Special Technology Supplement
Industrial gas turbines:
the perfect complement for
With the growth in
variable renewables,
energy storage is
expected to be the
key technology
for providing grid
support and shifting
renewable power to
when it’s needed.
Siemens Energy’s
Anders Stuxberg
explains to
TEI Times why
industrial gas
turbines will
be crucial in
storage in an
optimised system.
Junior Isles
prices for nal consumers, and the use
of smart meters.
“When you look at the demand for
balancing power, storage solutions
are efcient systems, with up to 80
per cent of the energy coming back
[from the storage]. But it is not eco-
nomical to design an energy storage
system for all possible situations. And
when you empty the storage, you
have to ll-in with something else,”
said Stuxberg.
That “something else”, he says, will
typically be (fuel red) thermal
plants, i.e. the backup power capacity
that must exist in the grid anyway to
ensure reliable supply when there is
no wind or solar for a long period.
There are several options as to
which technology, or group of tech-
nologies, can support renewables-
plus-storage depending on the sce-
nario. For example, arguments are
sometimes made for fuel cells while
other experts present compelling
cases for fast-start generating assets
such as gas turbines and reciprocating
Stuxberg believes industrial gas
turbines are currently the best all-
round option. He commented: “In a
deeply decarbonised energy system,
gas turbines will play a key role both
for mid-merit power supply and as
backup power. Although some argue
that fuel cells will take that role, that
can only happen if fuel cells for a
fully functional and installed power
generation plant become cheaper than
gas turbines. We are not there today
and I believe that if it happens, it will
take many decades. Fuel cells,
though, are already a good option for
microgrids and mobility applications.
The requirement that backup power
also should be fuel exible, e.g. use
both hydrogen and liquid renewable
ith the urgent need to combat
climate change, wind and
solar power are growing at a
phenomenal rate. According to the
International Energy Agency, renew-
ables will meet 80 per cent of global
electricity demand growth during the
next decade. Solar PV, for example,
dubbed by the IEA as “the new king”
of electricity supply, grows by an
average of 13 per cent per year between
2020 and 2030, meeting almost one-
third of electricity demand growth
over the period.
The variable nature of wind and
solar, however, presents challenges in
terms of grid stability and how best to
provide backup power for when the
wind is not blowing or the sun is not
With targets set for reaching zero
carbon emissions in the electricity
sector, clearly the goal must be to
support renewables as far as possible
with energy storage a zero carbon
source of grid exibility. The ques-
tion, however, is what generating as-
sets to deploy alongside storage, and
how to achieve the best mix of storage
and those assets in terms of cost and
Anders Stuxberg, Specialist in
Power Plant Process Integration at
Siemens Energy AB said: “Gas tur-
bines (GTs) will be the technology of
choice to be dispatched when storage
power capacities are insufcient for
the demand and also when the storage
becomes emptied. If you look at bal-
ancing supply and demand through
the grid in general, you have to look
at it over a number of different time-
frames. The system has to be man-
aged, second-by-second, minute-by-
minute, hour-by-hour, using different
technologies. You also have to look at
balancing over longer timeframes…
The question is how to optimise these
storage and generating resources.
Storage will handle the bulk of energy
for balancing, but there will not be a
business case to try to cover every-
thing with storage alone, you will
need to complement it with GTs.
“By implementing storage, the op-
erating prole for GT-based plants
will be signicantly changed. GTs
will be a cornerstone of the grid infra-
structure but with a new role in future
compared to what we have been used
to seeing. You will see a shift to
backup power instead of peaking
units and exible mid-merit com-
bined cycle plants instead of baseload
plants; this will favour industrial GTs
for new installations. Industrial gas
turbines are also suited to use hydro-
gen as fuel and fuels produced
through power-to-X schemes,” said
With storage expected to take cen-
tre-stage in maximising the integra-
tion of renewables and distributed
generating sources, the market for
the technology is forecasted to grow
exponentially over the next decade
(see box).
