GOOD PRACTICE GUIDE:

Calibration and use of humidity sensors

for hydrogen refuelling station
applications

This project has received funding from the EMPIR programme

co-financed by the Participating States and from the European
Union’s Horizon 2020 research and innovation programme.

[Link]
 Foreword
This guide provides good practice recommendations for the calibration and use of hygrometers for
the measurement of the humidity of hydrogen fuel at hydrogen refuelling stations (HRSs).
HRS operators using hygrometers for monitoring the quality of hydrogen in their stations will find the
guidance within helpful to consider when understanding the potential errors and measurement
uncertainty of online water content measurement. Guidance includes recommended methods of
calibration for the common hygrometer types used, including comparison against standards that can
replicate the industrial conditions of use, such as the Multi-Gas, Multi-Pressure Humidity Generator
at NPL [1].
The information contained within can also be of use to those involved in hydrogen storage and transfer
of hydrogen to the grid where online monitoring of water content is required to ensure quality
requirements are met. Hygrometer manufacturers may also find this guide useful.

Acknowledgements
This Good Practice Guide is jointly funded by the European Metrology Programme for Innovation and
Research and the National Measurement System Programme for Advanced Manufacturing Metrology
funded by the UK Government Department for Business, Energy and Industrial Strategy.

Many thanks are due to Stephanie Bell and Arul Murugan for their assistance with the content of the
guide.

This good practice guide was written as part of Activity 3.2.2 of the EMPIR Metrology for Hydrogen
Vehicles (MetroHyVe) project. The three-year European project commenced on 1st June 2017 and
focused on providing solutions to four measurement challenges faced by the hydrogen industry (flow
metering, quality assurance, quality control and sampling). For more details about this project please
visit [Link].

This guide was written by:

Paul Carroll National Physical Laboratory [Link]@[Link]

© NPL Management Limited, August 2019

Version 1.0

1
 Contents
1. Introduction ........................................................................................................................................ 3
2. Humidity theory .................................................................................................................................. 7
3. HRS humidity measurement best practice ....................................................................................... 10
4. Hygrometer calibration ..................................................................................................................... 12
5. Dew-point probe behaviour observed during MetroHyVe hygrometer study ................................. 14
6. Sources of uncertainty in humidity measurements .......................................................................... 16
7. Summary of good practice recommendations ................................................................................. 17
8. References ........................................................................................................................................ 18
9. Links to other sources of useful information .................................................................................... 19

2
 1. Introduction

Background
Increasing use of hydrogen fuel-cell vehicles is contributing to emission reductions and towards
meeting Europe’s target to be carbon-neutral by 2050. A large hydrogen fuel infrastructure for
vehicles is currently in development across Europe.
The hydrogen refuelling station (HRS) industry must meet certain measurement requirements set by
European legislation (DIRECTIVE 2014/94/EU [2]) that cannot currently be achieved due to the lack of
available methods and standards.
Measurement of the contaminant levels in the hydrogen produced from a HRS is required by
regulators to ensure the fuel delivered to vehicles is of an acceptable quality. The ISO 14687-2 [3]
standard specifies the limit for maximum water vapour content in hydrogen for fuel-cell vehicle use,
together with concentration limits of other contaminants.

Water vapour content level importance for hydrogen vehicle fuel

Hydrogen produced by electrolysis and other methods can contain a large amount of water vapour
which is removed during drying processes in HRSs.
The level of water vapour in hydrogen as supplied to the vehicle is not critical for the fuel cell itself,
where high operating relative humidity is controlled locally in-vehicle. There is, however, a need for
any water present in the refuelling station hydrogen to remain in the gas phase at -40 ºC (which is the
general pre-cooling temperature prior to dispensing) and at pressures in excess of tank pressures of
70 MPa (700 bar). Water is an issue for hydrogen dispensing systems, the onboard vehicle tank system
and fuel cell components, due to the possible formation of ice. Excess water in the liquid state can
also cause corrosion of metallic components.
When water is present as an aerosol it also provides a transport mechanism for water-soluble
impurities, especially as a solvent for cations such as Na+, K+, Ca2+, Cs+ and NH4+.

Water vapour regulation for hydrogen vehicle fuel

Reliable measurement of water vapour content is needed by the hydrogen vehicle refuelling industry
to ensure that the fuel being delivered meets the specified regulatory limits for contminants.
The maximum water vapour content limit set by the ISO 14687-2 regulation is for the hydrogen used
for fuel-cell vehicles to contain less than 5 µmol mol-1 of water vapour at the dispenser nozzle.
The water vapour content limit is expressed as an amount fraction – the ratio of the number of moles
of a water vapour in a humid gas to the total number of moles of humid gas present.

