Wednesday, July 15, 2015

Cost of solar power (54)

Yesterday I blogged about the DeGrussa mine PV/battery installation.  I made a few simple estimates about costs, but omitted to make an estimate of the Levelised Cost of Electricity.  That’s what I’ll provide today.

The cost of the DeGrussa project is clearly stated – AUD 40 million.  That buys a PV system with 10.6 MW peak power, one-axis tracking, and a battery installation that delivers 6 MW power for an unspecified amount of time. 

I can estimate the annual output of the DeGrussa installation from the diesel fuel usage that has been avoided.  It’s stated to be 5 million litres of fuel, thereby abating 12,000 t of CO2 emissions per year.

Here are the relevant properties of diesel fuel (data source):
  • density: 0.832 kg/litre
  • carbon content: 86.1%
  • energy density: 35.9 MJ/litre
So 5 million litres of diesel fuel contains 179.5 × 10^12 J of chemical energy.  If the diesel engine is 38% efficient, that would give 68.2 × 10^12 J = 18,947 MWh of electrical energy.

[To check the numbers, the 5 million litres of diesel would give rise to 5 × 10^6 × 0.832 × 0.861 × 44/12 kg of CO2, which I make to be 13,133 t CO2.  Not a perfect match with the 12,000 t of CO2 abatement that is claimed in the media releases, but good enough.]

I now proceed to calculate the Levelised Cost of Electricity using my standard 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), which means that the required rate of capital return is 9.4%,
  • all projects have the same annual maintenance and operating costs (2% of the total project cost), and
  • government subsidies are neglected.
For further commentary on my LCOE methodology, see posts on Real cost of coal-fired power, LEC – the accountant’s view, Cost of solar power (10) and (especially) Yet more on LEC.  Note that I am now using annual maintenance costs of 2% rather than 3% as in posts during 2011.

The results for the DeGrussa installation are as follows:

Cost per peak Watt              AUD 3.77/Wp
LCOE                                     AUD 240/MWh

The components of the LCOE are:
Capital           {0.094 × AUD 40×106}/{18,947 MWhr} = AUD 198/MWhr
O&M              {0.020 × AUD 40×106}/{18,947 MWhr} = AUD 42/MWhr

By way of comparison, LCOE figures (in appropriate currency per MWh) for all projects I’ve investigated are given below.  The number in brackets is the reference to the blog post, all of which appear in my index of posts with the title “Cost of solar power ([number])”:
(2)        AUD 183 (Nyngan, Australia, PV)
(3)        EUR 503 (Olmedilla, Spain, PV, 2008)
(3)        EUR 188 (Andasol I, Spain, trough, 2009)
(4)        AUD 236 (Greenough, Australia, PV)
(5)        AUD 397 (Solar Oasis, Australia, dish, 2014?)
(6)        USD 163 (Lazio, Italy, PV)
(7)        AUD 271 (Kogan Creek, Australia, CLFR pre-heat, 2012?)
(8)        USD 228 (New Mexico, CdTe thin film PV, 2011)
(9)        EUR 200 (Ibersol, Spain, trough, 2011)
(10)      USD 231 (Ivanpah, California, tower, 2013?)
(11)      CAD 409 (Stardale, Canada, PV, 2012)
(12)      USD 290 (Blythe, California, trough, 2012?)
(13)      AUD 285 (Solar Dawn, Australia, CLFR, 2013?)
(14)      AUD 263 (Moree Solar Farm, Australia, single-axis PV, 2013?)
(15)      EUR 350 (Lieberose, Germany, thin-film PV, 2009)
(16)      EUR 300 (Gemasolar, Spain, tower, 2011)
(17)      EUR 228 (Meuro, Germany, crystalline PV, 2012)
(18)      USD 204 (Crescent Dunes, USA, tower, 2013)
(19)      AUD 316 (University of Queensland, fixed PV, 2011)
(20)      EUR 241 (Ait Baha, Morocco, 1-axis solar thermal, 2012)
(21)      EUR 227 (Shivajinagar Sakri, India, PV, 2012)
(22)      JPY 36,076 (Kagoshima, Kyushu, Japan, PV, start July 2012)
(23)      AUD 249 (NEXTDC, Port Melbourne, PV, Q2 2012)
(24)      USD 319 (Maryland Solar Farm, thin-film PV, Q4 2012)
(25)      EUR 207 (GERO Solarpark, Germany, PV, May 2012)
(26)      AUD 259 (Kamberra Winery, Australia, PV, June 2012)
(27)      EUR 105 (Calera y Chozas, PV, Q4 2012)
(28)      AUD 205 (Nyngan & Broken Hill, thin film PV, end 2014?)
(29)      AUD 342 (City of Sydney, multiple sites, PV, 2012)
(30)      AUD 281 (Uterne, PV, single-axis tracking, 2011)
(31)      JPY 31,448 (Oita, PV?, Japan, to open March 2014)
(32)      USD 342 (Shams, Abu Dhabi, trough, to open early 2013)
(34)      USD 272 (Daggett, California, designed 2010)
(35)      GBP 148 (Wymeswold, UK, PV, March 2013)
(36)      USD 139 (South Georgia, PV, June 2014)
(37)      USD 169 (Antelope Valley, CdTe PV, end 2015)
(38)      AUD 137 (Mugga Lane, PV, mid 2014)
(39)      AUD 163 (Coree, fixed PV, Feb 2015)
(40)      AUD 298 (Ferngrove Winery, PV, July 2013)
(41)      USD 125 (Cerro Dominador, CST, mid 2017)
(42)      USD 190 (La Paz, PV, September 2013)
(43)      USD 152 (Austin Energy, PV, 2016)
(44)      AUD 304 (Weipa, PV, January 2015)
(45)      AUD 256 (Kalgoorlie-Boulder, PV, August 2014)
(46)      AUD 141 (new Moree Solar Farm, PV, one-axis tracking, December 2015)
(47)      AUD 184 (Brookfarm, PV, December 2015)
(48)      USD 110 (Amanecer, PV, June 2014)
(49)      USD 113 (DEWA, PV, April 2016)
(50)      USD 284 (Ashalim, solar thermal, 2017)
(51)      USD 256 (Xina Solar One, solar thermal, 2017)
(52)      AUD 129 (Barcaldine, PV, one-axis, March 2017)
(53)      AUD 139 (Nyngan & Broken Hill, fixed PV, late 2015)
(54)      AUD 240 (DeGrussa, PV/batteries, early 2016)


