TP003

Measurement Canada
Volumetric Engineering
Technical Paper 003
Gravimetric Calibration of
Volumetric Standards
Issued: June 2, 1999
Revised Dec 1
st
, 2009
By: Christian Lachance, P. Eng.
Senior Engineer - Liquid Measurement
Measurement Canada, Volumetric Engineering Page 2 of 25
Technical Paper 003
Gravimetric Calibration of Volumetric Standards Dec 1
st
, 2009
Table of Contents
1.0 The Gravimetric Calibration of Volumetric Standards . . . . . . . . . . . . . . . 3
2.0 Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Water Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Mass Determination of the Delivered Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.1 Method A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.2 Method B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.3 The Use of Apparent Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.0 Gravimetric Calibration Uncertainty Overview . . . . . . . . . . . . . . . . . . . . 12
4.0 Uncertainty of the Base Volume Determination . . . . . . . . . . . . . . . . . . . . . 13
4.1 Volumetric Standard Temperature Uncertainty, u(T
shell
) . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Cubical Coefficient of Expansion Standard Uncertainty, u(") . . . . . . . . . . . . . . . . . 15
4.3 Water Density u(D(T
w
)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3.1 Water Temperature Standard Uncertainty, u(Tw) . . . . . . . . . . . . . . . . . . . 16
4.4 Mass Determination Standard Uncertainty, u(Mw) . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.4.1 Air Density Standard Uncertainty, u(D
a
) . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.4.1.1 Air Temperature Uncertainty, u(Ta) . . . . . . . . . . . . . . . . . . . . . . 20
4.4.1.2 Relative humidity uncertainty u(RHa) . . . . . . . . . . . . . . . . . . . . . 20
4.4.1.3 Atmospheric pressure uncertainty, u(Pa) . . . . . . . . . . . . . . . . . . . 21
4.4.2 Reference Mass Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.4.3 Reference Mass Density Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.4.4 Comparator Uncertainty and Method Repeatability . . . . . . . . . . . . . . . . . . 21
4.4.5 Water Density Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.0 Volumetric Standard Reproducibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.1 Leveling Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.2 Scale Plate Reading Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.3 Wetting Variance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
6.0 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Technical Paper 003
st
, 2009
( ) ( )
V V T T
Tref T ref shell
= + 1 o
( )
( ) ( )
V
M
T T
Tref
w
Tw
ref shell
= +
o 1
1.0 The Gravimetric Calibration of Volumetric Standards
Common large volumetric standards are calibrated to deliver a known volume. Gravimetric
calibration of volumetric standards consist of weighing water delivered by the volumetric
standard and then deriving the volume from the mass and density of the water. Since the volume
of the standard is dependent on temperature, the reported value is corrected to a reference
temperature.
2.0 Mathematical Model
2.1 General
Volume is determined from mass and density by the following equation:
(2.1) V
M
T
T
=
where T = temperature
V
T
= volume @ T
M = absolute mass
D
T
= density @ T
The volume of the standard is a function of the temperature such that:
(2.2)
where V
Tref
= volume of standard @ Tref
V
T
= volume of standard @ T
T
ref
= reference temperature
" = cubical coefficient of expansion of the standard
T
shell
= temperature of the standard
Combining (2.1) and (2.2) we obtain:
(2.3)
Technical Paper 003
st
, 2009
( )
( ) ( )
( )
SMOW
Tw a
Tw a Tw a
a Tw a
kg m =
+ +
+
(
(
5
1
2
2
3 4
3
1
where M
w
= absolute mass of water
T
w
= temprature of water
= density of water @ T
w
( ) Tw
2.2 Water Density
Water density values are obtained from published formulation by Tanaka
1
where the density is
provided as a function of temperature and isotopic content. The formulation provides the density
values and their uncertainties for Standard Mean Ocean Water (SMOW) at a pressure of 101 325
Pa.
(2.4)
where a
1
= -3.983 035
a
2
= 301.797
a
3
= 522 528.9
a
4
= 69.348 81
a
5
= 999.974 950
D
SMOW
(Tw) = Standard Mean Ocean Water density
SMOW is air free water characterized by isotopic content. Most sources of water will have
isotopic content that differ from SMOW. When the highest accuracy is required, the isotopic
abundance of the water source must be determined and the density correction applied.
The expected difference between aerated and air free water density is calculated as per Bignell
2
with the following equation:
D
aerated
= -0.004612 + 0.000106 Tw (2.5)
Where Tw is in degrees C and D
aerated
is in kg/m
3
.
Water that has been distilled and not exposed to air is considered air-free water while water that
has been exposed to air will become air saturated. The gravimetric calibration process and
distilled water storage at MC laboratories are such that the water can be assumed to be fully
saturated. The density of the water at Tw is then:
Technical Paper 003
st
, 2009
( ) ( ) ( ) Tw Tw Tw
SMOW aerated
= + + A A
Where
)D(Tw) = is the deviation in the actual water density from SMOW with correction for
aeration.
2.3 Mass Determination of the Delivered Water
Mass can be determined by the single substitution method using an electronic single pan balance.
Two methods will be presented. Both methods require a container which can be the volumetric
standard or a separate container into which the water is delivered.
2.3.1 Method A
Method A is a general purpose method where the weight of the water is obtained by subtracting
the empty container weight from the weight of the filled container.
This method is a 4 step process:
1. Weighing of the empty container, record:
I
E
2. Weighing of reference weights close in value to that of empty
container, record:
I
RE
M
RE
D
R
Technical Paper 003
st
, 2009
3. Weighing of container and water, record:
D
a
I
F
D(Tw)
4. Weighing of reference weights close in value to that of container and the water, record:
I
RF
M
RF
D
R
Where
M
w
= absolute mass of water delivered or contained
M
E
= absolute mass of container empty
M
F
= absolute mass of container filled
M
RE
= absolute mass of reference weight for empty container
M
RF
= absolute mass of reference weight for filled container
D(Tw) = density of water @ Tw
D
R
= density of reference weights
D
a
= density of air during weighing
Technical Paper 003
st
, 2009
W gM
a
x
=
|
\
|
.
| 1

