Clean Energy, 2019, Vol. 3, No.

2, 114–125

doi: 10.1093/ce/zkz006
Advance Access Publication Date: 9 May 2019
Homepage: https://academic.oup.com/ce

Insight
HyDeploy: The UK’s First Hydrogen Blending

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

Deployment Project
Tommy Isaac*
Principal Engineer, Progressive Energy Ltd, Swan House – Bonds Mills, Stonehouse, Gloucestershire GL10 3RF, UK
*Corresponding author. E-mail: [email protected]

Abstract
The HyDeploy project is the UK’s first practical project to demonstrate that hydrogen can be safely blended into
the natural-gas distribution system without requiring changes to appliances and the associated disruption.
The project is funded under Ofgem’s Network Innovation Competition and is a collaboration between Cadent
Gas, Northern Gas Networks, Progressive Energy Ltd, Keele University (Keele), Health & Safety Laboratory and
ITM Power. Cadent and Northern Gas Networks are the Gas Distribution Network sponsors of the project. Keele
University is the host site, providing the gas-distribution network, which will receive the hydrogen blend.
Keele University is the largest campus university in the UK. Health & Safety Laboratory provides the scientific
laboratories and experimental expertise. ITM Power provides the electrolyser that produces the hydrogen.
Progressive Energy Ltd is the project developer and project manager. HyDeploy is structured into three distinct
phases. The first is an extensive technical programme to establish the necessary detailed evidence base in
support of an application to the Health & Safety Executive for Exemption to Schedule 3 of the Gas Safety
(Management) Regulations (GS(M)R) to permit the injection of hydrogen at 20 mol%. This is required to allow
hydrogen to be blended into a natural-gas supply above the current British limit of 0.1 mol%. The second phase
comprises the construction of the electrolyser and grid entry unit, along with the necessary piping and valves,
to allow hydrogen to be mixed and injected into the Keele University gas-distribution network and to ensure all
necessary training of operatives is conducted before injection. The third phase is the trial itself, which is due to
start in the summer of 2019 and last around 10 months. The trial phase also provides an opportunity to undertake
further experimental activities related to the operational network to support the pathway to full deployment of
blended gas. The outcome of HyDeploy is principally developing the initial evidence base that hydrogen can be
blended into a UK operational natural-gas network without disruption to customers and without prejudicing the
safety of end users. If deployed at scale, hydrogen blending at 20 mol% would unlock 29 TWh pa of decarbonized
heat and provide a route map for deeper savings. The equivalent carbon savings of a national roll-out of a
20-mol% hydrogen blend would be to remove 2.5 million cars from the road. HyDeploy is a seminal UK project for
the decarbonization of the gas grid via hydrogen deployment and will provide the first stepping stone for setting
technical, operational and regulatory precedents of the hydrogen vector.

Keywords:  HyDeploy; hydrogen blending; demonstration project

Received: 5 January, 2019; Accepted: 8 March, 2019

© The Author(s) 2019. Published by Oxford University Press on behalf of National Institute of Clean-and-Low-Carbon Energy
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://
creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, 114
provided the original work is properly cited. For commercial re-use, please contact [email protected]
 Isaac | 115

Introduction 350

Annual Generation (TWh)

The use of fossil fuels throughout the twentieth century 300
was the driving force behind the unprecedented change in 250
social structure and improvements in quality of life wit- 200
nessed. The power to harness such a high-energy-density, 150
storable and abundant energy source transformed every
100
aspect of society. No longer was personal movement lim-
50
ited by animal feed and no longer was mechanization lim-
ited by human strength. The global economy transformed –
2013 2014 2015 2016 2017
from a supply limited world to one of almost limitless
potential. It would be difficult to underestimate the impact Nuclear Renewables Coal Gas
that fossil fuels have had on societal development and to
Fig. 1  Electricity generation by source

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

simply characterize them as ‘bad’ or ‘evil’ would be myopic.
The benefits society reaps from the use of such a cheap
600
and convenient energy source are not without conse-
quences. The oxidation of hydrocarbons releases heat that 500