Regardless of which of the various
storage solutions is selected, however,
they are all generally limited by two
parameters: power capacity and en-
ergy capacity, i.e. duration of storage
at full power. Stuxberg noted that
when optimising storage solutions,
power plant owners will size for the
most frequent instances that give the
most energy trade volume and then
leave the residual load to some other
He said: “There will be many days
the energy in the storage is insufcient
for the demand and many days when
storage systems have less power ca-
pacity than needed, at least during
part of the dispatch duration. So other
technologies will be called for both at
surge of power and of energy, there
will be a play between different types
of storage solutions and capacity
He also noted: “Storage technolo-
gies that can shift operating mode af-
ter the storage is emptied continuing
power production by ring a supple-
mentary fuel – will also play a role in
backup supply, i.e. double benets to
the system. Examples are: power-to-
hydrogen-to-power where the hydro-
gen-to power unit (gas turbine) oper-
ates on e-methanol when the gas
storage is emptied, or a thermal stor-
age plant that also can run by ring of
e-ammonia when the thermal storage
is emptied.”
Stuxberg says there will also be
competition between storage and
demand response (DR). If altering
the time of energy use (e.g. smart
charging electric cars) does not dam-
age business, then DR will be more
efcient and cost competitive than
Many types of DR will, however, be
limited in much the same way as stor-
age. For example, duration mainly
limited by the nature of the demand
that has been put on hold will nor-
mally be limited to a number of hours.
The amount of DR that will be avail-
able naturally depends on the price
incentive, the volatility of energy
Power plant owners will optimise storage solutions size for the most frequent instances that give
the most energy trade volume and leave the residual load to another technology
Don’t let balancing power capacity get out of balance
Storage will handle the bulk of energy for balancing but it will
need to be complemented with GTs
slightly more expensive power than
the storage system. If the dispatch is
just based on a commercial energy
trade, then hybrid plants comprising a
combination of e.g. renewable power,
storage and GT may be a good busi-
ness as smarter dispatch can be
Typically, many gas turbines will be
installed in an electric grid to provide
the necessary backup power. The
dispatch order for these will be based
on cost or environmental footprint.
Since the requirement will be for a
fairly low dispatch rate, Stuxberg
says a large portion of dispatch may
be based on capacity auctions where
a xed compensation for just existing
as available backup is paid out.
If efciency is also credited, e.g. by
dispatch order, then a fair portion of
these cycling GTs will be congured
as combined cycle. However, the
bottoming steam cycle must then be
suited to frequent starts, i.e. fast and
with low start-up cost. Stuxberg notes
that in a future where these mid-merit
plants need to operate on renewable
fuel, which will be expensive, a bot-
toming cycle will be required for
many of these plants for the sake of
opex. The remaining plants, which
will have a low dispatch rate of, say,
less than 500 hours per year, will not
be so sensitive to efciency but will
need to have low capex and xed
standstill cost.
“So, for the power generation busi-
ness, we will see two typical types of
GT plants for the future: combined
cycle plants for cycling operation,
dispatching in a mid-merit pattern of
somewhere between 1000 and 3000
hours per year; and simple cycle
plants, with dispatch often less than
500 hours per year. The traditional
base load plant is thus replaced by a
very exible mid-merit plant, while
the traditional peaking plant is re-
placed by demand response and stor-
age solutions plus a large quantity of
backup power.”
His absolute conviction is that in-
dustrial gas turbines present the best
suitability to this type of future duty
for both these plant types. “They have
very high reliability due to simplicity
in design concept, high combined
cycle efciency, low price, low main-
tenance cost, good fuel exibility and
much better grid stabilisation charac-
teristics (by high inertia and strong
control response) than aeroderivative
GTs or recip engines,” he said.
For both these plant types, his ex-
pectation is that there will be an aver-
age of one start every one to four
days, most frequent for the mid-merit
type. Stuxberg predicts a wide operat-
ing regime for such gas turbine plants.
For demand response (DR) and for
energy storage systems, he noted that
they will dominate dispatch of bal-
ancing power for short duration and
during periods of low demand for re-
sidual power.
