1
The Base-Load Fallacy
Mark Diesendorf
Institute of Environmental Studies
University of New South Wales
Sydney 2052, Australia
Email: m.diesendorf@unsw.edu.au
Abstract
It is claimed by some that a large-scale electricity generation system cannot be based upon
renewable sources of energy, because the latter are alleged to be ‘intermittent’ sources that
cannot provide base-load (24-hour) power. This paper shows that there is actually a wide variety
of renewable energy sources with different types of time variability. Some of these have similar
variability to coal (e.g. bio-electricity, hot rock geothermal, solar thermal electricity with thermal
storage) and are therefore base-load. Although large-scale wind power has a different variability,
it can substitute for some base-load coal with the assistance of a small amount of peak-load
power plant (e.g. gas turbine). Together, a mix different types of renewable energy sources can
replace a conventional generating system and can be just as reliable.
Introduction
Opponents of renewable energy, from the coal and nuclear industries and from NIMBY (Not In
My Backyard) groups, are disseminating the fallacy that renewable energy cannot provide baseload
power to substitute for coal-fired electricity. Even Government Ministers and some ABC
journalists are propagating this conventional ‘wisdom’, although it is incorrect. The political
implications are that, if the fallacy becomes widely believed to be true, renewable energy would
always have to remain a niche market, rather than achieve its true potential of becoming a set of
mainstream energy supply technologies.
The refutation of the fallacy has the following key logical steps:
• With or without renewable energy, there is no such thing as a perfectly reliable power
station or electricity generating system.
• Electricity grids are already designed to handle variability in both demand and supply. To do
this, they have different types of power station (base-load, intermediate-load and peak-load)
and reserve power stations.
• Some renewable electricity sources (e.g. bioenergy, solar thermal electricity and geothermal)
have identical variability to coal-fired power stations and so they are base-load. They can be
integrated into electricity grids without any additional back-up, as can efficient energy use.
• Other renewable electricity sources (e.g. wind, solar without storage, and run-of-river hydro)
have different kinds of variability from coal-fired power stations and so have to be
considered separately.
• Wind power provides a third source of variability to be integrated into a system that already
has to balance a variable conventional supply against a variable demand.
ANZSEE Solar 2007 Conference, Alice Springs, October
2
• The variability of small amounts of wind power in a grid is indistinguishable from variations
in demand. Therefore, existing peak-load plant and reserve plant can handle small amounts
of wind power at negligible extra cost.
• For large amounts of wind power connected to the grid from several geographically
dispersed wind farms, total wind power generally varies smoothly and therefore cannot be
described accurately as ‘intermittent’. Thus, the variability of large-scale dispersed wind
power is unlike that of a single wind turbine. Nevertheless, it may require some additional
back-up.
• As the penetration of wind power increases substantially, so do the additional costs of
reserve plant and fuel used for balancing wind power variations. However, when wind
power supplies up to 20% of electricity generation, these additional costs are still relatively
small.
These steps are now discussed in more detail. First it is necessary to define ‘base-load’.
Base-load power stations
A base-load power station is one that is in theory available 24 hours a day, seven days a week,
and operates most of the time at full power. In practice, this is an ideal. In reality, even base-load
power stations break down from time to time and, as a result, can be out of action for weeks. In
mainland Australia, base-load power stations are mostly coal-fired – a few are gas-fired. Coalfired
power stations are by far the most polluting of all power stations, both in terms of
greenhouse gas emissions and local air pollution.
Overseas, some base-load power stations are nuclear. They produce little pollution during normal
operation, but much pollution (including carbon dioxide emissions) from mining, enrichment,
plant construction and decommissioning, reprocessing and waste management. They also
increase the risks of proliferation of nuclear weapons and have the capacity for rare but
catastrophic accidents.
