Least-cost electricity mix for South Africa by 2040
Scenarios for South Africa’s future electricity mix
CSIR Energy Centre
Cape Town, 3 November 2016
Jarrad Wright
Dr Tobias Bischof-Niemz
Joanne Calitz
Crescent Mushwana
Dr Tobias Bischof-Niemz
Chief Engineer
Background
The Integrated Resource Plan (IRP) is the expansion plan for the South African power system
In its most recent version, the IRP 2010 plans a doubling of power-generation capacity from 2010 to 2030
Since the date of its release in early 2011, two main assumptions have changed
• The demand forecast is now significantly lower than in IRP 2010
• The costs of solar PV and wind are significantly lower than predicted in IRP 2010
The CSIR has therefore conducted a study to re-optimise the South African power mix until 2040
Two scenarios were defined to quantify two different ways of expanding the South African power system
• “Business-as-Usual” – generally aligned with IRP 2010, updated demand forecast, no new optimisation
• “Re-Optimised” – least-cost re-optimisation of the demand/supply gap that widens from 2020-2040
An hourly expansion and dispatch model (incl. unit commitment) using PLEXOS
was run for both scenarios to test for adequacy and for economic feasibility
2
Agenda
Background
Approach and assumptions
Results
Conclusions
3
IRP 2010: expansion plan for South Africa’s power system until 2030
Installed capacity and electricity supplied from 2010 to 2030 as planned in the IRP 2010
Business-as-Usual
Re-Optimised
Electricity supplied
in TWh per year
Total installed
net capacity in GW
500
100
85.7
8.4
80
9.2
9.6
+1.8
60
42.2
40
2.1
2.4 1.8
20
35.9
400
1.2
4.8
300
7.3
2.4
200
41.1
100
0
2010
2015
2020
2025
2030
Renewables
Solar PV
CSP
4
Wind
Hydro
Nuclear
Peaking
Gas
Coal
436
Carbon free
CO2 emissions
[Mt/yr]
0
2010
5%
(12 TWh/yr)
2015
2020
2025
2030
14%
(62 TWh/yr)
10%
34%
(25 TWh/yr)
(149 TWh/yr)
237
275
Note: Renewables include solar PV, CSP, wind, biomass, biogas, landfill and hydro (includes imports); CO2 emission intensity moves from 912 kgCO2/MWh (2010) to 600 kgCO2/MWh (2030)
Sources: DoE IRP 2010-2030; CSIR Energy Centre analysis
Link between planning and real world needs to be established
In-principle process of IRP planning and implementation
Currently, no feedback loop from
procurement results to IRP planning
assumptions institutionalised
Planning /
simulation
world
Inputs
•
•
•
•
Demand forecast
Technology costs
assumptions
CO2 limits
Etc.
Inputs
Actuals /
real world
5
Sources: CSIR Energy Centre analysis
•
Ministerial
Determinations based
on capacity expansion
plan
Output
IRP model
•
Capacity expansion
plan
Installed capacity
Total installed
net capacity in GW
(least-cost optimisation)
100
85.7
80
8.4
1.2
9.2
60
9.6
+1.8
4.8
7.3
2.4
42.2
40
20
2.4
1.8 2.1
41.1
35.9
0
2010
Procurement
(competitive tender
e.g. REIPPPP, coal IPPPP)
2015
2020
2025
Outcomes
•
•
•
Preferred bidders
MW allocation
Technology costs
actuals (Ø tariffs)
2030
Solar PV
CSP
Wind
Hydro
Nuclear
Peaking
Gas
Coal
Actual solar PV tariffs now well below cost assumptions of IRP 2010
First four bid windows’ results (solar PV) of Department of Energy’s REIPPPP
Tariff in R/kWh
(Apr-2016-Rand)
4.0
