Content uploaded by Johannes L. Van Niekerk
Author content
All content in this area was uploaded by Johannes L. Van Niekerk on Nov 18, 2015
Content may be subject to copyright.
ROADMAP FOR THE DEPLOYMENT OF CONCENTRATING
SOLAR POWER IN SOUTH AFRICA
Adriaan J. Meyer1 and Johannes L. van Niekerk2
1 Mr AJ Meyer, BEng (Electric & Electronic) & MScEng (Mechanical), Research Engineer, Centre for Renewable and
Sustainable Energy Studies, Department of Mechanical and Mechatronic Engineering, Stellenbosch University, South
Africa. Private Bag x1, Matieland, 7602, South Africa. Telephone: +27 21 808 4945. E-mail: meyer@sun.ac.za.
2 Prof JL van Niekerk, Director: Centre for Renewable and Sustainable Energy Studies, Stellenbosch University, South
Africa.
Abstract
To identify suitable sites in the Republic of South Africa for short-, medium- and long-term implementation
scenarios for concentrating solar power (CSP) plants, geographical information system (GIS) technology was
used by the researchers. An area was assumed suitable if it receives a satisfactory amount of annual
irradiation; is close enough to suitable transmission substations; the transmission lines feeding these
substations adhere to certain criteria; the area is flat enough; there are sustainable water sources close by; and
it has a suitable landuse.
For immediate implementation, various sites near substations in the provinces of the Free State-, Northern
Cape- and North West- in South Africa were identified. DNI-levels at sites in the Northern Cape, for
instance, reached levels of up to 2900 kWh/m2/a.
In the medium-term, a number of high-voltage transmission line- and substation projects that are currently
under development or construction in the Northern Cape will strengthen the energy grid in this province
significantly, and will substantially increase the viable CSP-areas. It will further increase the current
generation limit of the transmission system in the Northern Cape from 1129 MW to an estimated figure up to
3000 MW.
For long-term planning, several additional areas were identified that are not currently viable. In this scenario
it is assumed that low-water consumption technologies, such as the Brayton Cycle Central Receiver, will
become commercially viable.
Keywords: CSP, South Africa, DNI, roadmap, solar, GIS.
1. Introduction
When the National Energy Regulator of South Africa (NERSA) announced the implementation of renewable
energy feed-in tariffs (REFIT) in March 2009 [1] and again in October 2009 [2], the interest of various
renewable energy participants were stimulated. Developers consequently became eager to construct
renewable energy power plants, including CSP-plants, in South Africa. For the first time in South Africa’s
history, independent power producers (IPPs) became viable - not simply to sell electricity into the grid under
the REFIT-programme, but also by means of wheeling (willing buyer-willing seller) agreements.
Developing countries are typically characterised by a midday electricity peak. South Africa is however
different in this regard and has a morning and more pronounced evening peak between 18:00 and 21:00; both
in summer and winter [3]. This characteristic, along with the country having some of the highest DNI-levels
in the world [4][5][6] and a general lack of natural gas as co-firing fuel, makes CSP with storage a very
attractive technology to generate electricity on a large scale to supply to the national grid.
In this paper the short-, medium- and long -term implementation scenarios of CSP-plants in South Africa are
discussed. The short- term scenario considers sites that are currently viable for CSP, whilst the medium-term
scenario examines sites that will become available by 2017 when the necessary and already approved grid
expansion projects will be completed. The long-term scenario investigates the effect that waterless CSP-
technologies will have on the implementation of CSP in South Africa.
