the wind/hydrogen demonstration system at utsira in norway ......the wind/hydrogen demonstration...

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 8 4 1 – 1 8 5 2

Avai lab le a t www.sc iencedi rec t .com

j ourna l homepage : www.e lsev ier . com/ loca te /he

The wind/hydrogen demonstration system at Utsira inNorway: Evaluation of system performance using operationaldata and updated hydrogen energy system modeling tools

Øystein Ulleberg a,*, Torgeir Nakken b, Arnaud Ete c

a Institute for Energy Technology, P.O. Box 40, NO-2027 Kjeller, Norwayb StatoilHydro Research Centre, NO-3908 Porsgrunn, Norwayc University of Strathclyde, Energy Systems Research Unit, Glasgow, Scotland G1 1XJ, UK

a r t i c l e i n f o

Article history:

Received 15 May 2009

Received in revised form

23 October 2009

Accepted 25 October 2009

Available online 15 January 2010

Keywords:

Wind

Hydrogen storage system

Water electrolysis

Stand-alone power system

Demonstration plant

Operational experience

System simulation

TRNSYS

HYDROGEMS

* Corresponding author.E-mail address: [email protected] (Ø

0360-3199/$ – see front matter ª 2009 Profesdoi:10.1016/j.ijhydene.2009.10.077

a b s t r a c t

An autonomous wind/hydrogen energy demonstration system located at the island of

Utsira in Norway was officially launched by Norsk Hydro (now StatoilHydro) and Enercon

in July 2004. The main components in the system installed are a wind turbine (600 kW),

water electrolyzer (10 Nm3/h), hydrogen gas storage (2400 Nm3, 200 bar), hydrogen engine

(55 kW), and a PEM fuel cell (10 kW). The system gives 2–3 days of full energy autonomy for

10 households on the island, and is the first of its kind in the world. A significant amount of

operational experience and data has been collected over the past 4 years. The main

objective with this study was to evaluate the operation of the Utsira plant using a set of

updated hydrogen energy system modeling tools (HYDROGEMS). Operational data (10-min

data) was used to calibrate the model parameters and fine-tune the set-up of a system

simulation. The hourly operation of the plant was simulated for a representative month

(March 2007), using only measured wind speed (m/s) and average power demand (kW) as

the input variables, and the results compared well to measured data. The operation for

a specific year (2005) was also simulated, and the performance of several alternative

system designs was evaluated. A thorough discussion on issues related to the design and

operation of wind/hydrogen energy systems is also provided, including specific recom-

mendations for improvements to the Utsira plant. This paper shows how important it is to

improve the hydrogen system efficiency in order to achieve a fully (100%) autonomous

wind/hydrogen power system.

ª 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1. Introduction hydrogen technology in remote areas is the notion that

Hydrogen can be used to store variable renewable energy,

such as solar and wind energy [1,2]. Previous studies [3–6]

show that there is a potential market for wind/hydrogen

energy systems in remote areas and/or weak grids, specifi-

cally for (1) Stand-alone power systems and (2) Hydrogen

vehicle fueling stations. The main motivation for applying

. Ulleberg).sor T. Nejat Veziroglu. Pu

locally produced renewable hydrogen will be able to

compete with traditional fossil fuels (e.g., diesel) sooner

here than in more densely populated areas. The situation

today is that remote communities are already experiencing

relatively high fuel costs. Remote areas with favorable

winds are therefore logical targets for wind energy based

hydrogen systems.

blished by Elsevier Ltd. All rights reserved.

mailto:[email protected]

http://www.elsevier.com/locate/he

Nomenclature

Acronyms:

AC, DC alternating current, direct current

BOP balance of plant

CO2, H2 Carbon dioxide, Hydrogen

EES engineering equation solver

FC fuel cell

HHV higher heating value

HYDROGEMS HYDROGen Energy ModelS

IFE Institute for energy technology

IGBT insulated-gate bipolar transistor

NiCd nickel cadmium

PEM proton exchange membrane

R&D research & development

RE renewable energy

SAPS stand-alone power systems

TRNSYS TRaNsient SYstem simulation program

WECS wind energy conversion system

Symbols & Units:

P power, kW

v speed, m/s

h hour

kVA kilovolt–Ampere

kW, MW kilowatt, Megawatt

ms millisecond

m2 square meters

Nm3 normal cubic meters

V euro


Several wind/hydrogen concepts and system designs have

been studied over the past decade [4–9], and a few physical

installations have been made [3,7,10–19]. An overview of the

most known wind/hydrogen demonstration systems and pilot

plants installed around the world the last decade is provided

in Table 1. Most of the installations listed here are small-scale

systems based on wind turbines with only few kWs and a DC-

busbar stabilized by a battery bank. The most notable excep-

tion is the autonomous wind/hydrogen system at the island of

Utsira (Norway), which provides power to 10 households via

a local AC mini-grid [12,18]. Another relatively large system is

the DC-based system serving the West Beacon Farm in

Table 1 – Overview of wind/hydrogen-systems installedworld-wide, 2000–2008.

Yeara Location Name of project Refs.

2000 ENEA Research

Centre, Casaccia, Italy

Prototype wind/electrolyzer

testing system

[7]

2001 University of Quebec,

Trois-Rivieres, Canada

Renewable energy

systems based on hydrogen

for remote applications

[10]

2004 Utsira Island, Norway Demonstration of

autonomous

wind/hydrogen-systems

for remote areas

[12,18]

2004 West Beacon Farm,

Loughorough, UK

HARI–Hydrogen and

Renewables Integration

[13]

2005 Unst, Shetland

Islands, UK

PURE–Promoting Unst

Renewable Energy

[3]

2006 IFE, Kjeller, Norway Development of a

field-ready small-

scale wind-hydrogen

energy system

[16,17]

2006 NREL, Golden,

Colorado, USA

Wind-to-hydrogen

(Wind2H2)

demonstration project

[14,15]

2007 Pico Truncado,

Argentina

Wind/hydrogen

demonstration plant

[11]

2007 Keratea, Greece RES2H2 wind-hydrogen

pilot plant, with H2-storage

in metal hydrides

[20]

a Approximate year of commissioning.

