because each insulation material has different charac-
— Cellular glass
teristics, two tables were generated to provide correc-
Thickness, in.

0.4 tion factors for estimating the optimal thickness. To cor-

0.3 rect for a material other than mineral wool, the factor FM
is used for a temperature at r1 other than 200°F (Table 3),
0.2 and FD is used to correct for pipe diameters other than
0.1 1 in. (Table 4).
Therefore, a final equation that can be used to find the
0
2 3 4 5 6 7 8
optimal insulation thickness is given by Equation (16):
Cost, $/million Btu
Figure 3. The optimal thickness for various insulation materials depends on
the energy cost when compared to the base case (16)
64 Chemical Engineering www.chemengonline.com october 2015
 4.00 mineral wool. From Authors
3.50 the graph we find Alejandro Anaya Durand (Parque
that the base case España No. 15b Col. Condesa,
3.00 C.P. 06140, Mexico,D.F. Email:
optimal thickness aanayadurand@hotmail.com) is a
2.50 is 0.16 in. professor of chemical engineering
Yp, $/ft/yr

2.00 Step 2. From Fig- at the National Autonomous Uni-

ure 4, determine versity of Mexico (UNAM), and has
1.50 over 56 years of experience in
1.00
the value of FT to projects and process engineering.
correct the pipe He retired from Instituto Mexicano
0.50 del Petroleo in 1998 after holding top positions. For 50
temperature (T1)
years he has been an educator in chemical engineering
0.00 from the base tem- at several universities in Mexico, and presently he is
0 0.1 0.2 0.3 0.4
Thickness, in.
0.5 0.6 0.7 perature of 200°F. also a consultant at several engineering companies. He
Figure 6. For the example problem, numerical calculations show that the For x = 350ºF, the has published over 280 articles related to engineering
and education; is a Fellow of the AIChE; a member of
total cost, yp, is a minimum at 0.464 in. intersection of the National Academy of Engineering; and has received the
line that corre- main chemical engineering awards in Mexico. He holds
To demonstrate the use of this sponds to the given cost of energy a M.S. in project engineering from UNAM.
shortcut method, the following ex- ($2/million Btu) is at FT = 1.75. Allan Ricardo Osornio Mal-
ample calculation is presented, and Step 3. From Figure 5, determine the donado (Tizoc No. 5, San Miguel
the results are compared to numeri- value of FN to correct the pipe diam- Xochimanga, Zip Code 52927,
cal calculations. eter from the base size of 1 in. For x= Atizapan Zaragoza, Estado Méx-
ico, México; Phone: +52 1
2 in. the intersection of the line that 5535745303; Email: alan_rich-
Example corresponds to the given cost of en- ard15@hotmail.com) is a ninth-
Consider a 2-in. dia. steel pipeline ergy ($2/million Btu) is at FN = 1.56. semester chemical engineering
honor student at UNAM.
that carries steam at 350°F. The am- Step 4. Use Table 3 to obtain the value
bient temperature is 70°F. Assume for FM. For mineral wool, FM = 1.
the price of the energy is $2/million Step 5. Use Table 4 to find the value Alexander Álvarez Holly
(Avenida Acequia 30-1, Colonia
Btu. The time factor for depreciation for FD. For mineral wool, FD = 1. Villa Coapa Zip Code 14390, Del-
and amortization is 5 years, annual Step 6. Use Equation (16), and mul- egación Tlalpan, D.F., México;
interest for amortization is 15% and tiply the factors found in the previ- Phone: +5219982133296;
Email: alexanderalvarezholly@
there is an additional cost of instal- ous steps by the base case optimal gmail.com) is a ninth-semester
lation per year of 4% (20% over 5 thickness to obtain the optimal thick- honor student of chemical engi-
years). Find the optimal thickness for ness = 0.16 × 1.75 × 1.56 × 1 × 1 neering at UNAM. Currently doing
an internship at Petroleos Mexi-
mineral wool. = 0.4368 in. canos Internacional.
Rigorous method. First we use Compared to the rigorous numeri- Itzayana Montesinos Aguilar
the numerical method discussed cal method, the shortcut method (Conejos 107 Lomas de Lindavista
above. Assuming an average ther- gives an error around 5.8%. And El Copal, Tlalnepantla de Baz, Edo.
mal conductivity of 0.025 Btu/h.ft.°F since insulation with 0.4369 in. thick- Mex., México; Phone: +52 1
5541867542; Email: montesinos.
and solving Equation (8) for r2 such ness is not available, we would need itzayana@gmail.com) is a chemi-
that Equation (8) equals Equation (9) to purchase the next available size, cal engineering honor student
gives an optimal value for r2 of 1.65 which is 0.5 in — the same result as from UNAM. She is currently doing
her ninth semester in Spain as an
in., and the temperature T2 is found calculated numerically. exchange student at Universitat
to be 125.7°F. That means the opti- de Barcelona. Process design and consulting are found
mal thickness is r2 – r1 = 0.464 in. For Concluding remarks among her interests.
these optimal values of r2 and T2, the The shortcut method presented here Adalberto Fuentes Martínez
heat loss, U, is equal to 95.3 Btu/h. has been shown to be much simpler (Calzada de las Arcadas No. 59
Int. 17, Colinas del Sur, Alvaro
ft. A plot of Equation (12) shows that and faster that the rigorous numerical Obregón, DF, México; Phone: +52
the total cost of insulation plus the method for determining the optimal 1 5532315778; Email: betofm8@
cost of heat loss, yp, is minimized to insulation thickness. The method can gmail.com) is a ninth-semester
honor student of chemical engi-
$2.63/ft at a thickness of 0.464 in. also be expanded to include other in- neering at UNAM and works as an
(Figure 6). sulation materials and a wider range intern at Petroleos Mexicanos In-
Since we cannot buy insulating of pipe sizes. n ternacional. He has personal inter-
material of 0.40 in.-thickness, we Edited by Gerald Ondrey est in the production of energy processes.
would need to purchase the next María Daniela Avalos Galván
available size, which is 0.5 in. References (Andador de los Esteros Módulo
Shortcut method. Next, we solve 1. Bird, R. B., Stewart, W. E. and & Lightfoot, E. N., “Trans- 40 Depto 401, Gustavo A. Madero,
DF, México; Phone: +52 1
the same example problem, but port Phenomena,” (2nd ed.), John Wiley and Sons, New
5522983447; Mail: maddy.
York, N.Y., 2006.
using the shortcut method. ag38@gmail.com), is a ninth-se-
2. Kern, D., “Procces of Heat Transfer,” McGrawHill, New
Step 1. Obtain the base case opti- York, N.Y., 2001.
mester chemical engineering
honor student at UNAM. She par-
mal thickness from Figure 3. For the 3. Turner, W. C. and Mallory, J. F. “Handbook of Thermal ticipated in the internship program
x-axis value use the given cost of en- Insulation Design Economics for Pipes and Equipment,” at the General Direction of Com-
ergy ($2 /million Btu) and intersect it 1st ed. , Krieger Pub. Co., 1980. puting and Technologies of Information and Communi-
with the line that corresponds to the 4. Welty, J., and others, “Fundamentas of Momentum, Heat cation (DGTIC) UNAM and now she’s part of the inter-
ship program at Agilent Technologies.
desired material, which in this case is and Mass Transfer,” (5th ed.), John Wiley and Sons, New
York, N.Y., 2007.

Chemical Engineering www.chemengonline.com october 2015 65