2. Short-term scenario (2011 – 2017)
2.1. Current status of CSP in South Africa
A number of CSP-projects are planned under the first phase of REFIT, in which 1025 MW was allocated to
renewable energy [7]. No CSP-project has yet been constructed and no power purchase agreement (PPA) has
been awarded. The CSP-projects under development in South Africa include (only projects where the
Environmental Impact Assessment (EIA) study and/or ground solar measurements have been commissioned
form part of the list) [8]:
150 MW, near Pofadder in the Northern Cape by Kaxu CSP;
75 MW, near Groblershoop in the Northern Cape by Solafrica Thermal Energy;
100 MW, near Kathu in the Northern Cape by Renewable Energy Investments South Africa (REISA)
125 MW, near Upington in the Northern Cape by Ilangalethu Solar Power;
110 MW, near Upington in the Northern Cape by Khi CSP South Africa;
100 MW, near Kimberley in the Northern Cape by Afri-Devo;
30 MW, near Daniёlskuil in the Northern Cape by Afri-Devo; and
100 MW, near De Aar in the Northern Cape byAfrican Clean Energy Developments, (ACED).
In addition to these plants, Eskom has been in the process of timeously planning a 100 MW Central Receiver
with salt storage near Upington in the Northern Cape for more than ten years[8]. This brings the total CSP-
projects under development to 890 MW.
2.2. GIS study
2.2.1 Background
In order to assess the CSP potential of South Africa, a study incorporating geographical information system
(GIS) technology was conducted. Through the study the researchers identify viable South African locations
for CSP-plants, and were able to calculate an estimated total installed capacity for CSP.
2.2.2 Method
A similar approach was taken to that of Fluri [9]. Appropiate screens for a GIS-analysis are defined and
applied to gathered data for the area of interest; allowing the identification of potential CSP- sites.
Satellite-derived DNI, obtained from the United States’ National Renewable Energy Laboratory (NREL) is
used. The spacial resolution is 40 km by 40 km. Annual sums based on the long-term averages from 1985 to
1991 is used. The data is freely available on the Solar and Wind Energy Resource Assessment (SWERA)
website [6], and has an estimated accuracy of 10 % [6]. In figure 1 a map derived from this data is depicted,
along with South Africa’s province names. Only sites with an average annual sum of 2400 kWh/m2 and
above are furthermore considered.
Fig. 1. DNI-map of South Africa, including main perennial rivers.
A digital elevation model (DEM), the Shutter Radar Topography Mission (SRTM) 90 m DEM with a 90 m
spatial resolution, was used to identify areas with a suitable slope. Only areas with a slope less than 2° (in any
direction) were considered. Developers and technology suppliers consider sites with a slope up to 3° as viable
for CSP-plants [10].
Transmission line- and substation data was obtained from the local utility, Eskom [11]. In this study
transmission lines are defined as lines with a voltage of 220 kV and above. Only areas within a 30 km radius
from transmission substations were considered, as portrayed by figure 2. Substations consisting of only series
capacitor banks, and therefore no transformers, were excluded; and substations that have a feed-in capability
of 50 MW and above alone were considered. Only substations that are fed by multiple transmission lines
(220 kV and above) were furthermore considered. In certain instances, if a 132 kV was present, the site was
included.
Fig. 2. Transmission network map of South Africa, including areas with
annual DNI above 2400 kWh/m2 and 30 km buffer around substations.
Built-up areas; water bodies and wetlands; mines and quarries; forests; parks and conservation areas were
excluded from a suitable CSP-site.
Only areas that are within 20 km from suitable water sources were furthermore considered. For the purposes
of the study a suitable water source was defined as selected perennial rivers with a suitable supply of water
[12].
After all these criteria were applied, parcels with an area of at least 2 km2 were considered. In table 1 the
criteria of this study is compared to that of Fluri [9]’s previous study. The final result is reflected in figure 3.
From this figure, a total area of 14 666 km2 is considered viable. This relates to 262 GW. CSP-plant land
requirements were based on that of parabolic through with seven hours of thermal storage - 56 km2/GW [13].
Only 51 % of the viable sites (134 GW) are in the Northern Cape due to the low number of viable substations
near water resources.
Criteria
Fluri [1]
Present Study
Slope
< 1°
<2°
Grid
<20 km from
transmission lines
<30 km from
transmission substations
DNI
>2555 kWh/m2/a
(>7 kWh/m2/d)
>2400 kWh/m2/a
(>6.58 kWh/m2/d)
Table 1. Criteria comparison
Fig. 3.Viable short-term CSP-sites.