Loughborough (UK) [13], which has a hydrogen storage

capacity of about 2850 Nm3, slightly larger than the 2400 Nm3

installed at Utsira. Most of the systems listed in Table 1 were

based on commercially available alkaline electrolyzers with

rated hydrogen production capacities and operating pressures

in the range of 0.2–10 Nm3/h and 7–20 bar, respectively. The

exceptions were the prototype PEM-electrolyzer at IFE

(0.3 Nm3/h and 15 bar) [16] and the prototype high-pressure

alkaline electrolyzer in the PURE-project (3.6 Nm3/h and

55 bar) [19]. In comparison, the hydrogen pressure and

production capacity for the alkaline electrolyzer installed at

Utsira was 12 bar and 10 Nm3/h.

The overall motivation for the Utsira project is to demon-

strate how renewable energy and hydrogen systems can

provide safe and efficient power supply to communities in

remote areas. The main goal for the project is to perform

a full-scale demonstration and testing of a wind/hydrogen

energy system [12]. After several years with concept devel-

opment, system design, and project planning undertaken by

several partners [21,22], the decision to build the Utsira plant

was made in April 2003. The installations in place were mainly

realized by Norsk Hydro (now StatoilHydro) and Enercon, the

two main partners in the project. The Utsira plant was inau-

gurated in July 2004, is still in operation, and continues to play

an important part in StatoilHydro’s R&D on hydrogen tech-

nology [23]. A significant amount of operational data and

experience has been collected since the commissioning of all

of the systems during the winter 2004/2005 [18]. The main

objective of this study is to evaluate the energy performance

of the Utsira plant using actual operational data and an

updated set of hydrogen energy system modeling tools. A

thorough discussion on issues related to the design and

operation of wind/hydrogen energy systems is also provided,

including specific recommendations for improvements to the

Utsira plant.

2. System description

The Utsira wind/hydrogen energy demonstration plant is

located at the island of Utsira, ca. 20 km off the west coast of

Haugesund in Norway. This island was picked because of its

Fig. 1 – Photo of the Utsira wind/hydrogen demonstration plant (Ulleberg, 2005).

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 1 8 4 1 – 1 8 5 2 1843

excellent wind conditions and small, but representative,

electrical power demand. The island also provides a great

opportunity to test autonomous renewable energy mini-grid

power systems, as back-up power from the main land is only

available through a subsea cable.

A photo and a schematic of the Utsira system are given in

Figs. 1 and 2, respectively. Two 600 kW wind turbines are

installed on the site, but only one of the wind turbines is

dedicated to the demonstration plant. The main components

in the hydrogen system are a 10 Nm3/h alkaline electrolyzer

(12 bar), an 11 Nm3/h hydrogen compressor (12–200 bar, 2-

stage, diaphragm), a 2400 Nm3 hydrogen gas storage (200 bar),

and a 55 kW hydrogen engine generator system (genset). A

10 kW PEM fuel cell is also part of the system, but this is

currently not in use. The characteristics of the main compo-

nents are summarized in Table 2, some of these are discussed

in more detail below. A more detailed description on how the

system is connected electrically is also provided below.

2.1. Local grid and customer connection

The components in the autonomous power system at Utsira

(Fig. 1) are electrically connected at 400 V and 50 Hz. A sepa-

rate transformer (315 kVA, 22/0.4 kV) is connected to a 1.5 km

long cable that transmits power from the autonomous system

to the customer substation. Here 10 households are connected

to the customer substation at 230 V, which is the standard

voltage level in Norway. The customer substation also

includes a 22 kV busbar circuit breaker to enable fast switch-

ing of the customers from autonomous power system mode to

grid-connected mode. This feature was added to safeguard

against power failure, and to simplify service, maintenance,

and modifications on the demonstration plant. Since the

circuit breaker could also be used in case of an emergency

(e.g., power failure), the requirement for redundancy in the

autonomous system was minimized [12]. The system at Utsira

is capable of 2–3 days of full energy autonomy for 10 house-

holds located on the island.

2.2. Wind turbines and grid stabilizing equipment

Two wind turbines (Enercon E-40), each with a maximum

power output of 600 kW, have been installed on the site

(Fig. 1), but only one of these is dedicated to the autonomous

power system. The wind turbine is connected to the autono-

mous system via a separate 300 kW one-directional inverter,

which draws electricity from the wind turbine DC-circuit and

feeds this to the autonomous system. (The other wind turbine

feeds power directly into the local grid, and is not part of the

autonomous system). In practice, this means that there is

a cut-off in power from the wind turbine at 300 kW, and

surplus power can be fed to the local grid (in parallel to the

other wind turbine). Grid stabilizing equipment was required

for the autonomous system, as this was connected to a weak

grid. The grid stabilizing equipment installed consists of

a flywheel (5 kWh) for frequency control, a master synchro-

nous machine (100 kVA) for voltage control and short circuit

power, and a NiCd battery (34 kWh, 100 kW) for redundancy.

