Scholarly article on topic 'Scenarios for a South African CSP Peaking System in the Short Term'

Scenarios for a South African CSP Peaking System in the Short Term Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
Keywords
{"Concentrating Solar Power (CSP)" / "Peaking CSP syst ems" / "Levelized Cost s of Energy (LCOE)"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — C. Silinga, P. Gauché

Abstract The South African Integrated Resource Plan is a policy document, which by law allocates the energy resources that will be built to meet the future electricity needs of South Africa. The current Integrated Resource Plan indicates the electricity generation types that will be built from 2010 to 2030. It states that most of the future peak load will be met by Open Cycle Gas Turbines which operate using diesel and represents an allocation of 4,930M W. Further, the Integrated Resource Plan does not identify CSP as a potential peaking solution and allocates 1,200M W of capacity to CSP. This represents less than 2% of total capacity in 2030. This paper investigates the feasibility of utilizing CSP Plants as peaking plants in the short to medium term based on a proposition that under certain scenarios, a fleet of unsubsidized CSP peaking plants could drop the LCOE of the current Integrated Resource Plan. This is done by modeling a contemporary CSP tower system with Thermal Energy Storage. The Gemasolar CSP plant is used as the reference plant in order to obtain operating parameters. Our analysis suggests that at current fuels costs, diesel powered Open Cycle Gas Turbines produce electricity in excess of 5.08 ZAR/kWh (∼0.63 US$/kWh), significantly above current CSP energy generating costs. This is the context that informed the undertaking of this study, to influence policy and provide technical evidence that CSP can guarantee and deliver energy at competitive costs in the short term. Two alternate scenarios show a lower LCOE for providing peak power. The most promising is a combined distributed CSP system wit h diesel powered Open Cycle Gas Turbine system as backup. The LCOE for this system is 2.78 ZAR (∼0.34 $/kWh) or a drop of 45% when no fuel price inflation is considered. This system also increases security of supply due to a lower dependence on fuel prices.

Academic research paper on topic "Scenarios for a South African CSP Peaking System in the Short Term"

Procedía

Scenarios for a South African CSP peaking system in the short term

C.Silinga3*, P. Gauche3

aSolar Thermal Energy Research Group, Dept of Mechanical and Mechatronic Engineering, Stellenbosch University, Private BagXl Matieland,

7602, South Africa

Abstract

The South African Integrated Resource Plan is a policy document, which by law allocates the energy resources that will be built to meet the future electricity needs of South Africa. The current Integrated Resource Plan indicates the electricity generation types that will be built from 2010 to 2030. It states that most of the future peak load will be met by Open Cycle Gas Turbines which operate using diesel and represents an allocation of 4,930 MW. Further, the Integrated Resource Plan does not identify CSP as a potential peaking solution and allocates 1,200 MW of capacity to CSP. This represents less than 2% of total capacity in 2030.

This paper investigates the feasibility of utilizing CSP Plants as peaking plants in the short to medium term based on a proposition that under certain scenarios, a fleet of unsubsidized CSP peaking plants could drop the LCOE of the current Integrated Resource Plan. This is done by modeling a contemporary CSP tower system with Thermal Energy Storage. The Gemasolar CSP plant is used as the reference plant in order to obtain operating parameters.

Our analysis suggests that at current fuels costs, diesel powered Open Cycle Gas Turbines produce electricity in excess of 5.08 ZAR/kWh (~0.63 US$/kWh), significantly above current CSP energy generating costs. This is the context that informed the undertaking of this study, to influence policy and provide technical evidence that CSP can guarantee and deliver energy at competitive costs in the short term. Two alternate scenarios show a lower LCOE for providing peak power. The most promising is a combined distributed CSP system with diesel powered Open Cycle Gas Turbine system as backup. The LCOE for this system is 2.78 ZAR (~0.34 $/kWh) or a drop of 45% when no fuel price inflation is considered. This system also increases security of supply due to a lower dependence on fuel prices.

© 2013 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/3.0/).

Selection andpeerreviewbythescientific conferencecommittee ofSolarPACES2013 underresponsibilityof PSEAG. Final manuscript published as received without editorial corrections.