K
x
x
a
=
|
\
|
.
| 1

I
E
= scale indication of empty container
I
RE
= scale indication of reference weights for container
I
F
= scale indication of a filled container
I
RF
= scale indication of reference weight for filled container
W
w
= weight of water
W
F
= weight of filled container
W
E
= weight of empty container
The mass of water can then be calculated from the recorded data.
The weight of an object is related to its mass by
(2.7)
where
W = weight of object
g = acceleration due to gravity
M = absolute mass of object
D
a
= density of air surrounding object
D
x
= density of object
let
(2.8)
For an electronic single pan balance the indication, I, is proportional to the weight applied, such
that:
I = $ M K (2.9)
Technical Paper 003
st
, 2009
( )
M K
I M K
I
E E
E RE RE
RE E
=
1 A|
I M K
E I E E
E
= |
I M K
RE I RE RE
RE
= |
M K
I M K
I M K
E E
E RE RE
RE RE RE E
=
A|
where $ is a calibration factor which includes g.
For the weighing method presented we have for the weighing of the empty container:
(2.10)
(2.11)
The ideal balance will have a value of $ that is constant for all weighings. In practice, however,
a number of factors will cause this parameter to vary. Of concern are the effects of the balance
repeatability, non- linearity, stability and off-center loading. These effects and their contribution
to the weighing uncertainty can be mitigated by proper weighing method. Other factors such as
ambient temperature, loading method and magnetism may be significant in certain applications
or with certain equipment. These should be investigated to ensure that they are accounted for in
the determination of the uncertainty if required.
In order to account for the effects of linearity, eccentricity and stability in the estimate of the
uncertainty, the difference between the sensitivity coefficients of the artefact and the subsequent
reference mass weighing for the empty container is written as:
(2.12) | | |
I I E
RE E
= A
Combining (2.10), (2.11) and (2.12) we obtain:
(2.13)
Since the )$ terms is a small value and M
RE
K
RE
is close in value to I
RE,
(2.13) can be
approximated by:
(2.14)
the same is true for M
F,
where:
Technical Paper 003
st
, 2009
M K
I M K
I
F F
F RF RF
RF F
=
( ) 1 A|
M K
I M K
I
I M K
I
W W
F RF RF
RF F
E RE RE
RE E
=