Intensity (gCO2/kWh)
can either be used directly or converted to work, although
400
the process of oxidation also produces carbon dioxide and
water vapour. The production of carbon dioxide is the sin- 300
gle largest consequence that results from the use of fos- 200
sil fuels. The production of carbon dioxide outside of the
natural ecological cycle allows infrared radiation to be 100
trapped within the atmosphere leading to a rise in global 0
temperatures. Temperature is the largest determinant of 2013 2014 2015 2016 2017
any physical process—too much and everything breaks
Fig. 2  UK electricity carbon intensity
down into subatomic soup, too little and nothing moves
or interacts. Organic life only exists within a relatively nar-
row temperature range, so processes that interfere with Unlike the electricity grid, the demands of the gas grid
the natural temperature range of the Earth in perpetuity are seasonal and diurnal, as shown by Fig. 3. This creates
are not consistent with the Earth’s ability to sustain life. peaks and troughs of demand, which can easily be accom-
Given that energy use is a fundamental requirement modated with pipes and gas pressure, but are very difficult
for social development, and carbon-dioxide production to accommodate with wires and transformers.
scales with fossil-fuel energy use, the generation of carbon The unsuitability of the electricity grid to handle the
dioxide is inherent within the current models of economic ‘peaky’ demand profile of the gas grid is the central reason
prosperity. It is this relationship that rests at the heart of why electrification of heat is unlikely to be the sole means
the decarbonization problem—how to live in a prosperous by which heat will be decarbonized. Notwithstanding the
world without promoting the production of carbon dioxide. fact that the gas grid provides over three times the energy
In most modern economies, energy is delivered by of the electricity grid each year [2, 4], replacement would
three systems: the electricity grid; the gas grid; and the oil mean construction on the scale of three electricity grids.
extraction, processing and transportation supply chain. Hydrogen has the potential to play a key role in the de-
Decarbonizing electricity has been the principal focus of carbonization of heat. As it is a gaseous molecule at atmos-
policy makers since climate change entered the public pheric conditions, it can be used as a direct replacement
lexicon. Within the UK, this is being achieved by closing for natural gas within the gas grid. This is demonstrated in
coal power stations and replacing them with gas, wind, Europe where hydrogen blending is commonplace and tri-
solar and biomass. This strategy has been very success- als are underway to increase the blending level to 20 mol%
ful and has led to a 50% reduction in the carbon intensity [3]. The gas grid is a highly sophisticated written-down na-
of electricity from 2013 to 2017 [1]. The outcome of these tional asset that extends across the entire country, con-
activities is to demonstrate that material change can be necting to all commercial and industrial sites, half of all
achieved with the right focus and regulatory mechanisms. power-generation sites and almost all domestic houses.
Figs 1 and 2 illustrate the results that have been achieved Therefore, the lowest-cost pathway to decarbonize heat
within the electricity market. will likely be predicated on the continued use of the gas
The decarbonization of heat remains one of the larg- grid. The introduction of hydrogen into the gas grid pro-
est unsolved climate-change problems. The reason heat vides a universal pathway to decarbonize all end users—
is so difficult to decarbonize is because the gas grid in its industrial and commercial, electricity generation and
current form is so well suited to the demanding needs of domestic homes. Hydrogen deployment could take many
its users. forms, from blending into the Local Transmission System
 116 | Clean Energy, 2019, Vol. 3, No. 2

4500 GWh

4000 GWh
BEAST FROM THE EAST
3500 GWh

3000 GWh
HEAT
GWh per day

2500 GWh TRANSPORT

2000 GWh

1500 GWh

1000 GWh
Power

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

500 GWh

0 GWh
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
2015 2016 2017 2018

Fig. 3  UK energy demands. Author: Dr Grant Wilson, Birmingham University. Source date: National Grid, Elexon and BEIS.

to conversion of industrial users and potentially full net- without prejudicing the safety of end users or modifying
work conversion. appliances. The structure of the project is such that cus-
The UK gas grid has a history of hydrogen in the form of tomers’ safety has been the leading focus of all project
Towns Gas. Prior to the discovery of the North Sea reserves, areas and technical lines of enquiry. The project has been
the UK’s gas supply came from gasified coal called Towns divided into three phases:
Gas. This gas contained up to 50 mol% hydrogen, although
(i) Phase 1: Securing regulatory permissions;
the carbon-monoxide content resulted in a safety risk
(ii) Phase 2: Construction and preparation;
to the public. The discovery of the North Sea reserves
(iii) Phase 3: The trial.
redefined the energy landscape of the UK as abundant,
­
low-cost natural gas replaced coal mining and gasification.
GS(M)R defined what could be transported within the natu-
ral-gas grid. The defined composition was based on the pre- 1.1 Phase 1
vailing make-up of natural gas from the North Sea, which Phase 1 has been an 18-month scientific programme
does not naturally contain hydrogen, so GS(M)R capped the designed to build the necessary evidence base to dem-
upper limit of hydrogen content to 0.1 mol%, primarily for onstrate that hydrogen is safe to blend into the Keele
measurement purposes. The rationale for this cap is there- University private gas network for the purposes of the trial
fore a regulatory expedience, as North Sea gas contains period. The programme consisted of three main elements:
essentially no hydrogen. Hydrogen represents a low-carbon laboratory work; network appliance testing; and risk anal-
alternative within the context of climate change so the 0.1- ysis—all leading to an exemption application to the Health
mol% limit is now rightly being challenged. & Safety Executive (HSE) to sanction the injection of hydro-
Beyond heat, hydrogen has the potential to aid the decar- gen up to 20 mol%.
bonization of both electricity and transport. Electricity The extensive laboratory work undertaken was led
decarbonization could be achieved either via hydrogen- by the Health & Safety Laboratory (HSL) and centred on
fuelled gas turbines or industrial fuel cells—providing both understanding the implications of a 20-mol% hydrogen
baseload and flexible supply to complement intermittent blend relative to natural gas. This work involved test-
renewables. Transport decarbonization could be achieved ing a range of domestic appliances, including gas boilers,
either via providing low-carbon electricity during charging cookers and fires. Alongside appliances, a materials pro-
times or directly in fuel cells. gramme of experimentation was undertaken to under-
Whichever technological pathways prevail within a low- stand the mechanical-property implications of soaking
carbon economy, it is clear that hydrogen has the potential known materials in a blend of hydrogen and natural gas.
to play a central cross-system role. The materials programme was informed by an asset sur-
vey of the Keele network to ensure all gas-seeing materials
on the network would be tested prior to the trial. This built
1  Project structure confidence that the network and all operating compo-
The purpose of HyDeploy is to demonstrate that hydrogen nents, such as control valves, would maintain operational
can safely be blended into the UK’s gas grid at 20 mol% integrity throughout the trial.
 Isaac | 117