Renewable energy can provide several different clean, safe, base-load technologies to substitute
for coal (Diesendorf 2007a):
• bioenergy, based on the combustion of crops and crop residues, or their gasification
followed by combustion of the gas;
• hot rock geothermal power, which is being developed in South Australia and Queensland;
• solar thermal electricity, with overnight heat storage in water or rocks or a thermochemical
store; and
• large-scale, distributed wind power, with a small amount of occasional back-up from
peakload plant.
It is obvious that the first three of these types of renewable power station are indeed base-load.
Efficient energy use, the natural companion of renewable energy, can also substitute directly for
base-load coal. However, the inclusion of large-scale wind power in the above list may be a
surprise to some people, because wind power is often described as an ‘intermittent’ source, one
that switches on and off frequently. Before discussing the variability of wind power, we
introduce the concept of ‘optimal mix’.
ANZSEE Solar 2007 Conference, Alice Springs, October
3
Optimal mix of base-load and peak-load
An electricity supply system cannot be built out of base-load power stations alone. These stations
are inflexible to operate. They take all day to start up from cold and in general their output
cannot be changed up or down quickly enough to handle the peaks and other variations in
demand. Base-load stations used as reserve cannot be started up quickly from cold. Base-load
power stations, especially coal-fired and nuclear, are generally cheap to operate, but their capital
costs are high. So they cannot be used just to handle peaks in demand. To pay back their high
capital costs, base-load power stations must be operated as continuously as possible. A faster,
cheaper, more flexible type of power station is needed to complement base-load and handle the
peaks.
Peak-load power stations are designed to be run for short periods of time each day to supply the
peaks in demand and to handle unpredictable fluctuations in demand on timescales ranging from
a few minutes to an hour or so. They can be started rapidly from cold and their output can be
changed rapidly. Some peak-load stations are gas turbines, similar to jumbo jet engines, fuelled
by gas or (rarely) by oil. They have low capital costs but high operating costs (mostly fuel costs).
Hydro-electricity with dams is also used to provide peak-load power. Because the amount of
water available is limited to that stored in the dam, the ‘fuel’ of a hydro power station is a scarce
resource and therefore a valuable fuel that is best used when its value is highest, that is, during
the peaks.
A third type of power station, intermediate-load, runs during the daytime, filling the gap in
supply between base- and peak-load power (see Figure 1). Its output is more readily changed
than base-load, but less than peak-load. Its operating cost lies between those of base- and peakload.
Sometimes intermediate load is supplied by gas-fired power stations and sometimes by
older, smaller, black coal-fired stations.
Clearly, if an electricity generating system has too much peak-load plant, it will become very
expensive to operate, but if it has too much base-load plant, it will be very expensive to buy. For
a particular pattern of demand there is a mix of base-load, intermediate-load and peak-load plant
that gives the minimum annual cost. This is known as the optimal mix of generating plant.
Figure 1 sketches how a mix of base-load, intermediate-load and peak-load generation combines
to meet the daily variations in demand in Summer and Winter.
Reliability of generating systems
Even an optimal mix of fossil-fuelled power stations is not 100% reliable. To achieve this would
require an infinite amount of back-up and hence an infinite cost. In practice, a generating system
has a limited amount of back-up and a specified reliability. This can be measured in terms of (a)
the average number of hours per year that supply fails to meet demand or (b) by the frequency
and duration of failures to meet demand.
ANZSEE Solar 2007 Conference, Alice Springs, October
4
Consider an electricity generating system comprising N thermal power generation units with
rated capacities ci , where i = 1, …, N, with total rated capacity
C = Σ ci
where the sum is over all values of i from 1 to N.
At a given time, the available capacity (i.e. that which is not undergoing planned or forced
outage) of unit i is a random variable ai and the total available capacity at a given time is
A = Σ ai
The load or demand at a given time is the random variable L. Measure (a) of the reliability of the
generating system (mentioned above) is the Loss of Load probability (LOLP), denoted by p0,
which is the average value of the fraction of time that the load L is greater than the total available
power A:
p0 = Average [Pr (A
ci, N and hence C. Ultimately the choice is political: how many hours per year of blackouts can a
government tolerate?