He noted, however: “When looking
at capacity it is hard to rule out rare
events with low probability, thus in-
stalled GT power capacity in the grid
will need to be large. The scale of
backup capacity needed depends
predominantly on the capacity factor
fuel is also a cost issue, if not a prob-
lem, for fuel cell plants.
“Reciprocating engines compared
with gas turbines have pros and cons.
In short, they are less efcient than
combined cycles and are more expen-
sive per capacity than simple cycle
GTs, with the exception of emergency
diesels gensets, which have a shorter
lifespan. For mid-merit operation,
maintenance cost is an important
factor to consider industrial GTs
have lower maintenance cost than,
e.g. recip engines or fuel cells.”
He also notes that conventional
boilers with steam plants are too in-
exible to handle the frequent starts
and stops to balance residual power
demand. Further, their efciency is
low, especially if designed for renew-
able fuels such as biomass.
Based on the shortcomings of these
technologies, Stuxberg believes the
focus for grid balancing should there-
fore be on a blend of industrial gas
turbines (IGTs) and storage solutions
and a probable future dispatch prole
for those assets.
IGTs in the range up to 70 MW are
typically used in a number of applica-
tions. CHP applications are common
across the whole range due to their
ability to meet heat demand. The
smaller machines may be deployed in
settings like hospitals, universities,
small industries and O&G, to provide
power in areas where the grid is not
completely stable or onsite generation
is required. Medium-sized machines
in the upper range of 30-70 MW may
be used by, independent power pro-
ducers (IPPs), industrial CHP asset
owners, the O&G industry, munici-
palities producing electrical power
for the grid and heat for district heat-
ing networks, as well as utilities.
Stuxberg believes the operating
prole of IGTs in the future will not
be same as the peaking units of today.
Units in the future he says might start-
up and shut down once a day during
parts of the year, be in standby other
periods and also occasionally run for
a longer period, as opposed to cycling
several times per day.
With storage expected to be the rst
option for supplying multiple daily
power peaks, operators must then
decide how gas turbines will operate
to complement this storage.
Stuxberg foresees gas turbines be-
ing dispatched when the energy re-
quired exceeds what is available in
the storage. This will likely be after
the large afternoon/early night peak
or possibly in the morning. Gas tur-
bines will also be called for when all
storage solutions are already provid-
ing near full power capacity, i.e. typi-
cally during the evening peak.
He explained: “If GTs are being
called on every day for one of the two
reasons, power surge or energy surge,
then that’s a signal to storage inves-
tors that here you have an attractive
business opportunity – just buy some
more capacity. It’s low-hanging fruit.
So my conclusion is that GTs will
typically start once every 2-4 days on
average; some days they might be
called on twice and many other days
not at all.
“Traditional peaking plants and
base load plants will no longer be
suitable for this kind of market. So if
we have a GT on the system to ensure
backup anyway, the question is:
should you operate it for more hours,
which means more fuel consumption,
or should you make the storage
slightly bigger?”
According to Stuxberg, that optimi-
sation determines how the gas turbine
is operated, the type of turbine se-
lected and whether the plant should
be simple cycle or combined cycle.
He explained: “Generally, each ad-
dition of duration for a storage tech-
nology comes at an added investment,
which needs to be paid for by less and
less events since long duration events
are less frequent than shorter events.
The marginal cost of longer operation
for a GT plant ring renewable fuel
on the other hand is constant as it just
adds fuel consumption (fuel storage is
relatively cheap). The duration at the
cross-over point between technolo-
gies depends on event probability, a
number of economic factors and
choice of technology. The decreasing
probability of long events explains
why even pumped hydro plants, at
present, often are sized to t just one
day cycles.”
He added: “Grid balancing of up to
a couple hundred megawatts would
be fairly common. This could be di-
vided across a number of machines so
you can follow demand better without
running machines at part-load.”
Such an installation would have to
be capable of meeting several require-
ments. Firstly, it should be capable of
starting “reasonably” fast.