( ) ( ) 1 1 A A | |
M
K
K
I M
I
I M
I
W
R
W
F RF
RF F
E RE
RE E
=
|
\
|
.
|
( ) ( ) 1 1 A A | |
M K M K M K
w w F F E E
=
W W W
w F E
=
K K K
RF RE R
= =
(2.15)
The weight of the water delivered is obtained by:
(2.16)
Combining (2.7) , (2.8) and (2.16) we obtain:
(2.17)
Combining (2.17), (2.14) & (2.15) we obtain:
(2.18)
If reference weights of the same density are used for the filled substitution and the empty
subsitution, then:
(2.19)
By combining (2.18) and (2.19), we obtain:
(2.20)
The mass of the water is obtained by assuming a value of 0 for the )$ terms.
Technical Paper 003
st
, 2009
I I M K
F E I I W W
F E
=

|
| | |
I I I I
RF E F E

= + A
I I M K
RF E I I R R
RF E
=

|
( )
M
I I
I I
K
K
M
W
F E
RF E
R
W
R
=

( )
( )
1 A|
2.3.2 Method B
Method B tares the weight of the empty container thus only requiring one substitution weighing.
It can be used when the substitution weight can be weighed with the empty container.
This method is a 3 step process:
1. Weigh the empty container, record Ie ( please note that this step is last when the standard test
measure is used as the container).
2. Weigh the filled container, record: I
F
, D
a.
3. Weigh the empty container and reference weight close in value to the water delivered, record:
I
RF ,
M
R ,
D
R
.
Because reference weights are added directly to the tare, the filled container indication less the
tare indication will be proportional to the reference weights such that:
(2.21)
and
(2.22)
In order to account for the effects of linearity, eccentricity and stability in the estimate of the
uncertainty, let:
(2.23)
combining (2.21), (2.22) and (2.23) and since the )$ term is a very small value and M
R
K
R
is
close in value to ( I
RF
- I
E
), we obtain the following:
(2.24)
The mass of the water is obtained by assuming a value of 0 for the )$ term.
Since the weighing is done at filled load, the error introduced by the scale linearity, stability
Technical Paper 003
st
, 2009
and eccentricity M
R
/(1-)$) can be approximated with M
F
/(1-)$
F
). This is convenient if the
estimate of )$
F
is available while )$ is not.
2.3.3 The Use of Apparent Mass
Apparent mass is related to absolute mass by the following:
(2.25) M M
apparent
a
apparent
a
1 1
|
\
|
.
|
|
=
|
\
|
.
|
Where: M = absolute mass

M
apparnt
= apparent mass
D
apparnt
= apparent density
D
a
= density of air
D = density of mass
The apparent mass of an object is thus a function of air density. The difference in the value of
the apparent mass due to a difference in air density can be obtained by:
(2.26)
A
A
M M
M
apparent
a
apparent
a apparent
c
c
~ ~
|
\
|
.
|
|
1 1
The difference of the apparent mass value, or resulting error, due to differing air density in
calibration and in use is:
(2.27) ( )
Error
apparent
air during cal ai during use
=
|
\
|
.
|
|