The economic optimal mix of thermal generating units, for a given value of p0, is the
configuration of base-load, intermediate-load and peak-load power stations that minimises the
cost function
F = Σ ciyi + eizi (2)
where the sum is again over all values of i. Here yi is the annualised capital cost per megawatt of
rated capacity ci; ei is the annual energy generated by unit i; and zi is the total operation,
maintenance and fuel cost per unit of energy generated. The cost function Equation (2) is
evaluated numerically under the constraint given by Equation (1), as shown by Martin and
Diesendorf (1982). The calculation is a non-trivial, since A and L are random variables (i.e.
described by probability distributions which are obtained from empirical data).
Wind power as base-load
To replace the electricity generated by a 1000 megawatt (MW) coal-fired power station, with
annual average power output of about 850 MW, a group of wind farms with capacity (rated
power) of about 2600 MW, located in windy sites, is required. The higher wind capacity allows
for the variations in wind power and is taken into account in the economics of wind power.
Although this substitution involves a large number of wind turbines (for example, 1300 turbines,
each rated at 2 MW), the area of land actually occupied by the wind turbines and access roads is
only 5–19 square km, depending upon wind speed. Farming continues between the wind
ANZSEE Solar 2007 Conference, Alice Springs, October
5
turbines. For comparison, the coal-fired power station and its open-cut coal-mine occupy
typically 50–100 square km.
Although a single wind turbine is indeed intermittent, this is not generally true of a system of
several wind farms, separated by several hundred kilometres and experiencing different wind
regimes. The total output of such a system generally varies smoothly and only rarely experiences
a situation where there is no wind at any site (Sinden 2007). As a result, this system can be made
as reliable as a conventional base-load power station by adding a small amount of peak-load
plant (say, gas turbines) that is only operated when required.
Computer simulations and modelling show that the integration of wind power into an electricity
grid changes the optimal mix of conventional base-load and peak-load power stations. The
method is to include wind power as a negative load in Equations (1) and (2). Empirical data are
used for the probability distribution of wind power (Martin and Diesendorf 1982).
The result is that wind power replaces base-load with the same annual average power output.
However, to maintain the reliability of the generating system at the same level as before the
substitution, some additional peak-load plant may be needed. This back-up does not have to have
the same capacity as the group of wind farms. For widely dispersed wind farms, the back-up
capacity only has to be one-fifth to one-third of the wind capacity. In the special case when all
the wind power is concentrated at a single site, the required back-up is about half the wind
capacity (Martin & Diesendorf 1982; Grubb 1988a & b; ILEX 2002; Carbon Trust & DTI 2004;
Dale et al. 2004; UKERC 2006).
Furthermore, because the back-up is peak-load plant, it does not have to be run continuously
while the wind is blowing. Instead the gas turbines can be switched on and off quickly when
necessary. Since the gas turbine has low capital cost and low fuel use, it may be considered to be
reliability insurance with a small premium.
Of course, if a national electricity grid is connected by transmission line to another country (for
example, as Western Denmark is connected to Norway), it does not need to install any back-up
for wind, because it purchases supplementary power from its neighbours when required and sells
excess wind energy to its neighbours. In practice it makes little difference whether a generating
system installs a little of its own back-up or purchases it from neighbours.
Solar electricity
Because it is still very expensive to store electricity on a large scale, grid-connected solar
electricity from photovoltaic (PV) modules is not stored. If and when advanced batteries become
less expensive, PV electricity would become base-load. Meanwhile, even without storage, a large
amount of solar PV can substitute for coal and/or gas combusted in intermediate-load power
stations. Furthermore, by orienting the solar collectors to the north-west instead of to the usual
north (in the southern hemisphere), the peak in solar generation overlaps to a large degree with
the broad daily peak in Summer demand (Figure 1b). Thus, statistically speaking, even solar
electricity without storage has a degree of reliability during the daytime.