Keywords: Concentrating Solar Power (CSP); Peaking CSP systems; Levelized Costs of Energy (LCOE)

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Energy Procedia 49 (2014) 1543 - 1552

SolarPACES 2013

* Corresponding author. Tel.: +27-21-808-4016; fax: +27-21-808-4933 E-mail address: 16835719 @sun. ac.za

1876-6102 © 2013 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/Kcenses/by-nc-nd/3.0/).

Selection and peer review by the scientific conference committee of SolarPACES 2013 under responsibility of PSE AG. Final manuscript published as received without editorial corrections. doi: 10.1016/j.egypro.2014.03.163

1. Introduction

South African (SA) electricity demand is characterized by a mid-morning and evening peak period. These are the periods when energy demand is significantly higher than baseload. The morning peak period typically occurs between 07:00 - 10:00. The evening peak typically occurs between 18:00 - 20:00. In order to meet the peak period energy demand, flexible energy systems with quick start-up time are needed. During the peak periods, any energy system that is able to supply energy at competitive prices is valuable.

Currently, the SA energy utility employs Open Cycle Gas Turbine (OCGT) systems and pumped storage systems for peaking electricity [1]. There is 1,400 MW of installed pumped storage capacity which contributed 1.2% to the final energy supply in 2012 [2]. There is a limited potential for pumped storage in SA due to the dry arid climate. There is 2,426 MW of installed OCGT capacity in SA, all of which are run on diesel fuel and contributed 0.2°% to the final energy supply in 2012 [2]. This is relatively small contribution to the energy supply. However, the current plans in the energy industry are to increase the installed capacity of the OCGT systems relative to increasing energy demand [1]. The IRP allocates 4,930 MW of capacity to OCGT to 2030 [1].

Coal currently accounts for 90°% of the final electricity generation in SA [2]. The IRP identifies the need to diversify the energy supply and adopt energy systems that emit less greenhouse gasses. The IRP allocates less capacity to coal and more capacity to a nuclear fleet and renewable energy systems. The IRP does not identify Concentrating Solar Power (CSP) systems as potential peaking solutions.

The CSP capacity allocation in the IRP is 1,200 MW until 2030 [1]. SA has one of the best solar resource in the world, with some areas reaching an annual DNI of 3,000 kWh/m2/yr [3]. Also, recent studies prove that SA has adequate suitable land for a CSP systems rollout [3]. A CSP system with the thermal energy storage (TES) provides dispatchable electricity.

This paper investigates the feasibility of utilizing CSP system as peaking plants in the short to medium term based on a proposition that under certain scenarios, a fleet of unsubsidized CSP peaking plants could drop the LCOE of the current IRP. This is done by modeling a contemporary CSP tower system with TES, obtaining the LCOE and comparing it with the OCGT systems LCOE.

The following section of this study gives a broad overview of OCGT and CSP technology. Subsequent sections describe the methodology of this study, the model used in the study, three scenarios that are developed and finally results from these scenarios and conclusions.

Nomenclature

A aperture area of the solar field r discount rate

a, receiver emitting area T a ambient temperature

Et electricity generation in the year t T 1 H receiver outlet temperature

F, receiver view factor T r / ri mean temperature receiver

Ft fuel expenditures in the year t W work done

h heat transfer coefficient receiver Vth thermal efficiency

It investment expenditure in the year optical solar field optical efficiency

Mt O&M expenditures in the year t Zr receiver emissivity

n life of the system a Stefan-Boltzmann constant

Ç^in / out heat flow receiver a receiver absorptivity

Acronyms

CES Conventional Energy Systems OCGT Open Cycle Gas Turbine

CSP Concentrating Solar Power O & M Operation and management costs

DNI Direct Normal Irradiation RE Renewable Energy Systems

HTF Heat Transfer Fluid SA South Africa

IRP Integrated Resource Plan TES Thermal Energy Storage

LCOE Levelised Costs of Energy ZAR South African Rand

2. Technology overview

2.1. Open Cycle Gas Turbine

The OCGT is characterized by low capital costs and high O&M costs because it utilizes fossil fuel as the source of energy and rejects high quality waste heat. These systems can be installed with short lead times, are light and can be deployed in a modular way. These systems provide a very good option for dispatch due to their ability to be electronically controlled and brought online fast. The records show that they can be synchronized with the grid and brought online from start in about 10 minutes [4]. The main challenge about utilizing the OCGT systems is the vulnerability to fluctuating fuel prices, particularly in the South African context where we currently rely on diesel as the fuel.