1 1

_ _ _ _
This error is small in relation to the uncertainty of the mass. For example for a situation where
the apparent density is 8000 kg/m3 but the actual mass density is 8600 kg/m3 and where the
difference between the air density during the calibration of the mass and it use differs by 0.1
kg/m3 the error will be less than 1 ppm. Therefore the following approximation can be made:
Technical Paper 003
st
, 2009
(2.28)
M M
use
apparent
use
apparrent
cal
1 1
|
\
|
.
| ~
|
\
|
.
|
|
,
Where M
apparnt
, D
cal
= apparent mass determined in calibration air density.
3.0 Gravimetric Calibration Uncertainty Overview
Gravimetric calibration gives rise to uncertainty due to two key factors: base volume
determination and the volumetric standard reproducibility.
The mathematical model for the base volume determination is provided by equation (2.3), it
covers the following influence factors:
1) uncertainty in the mass determination;
2) uncertainty in the water density determination;
3) uncertainty of the cubical coefficient of expansion; and
4) uncertainty of the temperature of the volumetric standard.
The uncertainty of the base volume determination is obtained by root sum square of the influence
factors such that:
(3.1) u u u u u
T M
shell W W Volume determination
2 2 2 2 2
= + + +
o ( ) T
The amount of volume delivered by a volumetric standard can be expected to vary, this
reproducibility characteristic is due to:
1) leveling of the standard (u
leveling
);
2) reading of the standard (u
reading
); and
3) wetting and drip variances.
The uncertainty due to the reproducibility of the standard will be the combined effect of the
above. Since these effects are independent they can be combined by a root sum square:
(3.2)
u u u u
reproducibility
leveling reading wetting
2 2 2 2
= + +
Technical Paper 003
st
, 2009
( ) u V
fV
u
c Tref
i
n
Tref
i
i
2
1
2
2
( ) =
|
\
|
.
|
=
c
c_
_
The combined standard uncertainty is the reproducibility of the standard combined with the
uncertainty of the base volume determination. These can be combined by root sum square as per
the following equation:
(3.3)
u u u u u
u u u
V
T M
leveling reading wetting
Tref
shell W W
2 2 2 2 2
2 2 2
= + + +
+ + +
o ( ) T
4.0 Uncertainty of the Base Volume Determination
The mathematical model for the base volume determination was given in (2.3). Although T
w
and
T
shell
are assumed to be the same value, we can also assume the error of the two values to be
independent since the error of T
shell
is mostly due to the measurement process while the error of
T
w
is mostly due to the thermometer reading uncertainty. We will thus consider all influence
factors independent. The combined standard uncertainty can then be calculated by the
following:
(4.1)
where:
u
c
= combined standard uncertainty
x
i
= influence quantities
u(x
i
) = standard uncertainties attributed to the influence quantities
The following table lists the influence quantities, their sensitivity coefficients and the square of
the uncertainty contribution of the influence factor. The uncertainty terms have been
approximated, where appropriate.
Technical Paper 003
st
, 2009
( )
+
|
\
|
.
|
|
M
T T
T
w
ref shell
w
1
2
o
( ) ( )
M
Tw
u T
w
w

2
2
( )
( ) A
|
\
|
.
|
c
c_
V
Tref
i
( ) u
V
u
x
Tref
i
i
i
2
2
2
~
|
\
|
.
|
c
c_
_
M
T T
Tw
w
ref shell
|
\
|
.
|
( )
( )
+
|
\
|
.
|
|
M
T T
T
w
ref shell
w
1
2
o
( )
( )
1+ T T
T
ref shell
w
o
( )
Influence quantity Sensitivity coefficient Square of the uncertainty
contribution of the
influence factor (the
sensitivity coefficient has
been replaced by an
equivalent approximation
where appropriate)
P
i
T
shell
M
Tw
w
o
( )
|
\
|
.
|

(V " u(T
shell
) )
2
"
( ) ( )
V T T u
ref shell
( ) o
2
( )
SMOW w
T
( )
M
Tw
u T
w
SMOW w