2.3. Hydrogen genset and fuel cell

The hydrogen engine generator (genset) was a diesel engine

genset rebuilt for using hydrogen as fuel. Its rated power

capacity (55 kW) is sufficient to supply the customers

without relying on the fuel cell. The hydrogen genset was

designed for black start and parallel operation with the fuel

cell. It also provides voltage control and inertia to the

autonomous system. One of the most challenging tasks in

this part of the project was to find a fuel cell stack supplier

to take responsibility for the overall fuel cell system inte-

gration. The fuel cell system had to comply with strict

requirements for autonomous power system and tough

10 Houses

∼=

H2-engine

Wind Turbine

~

~

H2-storage

Water

Electrolyzer

H2-Compressor

Rectifier

Flywheel

Battery

Low Voltage Mini-Grid

Fuel

Cell

∼=

Sync. Machine

Inverter

Auxiliaries:

Feed water & cooling pumps, gas drier,deoxidizer, ventilation, fans, heating, lighting

Fig. 2 – System schematic of the wind/hydrogen demonstration plant installed at Utsira.


climatic conditions on the site. A manufacturer was even-

tually found, and the fuel cell system was integrated inside

a heated container. The fuel cell system was also designed

to be capable of black start.

3. Operational data and system performance

Over the past 3 years a significant amount of data from stand-

alone operation has been collected from the system at Utsira.

This particular study was based on operational data (10-min

averages) measured at Utsira in the period of 1–30 March 2007

(Fig. 3). The top plots in Fig. 3 show the wind power production

Table 2 – Main components installed at the Utsira wind/hydrogen demonstration plant.

Systemcomponent

Ratedcapacity

Supplier/manufacturer

Wind turbine 600 kW Enercon, Germany

Hydrogen engine 55 kW Continental, Belgium

Fuel cell 10 kW IRD, Denmark

Electrolyzer 10 Nm3/h, 50 kW Hydrogen

Technologies,a

Norway

Hydrogen compressor 5.5 kW Andreas Hofer,

Germany

Hydrogen storage 2400 Nm3,

200 bar (12 m3)

Martin Larsson,

Sweden

Flywheel 5 kWh Enercon

Battery 50 kWh Enercon

Master synchronous

machine

100 kVA Enercon

a Hydrogen Technologies AS (previously Norsk Hydrogen Electro-

lyzers AS) is fully owned by StatoilHydro.

and user load power demand, while the bottom plots show the

power produced by the hydrogen engine genset and power

consumed by electrolyzer (including auxiliaries and

compressor) versus pressure in the hydrogen storage. A

decrease in the hydrogen pressure from 145 bar to 33 bar from

1 to 30 March can be observed; the pressure increases up to

165 bar during a period with high levels of wind energy (Fig. 3,

period 1), before it decreases at times with less wind energy.

Data from March 2007 shows that 100% stand-alone oper-

ation is achieved ca. 65% of the time. This is slightly better

than the long-term performance, which on average gives

stand-alone operation about 50% of the time. A closer look at

the data (March 2007, hour 420–720) shows that the electro-

lyzer frequently needed to operate on grid-electricity in order

to produce extra hydrogen and level out the hydrogen storage

pressure, which otherwise would have decreased very rapidly.

A significant part of the work in the Utsira project has been

dedicated to the development of the power conditioning

system. A flywheel, a synchronous generator, and a battery

system ensure that the voltage and frequency on the local

mini-grid (the 10 households and the balance of plant for the

hydrogen system) is kept within standard limits and toler-

ances. The performance of the power system at Utsira is best

illustrated by taking a closer look at data for a shorter time

period (5 March 2007).

The on/off-switching of the electrolyzer and hydrogen

engine genset, as a function of the net available power (wind

power–user load) is illustrated in some more detail in Fig. 4: (1)

The master control system is programmed to switch on the

electrolyzer only during time of excess power (wind-load). (2)

Once the excess power on the grid goes below a critical low

point, the electrolyzer is switched off (2a) and the H2-engine

genset is switched on (2b). (3) Short term increases in power is

used to charge the flywheel. (4) Short term decreases in power

is covered by stored energy in the flywheel. The flywheel is

constantly being charged and discharged, which explains the

Fig. 3 – Operational data (10-min averages) from Utsira, 1–30 March 2007. (Top: Power produced by wind generator versus

power supplied to users (mini-grid). Bottom: Power from hydrogen engine generator versus power required by water

electrolyzer).


sometimes highly cyclic behavior of the power (e.g., Fig. 4, hour

101–107). (5) In more prolonged periods of excess power the

battery gets fully charged. (6) If this period of excess power

continues and the battery is fully charged, the H2-engine genset

is switched off, and the electrolyzer is switched on again.

The stand-alone power system at Utsira has been working

properly for the past 3 years. The exception was the fuel cell

system, which right from the start gave technical difficulties.

First, there was an incident where glycol-cooling liquid

leaked into and damaged several cells. These cells were

replaced and the glycol was exchanged with water. Next, the

fuel cell voltage monitoring system had to be replaced

because its gold coating had been damaged during the

assembly procedure. Finally, when the fuel cell was con-

nected to the stand-alone grid, the power electronics and

control system frequently detected a ‘‘grid failure’’ alarm

that caused the system to shut down. However, the main

problem with the fuel cell was that the stack itself degraded

rapidly, independent of being in operation or not. Hence, very

few real operation hours on the fuel cell (<100 h) was

achieved. Lately, after more than 3 years of steady and reli-

able operation, the hydrogen engine has also run into some

technical problems, and is no longer operational. The pistons

on the engine have been damaged beyond repair, and a new

engine will be put in place. The damage on the pistons in the

engine may have been caused by hydrogen leakage and

consequent reactions with the lubricating oil and/or by water

resulting from the combustion emulsifying the oil. Both

effects would result in increased friction and wear in the

engine. Hence, the operation of the Utsira plant over the past

few years show that both the fuel cell and hydrogen engine

have their technical challenges that must be overcome. A

new hydrogen engine will be ordered and will be installed as

soon as possible. A new fuel cell is also being considered.