The laboratory work also included understanding the the largest of which was 600 kW, as well as catering appli-
gas characteristics of a 20-mol% hydrogen blend relative ances. Once again, every appliance that was safe on natu-
to natural gas. This work allowed recommendations to be ral gas (operational performance and leakage testing) was
formed within the context of operational procedures to also safe on a blend containing 20 mol% hydrogen.
produce supplementary guidance for operatives on the The final element of Phase 1 was to aggregate all of the
Keele network. This work stream provided a pathway to experimental results and operational testing to under-
translate basic science into operational implications and taken a Quantitative Risk Assessment (QRA) of the Keele
ensure procedures were robust and appropriate. network. The QRA was developed by gas industry safety
An important experimental work stream related to expert Dave Lander Consulting alongside Kiwa Gastec.
gas detection. Accurately detecting gas concentrations is The risk assessment involved building a fault tree of
a technical foundation of safely operating a network. All the gas-distribution network using referenced data. The
relevant detectors underwent a rigorous regime of testing model was then scaled to the Keele network to quantify
to ensure any implications resulting from the presence of the baseline risk level of the network. Keele was chosen as

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

hydrogen could be accounted for and incorporated into the the host site in part because it is a very safe site. Each input
detection instrument. parameter was meticulously reviewed and, if necessary,
In tandem with the technical work stream, a thorough modified to account for any hydrogen effect. Once each
programme of appliance testing on the Keele site was car- permutation and mitigation measure had been incorpo-
ried out. Local Gas Safe engineers worked in collabora- rated into the model, a comparison with the baseline risk
tion with the gas-testing experts Kiwa Gastec to conduct was carried out.
a series of safety and performance tests on appliances fed The QRA demonstrated that the conveyance of blended
from the network at Keele that will have hydrogen blended gas up to 20 mol% hydrogen did not prejudice the safety of
into. Access was gained to 95% of the 101 domestic proper- the end users.
ties and evidence gathered on the remaining properties, The evidence base generated throughout Phase 1
which provided the necessary confidence that the installa- was presented to the HSE. Through a rigorous challenge
tions and appliances would be safe and reliable. and review process, the HSE granted an exemption to
The house-to-house testing consisted of carrying out a the hydrogen limit in Schedule 3 of the GS(M)R to allow
landlord’s Gas Safe test on all boilers to understand the hydrogen to be blended into the Keele network up to 20
baseline level of operational performance for appliances mol% for the purposes of the trial. The granting of the ex-
on the network. Any appliance deemed unfit or unsafe was emption marked the end of Phase 1 and the beginning of
remedied, with all works paid for by the project. The appli- Phase 2.
ances were then tested with a series of test gases to under-
stand how each individual appliance responded to the
introduction of hydrogen. No appliance required remedial 1.2 Phase 2
work to be undertaken as a result of the blended gas. All
Phase 2 of HyDeploy is the construction and operational
necessary remedial work was due to the appliance being
preparation phase of the project leading up to injection
unsafe or unfit. Leak tests were performed on the instal-
into the Keele network (Fig. 4) during the summer of 2019.
lations whilst charged with natural gas and then blended
The physical equipment to be installed includes:
gas. All appliances that passed on natural gas also passed
on blended gas. Any appliance that failed on natural gas (i) the 0.5-MW electrolyser;
was repaired and made tight; following remedial works, (ii) a grid entry unit that will mix the hydrogen with in-
the appliance was retested to ensure integrity. coming natural gas;
The house-to-house testing also conducted bottle test- (iii) sample points around the network to allow monitoring
ing on the commercial boilers fed from the Keele network, during the trial.

101 residential houses

8 student halls
17 office blocks & labs
7 recreational & service facilitates

Fig. 4  Keele university G3 network

 118 | Clean Energy, 2019, Vol. 3, No. 2

Alongside this primary equipment, further experimenta- will ensure the network and end appliances always main-
tion equipment will be installed. The in-trial experimen- tain safe functionality. Surveys of the network along with
tation will consist of materials testing, to understand other monitoring techniques will be conducted and com-
how real-world conditions of blended gas affect mechan- pared to the pre-trial baseline to allow any incremental
ical properties of common materials. As well as materials effect due to the presence of hydrogen to be detected.
work, accelerated testing of specifically installed domestic Following a successful trial of injection, the physical
boilers will be conducted. equipment will be decommissioned and moved on. The
The process of design and construction has been to network will then be returned to natural-gas operation.
national and international design standards with As Low
As Reasonably Practicable (ALARP) principals being applied
throughout. This focus on rigor has resulted in robust engi- 2  Scientific results
neering solutions being developed. A  thorough Factory The Phase 1 scientific programme provided the technical
Acceptance Testing (FAT) programme for equipment has