At the low water content levels usually encountered in HRS applications, values are often expressed
as micromoles of water vapour per mole of humid gas (µmol mol-1). This unit is informally referred to
as parts per million per unit volume (ppmv) for gas mixtures where volume fraction is approximately
the same as amount fraction.

Table 1 shows the limits for contaminants at the HRS dispenser nozzle regulated by the ISO 14687-2
hydrogen fuel quality standard.

3
 Table 1: ISO 14687-2 contaminant limits for hydrogen fuel quality at HRS dispenser nozzle

Impurity Maximum concentration of indivdual

contaminants / µmol mol-1

Helium 300

Nitrogen + argon 100

Water 5

Oxygen 5

Carbon dioxide 2

Total hydrocarbons 2

Carbon monoxide 0.2

Formic acid 0.2

Ammonia 0.1

Total halogenated compounds 0.05

Formaldehyde 0.01

Total sulphur compounds 0.004

Online water vapour content measurement of hydrogen at HRSs

Hydrogen from the HRS dispenser nozzle must be periodically sampled into vessels and sent for offline
analysis of the concentration of the regulated contaminants at independent laboratories. It is,
however, standard practice for operators to also support quality control according to ISO 19880-8 [4]
by monitoring the water content of the hydrogen fuel they produce continuously in real time through
online humidity measurements made using hygrometers.
Although the ISO 14687-2 standard states that measurements of water vapour content should be
made from the dispenser nozzle, this is not practical for continuous online measurement due to the
extremely high pressure at this point in the process, in excess of 70 MPa (700 bar).
An example of a practice used is for an electrolysis HRS operators to make online measurements of
the water content of the hydrogen fuel using a metal oxide dew-point probe at elevated pressure,
typically 2 MPa (20 bar), at a stage in the process after drying and prior to compression of the gas to
the nozzle delivery pressure.

4
 Common hygrometer types used for online water vapour content
measurement in HRSs include:
• Metal oxide dew-point probes
• Electrolytic principle hygrometers (P2O5 based sensor)
ISO 14687-2 laboratory-based methods for analysing water content of
hydrogen sampled into vessels include:
• Chilled-mirror hygrometers
• Water vapour spectrometers
• Gas chromatography mass spectrometry (GC-MS)
• Fourier-transform infrared spectroscopy (FTIR)
• Quartz crystal microbalance

Hygrometer calibration
It is important for the HRS operator to make as accurate a humidity measurement as possible to
ensure that they are producing hydrogen with a water content within the ISO 14687-2 quality limit.
The measurement error of a hygrometer can be evaluated through calibration. Calibration of a
hygrometer can be made by comparison directly against a standard humidity generator or against a
calibrated reference hygrometer traceable directly, or in multiple steps, back to an authoritative
standard. Hygrometer measurement error found through calibration should be corrected for and the
measurement uncertainty applicable to corrected humidity values should be carefully considered.
Table 2 shows the effect dew-point hygrometer measurement error of different magnitudes could
have on the calculated amount fraction of water vapour in the test gas.

Table 2: Significance of hygrometer measurement error or drift of 5 °C or 10 °C at 2 MPa

If corrections are not applied to a hygrometer that over-reads in error, then the humidity
measurements may imply that the hydrogen being produced requires further drying before it can be
compressed for delivery to vehicles when this is not actually the case. This could lead to reduced
efficiency of the HRS system due to extra processing of the gas that is not required.
If corrections are not applied to a hygrometer that under-reads in error, then the measurements may
imply that the hydrogen being produced meets the quality requirements when this is not actually the
case. This could lead to problems with the HRS system and potentially the vehicles the fuel is delivered
to due to the issues mentioned earlier.