4. Modeling tools

Over the past decade a number of different wind/hydrogen

concepts have been investigated using various modeling and

Fig. 4 – Operational data (10-min averages) measured at Utsira on 5 March 2007.


system simulation tools [2,8,16,24–26]. Below is a brief descrip-

tion of the simulation tools used in this particular study.

4.1. Simulation and modeling tools

The system simulations in this study were performed with

TRNSYS [27] and a set of hydrogen energy models (HYDRO-

GEMS) specifically developed for TRNSYS16 [26]. The

HYDROGEMS-library consists of the following component

models: Wind energy conversion systems (WECS), photovol-

taic systems (PV), water electrolysis (advanced alkaline,

adaptable to PEM), fuel cells (PEM and alkaline), hydrogen gas

storage, metal hydride hydrogen storage, hydrogen

compressor, secondary batteries (lead–acid), power condi-

tioning equipment, and diesel engine generator (multi-fuels,

including hydrogen). The source code for all of these models,

except the metalhydride model and alkaline fuelcellmodel, are

available for registered TRNSYS16 users. IFE has also

developed a duplicate HYDROGEMS library for EES (www.

fchart.com) based on the exact same source code as that

developed for TRNSYS, but these are only made available

through collaborative R&D projects. The main objective with

the EES-models is to perform more detailed component and

sub-system modeling, and to verify and calibrate more

advanced models based on experimental data. Both simu-

lations platforms, TRNSYS and EES, where used in this

study.

The HYDROGEMS component models are based on ther-

modynamics, electrochemistry, and applied physics (e.g.,

electrical, mechanical and heat and mass transfer engi-

neering). As a result the models are fairly physical and

generic, but in order to make the models suitable for energy

system simulations some simplifications have been made.

This is done by using empirical relationships for current–

voltage characteristics (IU-curves) for solar cells or electro-

chemical cells (e.g., fuel cells, batteries, electrolyzers), power

curves for wind turbines, efficiency curves for electronic

equipment (e.g., AC/DC-inverters), etc. Typical system model

parameters are: (i) Component sizes (e.g., cell areas), (ii)

System sizes (e.g., number of cells in series or parallel per

module), and (iii) Limits (e.g., min/max of voltages, currents,

temperatures etc.). Typical forcing functions required for

renewable energy system simulations are wind speed (m/s),

solar radiation (W/m2), current (A) or power (W) demand,

while typical variables calculated by the models are temper-

ature (�C), pressure (bar), current (A), voltage (V), and flow rate

(Nm3/h). The empirical parts of the models are designed so

that it is possible to find default parameters and/or calibrate

coefficients based on data found in literature (e.g., product

data sheets, papers and articles). It is a challenge to calibrate

model parameters and coefficients. Hence, access to actual

data from the hydrogen demonstration systems being

modeled is essential to ensure the validity of the models [28].

HYDROGEMS have been used to evaluate several other

hydrogen installations around the world [29].

4.2. Model updates and modifications

Modeling and characterization of the electrolyzer system,

including its complete balance of plant (BOP), and hydrogen

engine genset was performed in this study (Fig. 5). The

energy efficiency calculations for the electrolyzer and

hydrogen genset were both based on the higher heating value

(HHV) for hydrogen. The models were calibrated based on

actual system performance data collected from the operation

of the plant in March 2007 and additional component

performance data (e.g., energy efficiency and fuel consump-

tion curves) and general technical specifications from the

suppliers of the equipment.

The overall hydrogen production system efficiency at

maximum production was around 53%, while the electrolyzer

stack efficiency at these conditions was around 73%. At

maximum hydrogen production the total power consumption

of the hydrogen production system was about 65 kW. The

electrolyzer (stack) consumed about 73% of this power

consumption, while the auxiliaries (feed water pump, drier,

deoxidizer, cooling water pump etc.), rectifier/transformer,

and hydrogen compressor each contributed to about 9% of the

total power consumption (at maximum capacity). It should be

noted here that the electrolyzer installed at Utsira can operate

http://www.fchart.com

http://www.fchart.com

Fig. 5 – Characterization of electrolyzer and hydrogen engine generator systems. (Plots are generated using via calibrated

simulation models).


at 50–100% of its rated capacity, but in practice it seldom

operated down to 50% of its capacity (Fig. 3).

The maximum hydrogen engine genset efficiency was

around 20% (Fig. 5), but on average this efficiency was around

17–18%, due to frequent part-load operation (Fig. 3).

5. System simulations

A set of wind/hydrogen system simulations (in TRNSYS) were

performed using calibrated models (HYDROGEMS) for the

wind turbine, electrolyzer system, and hydrogen engine gen-

set. Annual simulations of the Utsira plant (reference system)

and three alternative wind/hydrogen system designs were

performed, based on hourly wind speed (m/s) and power

demand (kW) measured at the site in 2005 (Fig. 6).

A monthly simulation (March 2007) of the reference

system confirms that it is not possible to achieve 100% stand-

alone operation with the existing system design at Utsira

(Fig. 6, plot 1). In the first alternative design a hydrogen

storage with a much larger capacity (7200 Nm3 instead of

2400 Nm3) was investigated. The results show a negative

total hydrogen mass balance (i.e., pressure decrease) for the

given time period (Fig. 6, plot 2). A similar result is found if

the H2-engine genset is replaced by a fuel cell with twice the

efficiency (Fig. 6, plot 3). However, if these two alternative

system design changes are made simultaneously (i.e., fuel

cell system with a larger H2-storage), the hydrogen mass

balance is positive for the given time period (Fig. 6, plot 4).

These results indicate that full autonomy can only be ach-

ieved by improving the overall efficiency of the hydrogen

production system (electrolyzer), by increasing the hydrogen

storage size, and/or by increasing power generating effi-

ciency of the unit (e.g., by switching from hydrogen engines

to fuel cells). This was then investigated in some more detail

for a complete year (see simulations based on 2005 input data

described below).