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

foundation of the evidence presented to the HSE in support
also been developed. of the exemption application. The evidence base generated
The operation of a gas-distribution network requires the spanned appliances, gas characteristics, gas detection and
use of procedures to respond to incidents and undertake materials. An outline of the key results within each area
work. During Phase 1, all relevant operating procedures follows.
were reviewed with experts from Cadent, Northern Gas
Networks (NGN) and Keele. This resulted in the generation
of targeted supplementary guidance and modification to 2.1 Appliances
gas-detection instrumentation to account for hydrogen The laboratory appliance testing had two overarching
effects. During Phase 2, all relevant operatives, both within objectives:
Cadent, Keele and third-party contractors, will be trained
against the developed guidance. (i) to understand the performance implications of
The training and verification of operatives are key out- introducing a hydrogen blend, across a wide range of
puts of Phase 2 and hydrogen will only be injected once appliances; and
verification of training has been received. Ensuring all rel- (ii) to understand the limit of operability for a select
evant operatives are fully competent to work on a network number of appliances with regard to hydrogen content
containing a hydrogen blend is a key safety measure that within the fuel.
must be achieved to ensure a safe and successful trial. The first work stream consisted of selecting a broad range
Baseline surveying and network monitoring will be of domestic gas appliances, including gas cookers, fires
conducted throughout Phase 2 to build a detailed picture and boilers. Each appliance was fed with 13 test gases at
of the business-as-usual network. This dataset will allow a constant test pressure, each designed to promote a dif-
accurate comparison with network and survey data taken ferent response or flame characteristic. The test gases con-
during the trial. sisted of both reference gas mixtures (G20/G21/G23/G222)
As part of Phase 2, further Gas Safe checks will be and hydrogen additions, the highest concentration being
offered to all residents on the affected Keele network. By G20 (methane) with 28.4 mol% hydrogen.
repeating the Gas Safe checks, the trial will build further Each appliance was fitted with instrumentation to pro-
confidence in the integrity of the network. vide accurate characterization of the combustion effects.
Construction, training and networking preparation will Thermocouples were installed on critical components
be carried out up until the summer of 2019, when the trial along with pressure measurement and flue-gas analy-
will officially commence. sis to ensure all of the appliance areas of interest were
monitored.
The gas-fire testing assessed the performance of the
1.3 Phase 3
appliances. A  number of fires were tested, representa-
Phase 3 is the trial itself, which is due to continue for tive of the fires present on the Keele network. All fires
10 months. The objective of the trial is to demonstrate that tested were capable of operating on all of the test gases.
hydrogen can be safely blended into a natural-gas grid at Thermocouple readings indicated no risk of overheating
20 mol% without prejudicing the safety of the end users. of components. Even when operating on a blend, the fires
The reason for starting the trial in the summer is to avoid retained the characteristic flame colour, albeit marginally
the winter peak gas demand—when demand for services is at more subdued. As hydrogen stabilizes flames due to the
its highest. By running the trial for 10 months, the summer low higher flame speed, testing was conducted on an oxygen
of gas demand along with the winter peak can be captured. deficiency sensor (ODS) that cut off fuel supply when the
Throughout the trial, a rigorous regime of monitoring, flame receded from a thermocouple due to lack of oxygen.
inclusive of the process equipment, network and appli- When tested under extreme conditions of a fully blocked
ances, will be undertaken. This level of operational scrutiny flue in a sealed room, all fires but one cut off the fuel supply
 Isaac | 119

below the 200 parts per million (p.p.m.) carbon-monoxide

requirement.
The fire that did not cut off the fuel below 200  p.p.m.
presented a baseline reading of just below the cut-off at
180  p.p.m.; it therefore was an outlier to industry stand-
ards. All of the fires on the Keele site fitted with ODS
devices were verified to cut off below the 200-p.p.m. limit.
However, the testing indicated that fires with ODS devices
will require further test work to fully understand the base-
line variability of the devices for wider roll-out.
The gas-cooker results indicated that all critical com-
ponent temperatures remained within acceptable limits
when hydrogen was introduced. Variation in thermocou-

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

ple readings was observed as a result of changing the
fuel composition, but at no point did the variation in
temperature results indicate overheating of components Fig. 6  Methane + 28.4 mol% hydrogen flame
or potential degradation. Pressure indication within the
oven always remained within the necessary performance
safety specifications. Figs 5 and 6 show the resulting flame of reduction did not compromise the protective safety
appearance of pure methane (Fig. 5) compared to methane function of the device. For the minority of appliances
with 28.4 mol% hydrogen (Fig. 6). The difference is mar- that use flame-ionization current to control the fuel/air
ginal, which is indicative of the wider gas-cooker results. ratio, there was an observed difference in control be-
The gas-boiler testing focused on flue-gas analysis haviour but not enough to be deemed a safety concern.
alongside internal-temperature measurements and flame- Consideration should be given for further test work on
ionization currents. The flame characteristics for each test such control schemes.
gas, including up to 28.4 mol% hydrogen, were stable, with Following the baseline operational testing to determine
complete combustion being achieved. As expected, the expected performance of appliances during the trial, limit
carbon-dioxide readings reduced by up to 0.5 mol% with testing was conducted to understand the operational lim-
the addition of hydrogen. This metric is monitored as part its of a sample of appliances. This testing involved vary-
of servicing; although still within manufacturers’ guide- ing the composition of the fuel by increasing the hydrogen
lines, servicing engineers need to understand that this content until operational issues arose.
reduction is to be expected. The principal effect that hydrogen promotes for gas
The critical-temperature readings assessed compo- appliances is an increase in flame speed, which could
nents such as the burner plate and the heat exchanger. lead to flame-out as a result of the flame travelling at a
Much like the gas cooker, variation was observed but higher velocity than the flue/air mixture. For the selection
not enough to indicate incremental degradation or per- of appliances tested, flame-out due to light back began at
formance issues. As expected, the flame-ionization cur- 80 mol%, although some appliances only flamed out at
rent reduced with the addition of hydrogen. The level 100 mol% hydrogen. The limit of operability for the appli-
ances tested was far beyond the operating limits of up to
20 mol%. This work is foundational and further limit test-
ing with a wider range of appliances would be beneficial
to better understand the operational limits of current gas
appliances.
Overall, the laboratory testing indicated that, across the
range of domestic appliances tested, they were capable of
operating safely on hydrogen concentrations up to 28.4
mol%. It should be noted that all gas appliances sold in the
UK are certified with reference gas G222, which contains
23 mol% hydrogen.