5
 The uncertainty applicable to corrected hygrometer measurements should also be taken into account
when determining whether the humidity value measured can be considered to represent gas of an
acceptable water vapour content with respect to quality limits. The result of the corrected measured
value in summation with the uncertainty should be below the quality limit for compliance with
regulation to be shown.
Manufacturer’s factory calibrations of hygrometers
It is common for HRS operators to use metal oxide dew-point probes supplied with a factory
calibration certificate from the manufacturer. Measurement errors reported in manufacturer’s factory
calibration certificates tend to be very low as the response of the instrument can be adjusted to read
close to the reference value during this process.
It is therefore a widespread practice to use the measured humidity values with no correction and to
apply the “accuracy” from the manufacturer’s specification as a blanket measurement uncertainty
throughout the service period of the hygrometer. Typical accuracy claims from metal oxide dew-point
probe manufacturer’s specifications for dew-point temperature are of the order of ±2 °C. For some
instruments of this type it is claimed that this accuracy is applicable to dew-point temperature
measurements across the range -100 °C to +20 °C at pressures ranging from atmospheric pressure to
35 MPa (350 bar).
The factory calibrations are completed with nitrogen or air as the background gas at close to
atmospheric pressure.
MetroHyVe hygrometer study
In Activity 3.2.1 of the EMPIR Metrology for Hydrogen Vehicles (MetroHyVe) project six commercially
available metal oxide dew-point probes from five different manufacturers were calibrated against the
NPL Multi-Gas, Multi-Pressure Humidity Generator [1]. Tests of the response of the hygrometer to
changes in background gas species from nitrogen to hydrogen and changes in pressure from close to
atmospheric pressure to 2 MPa (20 bar) were performed.
Testing of the hygrometers included conditions representative of their use in HRS applications
(background gas hydrogen, pressure of 2 MPa) over a range of dew-point temperatures (-65 °C to
-20 °C) that correspond to water vapour amount fractions from 0.5 μmol mol-1 to 50 μmol mol-1 at
2 MPa, spanning the ISO 14687-2 quality limit of 5 μmol mol-1.
Repeated testing of the hygrometers at six and twelve months after the initial calibrations enabled
evaluation of the long-term drift behaviour of instruments of this type.
The findings of this study are described in Section 5 and provide the basis for some of the calibration
good practice recommendations in this guide which are summarised in Section 7.

6
 2. Humidity theory

An introduction to some relevant background humidity theory is required to understand some of the
recommendations made in this guide.
Definitions
Dew-point / Frost-point temperature Tdp / Tfp: temperature at which the vapour pressure of the
vapour in a humid gas is equal to the saturation vapour pressure over pure liquid and at which
condensate forms as liquid on cooling the gas. This value is defined as a frost-point temperature below
0 °C when the condensate formed is ice. However, water can exist as a liquid at temperatures below
0 °C and when referring to supercooled liquid dew point it is normal to state this explicitly.
Saturation vapour pressure of water es : maximum pressure of water vapour that can exist at a given
temperature. Expressed in units of pressure (e.g. pascals (Pa)).
Vapour pressure e : the partial pressure of a gas contributed by the water vapour. Expressed in units
of pressure (e.g. pascals (Pa)).

Figure 1 Saturation vapour pressure curves for (a) water and (b) supercooled water (blue, upper curve) and ice
(green, lower curve)

It can be seen from Figure 1 that below 0 °C the same saturation vapour pressure exists for different
values of frost-point temperature (ice) and dew-point temperature (supercooled water).

Relative humidity  : Relative humidity describes how saturated a gas (or a space) is with water
vapour. This is the most commonly used measure of humidity and is usually expressed as a percentage,
with the symbol ‘%rh’. Relative humidity ( ) is calculated using the following formula:

𝑒
 = 𝑒 × 100 (% rh) (1)
𝑠

where:
 e is the actual water vapour pressure
 es is the saturation vapour pressure at the temperature of the gas

7
 Pressure P : Overall pressure exerted by all components of the humid gas. Absolute pressure values
should be used in humidity calculations and expressed in units of pressure (e.g. pascals (Pa)).
Water vapour enhancement factor f : factor by which a value of pure saturation vapour pressure is
multiplied by to obtain the actual saturation vapour pressure which takes into account the effect of
the presence and overall pressure of air or other background gases.
Mole fraction of water vapour x : ratio of the number of moles of a water vapour in a humid gas to
the total number of moles of humid gas present where:
𝒆𝒇
𝒙= ( mol mol-1) (2)
𝑷
At the low water content levels expected in HRS applications values are often expressed as micromoles
of water vapour per mole of total gas (µmol mol-1), often informally referred to as parts per million
per unit volume (ppmv) for gas mixtures where volume fraction is approximately the same as amount
fraction.
Introduction to humidity calculations
There are methods for calculating and converting between humidity values in terms of dew point,
amount fraction of water vapour, relative humidity, vapour pressure and other parameters. However,
most humidity calculations involve several steps.

Some hygrometer companies provide software for making humidity conversions. This is often a good
alternative to struggling with pen, paper and a calculator. Either way, humidity calculations need to
be made carefully, with some understanding. In this section some of the more commonly made
humidity calculations and any considerations for applying these to measurements of the water
content of hydrogen are discussed.

Accepted formulae for calculating vapour pressure and water vapour enhancement factor from
measurements of dew-point temperature and overall pressure in air and nitrogen can be found in
papers by Sonntag [5] (also available in A Guide to the Measurement of Humidity [6]) and Hardy [7].