The hourly annual performance of alternative system

configurations that have the potential to reach full (100%)

energy autonomy was also simulated, and cost-estimated

based on the general cost functions listed in Table 3. A

comparison of the performance of two alternative system

configurations, one based on a fuel cell (with an efficiency of

50%) and the other on a hydrogen engine (with an efficiency of

25%), is shown in Fig. 7. It is important to point out here that

Fig. 7 does not show the techno-economic optimal system

configuration, but merely shows two alternative designs at

approximately the same overall system cost. The results show

that at system with a fuel cell requires a relatively small

electrolyzer (10 Nm3/h) and hydrogen storage (4800 Nm3)

compared to a system with a hydrogen engine, which requires

a relatively large electrolyzer (20 Nm3/h) and hydrogen storage

(11000 Nm3). Further economic analysis (not shown here)

show that the total investment costs for the main components

in these two alternative systems are both in the same order,

ca. 1 million V if based on the specifications in Table 3. In

summary, for the case of Utsira, a wind/hydrogen system

based on a fuel cell will be more compact and efficient (but

also slightly more costly) than a system based on a hydrogen

engine, which will be more bulky and less efficient (but

slightly less costly).

6. Discussion

6.1. Specific design issues and operational experience

Several important experiences have been gained from the

planning, building, and operation of the Utsira plant. First of

all, it is important to have a well-defined design basis and

operational philosophy, taking into consideration climate,

signal quality, communication (control and regulation), and

all system interfaces. Since wind/hydrogen energy systems

are not plug-and-play, but still fall into the category of R&D, it

Fig. 6 – Wind/hydrogen system simulation results using input data from Utsira (March 2007).


is particularly important to select the right project partners,

preferably a few with technical expertise. In order for a project

of this nature ever to be realized, it is also important to have

a strong focus on safety, health, and the environment. Finally,

great care should be made in the selection of the site for the

installation. An appropriate location has the following

features: Good to excellent wind conditions, small but repre-

sentative load, back-up systems in place, not too remote,

a supporting community, and access to service personnel.

Some of these points are discussed in more detail below.

Utsira has a typical North Sea offshore climate with high

winds, strong waves, salty air, and temperatures down to

below zero (freezing). High waves, particularly during winter,

made it difficult to transport the largest components out to the

site. Since some of the components had long delivery times, it

became extremely important to plan with these weather

conditions in mind. The electronic equipment and housings

were all prepared for a saline environment, and the fuel cell

and electrolyzer containers were heated to avoid freezing.

The Utsira mini-grid (i.e., the stand-alone power system

created for 10 households on the island) mainly includes

inverters based on IGBT-technology, but some conventional

inverters were also installed. The influence these inverters

have on the (weak) grid needed to be carefully considered.

Variations in signal quality (voltage and frequency) in the mini-

grid at Utsira were observed in the beginning of the project.

Table 3 – General cost functions for wind and hydrogentechnologies [4].

Component Lifetime(years)

Capital Costs(V/kW or V/m3)

O&M Costs(% of Captial Costs)

Wind turbine 20 800 V/kW 1.5

Electrolyzer 20 2000 V/kW 2.0

H2-compressor 12 5000 V/kW 1.5

H2-engine 10 1000 V/kW 2.0

Fuel cell 10 2500 V/kW 2.0

H2-storage 20 4500 V/m3 2.5

Reactive power, resonance, over-harmonics can occur, and

must be dealt with. The power supply for the electrolyzer

installed at Utsira was one source for such problems. In

general, all equipment should be kept as simple and robust as

possible, and redundancy should be considered. Due to the

uncertainty in future wind power production and customer

power demand, a slightly oversized installation should also be

considered. However, as always, there is a trade off to be made

between plant availability and overall system cost.

The Utsira plant was designed to be remotely operated, and

a system for self-testing and automatic remote resetting of

individual components after shutdowns was therefore put in

place. Remote resetting for all of the components was not

possible to achieve from the start of the project, due to limi-

tations in the available technology (equipment was not

designed for remote operation). The key lesson learned here is

that it is important to specify or choose equipment with a high

degree of fail-safe and remote operation. Remote operation is

a necessity for safe operation. Suppliers of equipment often

have proprietary control systems. Nevertheless, communica-

tion and interfaces between components should be described

and standardized. This applies especially for the hydrogen

part of the plant.

Safety was prioritized in the design of the Utsira plant. It

would be detrimental to the development of hydrogen as an

energy carrier if a serious accident would occur in such

a highly profiled demonstration project. The Utsira plant,

which is compact and complex and contains explosive zones

and advanced equipment, regularly receives many unskilled

visitors. Safety has therefore been given the highest priority,

and there have been no accidents. The key for achieving this is

proper training of operators and other personnel, good

working instructions for all parts of the system, and clear

distribution of responsibility at the site.

Finally, it is important that these first wind/hydrogen

demonstrations systems that serve as the main power system

for real users is working without problems; to keep the

customers satisfied and positive is important for the public

acceptance of hydrogen. At Utsira the customers can, in case

Fig. 7 – Wind/hydrogen system simulation based on input data from Utsira for a complete year (2005). Comparison of two

alternative system configurations: fuel cell versus hydrogen engine.


of failures or testing in the stand-alone mini-grid, be con-

nected back to the regular local grid (1 MW subsea cable to the

main land). This electrical configuration is working well. The

installation of a ‘‘dummy load’’ has also proven to be useful, as

this can be used to emulate the power demand during testing

of the system before it is connected back onto the mini-grid.