2.2  Gas characteristics

The gas characteristics of natural gas containing up to
20 mol% hydrogen, in comparison to natural gas, was an
important area of understanding to allow appropriate sup-
plementary guidance for operational procedures to be pro-
Fig. 5  100 mol% methane flame duced, where needed.
 120 | Clean Energy, 2019, Vol. 3, No. 2

The focus of the gas-characteristics work stream was to self-detonate. It was determined that, much like natural
understand the pertinent safety-related characteristics of gas, blended natural gas containing up to 20 mol% does
the gas relative to natural gas; this included: not express the characteristic of self-detonation.
An important characteristic of any gas is its British
(i) dispersion characteristics in the event of a leak;
Standard gas group. The gas-group rating of a gas defines
(iii) flammability characteristics;
the specifications of any electrical equipment that could be
(iii) combustion characteristics in the event of ignition.
exposed to it, as described by the Atmospheric Explosion
The dispersion characteristics of natural gas containing up (ATEX) rating of the electrical equipment. Both natural
to 20 mol% hydrogen were found to be comparable with gas and natural gas containing up to 20 mol% hydrogen
natural gas. As laminar flow is a function of the viscosity are deemed as IIA gases. Therefore, any current electrical
of the fluid, the relative leak rate of a 20 mol% hydrogen equipment suitability-rated for natural-gas exposure is
blend was found to be equivalent to natural gas—due also sufficient for exposure to blended gas containing up
to the blend having a viscosity that is 99% of methane’s. to 20 mol% hydrogen.

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

Turbulent flow is a function of the density of the fluid; The results of the gas-characteristics work stream were
therefore, given that the density of the blend was 85% aggregated and provided a framework for supplementary
relative to methane, it was determined that a turbulent guidance to be generated for the operational procedures.
leak containing 20 mol% hydrogen could be up to 10 vol% This ensured any ‘hydrogen effect’ was properly accounted
greater. However, given the lower enthalpy of combustion for within the relevant operational procedures to maintain
of hydrogen relative to natural gas, the potential energy- the same level of operational integrity currently experi-
release rate was in fact lower for leaks. enced by the Keele network.
Hydrogen and methane are extremely miscible; there-
fore, separation of the gases was assessed as not being
realistic in the context of the trial. For example, an isother- 2.3  Gas detection
mal column would need to be in the order of 100s of metres Accurate gas detection is a fundamental requirement for
to enable enough gravitation potential to separate the two the safe operation of a gas-distribution network. An exper-
gases from each other. This is consistent with other inter- imental programme was therefore undertaken to test com-
national assessments. monly used gas detectors, including domestic detectors, at
The flammability limits of hydrogen are known to be hydrogen blends of varying concentrations and map the
wider than natural gas; therefore, understanding the result- output responses of the instruments.
ant implications of blending up to 20 mol% was of inter- The full suite of output measurements was recorded,
est. The limit of most interest is the Lower Flammability inclusive of flammable-gas readings from the p.p.m. range
Limit (LFL), as this defines the point at which a mixture of to vol% range through to carbon-monoxide measurements.
released fuel and air becomes a hazard. The variability of All commonly used instruments on both the Cadent and
the LFL resulting from the spectrum of natural-gas com- NGN networks were tested; a selection is shown in Fig. 7.
positions leads to an imprecise conclusion. The effect of A testing enclosure was constructed with inlet connec-
mixing methane with hydrogen at 20 mol% resulted in a tions of methane and hydrogen; this allowed any desired
reduction in the LFL limit from 5 to 4.75 vol%. Although atmospheric composition to be generated and multiple
this magnitude of change is comparable to the baseline instruments to be tested with an identical environment.
variability of natural-gas data, a conservative position was
adopted to recommend a reduction in the LFL limit from 5
to 4.75 vol%.
Because hydrogen has a lower density than natural gas,
the buoyancy of a blend is greater than the buoyancy of
natural gas. In an open environment, this would aid dis-
persion from a potential leak; however, within an enclosed
environment such as a room, there would be little expected
difference between the two gases.
The flame speed of hydrogen is known to be greater
than the flame speed of natural gas. Therefore, a 20-mol%
hydrogen blend exhibits slightly different combustion
characteristics compared to natural gas. Relevant research
on a 20-mol% hydrogen mixture that is applicable to real-
world situations is limited, so a conservative assessment
was carried out. It was found that the overpressure pro-
file of blended gas resulted in higher peak overpressures
relative to natural gas. The greatest potential risk of the
overpressure characteristic of a gas is the propensity to Fig. 7  Gas-detection instruments. (a) Survey detectors; (b) fixed detectors.
 Isaac | 121