Humidity calculation applicability to HRS water content measurements

A common practice is for operators to make online humidity measurements of the hydrogen fuel in
the HRS system at elevated pressure with a hygrometer measuring in units of dew-point temperature.
As the ISO 14687-2 standard water vapour limit is specified in terms of amount fraction, x this requires
the results of measurements of dew-point temperature and overall pressure to be used to calculate
values for e and f in equation 2 such that a value of x can then be realised.
Care should be taken that an absolute pressure value is used for the overall value of pressure, P, as
the use of gauge pressure values could lead to significant errors in the calculations of amount fraction,
x. For example, x calculated as 4 μmol mol-1 with an absolute value of pressure of 0.5 MPa (5 bar)
would result in an amount fraction of 5 μmol mol-1 if a gauge pressure 0.4 MPa gauge (4 bar) was used
in the calculation.
As there is limited accepted data for water vapour enhancement factor in mixtures with hydrogen
then any value of f approximated using formulae for air and nitrogen [5,6,7] should have a suitable
estimation of uncertainty applied. If this uncertainty is expressed in terms of percent of value of f then
this will transfer directly as a percentage uncertainty in the calculated value of amount fraction from
equation 2.
The dew-point temperature measurements made at HRSs in practice tend to be relatively low leading
to calculated values of x much lower than the quality limit. In such cases the relatively large resultant

8
 amount fraction uncertainties can therefore be accommodated and the measurements made would
still demonstrate hydrogen of acceptable quality.
Amount fraction of water vapour is unaffected by changes in temperature and pressure of the gas.
Therefore, hydrogen can be monitored at pressures lower than that of the nozzle delivery pressure at
stages in the system prior to the final compressor and the water content value can be assumed to be
the same at the point of delivery. However, this supposes there is no stray water influence or
absorption of water during the compression stage which is why the periodic sampling of hydrogen for
quality analysis must be made from the nozzle.
Dew-point temperature measurements are not pressure independent. Variations in the total pressure
of a gas lead to changes in the water vapour partial pressure resulting in a change in dew-point
temperature.

9
 3. HRS humidity measurement best practice

Sampling
In general, correct sampling is required to ensure that the humidity of the gas reaching the hygrometer
is representative of gas at the test location. In this regard, it is important to avoid water being
spuriously added to the measured gas, and to avoid unintended removal of water from the sample.
For measuring very dry gases (frost points around -40 °C, -50 °C, or below; water vapour fraction
around 100 μmol mol-1, or less), sample tubing and all materials in the gas flow path are critical. In this
range, even tiny amounts of stray water released from surfaces can significantly add to the gas
moisture content, and can give badly misleading results. A plot of the dew point measured after the
same dry gas has passed through a variety of tubing materials as a function of time is shown in Figure
2. To avoid moisture permeation through tubing walls and minimise desorption effects, clean
internally electropolised stainless steel; the smallest possible volume of pipework; and long flushing
times should be used when sampling gas to be measured by hygrometers.

Figure 2. Dew-point measurements of the same dry gas after passing through a variety of tubing materials

Measurement of dry gases using metal oxide dew-point probes

In the majority of HRS conditions measured, the dew-point temperature will be -40 °C or below. The
response time of metal oxide dew-point probes is much slower than some laboratory based
instruments such as water vapour spectrometers or chilled-mirror hygrometers. As well as use of the
correct pipework and vessel materials as described above to ensure accurate sampling of test gas,
enough time must be left for sample tubing equilibration and for the instrument to stabilise, ensuring
a representative measurement is made. In practice this can take many hours at trace moisture levels
and hysteresis (metal oxide hygrometers can measure different values at the same humidity
depending on whether the current condition has been reached following a rise or fall in humidity) can
lead to further errors in the measured value.