6.2. General system design issues

Hybrid hydrogen energy systems and refueling stations have

the potential to be both energy and cost efficient. The

discharge of hydrogen through a dispenser for a hydrogen

vehicle (preferably a fuel cell vehicle) increases the run-time

of the wind/electrolyzer system without having to invest in an

excessively large hydrogen storage system [30].

Techno-economic studies of wind/hydrogen mini-grid

systems based on today’s compressed hydrogen gas storages

(200 bar) and hydrogen internal combustion engines or fuel

cells (ICE and FC conversion efficiencies of 20 % or 40 %,

respectively), show that it is realistic to have a hydrogen

storage system capable of storing wind energy for about 1–2

weeks [6]. The exact size of the hydrogen storage will depend

on the wind energy profile and the load profile, as well as the

choice of hydrogen to power conversion device.

In general, renewable energy hydrogen systems should be

designed with input from several energy sources in order to

reduce the need for large and costly hydrogen energy storage.

However, if a hydrogen system is based only on wind energy,

it is particularly important to pick a location with steady and

good wind energy resources. Previous studies have demon-

strated that the need for seasonal storage of energy (weeks to

months) is more dependent on the annual wind speed profile,

than the average wind speeds [21].

For systems located in cold climates and high latitudes,

such as remote islands in the Nordic region, there is often

a positive correlation between the wind energy input and the

user load (thermal loads usually increase with increasing

wind speeds). Unfortunately, this correlation is normally not

so perfect that it can completely offset the need for some

kind of seasonal storage of electricity in the form of

hydrogen. Furthermore, if part of the electric load is used for

heating purposes the mismatch between the load and wind

energy input is even higher. (This is often the case for

communities in remote areas with limited access to other

heat sources, e.g., locally produced biomass). Hence,

designing fully autonomous RE/H2-power systems (e.g.,

systems with FC-power> 100 kWel) will always be

demanding, even though it is theoretically possible. In prac-

tice, full autonomy is most realistically achieved in smaller

RE/H2-power systems (e.g., systems with FC-power< 2 kWel)

that have insignificant heating/cooling demands and are

based on renewable energy sources with relatively small

seasonal variations, as such systems do normally do not

require large hydrogen storages [2].

6.3. Electrolyzer operation

Previous RE/H2-studies and experiments performed at IFE

[16,28,31] show how important it is to install a water electro-

lyzer with a quick response to varying power input (load-

following), particularly in wind energy based systems. The

more wind energy that can be captured, the more hydrogen

can be produced. Ideally, the electrolyzer should operate

continuously, and be able to operate down to only a few

percent of its rated capacity. Most existing commercial alka-

line electrolyzers today only allow for operation down to

25–50% of their rated capacity. In practice, this means that the

electrolyzer must be shut down when the input power is lower

than this limit (normally 25–50% of its rated capacity). This is

a severe disadvantage because once the electrolyzer has been

switched off, it takes a while (30–60 min) before it can be

switched on again (due to purging with nitrogen). Hence, the


possibility to produce hydrogen during this period of time (if

the wind picks up again) is lost.

Results from the testing of a new and advanced large-scale

(MW) high pressure (30 bar) alkaline electrolyzers from Sta-

toilHydro Hydrogen Technologies [23] show that this tech-

nology should be able to handle down to 5–10% of it’s rated

capacity with a regulation speed of 200 ms [32]. This new

alkaline electrolyzer has a cell electrode area of 0.8 m2,

making it particularly suitable for larger plants with high

hydrogen production needs and large amounts of excess wind

energy available (MW range). If such an electrolyzer was to be

used in a smaller system, such as Utsira, only a few cells

would be required to produce the required amount of

hydrogen (10 Nm3/h). However, the voltage across a stack

based on only a few cells in series would be very low, and not

suitable for dynamic power conversion between the wind

energy generator and electrolyzer, as an AC/DC-inverter with

a very large voltage difference would be required. Hence, an

electrolyzer with smaller cells (allowing for more cells to be

placed in series) should be used instead. The high-pressure

alkaline electrolyzer described above can be down-scaled, but

there are also other options.

PEM-electrolyzers, with their inherent simplicity and

potential for quick startup and wide operational window

(5–100% of rated capacity), may prove to be more suitable for

wind/hydrogen-systems than advanced alkaline electrolyzers

[31]. This would particularly be true for systems based on

small to medium range (50–250 kW) electrolyzers. The most

suitable approach here would be to connect PEM-electrolyzer

stacks with electrode areas from 0.25 to 0.50 m2 in series and/

or parallel in order to achieve the proper voltage level (as close

as possible to the nominal operating voltage on the AC-bus-

bar, e.g., 400 V). The main disadvantage with the PEM-elec-

trolyzers is the uncertainty about the lifetime of the PEM.

6.4. Hydrogen pressure

Increasing the pressure in the hydrogen storage from 200 bar

to 450 bar, or even 700 bar, would increase the overall energy

density of the hydrogen storage, thus making it possible to

store more wind energy on the same footprint. However, high-

pressure hydrogen storage systems are likely to be more costly

than low-pressure systems, both from an investment and an

operational point of view. The increase in investment costs

will mainly be associated with the need for stronger storage

tanks (thicker steel walls and/or use of composite materials),

while the increase in operational cost will mainly be associ-

ated with the increased energy consumption for compression

of hydrogen. There are also likely to be extra costs associated

with the required safety system, although this depends on the

type of system installed.

There are two basic techniques to produce high-pressure

hydrogen gas: (1) Low-pressure electrolysis with a large

compression stage or (2) High-pressure electrolysis without

compression. Theoretical comparisons of these two hydrogen

production techniques have shown that high-pressure elec-

trolysis without a compressor is slightly more efficient (ca. 5%

more efficient) than low-pressure electrolysis followed by

a compressor [33].