2
2
( )
( )
|
\
|
.
|
M
W
( ) u M
Tw
w
( )
|
\
|
.
|
|
2
)D(Tw)
4.1 Volumetric Standard Temperature Uncertainty, u(T
shell
)
The volumetric standard temperature, Tshell, is assumed to be the same as that of the water Tw.
The uncertainty of this value is influenced by the thermometer calibration uncertainty,
thermometer drift and, for the most part, the uncertainty due to the temperature measurement
method. The latter is a function of the following factors:
- the assumption that the shell temperature is the same as that of the water, this is influenced
by the temperature difference between ambient and the water;
- possible temperature non-uniformity in the standard.
Technical Paper 003
st
, 2009
The volumetric standard temperature uncertainty is:
(4.2) u T u calibration u drift u method
shell
2 2 2 2
( ) = + +
Where:
u calibration can be obtained from calibration certificate;
u drift can be obtained from manufacturer specifications, based on calibration history or based
on ice point verification tolerance;
u method is estimated based on expected ambient conditions and water temperature differential.
4.2 Cubical Coefficient of Expansion Standard Uncertainty, u(")
The uncertainty of the assumed cubical coefficient of expansion is estimated at +/- 10 %:
(4.3) ( )
u o
o
=
01
3
.
4.3 Water Density u(D(T
w
))
The uncertainty of the water density determination is a function of the water temperature
measurement uncertainty, the density model uncertainty and the level of purity of the water
sample.
The uncertainty of the water density due to the measured temperature is obtained as follows:
(4.4) ( ) ( ) u u T
SMOW w
SMOW w
T
w
w
c
c
( )
( )
T
T
= -
The uncertainty of the SMOW density formulation is reported at 0.9 ppm. However, certain
isotopic ratio must be measured to obtain this level of uncertainty. Where the isotopic ratios
have not been measured a value of 20 ppm is suggested
4
as appropriate.
The correction for aerated water density is approximately 2 to 3 ppm at room temperature. Since
the correction is small relative to the 20 ppm uncertainty of air free water mentioned above, the
uncertainty of this correction can be neglected.
Technical Paper 003
st
, 2009
( )
( ) u Tw
kg m
A =
0020
3
3
. /
In some instance, where large volumetric measure are calibrated and where the facilities do not
provide a large supply of purified water such as the Measurement Canada gravimetric calibration
laboratory, purified water can be re-circulated and re-used. In order to assess the quality level of
the purified water used for gravimetric calibrations, the water density can be monitored by
comparing the density of representative samples to that of ultrapure water such as that produced
at the NRC facility. Comparisons should be conducted using a high precision densitometer. For
very high accuracy applications the results can be used to estimate a correction for the water
density when the reference water sample is traceable.
Unless a density correction is used, the uncertainty due to deviations from SMOW ,with
correction for aeration, is assumed to be +/- 20 ppm so that:
(4.5)
If a correction is used, the uncertainty of the correction must be estimated from the reference
sample stated uncertainty and the densitometer measurement repeatability.
4.3.1 Water Temperature Standard Uncertainty, u(Tw)
The uncertainty of this value results from the thermometer calibration uncertainty and the
uncertainty due to the temperature measurement method. The latter is a function of the
following factors:
- water temperature uniformity in the standard;
- self-heating effect of the probe in static water.
The volumetric standard temperature and water uncertainty is obtained by the root sum square of
the above mentioned factors:
W
2 2 2 2
( ) = + +
Where:
Ucalibration can be obtained from the thermometer calibration certificate;
Udrift can be obtained from thermometer manufacturer specifications, calibration history or it
can be estimated from ice point check tolerance;
Umethod can be estimated from temperature gradients in the standard and self heating
characteristics of the thermometer.
Technical Paper 003
st
, 2009
( )
I M
I
K
K
F RF
RF
R
w
F
|
\
|
.
|
|
\
|
.
|
|
1
1
2
A|
( )
I M
I
K
K
E RE
RE
R
w
F
|
\
|
.
|
|
\
|
.
|
|
1
1
2
A|
( ) M u
E E
2 2
A |
c
Mw
i
i
=
c
c
_ ( ) ~
|
\
|
.
| -
c
c_
_
M
u
w
i
i
2
2
M
I
K
K
RF
RF
R
w
|
\
|
.
|
|
\
|
.
|
( ) u I
F
2
M I
K
K I
RF F
R
W RF
|
\
|
.
|
|
\
|
.
|
1
2
( ) u I
RF
2
M
I
K
K
RE
RE
R
w
|
\
|
.
|
|
\
|
.
|
( ) u I
E
2
M I
K
K I
RE E
R
w RE
|
\
|
.
|
|
\
|
.
|
1
2
( ) u I
RE
2
4.4 Mass Determination Standard Uncertainty, u(Mw)
The uncertainty of the mass determination is again determined as per equation (2.20) and (2.24)
for methods A and B respectively.
The following table lists the influence quantities and their associated sensitivity coefficients for
method A:
contribution of the
influence factor (the
sensitivity coefficient has
been replaced by an
equivalent approximation
where appropriate)
P
i
I
F
I
RF
)$
F
( ) M u
F F
2 2
A |
I
E
I
RE
)$
E
Technical Paper 003
st
, 2009
I
I
K
K
F
RF
R
w
|
\
|
.
|
( )
( )
I
I
M
I
I
M
E
RE
RE
F
RF
RF
w r w
r w a
|
\
|
.
|

