Tech Tip

Consistent Engineering Units In Finite Element Analysis

Consistent Engineering Units In Finite Element Analysis

Depending on the modelling software you use for your Finite Element Analysis (FEA), your modelling
process often requires a very high level of understanding of engineering units.
Note that although your FEA user interface may provide a convenient method of declaring which
unit you are applying at any instant, it is important to know that almost every FEA solver does not
show units in the "input file".
Therefore, it is critical to understand units if you are:

taking a model from somewhere else;

reviewing someone else's FEA work, or simply
using any FEA modeller which leaves the units up to you.

When the units are left up to the user, the important point to note is the necessity to use
"consistent units".

SI Units
The easiest set of consistent units is the SI system of metres (meters), kilograms and seconds.
These derive associated units that are commonly used by designers/engineers/analysts, such as
Newtons, Joules, Watts, Pascals, plus generically titled items such as density, acceleration,
velocity, viscosity and many more.
The important point about using consistent units is the necessity to stick with units that work
correctly together - not to mix units that do not have a correct relationship with each other.
For example, if using the basic SI units of metres, kilograms and seconds, then:

Density must be in kg/m ;

2
Acceleration must be in metres/sec ;
Force must be in Newtons, because a Newton is the force it takes to accelerate 1kg at the
2
rate of 1m/sec . (Force = mass x acceleration, ignoring relativity effects of course), and
Power must be in Watts (Newtons x metres) per second.

Many engineering companies instruct their engineers to use basic SI units so there are no
questions about unit conversion or consistent units. However, this has its downsides, particularly
when using FEA.
Many people find the use of metres inconvenient, and so the use of millimetres (millimeters) by
CAD designers is very popular. Similarly, there are some particular problems to be avoided when
using metres with the very popular Nastran "small field" format. With only 8 characters wide per
field, a location at -1.015mm will be described as -1.02e-3 (metres) which is effectively only 3
significant figures. For models with fine details or where precise contact is involved, precision
down to only 1/100 of a millimetre is not enough. In extreme cases, closely spaced nodes can be
written out at the same coordinate location, which will typically produce fatal error messages.
So, as an aside, if you are using metres and Nastran, we recommend you use large field format if it is
not the default in your modelling system. We also recommend using large field format if you are
modelling fine details which are "distant from the model origin" for similar reasons of precision.
Anyway, back to units. If you make the popular choice of millimetres, and prefer using Newtons for
force, then this has specific consequences in your use of all other units to preserve consistency.

www.cadimageplm.com

The obvious consequences of choosing Newtons and millimetres are:

Applied Pressures are Newtons / mm , which is MegaPascals (MPa)

Stress results are thus also MPa.
Youngs Modulus is also in units of pressure, and is thus also MPa.

What is less obvious is that in choosing millimetres and Newtons, you have also forced a change in
your mass units. This is because F=ma, and so the question which needs to be asked is:
2

"What mass accelerates at 1mm/sec when a force of 1N is applied?"

1mm/sec2 is very slow (1000 times less than 1m/sec2, of course) - the answer is 1 Tonne.
Therefore, density is in Tonnes per cubic mm - thus 1e-9 for water and 7.8e-9 for a typical steel.
So, if you are running an analysis where gravity is involved (having chosen millimetres and Newtons),
2
then the magnitude of acceleration is 9810 (mm/sec ), point masses are Tonnes and density is in
3
Tonnes / mm .
Additional consequences of choosing millimetres and Newtons occur for heat transfer:

Conductivity in standard SI: Watts per (metre x Kelvin)

Conductivity in your consistent units: milliWatts per (millimetre x Kelvin). (Therefore no change in
the numeric value input)
Heat Flux in Standard SI: Watts per square metre
Heat Flux in your consistent units: milliWatts per square millimetre. Therefore numeric input
would be reduced by a factor of 1000 compared to Standard SI.
2
2
The same conversion is required for a convection coefficient - ie. W/m K becomes mW/mm K
so numeric input reduces by a factor of 1000.
Specific Heat in Standard SI: Joules per (kilogram x Kelvin)
Specific Heat in your consistent units: milliJoules per (Tonne x Kelvin). Therefore numeric
input would be increased by a factor of 1e6 compared to Standard SI. eg. 4180 [J/kgK]
becomes 4.18e9 [mJ/TK]

Note that results such as heat flux will obviously be presented in the same consistent units you
have chosen to apply.