The IRP states that the current proposed capacity on OCGT uptake assumes that these systems will be run on diesel because of the established infrastructure [1]. SA has estimated natural gas reserves of 513 tcf at the end of 2011, which is about 7% of the world's reserves [5]. The utilization of natural gas in the OCGT systems would potentially lead to lower LCOE. However, when looking at the current SA energy policies, the utilization of natural gas in OCGT is a long term goal. This study considers the current available source of energy for the OCGT which is diesel.

2.2. Concentrating Solar Power

The CSP system concentrates sunlight to achieve high temperature thermal energy. The CSP system that is modeled in this study is the CSP tower system as shown by the process diagram in Figure 1 and is a contemporary technology implemented in the Gemasolar plant and under construction by others. This system comprises of heliostats that reflect the sun rays onto a receiver on top of the tower. The receiver allows for solar flux thermal energy to be transferred to the heat transfer fluid (HTF). This thermal energy is then used to drive the steam turbine.

Currently, there is no commercial operational CSP plant in SA. There is a 150 kW pilot linear Fresnel system operated by BBEnergy [6]. There are three CSP systems under construction: KHI Solar (50 MW tower), KaXu Solar One (100 MW Trough) and Bokpoort (50 MW Trough) [3]. The IRP allocates more capacity to the Renewable Energy (RE) systems to 2030, with 17.8 GW allocated to RE, 9.6 to nuclear and 6.3 GW to coal [1].

The challenge with the RE uptake is in dealing with intermittency and the coincidence between the energy demand patterns and the solar resource availability. The CSP system with TES is a dispatchable source of electricity. It provides flexibility services to the grid. Large conventional energy systems (CES) are usually restricted to 50% - 100°% operating range of full capacity [7]. This may result in RE to curtail in periods of high RE resource. Also, the operation of CES avoids ramping of the systems to avoid high O&M costs [7]. The CSP systems overcome this challenge by providing ability to shift the generation times to time of high value. TES unlocks the true value of CSP and cost optimization by considering LCOE of the entire electric systemshould reflect this.

Figure 1. CSP tower system configuration

3. Methodology

The two main components of the method are; 1) technical modeling of plants and 2) the financial analysis of the energy system. This study only considers technical analysis of CSP and OCGT technologies. For the OCGT systems, the assumption is that the proposed capacity will be distributed along the points of high demand. Further the method assumes that the diesel fuel that is needed to run the OCGT is available on site, hence not factoring the diesel infrastructure costs in the financial model. The fuel costs of electricity generated from the OCGTs are then used as inputs to the financial model to determine the electricity generation costs of the OCGT system The CSP system comprises distributed CSP plants in 10 sites that are situated along the high voltage, high capacity line shown in Figure 2. The idea is that it would be more feasible in the short term to be close to the existing transmission system that requires less infrastructure investment. The proposed CSP system capacity from the proposed sites is optimized for the electricity generation. The electricity output from all the CSP systems is then used as the input in the financial model to determine the energy generation costs ofthe CSP system.

Levelized Cost of Energy (LCOE) is used to determine the energy generation costs of each system. Utilizing the LCOE to compare the different energy generation technologies is adequate because it allows for technology comparison based on the weighted average costs basis. The LCOE does not capture the daily fluctuations in demand and supply, which are seen as true value of energy [8]. However, the LCOE allows different technologies to be compared or equated base on average costs basis [8].