2.3. Current grid
The total capacity for connections on the transmission level that can be fed into the national transmission grid
of the North Zone (comprising of the Northern Cape and parts of the Free State) is 1000 MW [14], as
indicated by figure 4b. A similar figure for the North Zone on sub-transmission level is 129 MW [14]. The
total feed-in capacity of the North Zone is therefore 1129 MW. The potential CSP capacity identified
(262 GW) far exceeds this number. In order to increase the feed-in capability, the Northern Cape grid must
be strenghened.
Fig. 4.Feed-in capabilities of a) substation capacity b) Eskom zone capacity [14].
3. Medium- term scenario (2018 – 2025)
3.1. Current grid expansion projects
A number of grid expansion projects are being implemented [15] and estimated to be in operation by between
2013 and 2017. As indicated by figure 5, the projects that are of significance for this study include:
North-West Strengthening Phase;
Northern Cape Reinforcement Ferrum–Garona–Nieuwehoop and Sishen-Saldanha traction upgrade; and
Namaqualand Upgrade.
Fig. 5. Grid expansion project in the North Zone (redrawn with data obtained from [15]).
3.1.1 North West Strengthening
This project will significantly increase the amount of power that can be generated in the North Zone and
transmitted to the Gauteng province (North East); and will include the contraction of a new substation near
Vryburg. The phases of this project include:
Construct the Ferrum-Vryburg first 400kV line (2013);
Construct the Mercury-Vryburg first 400kV line (2013); and
Construct the Vryburg 400/132kV substation (2013).
3.1.2. Northern Cape Reinforcement Ferrum–Garona–Nieuwehoop and Sishen-Saldanha traction upgrade
This project will link with the North West Strengthening project and Ferrum with Aries; and will result in
significant upgrading of Garona and a new substation South-East of Upington, called Nieuwehoop. The
project will consist of:
125 MVA 400/132 kV transformer at Garona (2017);
Ferrum – Garona 400 kV line (2017);
Garona – Nieuwehoop 400 kV line (2017); and
Nieuwehoop – Aries 400 kV line (2017).
3.1.3 Namaqualand Upgrade
This project will significantly strengthen the North-West corner of the country and will result in the increased
capabilities of a substation that previously had a low feed-in capability (refer to Fig 4a). This project will
consist of:
Paulputs second 220/132kV 125MVA transformer (2015);
Aggeneis-Helios 400kV line (2017);
Aggeneis-Nama-Gromis-Oranjemund second 220kV line (2017); and
Aggeneis-Paulputs second 220kV line (2017).
3.4. Results of grid expansion projects
The result of the grid strengthening project will be that about 2000 - 3000 MW - significantly more power
that can currently be generated in the Northern Cape and transmitted to the large load centres [16]. Figure 6
shows the new CSP locations that become viable due to the planned Eskom expansions. The CSP potential
for the Free State does not change. The Northern Cape potential increases with 2625 km2 (49 GW) to
183 GW (261 GW for the North Zone). One site that is without fresh water supplies but located less than
30 km from the ocean was included. Sea water can be utilised for the condenser and may be desalted for
make-up and mirror-cleaning purposes. There is also the possibility that off-shore natural gas may become
available on the West coast in the future [17].
Fig. 6. Viable CSP locations after Eskom grid expansions.
4. Long-term scenario (2026 and beyond)
4.1. Waterless CSP
Most CSP-plants currently rely on the availability of water, which is used mainly for evaporative cooling.
Water is used in the steam cycle as make-up water (8 %); for wet cooling (90 %); and for cleaning mirrors
(2 %) [18]. If dry (air-cooled) condensers are employed, the water consumption is reduced to about 10 % of
the water needs of a wet-cooled plant [19].