|
\
|
.
|
|

2
( ) M u
W
w R
a
2
2
2
1 1

|
\
|
.
|
I
I
M
I
I
M
K
E
RE
RE
F
RF
RF
a
R W
|
\
|
.
|
|
\
|
.
|
2
( ) u M
R W
a
R
2
2
2
|
\
|
.
|
( )
( )
I
I
M
I
I
M
E
RE
RE
F
RF
RF
R a
W a
|
\
|
.
|

k

2
( ) M u
W
a
w
w

2
2
2
|
\
|
.
|
( ) u M
RF
2
I
I
K
K
E
RE
R
w
|
\
|
.
|
( ) u M
RE
2
D
a
D
r

D
w
M
RF
M
RE
Technical Paper 003
st
, 2009
( ) I u
RF F
2 2
A |
c
c
_
Mw
i
( ) ~
|
\
|
.
| -
c
c_
_
M
u
w
i
i
2
2
M
I I
K
K
R
RF E
R
W
( )
( ) u I
F
2
M I I
I I
K
K
R F E
RF E
R
w
( )
( )
|
\
|
.
|
2
( ) u I
RF
2

|
\
|
.
|
M I I
I I
K
K
R RF F
RF E
R
w
( )
( )
2
( )
( ) I I
M
u I
RF F
RF
E
2
2
2
( )
( )
I I
I I
M k
F E
RF E
R R
w r w
r w a
|
\
|
.
|

|
\
|
.
|
|

2
( ) M u
R
w R
a
2
2
2
1 1

|
\
|
.
|
( ) u M
R RF
a
R
2
2
2
|
\
|
.
|
( )
M K
I I
I I
R R
F E
RF E
a
W a
|
\
|
.
|

2
( ) M u
R
a
W
W

2
2
2
|
\
|
.
|
I I
I I
F E
RF E
R
w
a
R
M
K
|
.
|
|
\
|
.
|
|
\
|
.
|
2
The following table lists the influence quantities and their associated sensitivity coefficients for
method B:
contribution of the
influence factor
P
i
I
F
I
RF
)$
F
see text section 2.3.2
I
E
D
a
D
r
D
w
Technical Paper 003
st
, 2009
I I
I I
K
K
F E
RF E
R
w
|
\
|
.
|
( ) u M
R
2
( ) u RHa
Tolerance
=

3
M
R
4.4.1 Air Density Standard Uncertainty, u(D
a
)
Air density is calculated using the equation provided by Davis
3
. The uncertainty of the air
density is a function of air temperature, humidity and barometric pressure.
The uncertainty of the air density can be estimated by the following equation:
(4.7)
u u T u P u RH
a a a a
2 2 2 5 2 2 4 2 2
000415 118 10 122 10 ( ) ( . ) ( ) ( . ) ( ) ( . ) ( ) = + +

The above influence factors are addressed individually below.
4.4.1.1 Air Temperature Uncertainty, u(Ta)
The uncertainty of the air temperature is a function of the following factors:
- u thermometer calibration is obtained form calibration certificate;
- u thermometer drift can be obtained from thermometer manufacturer specifications,
calibration history or it can be estimated from ice point check tolerance;
- u temperature measurement method is estimated from the expected temperature fluctuation
during the calibration.
a
2 2 2 2
( ) = + +
4.4.1.2 Relative humidity uncertainty u(RHa)
The magnitude of relative humidity uncertainty can be large without significantly affecting
results. It can be estimated from manufacturers specifications as a type B and or from calibration
process limits.
(4.9)
Technical Paper 003
st
, 2009
( ) u M
Tolerance
R
=