The definition of LCOE is shown by equation (1) and is commonly used in the electricity sector. It is adapted from the IRENA report on RE systems costs analysis [9]. The capital expenditure costs for OCGT were obtained from the Brinckerhoff report on OCGT costs analysis [10]. For OCGT the report assumes 4,746.30 ZAR/kW (~593.28 US$/kW). Fixed O&M costs are assumed to be 82.84 ZAR/kWh (~10.35 US$/kW) and variable O&M costs are 0.05 ZAR/kWh (~0.0065 US$/kWh). The diesel costs are based on the current value of 9.89 ZAR/kg with the energy value of 35 MJ/kg. The capital costs expenditure costs for the CSP system is obtained from the SANDIA report on CSP tower costs reduction plan [11], refer to Table 1. The LCOE model assumes 8% interest rate on the loan and 10%% discount rate. The predicted lifetime of the energy systems is 30 years. Table 1 shows the assumed capital costs ofthe CSP system[11].

™ It + Mt + Ft .Z-^tt=i

LCOE =

(1 + r )t

(1 + r )t

Table 1. CSP tower system capital expenditure costs estimates [11]

Item Value (ZAR) Unit Value (US $) Unit

Heliostat Field 1600 R/m2 200 $/m2

Receiver 1600 R/kWth 200 $/kWth

Thermal storage 240 R/kWhth 30 $/kWhth

Power block 8000 R/kWe 1000 $/kWe

Steam generation 2800 R/kWe 350 $/kWe

O&M 520 R/kWyr 65 $/kWyr

Figure 2. SA map with the proposed sites for the model

4. Model description

The CSP model in this study is a systematic model of the CSP tower system. The average hourly solar resource data is used as inputs to evaluate the plant performance. This type of modeling evaluates the plant performance by considering the optical to thermal energy conversion. The key inputs for the modeling purposes are: the DNI solar resource, the solar field configuration, ambient temperature, wind speed and the receiver operating temperatures. The model from Gauché [12] is adapted and used for this study. The model has been validated using the results from the eSolar Sierra tower plant in California and it matches the expected annual performance of the Gemasolar plant reasonably well [12]. The model aims to generally replicate the Gemasolar plant with the understanding that it is a real plant proving the ability to dispatch [13]. The model allows the turbine rating and storage reference hours to be modified.

4.1. Solar field optics

The first requirement of the plant model is the continuous determination of sun position. The solar time, which is based on the angular motion of the sun across the sky, is derived and it contains standard time, longitudinal corrections and the equation of time. The equation of time is derived by Spencer [12]. From the solar time, the hour angle which is the conversion of solar time into angle is derived. After that the zenith angle and the azimuth angles are derived. These angles provide the incidence ray to the heliostat module and the receiver atop of the tower provides the reflected incidence target. The implementation of the position of the sun as well as the remaining model description has been documented by Gauche [14] and is provided here in summary.

The circular-like heliostat field of the Gsmasolar plant reveals a very low dependence of the solar azimuth angle on the heliostat field efficiency. This makes it convenient to express the optical efficiency as a single polynomial for quicker and simpler analysis, an important consideration when running CSP models for scenario analysis. Equation (2) is thus only a function of zenith angle and is used for all plant models assuming that the heliostat field and tower remain unchanged.

Vop^ai = 0.42546>| -1.14805 + 0.3507^4 + 0.7556>| + 0.59186>| + 0.08166>Z + 0.832 (2)

4.2. Receiver losses

The energy balance on the receiver is performed to determine the energy that is transferred to the HTF and sent to the TES. Equations (3) and (4) are used to perform the energy balance of the receiver. The model utilizes a fixed output temperature of 565 °C for the receiver based on the operating temperature of the Gemasolar plant. The inlet and outlet temperature of the receiver are fixed and the radiation component is solved for this range[12].

q — DNI *n *n * A

z-'in I optical I other a (3)

(1 -a)* Q in = v*sr ¿A, * F * fa " T4 )+ h * Ar * (t; - Ta)+ Qmt (4)

4.3. TES

The thermal energy from the receiver is sent to the TES and it is delivered to the turbine's steam generator during the peak period. The commercially available TES show a round trip efficiency of 95%% or higher [7]. This gives an average loss of less than 5%/24 hours or 0.2%%/hour. The model for this study assumes a 90%% round trip efficiency for the TES. This gives an average loss of 10%/24 hours or 0.5%%/hour. The model is significantly more conservative than existing TES systems to account for operational contingencies and as a starting point for this study. The proposed storage capacities range from 1 043 MWh and 1 217 MWh both with 7 storage hours.