“If there is some kind of communi-
cation protocol (using new IT solu-
tions and advanced forecasting tools)
in the market telling GT operators to
start in fair time before stored energy
runs out, then very fast GT start is not
required, 20 minutes should sufce,”
said Stuxberg. “Also when power
capacity becomes the issue, it should
on most occasions be possible to
predict when to dispatch GTs. How-
ever, power peaks come faster than
drainage of energy, so here dispatch
centres can reserve some power in the
storage by starting the GTs a bit in
advance when a demand ramp-up is
expected. Here a fast GT start pays off
a little as there is less need to reserve
power from storage dispatch and thus
there is a bit less operation of the GTs,
which could be assumed to produce
Special Technology Supplement
IGTs such as the SGT-800 are
typically used in a number of
The dispatch order for
GTs in the grid for backup
will be based on cost or
environmental footprint
Stuxberg: in a deeply decar-
bonised energy system, gas
turbines will play a key role
both for mid-merit power sup-
ply and as backup power
through to morning and for the bal-
ancing duty that storage solutions
would otherwise provide, as there is
no surplus renewable power during
the day for charging the storage.
Here, high efciency storage is
charged from high efciency mid-
merit GT plants during the day, as
this limits the need of thermal plant
capacity during the peaks. The result
is that the required thermal plant ca-
pacity is about twice the capacity of
installed storage.
If DR is added, it would reduce the
required amount of storage as well as
the power capacity for storage charg-
ing/discharging during an average
wind day. In the low wind scenario, it
would also reduce the need for in-
stalled thermal capacity, as it attens
the thermal power supply.
“Naturally reality is more complex
than these simple scenarios, with sea-
sonal variations on both demand and
supply, effects of clouding, fast uc-
tuations, grid disturbances etc.,” noted
Fuel exibility also has to be a key
consideration. If a machine is oper-
ated for less than 1000 hours/year,
the impact of fuel consumption on
environment and economics is rela-
tively small. However, the goal is to
of wind and solar and level of long
distance power transmission. Up to
about 50 per cent of grid capacity may
be expected; in isolated grids or grids
with weak connection to other grids
one may even argue for 100 per cent.
When you also look at resilience and
tolerance for grid failures most of the
GT installations should be distributed
in the grid, this favours mid-sized gas
turbines as well as exible CHP. In
large, high capacity grids, large GTs
will also be attractive for backup
power capacity due to low specic
“When looking at energy supply
rather than the installed capacity, de-
mand response and storage will dis-
patch maybe 80 per cent of all energy
needed for grid balancing and GTs
only the remaining 20 per cent. Those
GTs should preferably operate on re-
newable fuel.” he added.
The gure below shows demand as
well as solar and wind supply in a
simplied ctitious medium size
grid. On the left, wind supply during
an average day, where energy fed into
storage covers about 85 per cent of
the balancing need. On the right,
where wind supply is low, thermal
power generation is needed to replace
lower wind supply during the evening
run turbines on renewable fuels, and
uncertain policy in the long-term out-
look in this area is a challenge.
Stuxberg said: “There are a number
of optional renewable fuels for use in
GTs, hydrogen being one of the top
candidates, but today we don’t know
which of these will be economical or
available in the future and obviously
it will always depend on the site loca-
tion and operating prole. But the
point is, industrial gas turbines are
The market for IGT-based grid
balancing assets is huge – anywhere
in the world where there is renew-
ables growth calling for day-to-day
renewables support, while offering
emergency backup for the grid.
There is also room for large frame
gas turbines, where countries have
large robust grids.
“In Sweden, we have a lot of hy-
dropower but when we close down
nuclear capacity and replace with
wind farms, there isn’t enough ca-
pacity to handle the residual power
peaks. There we will see a large de-
mand for [GT] backup power. Those
machines would probably operate
for less than 10 per cent of the time.
In many markets today, there is no
compensation for having capacity in
place and that is an issue.