10
 Measurement of trace water content of hydrogen at elevated pressure
As amount fraction, x, of water vapour remains the same independent of the total pressure of the test
gas, water vapour in hydrogen can be measured at pressures lower than that of the nozzle delivery
pressure at a stage in the system prior to final compression and the result assumed to be the same at
the point of delivery.
The applicability of using this measured value as a water content quality check for the nozzle delivery
gas assumes that no further water is spuriously added to or unintentionally removed from the
measured gas during compression.
Pressure regulator devices [8,9,10] exist that are used at the nozzle to reduce the supply pressure to
an acceptable level for sampling hydrogen into vessels for offline purity analysis. Spot checks of the
water content of the hydrogen fuel from the nozzle after the compressor stage could be performed
using a hygrometer if it could be connected to such a device. The instrument response time and the
time taken for the test gas to stabilise may be prohibitive for this to be a practical approach.
Analogue output scaling
Measurement results from industrial dew-point hygrometers are usually logged from DC current
outputs scaled for example from 4 mA to 20 mA for the full scale range of the instrument (e.g. -100 °C
to +20 °C). Depending on the specific hygrometer type this current range could be equivalent to dew-
point temperature ranges of -110 °C to +20 °C, -100 °C to +20 °C or -80 °C to 0 °C. It is therefore
important to read the manufacturer’s instructions to ensure the correct conversion of the current
output to dew-point temperature.
Temperature compensation of metal oxide dew-point probe measurements
Some older dew-point probes might exhibit some error due to sensitivity to ambient temperature,
and the specification from when they were purchased will give a temperature coefficient that should
be used to compensate for this.
Most manufacturers now include an internal temperature sensor within the dew-point probe and
temperature compensation is applied from the measured ambient temperature (over a specified
range) to improve the accuracy of the humidity measurement. If the ambient temperature the
hygrometer at the HRS is exposed to is likely to differ significantly from the laboratory conditions at
which its calibration was performed then it is recommended to request calibration of the hygrometer
at a range of ambient temperatures representing the conditions of use. (e.g if measuring in ambient
conditions of 0 °C in the field, a different hygrometer measurement error to that found during
calibration in the laboratory at +20 °C may result).

11
 4. Hygrometer calibration

Like all measuring instruments, hygrometers don’t always read exactly true. They can suffer from bias
and long-term drift. Therefore calibration is essential for accurate humidity measurement.
Calibration is the comparison of an instrument against a reference value. For a hygrometer, this
comparison might be against a calibrated reference hygrometer, using an environmental chamber or
other stable source of humid gas. The reference hygrometer should itself have a calibration traceable
directly, or in multiple steps, back to an authoritative standard. This is known as measurement
traceability.
Some users assume that calibration will leave an instrument ‘reading correctly’. However, this is not
necessarily true because calibration is not the same as adjustment. A calibration certificate reports
any instrument errors, and gives the uncertainty in these. Calibration corrections then need to be
applied to instrument readings (or else taken into account as a source of uncertainty). With any
measurement, the practical uncertainty in using the instrument is always more than that stated on
the calibration certificate.
Calibrations against reference hygrometers, using humidity generators, can be performed in
laboratories. A calibration laboratory may have ISO 17025 [11] accreditation (through UKAS in the
UK) giving assurance that calibrations are traceable and competently performed.

Calibration methods
The calibration of hygrometers is typically carried out by comparing their indication, in a stable
environment, against a humidity reference. Several methods that can be used to generate reference
humidities to calibrate hygrometers used in HRSs are described in the sections that follow.
Single-pressure dew-point generator
Dew-point generators are capable of producing test gas streams with a well-defined dew-point
temperature. This is typically achieved by passing the gas through, or flowing it over, water or ice at a
defined temperature in a device called a saturator. The result of this process is the production of a gas
with a dew point that is equal to the temperature of the saturator used, assuming saturation of the
gas is efficient. The test gas used is usually air or nitrogen and is applied to test hygrometers at close
to atmospheric pressure. Recent developments at national measurement institutes such as NPL [1]
have led to facilities where humidified hydrogen can be used as the test gas and can be supplied at
elevated pressures reflecting industrial conditions of use.
Two-pressure generators
The operating principle of a two-pressure generator is for a gas to be humidified at high pressure and
then expanded to a lower pressure whilst maintaining a non-condensing temperature. In accordance
with Dalton’s law, as the total pressure falls, so does the partial pressure of water vapour and,
assuming that the initial value of the humidity is known, the dew point after the expansion process
can be determined. Different values of dew point or amount fraction can be obtained using a two-
pressure generator through variation of the initial pressure and saturation temperature.

12
 Hygrometer field checks with reference gas mixtures certified for water content
A static reference cylinder with certified water content can be prepared by transferring pure gases,
pure liquids, or gravimetrically prepared mixtures of known composition quantitatively into a cylinder
in which the calibration gas will be contained. Such cylinders can be used for field checks of
hygrometers at a HRS site. A check can help in assessing whether a hygrometer is functioning as
expected, and whether repair or recalibration is needed.
HRS hygrometer calibration recommendations
The anticipated range of humidities that will be encountered in service should be considered when
purchasing a calibration. A calibration consisting of at least three values of dew point or water vapour
amount fraction including points above and below the maximum and minimum humidity in this range
should be requested to understand the error characteristics of the hygrometer.