4.4. Power block

The power block indicates the performance of the steam turbine. In order to determine the performance of the steam turbine, a theoretical Chambadal-Novikov, modified Carnot efficiency is used. This is because no specific turbine is selected for this study. The high temperature reservoir is the hot salt temperature and the low temperature reservoir is the ambient temperature, assuming dry cooling is used. The following two equations, (5) and (6) are used to determine the work done and the efficiency of the heat engine.

W = vth *

Vth =1 -

5. Scenario modeling

The analysis in this study compares the LCOE of various system scenarios using the technical model described for CSP and an equally simplified model for OCGT plants. The models need to satisfy the hourly demand pattern based on the hourly solar and weather data for the selected sites. The 2010 national load demand from Eskom is used to assume the peak load [16]. Solar and weather data is also for 2010 and is satellite derived data known to be accurate for the Southern African region [15]. For this initial study, peak load is defined as 10% of the daily required electricity generation capacity. Everything else is assumed to be covered by baseload or mid merit power generation. The proposed capacity of the OCGT system and the CSP systems are modeled against the assumed peak load demand. Three scenarios are presented in this study and they are; the OCGT system energy supply, the CSP system with grid energy supply and the CSP system combined with the OCGT system

5.1. OCGT system energy supply scenario

This scenario predicts that the OCGT system supplies all the energy for the assumed peak load demand. A network of OCGT capacity from different locations constitutes the OCGT system. The operating parameters of the modeled OCGT system resemble the current operational systems in SA and these parameters are adapted from the Eskom [4] OCGT report. The system that is adapted for this study runs on diesel and has thermal efficiency of 35%

The energy generation cost analysis for the OCGT systems is done by using the LCOE equation (1). The equation requires the external inputs of investment expenditure, O&M costs, and the fuel expenditure and the electricity generation. These input costs are based on the Brickerhoff costs estimates for the OCGT systems are adapted for this study [10]. The total generated energy output from the OCGT system is used as input of the LCOE analysis.

The LCOE of the OCGT system in this scenario is 5.08 ZAR/kWh (~0.63 $/kWh) for the assumed energy demand of 7 5 87.44 TWh. The proposed capacity of the modeled OCGT system is 5 000 MW. The operational flexibility that is offered by these systems is overshadowed by the high LCOE costs. This justifies the reason of using them as peaking systems and operated over shot period in regions with fossil fuel shortages. However, it is crucial and necessary to explore more optimal utilization of these systems and further utilize them with other energy systems. The significant cost input of the OCGT is the fossil (diesel) fuel. Any fluctuations in fossil fuel prices affect the operation and result in LCOE increase. Figure 5 shows the LCOE fuel sensitivity study. The scenario 1 line represents the OCGT scenario. The current LCOE analysis is based on current conditions. However, by a small diesel increase compounded over the lifetime of the system the LCOE will significantly increase, as shown in Figure 5.

5.2. CSP system and Grid electricity supply scenario

CSP system capacity from all the sites as indicated in Figure 2 supply energy to meet the assumed peak load demand which is the same as the OCGT peak load. Figure 3 shows two days (January) of the operational configuration of the CSP system. The proposed capacity of the entire network of CSP is 3 300 MW. The model assumes that the CSP capacity is available homogenously and the cost analysis is based on the proposed capacity.

The LCOE investment expenditure costs of the CSP system are theoretical costs that are obtained from the Kolb report on costs estimates of CSP systems [11]. The Table 2 below shows the results from the proposed CSP system.

Figure 3. CSP electricity supply vs. demand example with 3 300 MW CSP capacity

CSP system capacity from all the sites as indicated in Figure 2 supply energy to meet the assumed peak load demand which is the same as the OCGT peak load. Figure 3 shows two days (January) of the operational configuration of the CSP system. The proposed capacity of the entire network of CSP is 3 300 MW. The model assumes that the CSP capacity is available homogenously and the cost analysis is based on the proposed capacity.

The LCOE investment expenditure costs of the CSP system are theoretical costs that are obtained from the Kolb report on costs estimates of CSP systems [11]. The table below shows the results from the proposed CSP system.