3.65
Assumptions: IRP2010 - high
Assumptions: IRP2010 - low
Actuals: REIPPPP (BW1-4)
3.5
∑ = 2.8 GW
3.0
2.5
2.18
2.0
1.5
1.17
1.0
0.91
0.5
0.62
0.0
2010
2012
BW1
6
2014
BW 4 (Expedited)
2016
2018
2020
2022
2024
2026
2028
2030
Year
Notes: REIPPPP = Renewable Energy Independant Power Producer Programme; BW = Bid Window; bid submissions for the different BWs: BW1 = Nov 2011; BW2 = Mar 2012; BW 3 = Aug 2013;
BW 4 = Aug 2014; BW 4 (Expedited) = Nov 2015 Sources: StatsSA for CPI; IRP 2010; South African Department of Energy (DoE); DoE IPP Office; CSIR analysis
Actual wind tariffs equally well below cost assumptions of IRP 2010
First four bid windows’ results (wind) of Department of Energy’s REIPPPP
Tariff in R/kWh
(Apr-2016-Rand)
2.0
Assumptions: IRP2010
Actuals: REIPPPP (BW1-4)
1.52
∑ = 4.0 GW
1.5
1.19
1.0
0.87
0.69
0.5
0.62
0.0
2010
2012
BW1
7
2014
BW 4 (Expedited)
2016
2018
2020
2022
2024
2026
2028
2030
Year
Notes: REIPPPP = Renewable Energy Independant Power Producer Programme; BW = Bid Window; bid submissions for the different BWs: BW1 = Nov 2011; BW2 = Mar 2012; BW 3 = Aug 2013;
BW 4 = Aug 2014; BW 4 (Expedited) = Nov 2015 Sources: StatsSA for CPI; IRP 2010; South African Department of Energy (DoE); DoE IPP Office; CSIR analysis
Agenda
Background
Approach and assumptions
Results
Conclusions
8
Demand grows, existing fleet phases out – gap needs to be filled
Forecasted supply and demand balance for the South African electricity system from 2016 to 2040
Electricity
in TWh/yr
500
4732
450
400
All power plants considered for
Supply gap
“existing fleet” that are either:
Solar PV
1) Existing in 2016
2) Under construction
CSP
3) Procured (preferred bidder)
Wind
Other RE
Peaking
Other (incl. cogen)
Gas (CCGT)
Hydro (incl. PS)
Nuclear
Coal
3661
350
300
288
250
200
150
100
50
0
2016
9
Decommissioning of
Eskom’s coal fleet
2020
2025
2030
2035
2040
Notes: MTSAO demand forecasts are extrapolated from 2025 to 2040 using CAGR; IRP 2016 under development is using High Growth Low Intensity (CSIR) demand forecast as base case.
1. Peak demand = 53.2 GW 2. Peak demand = 68.7 GW Sources: DoE (IRP 2010); DoE (IRP 2013); Eskom MTSAO 2016-2021; StatsSA; World Bank; CSIR analysis
Two scenarios defined to fill the supply/demand gap until 2040
Forecasted supply and demand balance for the South African electricity system from 2016 to 2040
1
Electricity
in TWh/yr
500
4732
IRP 2010
450
400
3661
300
Supply gap
Solar PV
CSP
Wind
2 Scenario: “Re-Optimised”
Other RE
• Coal, nuclear, gas, RE are all
available as supply options
Peaking
Other (incl. cogen) • Supply candidates chosen
by least cost optimisation
Gas (CCGT)
to meet energy and
Hydro (incl. PS)
capacity requirement
Nuclear
Coal
1
350
2
288
250
200
150
100
50
0
2016
10
2020
2025
2030
2035
Scenario: “Business-as-Usual”
• Generally aligned with IRP
2010, but demand shifted
• Nuclear as per briefing to
Portfolio Committee on
Energy (11 October 2016)
• New coal, nuclear, some RE
• New capacities fixed as per
IRP 2010 (no optimisation)
2040
Notes: MTSAO demand forecasts are extrapolated from 2025 to 2040 using CAGR; IRP 2016 under development is using High Growth Low Intensity (CSIR) demand forecast as base case.