There was a broad pattern of results from the detectors, 2.4 Materials

indicating that the presence of hydrogen does have an
The interaction of hydrogen with materials is an area of
effect on detectors currently used on the network. For
critical importance that must be understood when eval-
flammable-gas measurements, which generally rely on
uating hydrogen injection. The initial stage in this work
a thermal-conductivity-based sensor, readings would be
stream was to undertake a thorough asset register of the
oversensitive. This does not present a direct safety con-
Keele University gas network to evaluate the spectrum of
cern, as the effect is conservative; however, issues con-
materials for laboratory testing.
cerning nuisance alarms and operative confidence in the
Once all of the relevant materials on the Keele net-
instruments are import concerns. For the purpose of the
work were understood, a process of three separate testing
trial at Keele, a linearly affected instrument was chosen
regimes was undertaken:
and the appropriate calibration calculated to ensure the
output measurement would be robust to any blend up to (i) assess hydrogen uptake into powder and rod samples
20 mol% hydrogen. The manufacturer will calibrate the de- up to 75°C and up to 9 weeks of soaking;

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

tectors to the desired level to certify the detectors are fit (ii) soak specimens in a chamber for up to 6 weeks
for purpose. followed by mechanical testing;
The carbon-monoxide sensor used in common detec- (iii) assess any implications for electrofusion jointing and
tion instrumentation is an electrochemical cell that pipeline ‘squeeze-off’ techniques.
oxides the carbon monoxide and measures the rate of
The soaking chamber is shown in Fig. 8.
hydrogen-ion production. When hydrogen encounters
The importance of the three testing regimes was to
the sensor, it breaks down and produces further hydro-
understand not only any material effects, but also implica-
gen ion, therefore providing a second signal—this is
tions for operational procedures.
interpreted as carbon monoxide and results in an output
The powder and rod sampling to measure hydrogen
measurement. Much like flammable-gas instrumenta-
absorption concluded that, for around half the materials
tion, the laboratory testing indicated that the presence
tested, no meaningful hydrogen take-up was observed for
of hydrogen generated conservative readings that could
either the powder or rod samples. The materials that did
result in response procedures being invoked. The poten-
absorb a level of hydrogen released it upon heating.
tial risk is therefore nuisance and operative confidence
The soaking of samples, as pictured in Fig. 8, was under-
in the detector readings.
taken at 1.5 barg in conditions of pure methane, pure
Following a review of available detectors on the market, a
hydrogen and reference gas G222 (containing 23 mol%
technical selection process was conducted to identify appropri-
hydrogen). The soaking time was up to 6 weeks, followed
ate instrumentation that would not result in nuisance alarms.
by testing of mechanical properties.
A  detector was identified that contains a carbon-monoxide
The tensile testing measured:
hydrogen-compensated sensor that provides the necessary
level of compensation to not produce nuisance alarms due to (i) total elongation at failure;
a blend of up to 20 mol% hydrogen. The detector also has the (ii) modulus of elasticity;
appropriate ATEX rating and portability requirements. (iii) ultimate tensile strength;
The identified carbon-monoxide detector will be used (iv) proof strength.
in conjunction with an appropriately calibrated flamma-
ble-gas detector for the purposes of the trial; this strategy
was reviewed and agreed as part of the exemption pro-
cess. The two detectors in combination provide the nec-
essary level of measurement accuracy to allow all current
action levels to be maintained. A dedicated training pack-
age of familiarization has been developed for all opera-
tives and contractors associated with the operation of the
Keele network. Following delivery of the training package,
operatives will be deemed competent in the use of related
procedures and gas detection equipment.
Following the successful identification of available
detectors and the development of the associated training,
a process of industrial engagement with manufacturers
began. This process seeks to engage with gas-detection-
instrumentation manufacturers to provided a framework
of collaboration and facilitate the development of a single
detector capable of being deployed on both a natural-gas
network and blended network containing up to 20 mol%
hydrogen. Fig. 8  Materials-soaking chamber
 122 | Clean Energy, 2019, Vol. 3, No. 2

The results of the tensile testing for the suite of materials is expected due to exposure to blended natural gas con-
tested showed no noticeable effects on the tensile proper- taining up to 20 mol% hydrogen.
ties of materials on the network resulting from exposure
to the hydrogen blends at the operational pressures.
3 Equipment design
Alongside assessments of identified specific components,
these results provided the confidence that no incremental The process design of HyDeploy consists of the instal-
reduction in Keele network integrity will be experienced lation of a 0.5-MW electrolyser, supplied by ITM Power,
due to the introduction of a hydrogen blend at 20 mol%. along with a hydrogen grid entry unit (H2GEU), supplied by
The electrofusion and squeeze-off testing is relevant for Thyson. The process of hydrogen production and injection
operations associated with polyethylene (PE) pipelines. Fig. 9 is graphically represented in Fig. 10.
shows the process of squeezing off a PE pipe—a procedure The totality of the installation, which will be con-
used to isolate pipeline sections. structed at Keele, is inclusive of:
The piping samples were soaked in pure hydrogen for (i) a take-off line redirecting the main feed into the net-