You can compare results with my LCOE graphic.

On these numbers, the LCOE for the DeGrussa installation is quite expensive.  A solar thermal installation for comparison is Cerro Dominador, number 41 on the list above.  That has 110 MW peak power output from a heliostat/tower configuration with dry condensers, 17.5 hours thermal storage in molten salt, and an alleged Capacity Factor of 95%.  I calculated the LCOE for Cerro Dominador to be USD 125 per MWh (details). 

I calculate the Capacity Factor for DeGrussa to be 18,947/(10.6 × 24 × 365) = 0.204, which is not at all high for a system with one-axis tracking and a good solar resource.  With PV panels, the Capacity Factor is not improved by inclusion of storage, which is a major difference to solar thermal power generation with thermal storage.

We should of course make allowance for the fact that the Australian dollar is falling, currently at 0.743 USD.   Even so, the estimated LCOE for DeGrussa is perhaps 30% more than I expected.

Tuesday, July 14, 2015

DeGrussa mine installation

Big announcement today!

Sandfire Resources NL, an Australian mining company, has taken the plunge and invested in a major PV installation with battery storage at its DeGrussa copper and gold mine.  This project will set new worldwide benchmarks for integration of renewable energy into mining operations.

In this post, I’ll do a little detective work to look behind the media reports that are available here.  (At the link, you’ll find reports from ARENA, CEFC, juwi, Neoen, OTOC and Sandfire Resources.)

But first, some facts.  The DeGrussa mine is remote (900 km from Perth) and currently served by a 19 MW diesel generator.  The new PV installation will be integrated seamlessly with the existing diesel plant; it will have peak capacity 10.6 MW with “short-term” battery storage to deliver 6 MW.  There will be 34,080 PV panels with single-axis tracking installed on a 20 Ha site.  The project will be complete in early 2016.  Each year, the installation will offset 5 million litres of diesel fuel usage and 12,000 t CO2 emissions.

The actual owner of the project is Neoen, a French renewable energy firm.  juwi Renewable Energy Pty Ltd will develop and operate the project, with OTOC Ltd responsible for procurement and construction.  Sandfire Resources will purchase the electricity “at a fixed rate that is lower than the historical cost of diesel-generated power”.

The project is financed by CEFC (debt, AUD 15 million), ARENA (grant, AUD 20.9 million) and equity (mainly from Neoen, “less than $1 million” from Sandfire Resources).  The total project cost is AUD 40 million.