Imperial Units
For those of you who use Imperial Units, you have our sincerest condolences, however the same
set of principles apply. A consistent set of units must be used in order to produce a valid analysis.
For engineering, a basic set of imperial units is slug (mass), foot (length) and seconds (time).
It is critical to note that a pound is not a unit of mass for engineers and should not be used as a
unit of mass.
A pound is a unit of force

www.cadimageplm.com

A pound is the unit of force required to accelerate a slug at "1g", which is 32.19 feet/sec . As most of
us do not size each other up by estimating a person's "weight" in slugs, it is thus useful to know that
a mass of one slug weighs about 32 pounds force on earth (from the definition above).
Again, for many people, the use of feet for length is inconvenient, so inches are often used
instead, which thus means some conversions are required to preserve a consistent set of units.
The obvious consequence of using inches is that applied pressures, stress results and Youngs
Modulus are all in psi (pounds per square inch). This is common, as many sources supply material
constants in these units.
Not so obvious is that mass is now in units of "12 x slugs". Although some have named this mass unit
the "slinch" or "blob", we prefer the colloquially applied name, the "snail". (Thanks to Dr Tim Coates of
GKN Aerospace for supplying this useful name for the unit - we are unsure as to whether the term has
earlier origins).
-4

Thus density must be in units of snails per cubic inch, which for a typical steel would be 7.3e . A
snail weighs about 386 pounds force on earth (it is a large snail).
If doing thermal analysis using Imperial Units (in inches), then consistent units could be applied by
using:

inch pounds for Work

inch pounds per second for Power
(inch pounds per second) per square inch for Heat Flux
inch pounds per (snail x degree) for Specific Heat.

As most texts do not conveniently supply material constants or coefficients or equations using
these units (even densities are often misquoted as weight densities), then some additional
conversion and care is required.
Below is a table of some common mechanical and thermal units with their equivalent values in
four columns of commonly used consistent unit sets for FEA. Note that it is possible to use
"hybrid" combinations of consistent units, where certain properties (eg. energy, power, temperature)
may use different self-consistent units values compared to the remainder of the units set. This
requires caution if the physics is coupled and our conservative recommendation is not to use such
an approach. On that basis, the table ignores the possibility of doing so.
Note that due to some minor variations in the definitions of some values (such as g, BTU and
derived properties), other sources may show a difference of a few units variation in the 4th
significant figure for some of the values in the table. For the approximations associated with FEA,
such variations are insignificant.

www.cadimageplm.com

Each column represents a typical consistent unit set for FEA. The length units of your geometry dictate which single
consistent column set to use.
SI, Standard:
N,m,kg,s,K

SI, mm: N,mm,T,s,K

Imperial, inches: Lbf, Imperial, feet:

in, snail,s,F
Lbf,ft,slug,s,F Slug
snail=slinch,blob
= 32.19 "lb mass"
=386.2 "lb mass"

1Property/value
pound (Force)

4.449 Newton

4.449 N

1.000 lbf

1.000 lbf

1 Newton (Force)

1.000 N

1.000 N

0.2248 lbf

0.2248 lbf

1 kg (Mass)

1.000 kilogram

1.000e

-3
-2

1 slug (Mass)

14.60 kg

1.460e

1 snail (Mass)

175.2 kg

0.1752 T

Tonne

5.709e

8.333e

-3
-2

6.851e

snails

1.000 slug

1.000 snail

4.535e-

1 metre (Length)

1.000 m

1000 mm

39.37 in

3.281 ft

1 inch (Length)

0.0254 m

25.4 mm

1.000 in

8.333e

1 foot (Length)

0.3048 m

304.8 mm

12.00 in

1.000 ft

1 Pascal (Pressure)

1 N/m

1 psi (Pressure)

6896 Pa

1 MPa (Pressure)

1.000e

1 g (Acceleration)

9.810 m/sec

1 m/sec (Accel.)
2

1 ft/sec (Acceleration)
3

1kg/m (Density)
3

1 lbm/ft (Density)
3

1.000e
6.896e
Pa

1.000 m/sec

16.02 kg/m
4

1.000e

145.0 psi

20885 lbf/ft

2.768e

-12
-11
-8

39.37 in/sec

12.00 in/sec

T/mm

T/mm

9.356e

1.498e

-6
-3

32.19 ft/sec

3.281 ft/sec

snails/in

1.000 ft/sec
3

snails/in

3.107e

-3
-2

2
3

slugs/ft

slugs/ft

8.849 in.lbf

0.7374 ft.lbf

1 BTU (Energy)

1000 N.mm
ie.
mJ 6 mJ
1.055e

9334 in.lbf

777.8 ft.lbf

1 Calorie (Energy)

4.186 Joules

4186 mJ

37.04 in.lbf

3.086 ft.lbf

1 Watt (Power)

1.000 N.m/sec ie.

Watt
0.2930 Watts

1000 N.mm/sec ie. mW 8.849 in.lbf/sec

0.7374 ft.lbf/sec

293.0 mW

2.593 in.lbf/sec

0.2161 ft.lbf/sec

mW

9334 in.lbf/sec

777.8 ft.lbf/sec

mW

6600 in.lbf/sec

550.0 ft.lbf/sec

37.04 in.lbf/sec

3.086 ft.lbf/sec

1.055e

1 HP (Power)

745.8 Watts

7.458e

1 Calorie/sec (Power)

4.186 Watts

4186 mW

1 Watt/m (Heat Flux)

2

1.000 W/m

1 BTU/hr/ft (Heat Flux) 3.155 W/m

2
2

1.000e
3.155e

-3
-3

mW/mm
mW/mm

-3

53.69 slugs/ft

1.000 N.m ie.

Joule
1,055 Joules

1055 Watts

snails/in

1.940e

1 Joule (Energy)

1 BTU/sec (Power)

2.589e

-8

2.768e

T/mm

386.2 in/sec

lbf/ft

1 lbm/in (Density)