Calibrations need to be repeated, because hygrometers can drift due to conditions of use, or just due
to sensor ageing. The frequency of recalibration is dependent on usage, risk of drift, and how critical
the measurement is. Generally metal oxide dew-point probe manufacturers recommend a six to
twelve month interval between calibrations.
It is recommended that hygrometer long-term drift is assessed following the service period in a HRS
station. A common practice followed by HRS operators is to apply corrections from the factory
calibration certificate (or to use the output of the hygrometer with no correction) and assume this is
valid for the entire service period of the hygrometer. At the end of the service period the hygrometer
is returned to the manufacturer for replacement, usually via an exchange programme. It is not
common practice for users to request an “as-found” calibration of the returned hygrometer in order
to assess the measurement error drift during service. Such a calibration could reveal important
information about the performance of the hygrometer used and could explain trends in the HRS
hydrogen water content level data which may be due to instrument drift over time rather than HRS
system issues.

Recommendations when requesting calibrations of hygrometers for HRS use:

• Use a laboratory with ISO 17025 accreditation for humidity calibration in the
relevant range

• Request that the calibration is in the working units of the hygrometer e.g. dew-
point temperature (°C) or volume fraction (ppmv)

• Request at least three points covering minimum and maximum expected

humidity values of the process being monitored

• If possible, request that calibration is performed at the range of ambient

temperatures the hygrometer would experience during service

• Request an “as-found” calibration of the hygrometer at the end of service to

assess long-term drift

13
 5. Dew-point probe behaviour observed during MetroHyVe
hygrometer study

In Activity 3.2.1 of the EMPIR Metrology for Hydrogen Vehicles (MetroHyVe) project, six commercially
available metal oxide dew-point probes from five different manufacturers were calibrated against the
NPL Multi-Gas, Multi-Pressure Humidity Generator [1]. Conditions were applied that tested the
response of the hygrometers to changes in background gas species from nitrogen to hydrogen and
changes in pressure from close to atmospheric pressure to 2 MPa (20 bar). Calibrations were
completed across a range of dew-point temperatures (-65 °C to -20 °C) that correspond to water
vapour amount fractions from 0.5 μmol mol-1 to 50 μmol mol-1 at 2 MPa, spanning the ISO 14687-2
limit of 5 μmol mol-1.
The calibrations across this dew-point temperature range completed at close to atmospheric pressure
and with nitrogen as the background gas enabled testing of potential differences in instrument
response between the manufacturer’s factory calibration and a calibration completed in conditions
reflecting industrial use in a HRS.
One of the hygrometers tested in the study was owned by ITM Power, operators of the NPL onsite
HRS. This hygrometer was calibrated alongside the other hygrometers loaned to the project before it
was installed for its one year service period in the NPL onsite HRS. Subsequent calibrations of the HRS
hygrometer were performed six months into the service period during an HRS maintenance shutdown
and again at the end of a year of service. The five other hygrometers in the study were also re-
calibrated at these times and were kept in the NPL humidity laboratory measuring dry air in the
intervening periods.
Repeated calibrations of the test hygrometers at six and twelve month periods after the initial
calibration enabled evaluation of the long-term drift behaviour of instruments of this type.
The findings of this study [12] and resultant calibration recommendations are summarised below and
should be considered together with the measurement and calibration recommendations from
Sections 3 and 4.

Background gas dependence of instrument reading

Finding:
Of the six metal oxide probes tested in the study, four showed no noticeable change in reading when
the background gas carrying the water vapour was changed from nitrogen to hydrogen at a fixed
humidity condition. However, one of the hygrometers showed a downward shift in reading of
approximately 4 °C, whilst another showed an upwards shift of approximately 1°C, due simply to the
change in background gas species.
Recommendation:
Where available, a calibration in a background gas of hydrogen for hygrometers used in HRS
applications is recommended to address any gas-dependency not identified in the factory calibration
performed in a background gas of air or nitrogen. It cannot be assumed that corrections based on the
results of factory calibration certificates are applicable to a hygrometer used to measure the humidity
of hydrogen.

14
 Pressure dependence
Finding:
A dependence on background pressure of readings was observed for three of the six metal oxide
probes tested in the study. This effect was largest at -65 °C dew-point temperature, the lowest
humidity tested. The fixed humidity condition was applied at two pressures: 0.17 MPa and 2 MPa.
Those hygrometers affected responded in different ways, ranging from an upward shift of 1 °C to a
downward shift of 4 °C.
Recommendation:
Where available a calibration in hydrogen at the elevated pressure reflecting hygrometer use in HRS
applications should be performed to address any pressure-dependency not identified in the factory
calibration performed at atmospheric pressure in air or nitrogen.