We all know about the trend with PV prices – namely down, with an experience curve that shows 22% price reduction in module costs per doubling of global installed capacity.  That trend has been in place for more than 30 years.  Data is also starting to emerge about the cost of battery storage, see for example my blog post on recently published data.

What can we infer about the cost of the batteries?

Let’s work backwards from an estimate for the cost of the PV component.  According to my recent analysis of the Nyngan & Broken Hill project, the capital cost there was AUD 2.84 per peak Watt.  At the same capital cost, the PV component of the DeGrussa project would be AUD 10.6 × 10^6 × 2.84 = AUD 30.1 million.

That leaves AUD 9.9 million for the battery component, which is to deliver a peak power of 6 MW.  If the “short term” storage is for 1 hour, the battery storage is priced at AUD 9.9 × 10^6 /6,000 kWh = AUD 1,650 per kWh.  If the storage is for 2 hours, the cost is AUD 9.9 × 10^6 /12,000 kWh = AUD 825 per kWh.  For 3 hours it would be AUD 9.9 × 10^6 /18,000 kWh = AUD 550 per kWh.  You can make your own calculation based on your interpretation of “short term”.

Presumably other companies bid for the DeGrussa project, amongst them companies using Concentrated Solar Thermal power generation with storage.  So, it’s interesting that PV/battery technology has won the contract.

The other salient feature that emerges from the media reports is that this is a big market segment.  According to ARENA’s media release, “remote industries in Australia currently rely on 1.2 GW of power from diesel fuel”.  Based on the price tag for the DeGrussa mine, I estimate the market niche to be AUD 1.2 × 10^9 × 40 × 10^6 / (10.6 × 10^6) = AUD 4.5 billion.  Worth pursuing, and perhaps the DeGrussa project represents a pivotal moment.

Whatever, I congratulate all those involved in this development.

Note added one day later: 
I was also able to make an estimate for the Levelised Cost of Electricity for the DeGrussa installation.  See, post for 16 July 2015.

Wednesday, July 8, 2015

Cost of solar heating (1)

RenewEconomy has an interesting story today about a planned solar thermal installation in Oman.  The project name is Miraah, which is Arabic for mirror.  Other sources with details of the Miraah project are here (WSJ) and here (the California-based project developers, GlassPoint).

But Miraah is not for power generation, rather for process heat in the form of steam to be used for enhanced oil recovery.  In brief this involves forcing steam into an oil field so that heavy oil is easier to extract.  At present, the operator of the oil field generates necessary steam from natural gas, so the solar project will free up 5.6 trillion BTU (5.9 billion MJ) of natural gas per year for other use.  If those other uses don’t involve combustion of the natural gas or other release of CO2, then CO2 emissions of 300,000 tons per year would be abated.

The cost of the Miraah project is reported as USD 600 million for a power output of 1,021 MW thermal.  The area of the plant is 3.0 km^2 and 6,000 tons of steam will be produced per day.

The Miraah technology involves parabolic trough mirrors with one-axis tracking.  Steam will be generated directly.  The only unusual feature about the installation is that the parabolic troughs will be deployed inside 36 large glasshouses, which will be used to shelter the mirrors from harsh desert conditions.  That way, the mirrors, presumably the most expensive component of the installation, do not have to be as robust as otherwise.

What can we say about the cost of this heat collection system?  Here’s a back-of-the-envelope calculation …

Suppose the efficiency of heat collection is 65% and that peak insolation is 1 kWth/m^2.  Then the area of the mirrors is 1,021,000 kW/(1kW/m^2 * 0.65) = 1,570,769 m^2.  The cost of the project is USD 600 million, which equates to 600,000,000/1,570,769 = USD 382/m^2.

Another useful metric is the cost per GJ.  Using my normal methodology for calculating CAPEX plus OPEX (see any post entitled “cost of solar power” on this blog), I estimate USD 600 × 106 × (0.094+0.02)/ (5.9 × 106 GJ) = USD 11.6/GJ.  More expensive than natural gas, but maybe the Omani investors have a cheaper cost of capital than the 8% on which my estimate is based. 

None of the press reports that I saw mention the collection temperature, but let's say they want an operating steam pressure of 20 bar (2.0 MPa), at which pressure the boiling point is 212°C.

This project is remote (800 km from the nearest town and 160 km from the Yemen border), so construction costs would not be the cheapest on the planet.  Still, at USD 382 per square metre, the cost of the plant does seem rather high.  I’ve heard lots of enthusiastic solar thermal proponents talking about costs of around USD 100 per square metre.

I’m left pondering the question – what is world-best cost (USD per square metre) for large-scale solar heating?