Tuesday, April 26, 2011

LEC - the accountant's view

A reader with a background in accountancy would be so horrified by the way I calculate the cost of electricity that I might be accused of mental decrepitude.  But, as Shakespeare would say, there is method in my madness; I am only looking to compare different methods of generation, not to present a bulletproof document to the Taxation Office.

And, as I’ll show in this post, my conclusions about the cost of power are robust even if I do the calculations according to accounting standards, provided any tax applied to carbon emissions is modest.  Let me illustrate by making a continuation of my last post – “Real Cost of Coal-Fired Power”.

Recall that I was looking at the Levelised Electricity Cost (LEC) for production from a Rankine-cycle steam plant with the following attributes:
·         specific capital cost: AUD 1,500,000/MW
·         thermodynamic efficiency: 0.39
·         cost of coal: AUD 120/t
·         fraction of coal that is carbon: 0.8
·         energy content of coal: 27.9 GJ/t
In addition, let me make my usual assumptions
·         interest rate is 8%
·         non-fuel operations and maintenance costs are 3% of the project cost
·         nominally the payback period for loans is 25 years, which in this case I take to mean the generation equipment is depreciated over 25 years
and the following specific additional assumptions
·         the plant has capacity 1,000 MW
·         the project is 50% funded by debt
·         the rate of taxation is 30%
·         the Social Cost of Carbon (SCC) varies between AUD 0 and AUD 100
·         the Capacity Factor (CF) varies between 0.70 and 1.00
·         the required after-tax rate of return on capital is 5%

The LEC is selected so that the required after-tax rate of return is achieved.  In the case where the plant is used 100% of the time (CF = 1) and the social cost of carbon is ignored (SCC = 0), this in fact requires the LEC to be AUD 64.65/MWhr, and the Profit and Loss statement for the year is given in the table below (all figures in AUD millions):

less fuel costs
less non-fuel O&M
less depreciation
earnings before interest and taxation
less interest payments
pre-tax profit
less taxation
less SCC
nett profit after tax

On those figures, the after-tax rate of return is 37.5/750, or 5.00%, as required.

The various acronyms used are:
·         CF is the capacity factor, the fraction of time the generator is active
·         EBITDA is earnings before interest, taxation, depreciation and allowances
·         LEC is levelised electricity cost
·         O&M is operations and maintenance
·         SCC is the social cost of carbon expressed in AUD/ t CO2 emitted

Thus, on these assumptions, the conventionally accounted LEC is AUD 65/MWhr, which is close to the figure AUD 61/MWhr obtained with my usual methodology (see previous post).

I can now repeat the calculations for the range of SCC and CF values to give the Figure below (click to see original).  The dotted lines are LEC values based on the conventional accounting methodology, as above, and the full lines are the results from my last post (i.e. using a simplified set of assumptions).

If the Social Cost of Carbon is zero, then it does not matter much which method is used to assess the LEC.  As the Social Cost of Carbon is increased, then my simplified method underestimates the LEC as calculated by conventional accounting.

These LEC calculations involve a large number of parameters, more so under the conventional accounting treatment, less so under my simplified method.  By tweaking important parameters downwards (particularly the interest rate, the capital cost and the O&M costs), it is possible to strongly affect the calculated LEC.  That is why the UMPNER review’s LEC for nuclear power needs to be treated with suspicion (see post “Cost of Nuclear Power”, 14 March 2011).

My conclusions from this exercise:
·         For comparison purposes, I’m happy with the simplified methodology I have been using to estimate the LEC.
·         A tax for the Social Cost of Carbon increases the LEC substantially, particularly under conventional accounting methodology.

Friday, April 22, 2011

Real cost of coal-fired power

Please note:  see my blog post on 27 January 2015 for an update of this report.

If you’ve visited this blog previously, you’ll know that I report on the cost of electricity produced from solar power.  I do this on a level playing field, with all projects assessed under the same assumptions:
·         there is no inflation,
·         taxation implications are neglected,
·         projects are funded entirely by debt,
·         all projects have the same interest rate (8%) and payback period (25 years),
·         all projects have the same annual maintenance and operating costs (3% of the total project cost), and
·         government subsidies are neglected.

Typically, my estimates for the Levelised Electricity Cost (LEC) from solar power are disconcerting – the solar LEC is, in Australia at least, significantly more than the accepted cost of coal-fired power.  But what happens if the cost of coal-fired power is assessed according to my standard assumptions?  And what happens if a cost is applied to externalities due to coal production?

A recent authoritative report on externalities associated with coal-fired power is due to Epstein et al. “Full cost accounting of the life cycle of coal”, Annals of the NY Academy of Sciences 1219 (2011), 73-98.  Issues they take into consideration are
·         land disturbance
·         methane emissions from mines
·         carcinogens
·         public health burden of communities
·         fatalities due to coal transport
·         emissions of air pollutants from combustion
·         effects of mercury emissions (lost productivity, mental health, cardiovascular)
·         climate damage from combustion emissions (CO2, N2O, soot)
Epstein et al.’s best estimate for the additional monetised cost of coal production on the LEC was USD (2008) 178 per MWhr, with a range of USD 94-269 per MW hr.  This is in addition to other costs of production such as cost of fuel, capital costs, operations, maintenance etc.

Another way to account for the externalities is to apply a “Social Cost of Carbon” (SCC), expressed in monetary units per tonne of CO2 emitted.  This topic is the subject of vigorous economic research, as can be seen by a simple web search, and a range of SCC values is usually quoted.  It is widely agreed that the SCC will increase in time.  A recent commentary on the SCC from a US context is due to Frank Ackerman: “What is the social cost of carbon?” on Real-World Economics Review Blog, 27 April 2010.  He concludes:

“There are too many open questions in the SCC calculation to recommend a precise alternate value based on the information now available; there is a need for more extensive research, examining the full range of available studies of climate damages and costs, and analyzing assumptions about the risks and magnitudes of potential climate catastrophes.  In the United Kingdom, where carbon pricing and cost calculations have a longer, better-researched history, the latest estimate is a range of $41 to $124 per ton of CO2, with a central case of $83.”