1. Peak demand = 53.2 GW 2. Peak demand = 68.7 GW Sources: DoE (IRP 2010); DoE (IRP 2013); Eskom MTSAO 2016-2021; StatsSA; World Bank; CSIR analysis
Key assumptions: pessimistic regarding solar PV and wind cost,
optimistic regarding nuclear cost
Technology
Costing Logic
Compared to IRP 2010
Solar PV
Same as IRP 2010 by 2030
Slightly lower until 2030
Wind
Bid Window 4 Expedited tariff Lower
kept constant until 2040
CSP
Same as IRP 2013
Slightly higher
Coal
Coal IPP
Higher
Nuclear
as per IRP with Rosatom lowestimate CAPEX
Similar
Gas
as per IRP with fuel updates
Higher
All other assumptions and methodology fully aligned with IRP 2010, for example:
• Discount rate of 8% (real)
• PLEXOS software package used for long-term optimisation & production cost modelling
• Decommissioning schedule of existing Eskom fleet
• Demand forecast using MTSAO 2016-2021 (extrapolated until 2040),
reaches the IRP 2010 assumed 2030 level just before 2040
11
Key input cost assumptions for new supply technologies
Lifetime cost
per energy unit1
R/kWh
(Apr-2016-R)
Bid Window 1
Actual new-build
tariffs
Assumptions based
new-build cost
3.10
High-priced gas
at 150 R/GJ
1.51
Bid Window 1
1.03
0.62
0.62
Solar PV
Wind
Baseload
Coal (IPP)
Typical capacity factor2
1
1.05-1.16
1.17
Baseload
Coal (Eskom)
Nuclear
82%
92%
1.24
Gas (CCGT) Mid-merit Coal Gas (OCGT)
50%
50%
10%
Diesel (OCGT)
10%
Lifetime cost per energy unit is only presented for brevity. The model inherently includes the specific cost structures of each technology i.e. capex, Fixed O&M, variable O&M, fuel costs etc.
Changing full-load hours for conventional new-build options drastically changes the fixed cost components per kWh (lower full-load hours higher capital costs and fixed O&M costs per
kWh);
Assumptions: Average efficiency for CCGT = 55%, OCGT = 35%; nuclear = 33%; IRP costs from Jan-2012 escalated to May-2016 with CPI; assumed EPC CAPEX inflated by 10% to convert EPC/LCOE
into tariff; Sources: IRP 2013 Update; Doe IPP Office; StatsSA for CPI; Eskom financial reports for coal/diesel fuel cost; EE Publishers for Medupi/Kusile; Rosatom for nuclear capex; CSIR analysis
2
12
2.40
Future cost assumptions for solar PV aligned with IRP 2010
Tariff in R/kWh
(Apr-2016-Rand)
4.0
3.65
Assumptions: IRP2010 - high
Assumptions: IRP2010 - low
Assumptions for this study
Actuals: REIPPPP (BW1-4)
3.5
∑ = 2.8 GW
3.0
2.5
2.18
2.0
1.5
1.17
0.91
1.0
0.49
0.5
0.49
0.62
0.0
2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040
BW1
13
BW 4 (Expedited)
Year
Notes: REIPPPP = Renewable Energy Independant Power Producer Programme; BW = Bid Window; bid submissions for the different BWs: BW1 = Nov 2011; BW2 = Mar 2012; BW 3 = Aug 2013;
BW 4 = Aug 2014; BW 4 (Expedited) = Nov 2015 Sources: StatsSA for CPI; IRP 2010; South African Department of Energy (DoE); DoE IPP Office; CSIR analysis
Future cost assumptions for wind aligned with results of Bid Window 4
Tariff in R/kWh
(Apr-2016-Rand)
2.0
1.52
∑ = 4.0 GW
1.5
Assumptions: IRP2010
Assumptions for this study
Actuals: REIPPPP (BW1-4)
1.19
1.0
0.87
0.5
0.69
0.62
0.0
2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040
BW1
14
BW 4 (Expedited)
Year
Notes: REIPPPP = Renewable Energy Independant Power Producer Programme; BW = Bid Window; bid submissions for the different BWs: BW1 = Nov 2011; BW2 = Mar 2012; BW 3 = Aug 2013;