3
4.4.1.3 Atmospheric pressure uncertainty, u(Pa)
The uncertainty of the air pressure is a function of the following factors:
- u calibration;
- u drift can be obtained from manufacturer specifications or calibration history;
- u pressure measurement method is estimated from the expected barometric pressure
fluctuation during the calibration.
Combining the above by root sum square we obtain:
(4.10) u P u calibration u drift u method
a
2 2 2 2
( ) = + +
4.4.2 Reference Mass Uncertainty
Reference mass uncertainty can be obtained from the calibration certificate. Otherwise, the
grading of the reference mass can be used to estimate the uncertainty.
Most often the gravimetric calibration will be performed with class F1 weights which have a
tolerance of +/- 0.0005 %. Assuming a square distribution this can be converted to uncertainty
by:
(4.11)
4.4.3 Reference Mass Density Uncertainty
The reference mass density uncertainty component may be included in the uncertainty statement
of the reference mass. In this case, this uncertainty term need not be calculated. When stainless
steel reference mass is used, a tolerance of +/-200 kg/m
3
can be used as the density uncertainty.
4.4.4 Comparator Uncertainty and Method Repeatability
When an electronic balance is used as a comparator, the uncertainty of the balance indication is
due to its repeatability, eccentricity, stability and linearity which are dependent on the weighing
method. Linearity errors can usually be reduced to a negligible amount by using reference
weights which are close in value to the object weighed otherwise it must be accounted for.
Stability errors are usually reduced by the appropriate use of a substitution weighing method.
The uncertainty due to the scale eccentricity, stability and linearity must be evaluated in
conjunction with the weighing method.
Technical Paper 003
st
, 2009
The uncertainty due to the repeatability of the scale can be established as a type A uncertainty
after an analysis of the scale characteristics and the repeatability conditions under which it will
be used. This will provide the value of u(I). This should be used when a limited number of
calibration runs is carried out.
Alternatively, if many calibrations runs are carried out then it is possible to lump all repeatability
components under one factor which is estimated as a type A uncertainty from the standard
deviation of the calibration runs. Since the average of the determinations is reported then the
contribution of the repeatability is calculated as per:
u
repeatability
= Standard deviation of volume determination/ sqrt(n)
Where n = number of runs
This last approach will account for the scale repeatability error and therefore u(I) factors are not
counted.
4.4.5 Water Density Uncertainty
The density of the water during the weighing is very well known. As a result the uncertainty of
the mass determination due to the water density uncertainty is very small in relation to the
overall mass uncertainty. It can be neglected.
Technical Paper 003
st
, 2009
x
2
y = tan( ) u _
AV y
d
=
|
\
|
.
| t
2
2
5.0 Volumetric Standard Reproducibility
5.1 Leveling Error
When open neck volumetric measures are used, it is necessary to obtain a reading of the water
delivered in the measure. Two principal factors affect the accuracy of the reading, the operators
ability to read the scale and the proper leveling of the measure.
Improperly leveling of the volumetric measure will result in a reading error. This error is
dependent on the distance between the center of the measure to the sight tube and is dependent
on the neck diameter of the measure.
The following equations apply:
(5.1)
(5.2)
Technical Paper 003
st
, 2009
Where:
2 = leveling angle error
x = distance between center of neck and scale plate
)V = volume error
d = internal neck diameter of the volumetric standard
as a result the leveling uncertainty is:
(5.3) u leveling
d
( )
tan( )
=

|
\
|
.
|
u _
t
3 2
2
5.2 Scale Plate Reading Error
When an open neck volumetric measure is used by filling to the nominal mark, the reading
accuracy is assumed to be +/- 0.5 mm of scale such that:
(5.4) u reading
d cm
( )
.
=
|
\
|
.
| t
2
005
3
2
5.3 Wetting Variance
Volumetric standards that are calibrated to deliver a known volume will be subject to variances
in the amount of liquid volume which is delivered. This will be influenced by the liquid
properties, construction of the standard, drip time and method. Wetting variance is estimated
based on the standards calibration history or know characteristics for typical standard design and
construction.
Technical Paper 003
st
, 2009
6.0 References
1.M. Tanaka, G. Girard, R.Davis, A. Peuto and N. Bignell, Recommended table for the density of
water between 0C and 40 C based on recent experimental reports, Metrologia, 2001,38,301-
309
2.N. Bignell, Metrologia 19, 1983, 57
3. R.S. Davis, Equation for the Determination of the Density of Moist Air, Metrologia, 29, 1992,
67-70
4. Elsa Batista, Richard Paton, The Selection of Water Property Formulae for Volume and Flow
Calibration, Metrologia 44, 2007, 453-463

TP003

Uploaded by

Copyright:

Available Formats

TP003

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

TP003

Uploaded by

Copyright:

Available Formats

Measurement Canada

Where: M = absolute mass

You might also like