With that preamble, now follows my contribution to this debate.

The following assumptions (additional to my standard assumptions) are made in order to calculate the LEC from a Rankine-cycle steam plant:
·         specific capital cost: AUD 1,500,000/MW
·         thermodynamic efficiency: 0.39
·         cost of coal: AUD 120/t
·         fraction of coal that is carbon: 0.8
·         energy content of coal: 27.9 GJ/t
The LEC is then calculated for
·         SCC in the range AUD 0 to AUD 100 per tonne
·         Capacity Factors in the range 0.7 to 1.0 (the Capacity Factor is the fraction of time that the generator is active)

The results are shown in the figure below.

For these parameters, 0.97 t CO2 is emitted per MWhr.  So, as a rule of thumb, every $ increase in the SCC translates to an extra $ on the LEC when expressed in $/MWhr.

The cost of solar power no longer looks so expensive, particularly if you care to prognosticate on likely price increases for coal.  In that context, Bloomberg for example gives a time series for the spot price of coal from Newcastle, Australia; see CLSPAUNE:IND. One could reasonably expect that the cost of coal will increase significantly in the next decade.


Thursday, April 14, 2011

Cost of solar power (7)

Yesterday, the Australian media carried details of a new and unusual solar installation to be built in the cutely named town of Chinchilla in the state of Queensland.  This can strictly be classified as Concentrated Solar Thermal, but it’s unusual since the solar energy will be deployed in pre-heating water in an otherwise conventional coal-fired power station.

Here is an extract from a press release issued by the owner of the plant, CS Energy (www.csenergy.com.au):

“The Commonwealth Government and Queensland Government have given the green light to one of the world’s largest solar projects, announcing today that approval had been given for the $104.7 million Kogan Creek Solar Boost Project.  CS Energy’s 750 megawatt coal-fired Kogan Creek Power Station near Chinchilla in South West Queensland will soon become home to a 44 megawatt solar thermal addition representing the largest solar project in the Southern Hemisphere and the world’s largest solar integration with a coal-fired power station.”

According to Chief Executive, David Brown, the project will

“…  [avoid] production of 35,600 tonnes of greenhouse gases annually …  [and] increase the amount of electricity generated by up to 44 megawatts during peak solar conditions, providing an additional 44,000 megawatt hours of electricity per year.”

“The project will use AREVA Solar’s Australian-pioneered Compact Linear Fresnel Reflector (CLFR) technology to supply additional steam to the power station’s turbine, supplementing the conventional coal-fired steam generation process.”

Normally in Rankine-cycle steam turbines, water pre-heating is achieved using heat energy in steam that would otherwise be used to spin the turbines.  The solar pre-heater makes more steam available for the electricity generation process.  In principle, solar pre-heating should be very effective when used in conjunction with existing Rankine-cycle turbines. 

CS Energy’s web site reports that of the total project cost of AUD 104.7 million, AUD 34.9 million came from the federal government’s Renewable Energy Demonstration Program.  CS Energy also reports that

“CS Energy has received support from the Queensland Government through a contribution of $35.4 million to CS Energy’s Carbon Reduction Program, which has enabled the company to direct funds to the Kogan Creek Solar Boost Project.

So a significant chunk of the project’s cost has come from government support.  But let’s calculate the metrics of peak cost and LEC (Levelised Electricity Cost) under my usual assumptions:
·         there is no inflation,
·         taxation implications are neglected,
·         projects are funded entirely by debt,
·         all projects have the same interest rate (8%) and payback period (25 years),
·         all projects have the same annual maintenance and operating costs (3% of the total project cost), and
·         government subsidies are neglected.

That gives for the Kogan Creek project:

Cost per peak Watt     AUD 2.38/Wp
LEC                            AUD 295/MWhr

The components of the LEC are:
CAPEX           {0.094× AUD 105×106}/{44×103 MWhr} = AUD 224/MWhr
OPEX             {0.030× AUD 105×106}/{44×103 MWhr} = AUD 71/MWhr

The cost of per tonne of CO2 emissions reduced is also of interest.  I make that
{0.094 + 0.03} × AUD 105×106/35600 tonnes = AUD 365 / tonne.

By way of comparison, here are my LEC figures for all projects I’ve investigated:

Cost of solar power (2): AUD 199/MWhr (Nyngan, Australia, PV)
Cost of solar power (3): EUR 547/MWhr (Olmedilla, Spain, PV)
Cost of solar power (3): EUR 205/MWhr (Andasol I, Spain, trough)
Cost of solar power (4): AUD 257/MWhr (Greenough, Australia, PV)
Cost of solar power (5): AUD 432/MWhr (Whyalla, Australia, dish)
Cost of solar power (6): USD 177/MWhr (Lazio, Italy, PV)
Cost of solar power (7): AUD 295/MWhr (Kogan Creek, Australia, CLFR pre-heat)

As a big fan of solar thermal power generation, I’m disappointed by these latest estimates.  I had expected that the LEC for this sort of installation would be better.

The other comment I’d make concerns the cost to be applied to CO2 emissions in any emissions reduction scheme.  Here in Australia the discussions seem to focus on a typical cost of perhaps AUD 25 /tonne.  Because of very difficult political processes, that number might be accurate only within a factor of two.  Whatever, it’s a long way short of the AUD 365/tonne I’ve estimated for the Kogan Creek project.