BW 4 = Aug 2014; BW 4 (Expedited) = Nov 2015 Sources: StatsSA for CPI; IRP 2010; South African Department of Energy (DoE); DoE IPP Office; CSIR analysis
CO2 emissions constrained by RSA’s Peak-Plateau-Decline objective
PPD that constrains CO2 emission from electricity sector
CO2 emissions
(electricity sector)
[Mt/yr]
300
275
275
250
250
200
150
100
50
0
15
2010
2015
PPD = Peak Plateau Decline
Sources: DoE (IRP 2010-2030 Update); StatsSA; CSIR own analysis
2020
2025
2030
2035
2040
Agenda
Background
Approach and assumptions
Results
Conclusions
16
Least-cost: 70% RE energy in South African electricity sector by 2040
Comparison of energy supply for Business-as-Usual and a Re-Optimised scenario
1 Business-as-Usual
2 Re-Optimised
Electricity supplied
in TWh per year
Electricity supplied
in TWh per year
472 TWh
500
450
450
361 TWh
Solar PV
366 TWh
400
CSP
350
350
Wind
300
300
Other RE
250
250
200
200
150
150
100
100
50
50
400
0
2016
7%
(18 TWh/yr)
17
474 TWh
500
217 Mt/yr
Sources: CSIR analysis
2020
2025
2030
18%
(64 TWh/yr)
2035
0
2016
2040
19%
(91 TWh/yr)
Renewables
7%
(18 TWh/yr)
250 Mt/yr
CO2
217 Mt/yr
Peaking
Other
Gas (CCGT)
Hydro & PS
Nuclear
Coal
2020
2025
2030
2035
40%
(146 TWh/yr)
2040
71%
(332 TWh/yr)
100 Mt/yr
Significant solar PV and wind capacities rolled out until 2040
Comparison of generation capacity for Business-as-Usual and a Re-Optimised path to 2040
1 Business-as-Usual
2 Re-Optimised
Total installed
net capacity in GW
Total installed
net capacity in GW
200
200
175
175
150
150
125
125
101
11
1
12
93
100
75
53
63
50 1
2
25
40
80
73
7
8
2 6 2
8
8
13
42
8
13
37
36
184
151
1
113
18
100
83
2040 Peak: 68.7 GW 75
53
2030 Peak: 53.2 GW 50 1
2
25
40
0
77
1
32
63
18
2
2
27
2
10 5
17 2
35
0
2016
18
Sources: CSIR analysis
2020
2025
2030
2035
2040
43
2016
2020
2025
2030
Solar PV
Wind
Peaking
Gas (CCGT)
Nuclear
CSP
Other RE
Other (incl. cogen)
Hydro (incl. PS)
Coal
2035
2040
68.7 GW
53.2 GW
1
Business-as-Usual: Coal and nuclear dominate the 2040 energy mix
Exemplary Week under Business-as-Usual in 2040
Demand and
Supply in GW
60
40
20
0
Monday
Tuesday
Wednesday
Thursday
Solar PV
CSP
Hydro
Other RE
Coal
Wind
Peaking
Gas (CCGT)
Other (incl. cogen)
Nuclear
19
Sources: CSIR analysis
Friday
Saturday
Demand
Sunday
2
Re-Optimised: Wind and solar PV dominate the 2040 energy mix
Exemplary Week under Re-Optimised in 2040
Demand and
Supply in GW
60
40
20
0
Monday
Tuesday
Wednesday
Thursday
Solar PV
CSP
Hydro
Other RE
Coal
Wind
Peaking
Gas (CCGT)
Other (incl. cogen)
Nuclear
20
Sources: CSIR analysis
Friday
Saturday
Demand
Sunday
Re-Optimised scenario creates a steady, significant & increasing market
Roadmap of investment for wind and solar PV to 2040
1 Business-as-Usual
2 Re-Optimised
2030-2040
Wind: 4.5 GW/yr
120
Solar PV: 2.5 GW/yr
2030-2040
2020-2030
Wind: 0.4 GW/yr
2020-2030
Solar PV: 0.4 GW/yr
Wind: 2.8 GW/yr
Wind: 0.4 GW/yr
Solar PV: 1.5 GW/yr
77
49
Solar PV: 0.4 GW/yr
32
24
2
2016
21
BW1
7
12
11
2
2020
2030
2040
2016
BW 4 (Expedited)
Sources: CSIR analysis
43
15
8
7
BW1
7
18
2020
2030
BW 4 (Expedited)
2040
Re-Optimised R87 billion/year cheaper by 2040 (without cost of CO2)
Total cost of
power generation in
bR/yr (constant 2016)
500
450
400
350
488
1
2
Business-as-Usual
Re-optimised
Delta (BAU - Re-optimised)
401
314
300
250
292
Total Present Value of Delta
= R330 billion in 2016 Rand
200
150
87
100
50
22
0