Practical trade-off studies on high-pressure alkaline elec-

trolyzer systems show that high-pressure electrolysis might

be difficult to achieve, as increasing the operating pressure of

the electrolyzer is likely to come at a relatively high cost [34].

These increased costs are related to increased material and

system engineering cost associated with pressurizing the

electrolyzer stack itself, as well as extra costs associated with

the required safety and control systems.

In conclusion, it seems to make most sense to operate the

electrolyzers at lower pressures (ca. 10–40 bar), followed by

a hydrogen compressor. The optimal hydrogen storage pres-

sure will be a trade-off between extra investment cost of the

hydrogen storage on one side, and the increased investment

and operating cost associated with the hydrogen compressor

on the other side.

7. Conclusions and recommendations

The autonomous wind/hydrogen system at Utsira has

demonstrated that it is possible to supply remote area

communities with wind power using hydrogen as the energy

storage medium. However, further technical improvements

and cost reductions need to be made before wind/hydrogen-

systems can compete with existing commercial solutions, for

example wind/diesel hybrid power systems. Several areas of

improvements have been identified.

In general, the overall wind energy utilization must be

increased (at Utsira only 20% of the wind energy is utilized).

This can best be achieved by installing more suitable and

efficient electrolyzers that allow for continuous and dynamic

operation (load-following electrolyzers). Analysis of the

operational data from Utsira confirms that more efficient

electrolyzer systems (entire balance of plant, including

hydrogen compression) need to be developed. New and more

efficient and dynamic electrolyzers are currently being

developed in Norway.

The surplus wind energy should also be used to meet local

heating demands, both at the plant and in the households. In

addition, one could consider to utilize the hydrogen produced

(and possibly the oxygen) for other purposes (e.g., for trans-

portation). A more sophisticated wind energy forecasting and

load management system could also be implemented,

particularly if more complexity is added to the system (e.g.,

multi-purpose hybrid systems).

The hydrogen storage installed at the Utsira plant allows

for only 2–3 days of autonomous operation. A larger electro-

lyzer and hydrogen storage could improve this, but would lead

to a larger and more costly system. Another alternative would

be to improve the hydrogen–electricity conversion efficiency

by replacing the internal combustion engine with a fuel cell,

but this would lead to an equally costly system, as demon-

strated in this study. Hence, more compact hydrogen storage

systems and/or more robust and less costly fuel cell systems

need to be developed before wind/hydrogen-systems can be

technically and economically viable. Meanwhile, hybrid

system solutions based on more than one energy source (e.g.,

wind, solar, and bioenergy, depending on the location) should

be developed. The use of a diesel engine genset for back-up

power (redundancy) and/or to cover a small part (e.g., 5–10%)


of the annual load could also be an option, particularly for

systems located in areas where natural resources are scarce.

Fortunately, there are a few general trends that could make

renewable energy hydrogen system concepts similar to the

one at Utsira cost-effective in the not too distant future: The

introduction of ‘‘green’’ incentives (e.g., green certificates and

CO2-tax), increases in oil and gas prices, and capital appreci-

ation of green image and security of supply. All of these

factors justify and motivate the construction and further

development of full-scale demonstration system such as the

wind/hydrogen plant at Utsira. It is believed that such

demonstrations will help improve public awareness and

acceptance, improve cost competitiveness of renewable

energy, and reduce market barriers for new energy and tech-

nology solutions in general, and hydrogen technology in

particular.

Acknowledgments

The authors would like to acknowledge the long-term efforts

made by former colleagues in Norsk Hydro, and new

colleagues in StatoilHydro, in the design, building, and oper-

ation of the Utsira plant. The technical collaboration with

Enercon, the financial contributions made by the Research

Council of Norway and Enova, the public acceptance at Utsira,

and the scientific input from IFE and other research institu-

tions in Norway is also greatly appreciated.

r e f e r e n c e s

[1] Ghosh PC, Emonts B, Janßen H, Mergel J, Stolten D. Ten yearsof operational experience with a hydrogen-based renewableenergy supply system. Solar Energy 2003;75:469–78.

[2] Ulleberg Ø. Stand-Alone Power Systems for the Future:optimal Design, Operation & Control of Solar-HydrogenEnergy Systems, PhD thesis. Trondheim, NorwegianUniversity of Science and Technology; 1998.

[3] Gazey R, Salman SK, Aklil-D’Halluin DD. A field applicationexperience of integrating hydrogen technology with windpower in a remote island location. Journal of Power Sources2006;157:841–7.

[4] Glockner R. HSAPS Market Analysis Project. Final report forEU/ALTENER-project (Contract No. 4.1030/Z/01-101/200), IFE/KR/E-2005/003, Kjeller, Institute for Energy Technology; 2005.NO-2027.

[5] Korpas M, Greiner CJ. Opportunities for hydrogen productionin connection with wind power in weak grids. RenewableEnergy 2007;33:1199–208.

[6] Ulleberg Ø, Mørkved A. Renewable energy and hydrogensystem concepts for remote communities in the west nordicregion–the nolsoy case study. IFE/KR/E-2008/002 (electronic).Norway: Institute for Energy Technology; Kjeller, ISBN 978-82-7017-697-7; 2008.

[7] Dutton AG, Bleijs JAM, Dienhart H, Falchetta M, Hug W,Prischichd D, et al. Experience in the design, sizing,economics, and implementation of autonomous wind-powered hydrogen production systems. InternationalJournal of Hydrogen Energy 2000;25:705–22.

[8] Levene J, Kroposki B, Sverdrup G. Wind energy andproduction of hydrogen and electricity–opportunities for

renewable hydrogen. In: 2006 POWER-GEN Renewable Energyand Fuels Technical Conference, Las Vegas, Nevada; 10–12April 2006.