ANZSEE Solar 2007 Conference, Alice Springs, October
6
Solar energy can be stored at low cost as heat in water, rocks or thermochemical systems.
Therefore, solar thermal electricity with thermal storage can supply base-load and can be just as
reliable as base-load coal.
New technological developments in solar electricity, coupled with expanding overseas markets,
will gradually bring down prices.
Conclusion
Combinations of efficient energy use and renewable sources of electricity can replace electricity
generating systems based on fossil fuels and nuclear power. With renewable sources, base-load
electricity can be provided to the grid by bioenergy, hot rock geothermal, solar thermal
electricity with thermal storage in water, rock or thermochemical systems, and wind power with
a little back-up from gas turbines. Natural gas and coal seam methane can also substitute for
some base-load coal-fired power stations, although supplies of these gases are limited in eastern
Australia. Intermediate load can be supplied by solar PV electricity without storage, when it
becomes less expensive. When natural gas supplies become scarce, gas turbines used for peakload
supply can be fuelled by liquid or gaseous fuels produced from biomass.
By 2040 renewable energy could supply over half of Australia’s electricity, reducing CO2
emissions from electricity generation by nearly 80 per cent (Saddler, Diesendorf & Denniss
2004; Diesendorf 2007a & b). In the longer term, there is no technical reason to stop renewable
energy from supplying 100 per cent of grid electricity. The system could be just as reliable as the
greenhouse-intensive fossil-fuelled system that it replaces. Taking account of the high costs of
greenhouse impacts (Stern 2006), the barriers to a sustainable energy future are neither
technological nor economic, but rather are the immense political power of the big greenhouse
gas polluting industries: coal, aluminium, iron and steel, cement, motor vehicles and part of the
oil industry.
Actually, there is one constraint on a renewable electricity future. Growth in demand has to be
levelled off, or there will not be enough land for wind and bioenergy. In the long run, this would
entail a change in the national economic structure and the stabilisation of Australia’s population.
References
Carbon Trust and DTI (2004) Renewable Networks Impact Study: Annex 1 – Capacity Mapping and
Market Scenarios for 2010 and 2020.
www.carbontrust.co.uk/Publications/publicationdetail.htm?productid=CT-2004-03
Dale, L, Milborrow, D, Slark, R & Strbac, G (2004) Total cost estimates for large-scale wind scenarios in
UK, Energy Policy 32: 1949–956.
Diesendorf, M (2007a) Greenhouse Solutions with Sustainable Energy, UNSW Press, Sydney.
Diesendorf, M (2007b) Sustainable Energy for Australia, fact sheet no. 5,
Grubb, MJ (1988a) The potential for wind energy in Britain, Energy Policy 16: 594-607.
Grubb, MJ (1988b) The economic value of wind energy at high power system penetrations: an analysis of
models, sensitivities and assumptions, Wind Engineering 12: 1–26.
ANZSEE Solar 2007 Conference, Alice Springs, October
7
ILEX (2002) Quantifying the System Costs of Additional Renewables. ILEX/UMIST,
Martin, B & Diesendorf, M (1982) Optimal thermal mix in electricity grids containing wind power,
Electrical Power & Energy Systems 4: 155–161.
Saddler, H, Diesendorf, M & Denniss, R (2004) A Clean Energy Future for Australia, Clean Energy
Future Group, Sydney. Full report (1.24 MB) available on _
Sinden G (2007) Characteristics of the UK wind resource: long-term patterns and relationship to
electricity demand. Energy Policy 35: 112-27.
Stern N (2006) Stern Review: The Economics of Climate Change, October,
UKERC (2006) The Costs and Impacts of Intermittency, UK Energy Research Centre,
Figure 1: Typical power demand (load) by time of day in (a) winter and (b) summer
In Winter the two peaks occur at breakfast and dinner time. In Summer the single broad peak occurs in early to midafternoon.
No comments:
Post a Comment