Table 2. CSP system financial model results

Item Value Unit

CSP capacity 3 300 MW

LCOE 1.89 ZAR/kWh

Fulfilment coefficient 0.82

Curtailment coefficient 0.06

The CSP system achieves an LCOE of 1.89 ZAR/kWh (~0.23 $/kWh) and has a fulfillment coefficient of 0.82. Figure 3 shows the operation of CSP system over two days. On the first day, the CSP fulfills the load demand. However, on the second day, the CSP does not fulfill all the load demand. This results in gap energy demand.

Energy from the national grid is used for the gap demand during the operation of the CSP system. This is done by charging the hot salt tank during periods of inadequate solar resource. The operation of charging the hot tank is based on proxy weather model that is assumed specifically for this study. When the proxy weather model predicts inadequate solar resource in the next day, the predicted amount of energy is drawn from the grid to charge the hot tank. Energy is bought from the grid during the off-peak periods when the tariffs are low and it is sent back to the grid during peak period. During the charging of TES, the following losses are considered: the modified Carnot efficiency losses; TES charge losses and the TES hourly thermal losses. Due to weather forecasting uncertainty, it is generally necessary to purchase more grid electricity that would be required.

The amount of energy that is bought from the grid is 808 371.07 MWh. It has a fulfillment coefficient of 0.58 of the total gap demand. The combination of the grid energy and the CSP generated energy has a fulfillment coefficient of 0.92 meaning that peaking needs are not fully guaranteed in this scenario.

The costs of the grid energy used to charge the TES are then factored into the LCOE calculation. The LCOE of the combined system increases from the initial 1.89 ZAR/kWh (0.23 $/kWh) to 3.00 ZAR/kWh (0.37 $/kWh).

A system increasing fulfillment to 1.0 would be more expensive yet but this has not been fully explored. Alternatively, a small fossil backup system could supply this gap.

5.3. CSP system and OCGT energy supply scenario

3.000 2.500 2.000 1.500 1.000 500 0

Eskom 2010 demand (MW)

12 16 20 24 28 32 36 40 44 48 Year hour

CSP Supply total

■ Demand gap total

Figure 4. CSP system and OCGT system operational configuration

This scenario assumes that the CSP capacity that is proposed earlier is installed. Further, the OCGT system capacity that is proposed in the IRP is also installed. The CSP system performs as described previously to deliver energy. The OCGT capacity is only used to deliver gap demand energy. Figure 4 shows the operation of the CSP and the OCGT system over a two day period (February). On the first day, the CSP system fulfills the demand. The CSP system does not fulfill the demand on the second day. The CSP system delivers energy at its full capacity whenever there is adequate solar thermal energy and the OCGT delivers the demand gap implying that the OCGT system delivers a smaller amount of energy relative to installed capacity.

ce 14,00

ä 12,00

LU O 10,00

U 8,00

Seena rio 3 OCGT :omponent

- Seenari o 1

i- - ~ ~ ~

Scenario 3

IE-- A " Scenari o 3 CSP com ponen

2% 4% 6%

Diesel costs annual increase

Figure 5. OCGT LCOE fuel sensitivity analysis

The OCGT system is only utilized to meet the gap energy demand that the CSP is unable to meet. Even though the installed capacity of the OCGT is able to deliver the assumed peak load demand but a small amount of it is utilized. The negative impact of the underutilization of the OCGT is the resultant high relative LCOE of the OCGT system. The scenario 3 OCGT component shows the fuel sensitivity when using the OCGT to supply gap demand. That results in higher LCOE from OCGT due to lower energy generation. The LCOE of the OCGT system alone that functions as the gap demand filler to the CSP system is 6.67 ZAR/kWh ~(0.83 $/kWh). This is 15% more than the scenario 1 LCOE. For the CSP, the LCOE is 1.89 ZAR/kWh (0.23 $/kWh).