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

6 weeks and then squeezed off, followed by hydrostatic work to the compound containing the equipment;
testing 1 and 6 weeks after. The pipework passed both (ii) the 0.5-MW electrolyser, which will split water using elec-
hydrostatic tests, indicating that exposure to hydrogen did tricity to produce hydrogen and oxygen; the oxygen will
not compromise the pipeline’s integrity to be isolated and be vented and the hydrogen passed into a buffer tank;
sequentially returned to service. (iii) the control scheme of the H2GEU will measure the in-
During Phase 3, further materials work will be undertaken coming natural-gas Wobbe Index and calculate how
via soaking 42 sample specimens in the actual blended gas much hydrogen can be blended into the natural gas
over the course of the 10-month trial. Twenty-one samples whilst remaining within the GS(M)R Wobbe range;
will be removed half way through the trial and the other (iv) up to 20 mol% will be mixed with the natural gas; the
half removed at the end. Each batch will undergo the same blended gas will then pass through a ‘volume loop’ to
mechanical-property testing as was conducted in Phase 1. The allow a final check of the composition to take place before
purpose of testing two batches is to establish whether steady- returning the blended gas to the main feed line into the
state effects have been achieved and to build a more detailed Keele network;
picture of expected long-term effects in real-world conditions. (v) if, at any time, the blended gas is out of specification, the
Further testing is recommended for materials with network supply will be automatically returned to natural
corroded surfaces, as surface effects play a role in the gas and the out-of-spec gas purged from the system;
absorption mechanics of hydrogen. To facilitate the wider (vi) if, for any reason, there is a process upset that prevents
deployment of hydrogen blending, further materials test operation of the electrolyser or H2GEU, the network
work would be needed to test materials not present on the supply will be automatically returned to natural gas.
Keele network. This additional testing would be needed to
extend the evidence base beyond the Keele network and Within the volume loop of the H2GEU, there will be a
out onto the public network. housing containing material specimens. These material
It was concluded that, for the purposes of the trial on the specimens will be exposed to real-world blended gas
Keele network, no incremental degradation of materials for the course of the trial. Half of them will be removed
halfway through the trial for tensile testing, whilst the
other half will be removed at the end of the trial for the
same tensile testing.
The process of designing the equipment has been a col-
laborative effort with all relevant international and British
Standards applied. A  full Hazard and Operability Study
along with a Layers of Protection Analysis have been car-
ried out to ensure rigorous application of ALARP principals
was performed throughout the design process.
All equipment to be installed will undergo a FAT, which
will ensure the functional specification of the equipment
has been achieved through the manufacturing process.
Only once a positive result has been obtained through the
FAT will equipment be certified to be installed on the Keele
network.
Once the trial at Keele has concluded, the equipment
will be decommissioned and the network returned to
operation on natural gas. The electrolyser and H2GEU will
be transported to the first trial location for HyDeploy2 to
Fig. 9  Squeeze-off demonstration commence public trials.
 Isaac | 123

Compound

Power
New Assets
Water
Electrolyser
Existing Assets Drain

Keele Assets
H2

Natural Gas
Power
H2GEU
Natural Gas + H2 Blend

Volume Loop

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

Experimental Equipment

G3 Governor

Fig. 10  HyDeploy process design flow chart

4  Blending project pipeline

Full deployment plan & roadmap
4.1 HyDeploy2
HyDeploy is the foundational project for hydrogen-blend- Public Trial 1
ing deployment in the UK. It will create the initial evidence Address evidence
base to demonstrate that hydrogen can be blended into gap for public trials
a live operating gas-distribution network at 20 mol% in a Public Trial 2
safe and non-disruptive manner.
The next step is to replicate HyDeploy on a larger public Address evidence gap for wider deployment
network. HyDeploy2 is the connecting step between trial-scale
deployment and commercialization—further building the neces- Fig. 11  HyDeploy2 programme structure
sary evidence to demonstrate hydrogen blending. HyDeploy2 will
consist of the same project development team as HyDeploy, with
each trial. The purpose of this work will be to broaden the
Cadent and NGN as the network sponsors, Progressive Energy as
initial evidence base developed through HyDeploy within
the project developer and with the scientific programme lead by
the context of public networks. For example, further ma-
HSL. HyDeploy2 will be the largest gas project ever funded through
terials work will be needed to assess all potential materials
the Ofgem Network Innovation Competition, with confirmation
on the Local Transmission System (LTS) and understand
of funding received in November 2018. The strategic objective
any implications of blended-gas contact.
of HyDeploy2 is to bring hydrogen injection to the same level of
Successful delivery of HyDeploy2 will lay the founda-
regulatory approval as biomethane injection. This would create
tion for commercial deployment and injection of hydrogen
the regulatory framework for commercial-scale production to
blends at a material scale.
be developed and injected into the gas grid without undertaking
network checks or further evidence gathering.
The route map for HyDeploy2 (Fig. 11) is to carry out two
4.2 HyNet
further trials over the course of 4  years, starting in 2019— HyNet is the commercial realization of hydrogen blending
in total supplying ~1500 households with blended gas. The into the natural-gas network and is being developed in col-
magnitude of pre-trial appliance testing on blended gas will laboration between Cadent and Progressive Energy in the
be incrementally rolled back, through separate exemptions north-west of England.
for each trial. This incremental approach through two trials is HyNet will be the first project on a national scale to develop
required, as, ultimately, sufficient confidence must be gener- hydrogen production for the purposes of decarbonizing both
ated to justify deployment without any pre-injection appli- industrial clusters and domestic heat. Using Auto Thermal
ance testing. Reformer Technology (ATR) in conjunction with a Carbon
Throughout HyDeploy2, a large body of laboratory work Capture and Storage infrastructure, as shown in Fig. 12, the
will be undertaken to support the exemption process for output of the initial HyNet project will be to achieve:
 124 | Clean Energy, 2019, Vol. 3, No. 2

(ii) gas detection to support network operation is possible

with current technology and is further enhanced by
the operational attributes of the Keele network;
(iii) no materials-related issues have been identified on

Keele’s network at operational pressures for natural
gas containing up to 20 mol% hydrogen;
(iv) properly installed and maintained domestic appliances
are just as safe whilst operating with a blended gas of
up to 20 mol% hydrogen.