Only one CSP-plant has been fitted with a dry air-cooled condenser up to date, namely the 1.4 MW linear
Fresnel PE1 plant in Spain [20]. Developers of 33 % of the 5 934 MW of CSP-plant that are under
construction worldwide have indicated that they will employ air-cooled condensers [21]. Air-cooled
condensers reduce the annual efficiency of a CPS-plant by about 5 % [19]. In the warm, arid areas of South
Africa, with high DNI levels, this is not an attractive option, but may still be employed. An alternative is
hybrid cooling, which functions primarily as an air-cooled condenser but uses water during midday periods.
Nearly waterless mirror cleaning has been employed on the 1.4 MW PE1-plant. A “robot” that is placed on
the mirrors operates with a brush to clean the surfaces [22].
The next generation CSP-plants cycles may be waterless. If a gas turbine (Brayton cycle) is employed rather
than a steam turbine (Rankine cycle), the water required for both the make-up water and the condenser
becomes absolute. A 100 kW demonstration has been commission in Israel in 2009 [23]. If a waterless
mirror-cleaning device can be employed, the water requirement for this type of CSP-plant becomes virtually
zero.
4.2. Locations with high DNI
The possibility of a waterless CSP-plant will unlock areas that were previously not viable for CSP-plants,
depicted in figure 7.
Fig. 7. Viable CSP locations for waterless CSP-plants.
Large parts of the country that currently have a low population density; are viable for waterless CSP-plants. It
has been noted that shale gas could in the future be present [24] in areas that overlap with viable CSP
locations indicated in figure 7 [25]. If the gas may be produced in an environmentally-accepted method, it
will increase the viability of this area for CSP-plants and will guarantee base load capabilities for the long-
term replacement of coal.
One possible constraint to the area identified is the implementation of a Radio Frequency Interference (RFI)-
free reserve that was promulgated in 2007 in terms of the Astronomy Geographic Advantage Act [26]. This
reserve requires low levels of RFI for the KAT7 radio telescope and the construction of the MeerKAT radio
telescope and the bid for a larger similar project, the Square Kilometer Array (SKA).
In addition to the area identified in figure 7, a large portion of southern and central Namibia as well as large
parts of southern Botswana would also be viable using the same criteria. This may result in the export of
large quantities of electricity from these countries to other African countries, including South Africa.
4.3. Strategic grid planning
The area identified in figure 7 will require new grid infrastructure to export the electricity to the load centres
that are mainly located in the northern parts of the country. If a viable thermal storage technology can be
incorporated with the waterless CSP-technology, the area identified could potentially power the entire
Southern Africa.
5. Conclusion and Recommendations
One of the large technical limits for the implementation of CSP plants in South Africa is the weak
transmission grid in the areas with high DNI values. The Eskom grid expansion project that was discussed in
this paper deserves prioritisation in order to increase the amount of power that can be generated and fed into
the grid in the viable CSP areas.
The immediate implementation of CSP under REFIT Phase 1 is an important field that deserves government
support. CSP with storage should be encouraged in order to further develop these technologies and to
demonstrate that it is able to supply energy efficiently during the evening peak demand. Air-cooled and
hybrid condensers should be encouraged in order to minimise the consumption of fresh water; a commodity
that is under substantial pressure in South Africa.
Waterless CSP-technologies should be a research focus and strategic grid planning should be made in order
to expand the grid to areas that will become viable once this technology is commercially available.
Acknowledgements
The authors would like to thank the Solar Research Energy Research Group (STERG) at Stellenbosch
University, which funded all costs involved in attending this conference through the Hope Project, as well as
Mr Crescent Mushwana of Eskom grid planning department who supplied grid capacity information.
References
[1] South African Renewable Energy Feed-in Tariff (REFIT) – Regulatory Guidelines 26 March 2009,
obtained from the Nersa website, accessed April 2009: www.nersa.gov.za.
[2] National Energy Regulator, in the matter regarding Renewable Feed-in Tariffs Phase II, 29 October 2009,
obtained from the Nersa website, accessed November 2009: www.nersa.gov.za.
[3] Eskom Annual Report 2010, available online, accessed June 2011: www.eskom.co.za/c/84/annual-report.