[9] Ulleberg Ø, Miland H. HYTREC wind/electrolysis system. IFE/KR/F-2006/125. Kjeller: IFE; 2006.

[10] Agbossou K, Chahine R, Hamelin J, Laurencelle F, Anouar A,St-Arnaud J-M, et al. Renewable energy systems based onhydrogen for remote applications. Journal of Power Sources2001;96(1):168–72.

[11] Aprea JL. Hydrogen energy demonstration plant in Patagonia:description and safety issues. International Journal ofHydrogen Energy 2009;34(10):4684–91.

[12] Eide PO, Hagen EF, Kuhlmann M, Rohden R. Construction andcommissioning of the Utsira wind/hydrogen stand-alonepower system. In: EWEC 2004-European Wind EnergyConference, London; 20–22 November 2004.

[13] Gammon R, Roy A, Barton J, Little M. HARI–hydrogen andrenewables integration. Case study report. IEA HydrogenImplementing Agreement, http://www.ieahia.org/pdfs/HARI.pdf; 2006. Annex 18.

[14] Harrison KW, Kroposki B, Pink C. Characterizing electrolyzerperformance for use in wind energy applications. In:American Wind Energy Association’s (AWEA) WindpowerAnnual Conference, Pittsburgh, Pennsylvania; 4–7 June 2006.

[15] Harrison KW, Martin GD. Renewable Hydrogen: integration,Validation, and Demonstration. In: 2008 NHA AnnualHydrogen Conference, Sacramento, California; 30 March–3April 2008.

[16] Miland H. Operational Experience and Control Strategies fora Stand-Alone Power System based on Renewable Energyand Hydrogen. PhD thesis, Norwegian University of Scienceand Technology, Trondheim; 2005.

[17] Miland H, Glockner R, Taylor P, Jarle Aaberg R, Hagen G. Loadcontrol of a wind-hydrogen stand-alone power system.International Journal of Hydrogen Energy 2006;31(9):1215–35.

[18] Nakken T, Frantzen E, Hagen EF, Strøm H. Utsira-demonstrating the renewable hydrogen society. In: WHEC16-World Hydrogen Energy Conference, Lyon; 2006.

[19] Baker JN, Carter S. Evaluation of hydrogen demonstrationsystems. Report No. F/04/00279/00/REP (DTI Contract No. F/04/00279/00/00). UK: EA Technology Limited; 2005.

[20] Varkaraki E, Tzamalis G, Zoulias E. Operational experiencefrom the RES2H2 wind-hydrogen plant in Greece. In:WHEC17-World Hydrogen Energy Conference, Brisbane;2008.

[21] Glockner R, Kloed C, Nyhammer F, Ulleberg Ø. Wind/hydrogen systems for remote areas–a Norwegian case study.In: WHEC 2002–14th World Hydrogen Energy Conference,Montreal; 9–14 June 2002.

[22] Ulleberg Ø, Glockner R. Development of renewable energy/hydrogen systems: from concepts to actual demonstrations.In: Hydrogen and fuel cells futures conference, Perth; 12–15September 2004.

[23] StatoilHydro Hydrogen Technologies, http://www.electrolysers.com/; 2009.

[24] Kelouwani S, Agbossou K, Chahine R. Model for energyconversion in renewable energy system with hydrogenstorage. Journal of Power Sources 2005;140(1):392–9.

[25] Korpas M. Distributed Energy Systems with Wind Power andEnergy Storage, PhD thesis. Norwegian University of Scienceand Technology, Trondheim; 2004.

[26] Ulleberg Ø, Glockner R. HYDROGEMS–hydrogen EnergyModels. In: WHEC 2002–14th world hydrogen energyconference, Montreal; 9–14 June 2002.

[27] Klein SA, Beckman WA, Mitchell JW, Duffie JA, Duffie NA,Freeman TL, et al. TRNSYS–a transient system simulationprogram. v16.01.0003. Madison: Solar Energy Laboratory,University of Wisconsin; 2007.

http://www.ieahia.org/pdfs/HARI.pdf

http://www.ieahia.org/pdfs/HARI.pdf

http://www.electrolysers.com/

http://www.electrolysers.com/


[28] Ulleberg Ø. Modeling of advanced alkaline electrolyzers:a system simulation approach. International Journal ofHydrogen Energy 2002;28(1):7–19.

[29] Ulleberg Ø, Ito H, Maack MH, Ridell B, Miles S, Kelly N, et al.Hydrogen implementing agreement (HIA), task 18-integratedsystems evaluation, subtask B–demonstration projectevaluations. Final report ISBN 978-0-9815041-0-0. Paris:International Energy Agency; 2007.

[30] Padro CG. IEA Hydrogen implementing agreement Annex 13-design and optimization of integrated systems. Final report.Golden, Colorado: NationalRenewableEnergyLaboratory;2004.

[31] Vie PJS, Ulleberg Ø, Fell HJ, Rasten E. Development ofa Prototype PEM-electrolyser for distributed renewable

energy systems: testing under Emulated wind powerconditions. In: WHEC2008-World Hydrogen EnergyConference, Brisbane; 15–19 June 2008.

[32] D’Erasmo P. HYSTRUC–demonstration and testing of aninnovative 30 bar, low cost, small size pressure electrolyser,in the MW power range, for cost efficient production ofelectrolytic hydrogen. Personal Communication; 22November 2005.

[33] Onda K, Kyakuno T, Hattori K, Ito K. Prediction of productionpower for high-pressure hydrogen by high-pressure waterelectrolysis. Journal of Power Sources 2004;132:64–70.

[34] Ibrahim S, Cohen S. Alkaline high pressure electrolysis. 2005DOE Hydrogen Program Review, 2005.

the wind/hydrogen demonstration system at utsira in norway ......the wind/hydrogen demonstration...

Documents