The extent of applicability of the above conclusions is cur-

rently bounded to the Keele network, as the evidence base
Fig. 12  HyNet infrastructural model generated was necessarily focused on undertaking a safe

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019

trial at Keele University.
Through the process of HyDeploy2, the underlying evi-
(i) carbon savings of +1 MTCO2/yr;
dence base will be broadened to be representative of the
(ii) non-disruptive decarbonization of 2 million homes in
wider public network. This will build regulatory confidence
the Manchester–Liverpool region;
in hydrogen blending to ultimately facilitate commercial
(iii) decarbonization of industrial clusters via hydrogen

deployment through HyNet and beyond. For further read-
conversion.
ing on the topic of hydrogen blending please see [6–12].
The HyNet programme will create an underlying carbon-
dioxide transport and storage infrastructure to enable
further carbon savings. The design of the project will be
Acknowledgements
inherently expandable, allowing further development of
Acknowledgement should be given to the Office of Gas &
hydrogen production and carbon capture to accommodate
Electricity Markets for providing the funding for both HyDeploy
geographical expansion.
and HyDeploy2 to enable the programme described to be delivered.
Operation of the hydrogen-production facility is sched- Acknowledgement should also be given to the project partners:
uled for the mid-2020s, which will allow carbon savings to Cadent Gas, Northern Gas Networks, Health & Safety Executive—
be realized in line with the 2018 Committee on Climate Science Division, Keele University and ITM Power, as well as the
Change report on the UK decarbonization progress [5]. The technical contributors Otto Simon Limited, Kiwa Gastec and Dave
project scope will be inclusive of ATR technology generat- Lander Consulting, along with advisory organizations who have
contributed to HyDeploy.
ing 800 MW(th) of hydrogen, to be distributed in a purpose-
built hydrogen network feeding directly into industrial
clusters and injecting into the LTS at four strategic loca-
tions to allow a blend to be provided to 2 million homes. References
The strategic objective of HyNet is to build upon the HyDeploy [1] National Grid. Future Energy Scenarios Workbook. July 2018.
project series to deploy hydrogen in homes on a national scale, http://fes.nationalgrid.com/fes-document/ (21 February 2019,
in a non-disruptive manner. By avoiding the need to modify date last accessed).
[2] Department for Business, Energy & Industrial Strategy.

existing domestic appliances, a large barrier to deployment
Electricity Fuel Uses, Generation and Supply (DUKES 5.6). July
of hydrogen is removed. Once deployed, HyNet will create the 2018.
foundational infrastructure necessary to facilitate and promote [3] ENGIE. The GRHYD Demonstration Project. 2019. https://www.engie.
deeper carbon savings via the hydrogen vector. com/en/businesses/gas/hydrogen/power-to-gas/the-grhyd-
demonstration-project/ (21 February 2019, date last accessed).
[4] Department for Business, Energy & Industrial Strategy. Supply
5 Conclusions and Consumption of Natural Gas and Colliery Methane (DUKES
4.2). July 2018.
HyDeploy has successfully reached the end of Phase 1, [5] Stark C, Gault A, Joffe D, et al. Economy-wide performance in
with the granting of the UK’s first ever exemption for reducing emissions. In: Reducing UK Emissions: 2018 Progress
hydrogen injection to a gas network by the HSE. The Report to Parliament. London: Committee on Climate Change,
exemption is specifically for the trial at Keele for the 2018.
[6] Hodges  JP, Geary  W, Graham  S, et  al. Executive summary:

duration of the trial, although HyDeploy has demon-
main findings. In: Injecting Hydrogen into the Gas Network—a
strated that regulatory approval for hydrogen blending
Literature Search (Research Report 1047). Buxton: Health and
up to 20 mol% is possible. Safety Laboratory, 2015.
The current conclusions from the undertakings of [7] Melaina MW, Antonia O, Penev M. Executive summary: impact
Phase 1 indicate that: on end-user systems. In: Blending Hydrogen into Natural Gas
Pipeline Networks: A Review of Key Issues. Oak Ridge, TN: National

(i)
through meticulous review and consideration, Renewable Energy Laboratory, 2013.
procedures for the operation of a hydrogen-blended [8] Korner  A, Cecilia  Tam, Simon  Bennett. Technology Roadmap:

network can be developed; Hydrogen and Fuel Cells. Paris: International Energy Agency, 2015.
 Isaac | 125

[9] Ajayi-Oyakhire  O. Hydrogen—Untapped Energy? Kegworth: [11] Joffe D, Livermore S, Hemsley M, et al. Hydrogen in a Low-carbon
Institution of Gas Engineers & Managers (IGEM), 2013. Economy. London: Committee on Climate Change, 2018.
[10] HyDeploy website. https://hydeploy.co.uk/ (21 February 2019, [12] Wood C, Brown A. The Future of Gas: Transition to Hydrogen in the
date last accessed). Gas Grid. London: Dentons, 2019.

Downloaded from https://academic.oup.com/ce/article-abstract/3/2/114/5487479 by guest on 26 October 2019