Long-term drift of measurement error

Finding:
The measurement error of five of the six metal oxide hygrometers tested showed large drift towards
under-reading during the one-year period of the study. Values of long-term drift ranged between
-7.5 °C and -12.5 °C dew-point temperature.
The calibration results after six months of service showed that this was an approximately progressive
change for three of the affected hygrometers, with more drift occurring in the second six months of
the study. The two remaining hygrometers drifted during the first six months of service, but no further
drift was observed at the twelve-month calibration stage. Therefore, it is not possible to predict
whether hygrometer drift will occur at start of service, or towards the end, or uniformly with time.
Recommendation:
Methods could be developed for periodic field checks of hygrometers onsite at HRSs using cylinders
containing reference gas mixtures of water vapour in hydrogen. A check would help in assessing
whether a hygrometer has drifted significantly during service and whether recalibration or
replacement is needed.
It is recommended to obtain “as-found” calibration results at the end of the hygrometer service period,
in order to gain information about sensor drift. This would enable assessment of the extent of the in-
service drift of the hygrometer since the initial factory calibration.
Alternatively, if it is possible to have a second port available at the HRS a new hygrometer could be
installed in parallel with the hygrometer it is replacing (before this is removed). The extent of the
measurement drift during use could then be evaluated in-situ before the hygrometer due for
replacement is returned to the manufacturer.

15
 6. Sources of uncertainty in humidity measurements

As we have seen throughout this guide, every humidity measurement is subject to a degree of
uncertainty. The sources and magnitude of uncertainty involved in the use of a hygrometer will
depend on the instrument being used and on the specific measurement being made. However, as a
general guide, the following sources of uncertainty should be considered in humidity measurements
of hydrogen at a HRS:

 The calibration uncertainty of the hygrometer

 Long-term drift in hygrometer error
 Pressure sensitivity of the hygrometer
 Sensitivity of hygrometer to background gas
 Equilibration (settling time, response time)
 Temperature dependence of the hygrometer (if not temperature compensated)
 Irreproducibility of the hygrometer
 Hysteresis of the hygrometer
 Non-linearity of the hygrometer
 Short-term stability (noise)
 The resolution of the display or electrical output

16
 7. Summary of good practice recommendations

Calibration good practice:

✓ Ensure that the hygrometer calibration is traceable to national standards.

✓ Request that the calibration is in the working units of the hygrometer e.g. dew-point
temperature (°C) or water vapour volume fraction (ppmv).

✓ Request at least three calibration points covering minimum and maximum expected
humidities.

✓ If using corrections from a hygrometer’s factory calibration certificate (performed in nitrogen

or air at atmospheric pressure), ensure that possible measurement errors due to use in
hydrogen background gas at elevated pressure have been considered.

✓ Where possible obtain a calibration of the hygrometer with hydrogen as background gas at
the anticipated pressure of use to replicate operating conditions.

✓ Ensure that the hygrometer manufacturer has provided an estimate of long-term drift (or if
not available perform independent checks at regular periods during service).

✓ Measurements of humidity using a new hygrometer installed in parallel with the end of service
hygrometer it is replacing (before this is removed) can give an in-situ check of long-term error
drift.

✓ If an older dew-point probe type that may not have temperature compensation is being used,
request that calibration is performed at a range of ambient temperatures the hygrometer
would experience during service.

Measurement good practice:

✓ Ensure a representative humidity measurement from the hygrometer used. Avoid moisture
permeation through tubing walls and minimise desorption effects by using clean internally
electropolised stainless steel and the smallest possible volume of pipework when installing
hygrometer.

✓ Leave enough time for tubing material and hygrometer to stabilise.

✓ Apply corrections from calibration certificates to measured humidity values.

✓ Measurement uncertainty factors applicable to corrected humidity values should be carefully

considered.

17
 8. References

International Standards are subject to change. Reference should be made to the current published version of the
standard where applicable.

[1] Carroll PA, Bell SA, The NPL Multi-gas, Multi-pressure Humidity Generator. Paper in preparation,
Metrologia.

[2] DIRECTIVE 2014/94/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 22 October
2014 on the deployment of alternative fuels infrastructure.

[3] ISO 14687-2:2012. Hydrogen fuel product specification part 2: proton exchange membranes
(PEM) fuel cell applications for road vehicles.

[4] Draft BS ISO 19880-8 1, Gaseous hydrogen - Fueling stations - Part 8: Hydrogen quality control.
[5] Sonntag D, “Important new values of the physical constants of 1986, vapor pressure formulations
based on the ITS-90, and psychrometer formulae” Z. Meteorol. 70, 340 (1990).