2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040
22
Business-as-Usual incurs large cost from building new coal and nuclear
Comparison of total electricity system costs average electricity tariff of BAU and Re-Optimised mix
1 Business-as-Usual
Total system cost (real)
(Apr-2016 Rands) in bR/yr
2 Re-Optimised
517 (w/ CO2)
488 (w/o CO2)
550
Total system cost (real)
(Apr-2016 Rands) in bR/yr
500
500
450
450
400
350
300
300
250
250
200
200
150
150
100
100
50
50
23
Sources: CSIR
2020
2025
2030
401 (w/o CO2)
400
314 (w/o CO2)
350
0
2016
412 (w/ CO2)
550
2035
2040
292 (w/o CO2)
0
2016
2020
2025
2030
Cost of CO2
CSP
Other RE
Other (incl. cogen)
Hydro (incl. PS)
Nuclear (existing)
Solar PV
Wind
Peaking
Gas (CCGT)
Nuclear (new)
Coal (new)
2035
2040
Coal (existing)
Sensitivity on cost difference: Even if RE were 50% more expensive
than assumed, Re-Optimised is still cheaper than Business-as-Usual
Annual cost
delta of
BAU – Re-Optimised
by 2040
in bR/yr
1.0
0.9
-40
-20
0
20
40
0.8
60
80
0.7
100
Relative
RE/nuclear cost
120
140
0.6
160
Today (2016)
180
0.5
200
Actual (2020-2040)
220
0.4
240
260
Relative
RE/coal cost
0.3
24
0.3
Sources: CSIR analysis
0.4
0.5
0.6
0.7
0.8
0.9
1.0
280
87
Unit cost of power generation:
Re-Optimised case is almost 20 cents/kWh cheaper than BAU by 2040
Average cost of
power generation in
R/kWh (constant 2016)
1.1
1
2
Business-as-Usual
Re-Optimised
1.03
1.0
0.9
0.87
-18%
0.85
0.8
0.7
0.80
0.6
0.5
0.4
0.3
0.2
0.1
0.0
25
2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040
Factoring in cost of CO2 emissions:
Re-Optimised case is 23 cents/kWh cheaper than BAU by 2040
Average cost of
power generation in
R/kWh (constant 2016)
1.1
1.0
1
2
Business-as-Usual
Re-Optimised
0.94
-21%
0.87
0.9
0.8
1.10
0.86
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
26
2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040
Re-Optimised: CO2 emissions and water use significantly lower
Comparison of CO2 emissions and water use for BAU and a Re-Optimised scenario to 2040
Electricity sector
CO2 emissions
in MtCO2/yr
300
250
Electricity sector
water use
in billion litres/yr
275
70
275
61
250
60
231
200
216
195
56
49
50
-150 Mt/yr
(-60%)
55
-40 billion litres/yr
(-60%)
40
150
30
100
20
99
50
10
0
2015 2020 2025
27
0
2030 2035 2040
BAU
Sources: CSIR analysis
22
2015 2020 2025
Re-optimised
CO2 Cap
2030 2035 2040
Agenda
Background
Approach and assumptions
Results
Conclusions
28
South Africa can get 70% renewable energy share by 2040 at least cost
Solar PV, wind and natural gas is the cheapest new-build mix for the South African power system
It is the cost-optimal expansion to aim for a 70% renewable energy share by 2040
This “Re-Optimised” mix is almost R90 billion per year cheaper by 2040 than the Business-as-Usual
scenario (without factoring in cost of CO2 emissions – difference is > R100 billion per year with CO2)
The Re-Optimised mix will furthermore reduce South Africa’s CO2 emissions by 60% compared to BAU
Avoiding CO2 emissions and least-cost is not a trade-off anymore – South Africa can de-carbonise its
electricity sector at negative carbon-avoidance cost
Building out the required capacities until 2040 will provide a steady anchor offtake for a South African
solar PV and wind manufacturing industry
29
Ha Khensa
Re a leboha
Enkosi
Siyathokoza
Thank you
Re a leboga
Ro livhuha
Siyabonga
32
Dankie