The CSP LCOE is constant at 1.89 ZAR/kWh (~0.23 $/kWh). Coupling the CSP system with the OCGT system ensures that the system guarantees the peak load demand. The LCOE of the combined systems is 2.78 ZAR/kWh

(~0.34 $/kWh). The lower CSP LCOE lowers the OCGT LCOE to achieve this combined LCOE. Figure 5 shows the LCOE fuel sensitivity analysis. It is clear how the fluctuation in fuel prices affect the LCOE of the OCGT. On the other hand, the CSP LCOE stays constant because of low running costs. The main advantage of linking the two systems is that the CSP system cushions the high OCGT LCOE and achieves a lower combined LCOE.

6. Conclusion

Feasibility of peaking CSP systems in SA is conducted in this study. Three scenarios of the peaking systems in SA have been developed and presented. The OCGT system scenario show that the OCGT systems generate energy in excess of 5.08 ZAR/kWh (~0.63 $/kWh). Further, this scenario shows that the OCGT systems are vulnerable to fluctuating diesel prices. The CSP scenario established that the CSP systems generate the peak energy at 1.89 ZAR/kWh (~0.23 $/kWh). This is significantly lower cost than the currently used OCGT systems in SA. The utilization of grid electricity by CSP systems to supply the gap energy demand results in in LCOE increase of 37%. LCOE increases from 1.89 ZAR/kWh to 3.00 ZAR/kWh. This is still financially better than the OCGT LCOE. The scenario that combines the CSP system and the OCGT system results in a combined LCOE of 2.78 ZAR/kWh (~0.34 $/kWh). This scenario results in the lowest LCOE and it guarantees electricity generation.

The consequence of this is that SA does not need to gratuitously invest in CSP. A fleet of CSP plants optimized to be used with the OCGT fleet appear to drop the net cost of electricity while showing impressive resilience to fuel price fluctuations.

Acknowledgements

The authors would like to thank Centre for the Renewable and Sustainable Energy Studies (CRSES) and the Solar Thermal Research Group (STERG) at Stellenbosch University for funding the resources to perform this work and present it at SolarPACES. SolarGIS data © 2012 was generously supplied by GeoModel Solar s.r.o. making it possible to complete this project. The advice and support from Prof Wikus van Niekerk (CRSES). STERG researcher, Christina Auret is thanked for suggesting scenario 3 which is the scenario showing the ideal solution.

References

[1] IRP 2010. Integrated Resource Plan for electricity 2010 - 2030. Government Gazette, Republic of South Africa; 2011.

[2] Eskom. Shift performance, grow sustainably. Integrated Report; 2012.

[3] A. J. Meyer and J. L. Van Niekerk. Roadmap for the development of concentrating solar power in South Africa. 2012.

[4] Eskom Open Cycle Gas Turbine. Ankerlig and gourikwa gas turbine power stations. 2009.

[5] S. Moolman. Plans a foot to increase the role of gas in SA 's energy mix. Article; 2013.

[6] B. Gonzalez and C. Nell. BBenergy solar based energy project at Rosherville; 2013.

[7] P. Denholm and M. Mehos. Enabling Greater Penetration of Solar Power via the Use of CSP with Thermal Energy Storage; 2011.

[8] C. D. Kost, T. Schegl, J. Thomsen, S. Nold, and J. Mayer. Levelized Cost of Electricity Renewable Energies; 2012.

[9] IRENA. Renewable Energy Technologies: cost analysis series. International renewable Energy Agency; 2012

[10] P. Brinckerhoff. Cost estimates for Thermal peaking plant;2008

[11] G. J. Kolb, C. K. Ho, T. R. Mancini, and J. A. Gary. Power Tower Technology Roadmap and Cost Reduction Plan: SANDIA report. 2011.

[12] P. Gauche, S. Pfenninger, A. J. Meyer, and A. C. Brent. Modeling dispatchability potential of CSP in South Africa. SASEC 2012; May 2012

[13] J. I. Burgaleta. Gemasolar Thermosolar Plant. NREL Concentrating Solar Power Projects - Gemasolar Thermosolar; 2011

[14] P. Gauche, T. W. Von Backstrom, and A. C. Brent. Modeling Methodology for Macro Decision Making - Emphasis on the Central Receiver Type, SolarPACES. Granada, Spain; 2011

[15] GeoModel. SolarGISdata © 2012 GeoModel Solar s.r.o; 2010.

[16] Eskom, source unknown, 2012.