[6] A Guide to the Measurement of Humidity Published by the Institute of Measurement and
Control, 1996, 68 pp, ISBN 0-904457-24-9 [Link]/publications/guide-to-the-measurement-
of-humidity.
[7] Hardy B (1998) “ITS-90 Formulations for Vapor Pressure, Frost point Temperature, Dew point
Temperature, and Enhancement Factors in the range –100 °C to +100 °C”, Papers and Abstracts of
the Third International Symposium on Humidity and Moisture.

[8] SAE J2600. Compressed hydrogen surface vehicle fuelling connection devices. 2012.

[9] SAE J2799. 70 MPa compressed hydrogen surface vehicle fuelling connection device and optional
vehicle to station communications. 2007.

[10] ISO 17268:2012. Gaseous hydrogen land vehicle refuelling connection devices. 2012.

[11] ISO 17025 General requirements for the competence of testing and calibration laboratories.
2017.

[12] Carroll PA, Bell SA, A study of the measurement response of hygrometers typically used to
measure humdiity at hydrogen refuelling stations. Paper in preparation, Int. J. Hydrogen Energy.

18
 9. Links to other sources of useful information

International Standards are subject to change. Reference should be made to the current published version of the
standard where applicable.

Further reading
DIRECTIVE 2014/94/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 22 October 2014
on the deployment of alternative fuels infrastructure.
Brown AS, Vargha GM, Downey ML, Hart NJ, Ferrier GF, Hall KL. NPL report AS 64-methods for the
analysis of trace-level impurities in hydrogen for fuel cell applications. 2011.
Murugan A, Brown AS, Review of purity analysis methods for performing quality assurance of fuel
cell hydrogen. Int. J. Hydrogen Energy. 2015, pp. 1–15.
Standards – Hydrogen
ISO 14687-1, Hydrogen fuel — Product specification — Part 1: All applications except proton
exchange membrane (PEM) fuel cell for road vehicles.
ISO 14687-2:2012. Hydrogen fuel product specification part 2: proton exchange membranes (PEM)
fuel cell applications for road vehicles.
ISO 14687-3:2014. Hydrogen fuel product specification part 3: proton exchange membrane (PEM)
fuel cell applications for stationary appliances.
BS ISO 21087:2019 Gas analysis — Analytical methods for hydrogen fuel — Proton exchange
membrane (PEM) fuel cell applications for road vehicles.
BS ISO 19880-1 Gaseous hydrogen - Fuelling stations. Part 1: General requirements.
BS ISO 19880-8 1, Gaseous hydrogen - Fueling stations - Part 8: Hydrogen quality control.
BS EN 17124:2018 Hydrogen fuel - Product specification and quality assurance - Proton exchange
membrane (PEM) fuel cell applications for road vehicles.
SAE J2719. Hydrogen fuel quality for fuel cell vehicles. 2011.
ASTM D7649-10. Standard test method for determination of trace carbon dioxide, argon, nitrogen,
oxygen and water in hydrogen fuel by jet pulse injection and gas chromatography / mass
spectrometer analysis. 2010.
ISO/TR 15916 Basic considerations for the safety of hydrogen systems. 2015.
ISO/IEC 80079 (all parts), Explosive atmospheres.

19
 Humidity
Bell, S A (2013) The beginner's guide to humidity measurement. Measurement. Good Practice Guide.
124.
A Guide to the Measurement of Humidity Published by the Institute of Measurement and Control,
1996, 68 pp, ISBN 0-904457-24-9 [Link]/publications/guide-to-the-measurement-of-
humidity.
BS1339-1 (2002) Humidity. Terms, definitions and formulae.

BS1339-2 (2009) CD-ROM. Humidity calculation functions, tables and user guide.

BS1339-3 (2004) Guide to the measurement of humidity.

General metrology
Bell, S A (2001) A beginner’s guide to uncertainty of measurement. Good Practice Guide 11 (Issue 2).
JCGM 200 (2008) International vocabulary of metrology. Basic and general concepts and associated
terms (VIM).

ISO/IEC GUIDE 98-3(2008) Ed 1 Uncertainty of measurement. Guide to the expression of uncertainty

in measurement.

BS PD 6461-4 (2004) General metrology. Practical guide to measurement uncertainty.

Other useful links

National Physical Laboratory,
Hampton Road,
Teddington,
Middlesex
TW11 0LW
+44 (0)20 8977 3222
[Link]/temperature-humidity
humidity_enquiries@[Link]

British Standards Institute (BSI) [Link]

International Standards Organisation (ISO) [Link]

20