1 BTU/hr (Power)

kg/m

1.602e

144.0 lbf/ft

304.8 mm/sec

-2

1.000 psi

1000 mm/sec

2.088e

ft

MPa
2

psi

-2

slugs

1.450e

9810 mm/sec

3.107e

MPa

1.000 MPa

0.3048 ft/sec
1.000 kg/m

-3

-4

snails

-2

0.4535 kg

-6

2.589e

slugs

12.00 slugs

1 "pound mass" lbm

-3

-2

snails

-2

5.709e
2
-2
in.lbf/sec.in
1.801e

6.851e
2
ft.lbf/sec.ft
0.2161

in.lbf/sec.in

ft.lbf/sec.ft

Continued overleaf

www.cadimageplm.com

SI, Standard:
N,m,kg,s,K

SI, mm: N,mm,T,s,K

Imperial, inches: Lbf, Imperial, feet:

in, snail,s,F
Lbf,ft,slug,s,F Slug
snail=slinch,blob
= 32.19 "lb mass"
=386.2 "lb mass"

1
HP/ft2 (Heat Flux)
Property/value

8028 W/m2

8.028 mW/mm2

45.83 in.lbf/sec.in2

550.0 ft.lbf/sec.ft2

1 calorie/sec/in2
(Heat Flux)

6488 W/m2

6.488 mW/mm2

37.04 in.lbf/sec.in2

444.5 ft.lbf/sec.ft2

Temperature given in Kelvin K = #F x 5/9 + 255.37 K = #C x 1 + 273.15

(#K)
C = #F x 5/9 - 17.778 C = #K x 1 - 273.15
Celsius (#C) Fahrenheit (#F)

F = #K x 1.8 - 459.67 F = #C x 1.8 + 32

1 J/kg.K (Specific Heat)

1.000 N.m/kg.K

1.000e6 N.mm/T.K

861.1
in.lbf/snail.F

5.980
ft.lbf/slug.F

1 BTU/lbm.F (Specific
Heat)

4186 J/kg.K

4.186e9 mJ/T.K

3.605e6
in.lbf/snail.F

2.503e4
ft.lbf/slug.F

1 kcal/lbm.F (Specific Heat) 1.662e4 J/kg.K

1.662e10 mJ/T.K

14.31e6
in.lbf/snail.F

9.936e4
ft.lbf/slug.F

1 W/m.K (Conductivity)

1.000 W/m.K

1.000 mW/mm.K

0.1249
in.lbf/sec.in.F

0.1249
ft.lbf/sec.ft.F

1 BTU/hr.ft.F
(Conductivity)

1.730 W/m.K

1.730 mW/mm.K

0.2161
in.lbf/sec.in.F

0.2161
ft.lbf/sec.ft.F

1 W/m2.K
(Convection)

1.000 W/m2.K

1.000e-3
mW/mm2.K

3.172e-3
in.lbf/sec.in2.F

3.806e-2
ft.lbf/sec.ft2.F

1 BTU/hr.ft2.F
(Convection)

5.678 W/m2.K

5.678e-3 mW/mm2.K

1.801e-2
in.lbf/sec.in2.F

0.2161
ft.lbf/sec.ft2.F

1 kcal/hr.ft2.F
(Convection)

22.53 W/m2.K

2.253e-2 mW/mm2.K

7.146e-2
in.lbf/sec.in2.F

0.8576
ft.lbf/sec.ft2.F

Our recommendations:
ALWAYS use consistent units. That means the base units of Mass, Length, Time, Temp must
comply with any/all derived associated units you are using in your analysis (such as pressure, force,
velocity, density, acceleration, convection etc.).
Use metric if you can - the availability of material data and constants in workable imperial units
(particularly in thermal and fluids analysis) typically requires arduous conversion to maintain a
consistent set of units for analysis. However, units conversion of any sort is best minimised, so you
do not become the person blamed for losing the next Mars Climate Orbiter!
PARAM, WTMASS is available in NASTRAN so analysts can use force units as masses (eg. nonstructural mass in pounds per square foot). We believe this practice should be rigorously
avoided (except by analysts of the highest expertise), and that proper consistent units are always
used.
These guidelines have been assembled by and for Nastran users; however, the information is
equally applicable to users of Femap, Ansys, Abaqus, MSC.Software, MSC Marc, Patran, Ideas,
Strand, Algor, Cosmos, CosmosWorks and many others.

www.cadimageplm.com