[4] Measured DNI data comparison based on Baseline Surface Radiation Network obtained from the World
Radiation Monitoring Center (WRMC), accessed April 2011: www.bsrn.awi.de/en/home/.
[5] Satellite data and maps obtained from GeoModel Solar, www.solargis.info. Personal communication with
Dr. Marcel Suri, Director of GeoModel Solar, March 2011. E-mails/ notes in possession of corresponding
author.
[6] SWERA, (2008). Solar irradiation data sets for the use with geographic information systems. Solar and
Wind Energy Resource Assessment. Available online, accessed Aug 2010: http://swera.unep.net.
[7] Guide to Independent Power Producer (IPP) processes, available online, accessed July 2011:
www.eskom.co.za/c/73/ipp-processes/.
[8] Information obtained from the websites of the following EIA consultants: SavannahSA and Bohlweki, as
well as personal communication with Mr Tinashe Mugabe of Afri-Devo, July 2011. E-mails/notes in
possession of corresponding author.
[9] Fluri, TP, (2009), The potential of concentrating solar power in South Africa, Energy Policy 37 (2009)
p5075–5080.
[10] Ummel, Kevin, (2010), Global Prospects for Utility-Scale Solar Power: Towards Spatially Explicit
Modeling of Renewable Energy Systems, available online, accessed July 2011:
www.cgdev.org/content/publications/detail/1424669.
[11] Eskom, private communication with Mmbengi Makungo of Eskom, December 2008.
[12] River flow data obtained from the Department of Water Affairs, accessed May 2011: www.dwa.gov.za.
[13] Land requirements based on the Andasol 3 CSP plant in Spain adapted to South Africa DNI conditions
by the authors.
[14] Eskom Capacity Assessment of the 2012 Transmission Network (GCCA-2012), available online,
accessed January 2011:
http://simbalism.co.za/upload/Eskom_Generation_Connection_Capacity_Assessment_for_the_2012_Ne.pdf.
[15] Eskom Transmission Ten-Year Development Plan – Public Version, available online, accessed May
2011: www.eskom.co.za/content/TDP_v6_lowres%5B2%5D.pdf.
[16] Personal communication with Mr Crescent Mushwana of Eskom grid planning department, July 2011.
E-mails/ notes in possession of corresponding author.
[17] “BHP, PetroSA, Sasol awarded West Coast exploration rights“, article, Engineering News, 21
September 2010, available online, accessed July 2011: www.engineeringnews.co.za/article/bhp-billiton-
petrosa-sasol-awarded-west-coast-exploration-right-2010-09-21-1.
[18] CSP water consumption data obtained from NREL, available online, accessed May 2011:
www.nrel.gov/csp/troughnet/power_plant_systems.html.
[19] Concentrating Solar Power Commercial Application Study: Reducing Water Consumption of
Concentrating Solar Power Electricity Generation - Report to Congress , U.S. Department of Energy,
available online, accessed July 2011: www.eere.energy.gov/solar/pdfs/csp_water_study.pdf
[20] NREL list of CSP plants in operation, available online, accessed July 2011:
www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=6.
[21] NREL list of CSP plants under construction, available online, accessed July 2011:
www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=16.
[22] Information obtained on PE1 from Novatec Solar, available online, accessed July 2011:
www.novatecsolar.com.
[23] Information obtained from Aora-Solar, available online, accessed: July 2011: www.aora-solar.com
[24] “Shale gas could play key role in clean energy future – Shell“ – Engineering News, 10 March 2011,
available online, accessed August 2011:
www.engineeringnews.co.za/article/shale-gas-could-play-key-role-in-clean-energy-future-shell-2011-03-10
[25] Map obtained online, accessed August 2011: http://kalahari.iblog.co.za/2011/03/15/
[26] Act No. 21 of 2007: Astronomy Geographic Advantage Act, 2007, as published in the Government
Gazette, Volume 516, Cape Town, 17 June 2008, No. 31157, available online, accessed June 2011:
www.dst.gov.za/publications-policies/legislation/Astronomy%20Geographic%20Advantage%20Act.pdf.