“Grid integrity and resilience via
sufcient backup should mainly be
seen as part of the grid infrastructure
rather than energy trade. Solving
backup power supply with existing
coal red plants is a route that has
already proven a failure as it counter-
acts the greenhouse gas savings from
renewable power, i.e. incentives for
investment in more suitable backup
technology is needed” said Stuxberg.
He concluded: “Renewables and
storage systems will play the major
future role for energy supply but that
requires a lot of exible backup and
for that gas turbines are the most cost
effective today – if you need to build
capacity today; it’s gas turbines.
“We can only speculate on what
will happen in the future through
development of other technologies.
But we need to change the energy
system now. With the environmental
challenge, we cannot wait 30 years;
so we have to base it on the technol-
ogy we have today and industrial gas
turbines is an available technology
well t for the purpose. Backup
power also needs to be installed
ahead of renewable implementation
to ensure grid resilience, so the need
is urgent.”
Demand, solar and wind
supply in a simplied
ctitious medium size grid
Special Technology Supplement
The energy storage market is forecasted to grow exponentially
All storage technologies can store surplus renewable energy and return it to the
grid later, thus avoiding curtailment and increasing the use of renewable power.
According to analysis from IHS Markit, annual installations of energy storage
capacity globally will exceed 10 GW in 2021, more than doubling the 4.5 GW in-
crease in 2020. The existing capacity in stationary energy storage is dominated
by pumped-storage hydropower (PSH), but because of decreasing prices, new
projects are generally lithium-ion (Li-ion) batteries.
PSH capacity additions are predicted to remain constant at 5-10 GW per year,
while battery capacity is expected to grow from 2.3 GW/year in 2018 to above
30 GW/year in 2050. Total installed storage capacity was around 170 GW in
2019, a gure that is expected to reach 950 GW by 2050, according to IHS
Another report – ‘The Energy Storage Grand Challenge Energy Storage Mar-
ket Report 2020’ – published by the US Department of Energy forecasts a 27
per cent compound annual growth rate (CAGR) for grid-related storage through
to 2030. It projects annual grid-related global employment to increase about 15
times from around 10 GWh in 2019 to almost 160 GWh in 2030.
The type of storage deployed will depend on grid design and the distribution
of generating plants and loads unique to each grid. The technology selected
depends on which offers the best economic and operational capability according
to the services, range of capacity and energy discharge duration needed.
Super-capacitors and rotating grid stabilisers (ywheels and synchronous
condensers) provide instantaneous system responses and grid control. Both
technologies are aimed at applications in the range of approximately 1-100 MW.
Pumped storage hydro is the most dominant energy storage solution in terms
of globally installed megawatt capacity, representing some 93 per cent of the
operating system. It is a gigawatt-scale technology mostly used for energy shifting
and high-capacity rming with storage durations of around days or weeks with mini-
mal energy losses.
Further, capacity and operating reserve is provided when the asset is connected to
the grid. But although a mature and widespread technology, its main drawback is the
required topology of the site (large height differences are needed) and its physical
impact on the environment.
Thermal energy storage (TES) can improve utilisation of waste heat, assist in the
electrication of process heat supply, or store renewable energy for re-electrication
using a steam turbine. TES can also be integrated with thermal generation plants,
e.g. a combined cycle plant. A wide variety of heat storage media are available,
including liquids such as molten salt and pressurised water, or solids such as stone,
steel, concrete, or sand.
Liquid air energy storage (LAES) and compressed air energy storage (CAES) are
further technologies aimed at gigawatt-scale applications. LAES is based on the
cryogenic liquefaction of air when it is compressed with the use of (preferably)
renewable electricity. The liquid and the produced heat can be easily stored and
discharged when needed for re-electrication. CAES works similar but stores com-
pressed air. By adding a thermal storage to this technology, the overall efciency is
Li-Ion batteries are currently the technology of choice driven by their cost-effective-
ness and speed characteristics. They offer several applications, such as frequency
response, exibility enhancements of conventional power generation assets, black
start capabilities or energy arbitrage. Their sweet spot is up to around 250 MW and 5
hours of duration.