Medium-Term System Adequacy Outlook 2017 to 2021

Medium-term System Adequacy Outlook 2017 to 2021

31 July 2017

Contents

1. Overview ...... 3 2. Introduction ...... 3 3. Methodology and approach ...... 3 4. Assumptions ...... 5 4.1 Demand forecast ...... 5 4.2 Existing and committed supply resources ...... 7 4.2.1 Eskom installed capacity...... 7 4.2.2 New build commercial operation dates ...... 8 4.2.3 Non-Eskom capacity without the REIPPP...... 9 4.2.4 REIPPP...... 10 4.2.5 Demand-side management...... 10 4.3 Eskom plant performance ...... 10 5. Modelling scenarios ...... 12 6. Results and conclusion ...... 13 7. Glossary and abbreviations ...... 14

List of figures

Figure 1: MTSAO methodology ...... 5 Figure 2: Energy demand forecast ...... 6 Figure 3: Comparison of historical and forecast demand ...... 7 Figure 4: Plant performance for July 2017 MTSAO ...... 11 Figure 5: April 2016 and July 2017 MTSAO, % EAF Comparison ...... 11 Figure 6: Scenarios Considered ...... 12 Figure 7: Forecasted excess capacity (MW) from 2017 to 2021 ...... 13

List of tables

Table 1: Adequacy metrics ...... 4 Table 2: Energy demand forecast comparison in GWh ...... 6 Table 3: Eskom installed capacity ...... 7 Table 4: Medupi Power Station unit commercial operation dates ...... 8 Table 5: Kusile Power Station unit commercial operation dates ...... 9 Table 6: Non-Eskom supply sources, including imports (MW) ...... 9 Table 7: REIPPP committed capacity...... 10

1. Overview The South African Grid Code – System Operation Code, Version 9.0, requires “on or before 30 October of each year, the System Operator shall publish a review (called the ‘Medium Term System Adequacy Outlook’) of the adequacy of the Interconnected Power System (IPS) to meet the long term (5 year future) requirements of electricity consumers”. This review will be limited to the adequacy of the generation system for the Republic of South Africa. At the request of NERSA, Eskom System Operator has produced a mid-year view of the Medium-term System Adequacy Outlook (MTSAO) for the period 2017 to 2021. This is not a comprehensive study, but only investigated a limited number of future scenarios. A more comprehensive MTSAO covering the period 2017 to 2022 as per the Grid Code requirements will be published in October 2017. The October MTSAO will consider inputs and comments by NERSA and other interested parties.

2. Introduction The MTSAO provides a statement of generation adequacy to meet the expected electricity demand for the next five years (calendar years 2017 to 2021). The adequacy to transmit and distribute electricity does not form part of the MTSAO. The study determines the adequacy of the system to meet the expected demand of the country, made up of a combination of local consumption and exports. This demand is satisfied by a combination of all generation resources licensed by NERSA, imports, and demand-side management resources.

3. Methodology and approach The South African IPS is assessed based on the system adequacy metrics, as shown in Table 1 below. The adequacy metrics are chosen to provide information on the operational capacity and energy adequacy of the generation system to meet expected demand. The threshold for each of the metrics is set at the point of least total cost to the consumer. The adequacy metrics, as shown in Table 1 below, indicate both capacity and energy contingencies. Capacity-type contingencies, on the one hand, look at unexpected load increases, short-duration events (hours), as well as just sufficient capacity to supply demand, and lead to imbalances between supply and demand. Energy-type contingencies, on the other hand, look at higher-than-forecast load growth or loss of a large supply source, longer duration (weeks/months), as well as just sufficient baseload plant to supply load on a continuous basis, and lead to imbalances between supply and demand.

Table 1: Adequacy metrics

The system is deemed adequate only if all system adequacy metrics in Table 1 above are satisfied. Should any of the adequacy metrics not be met, additional resources are added, as shown in Figure 1; these resources are quantified in terms of baseload, mid- merit, and peaking capacity in MW.

Figure 1: MTSAO methodology

Available generation and demand response resources, both existing and new when in commercial operation, are dispatched to meet expected demand on an hourly basis for all hours in the study period. The dispatch is done on a least-cost basis and adheres to all system requirements (demand and reserves), resource constraints (generator capabilities), and the generator/demand response owner dispatch regime.

4. Assumptions Key assumptions are based on demand forecast, existing and committed supply resources, and plant performance. 4.1 Demand forecast Two country-level demand forecasts were developed internally in Eskom, namely, the moderate growth demand forecast and the high demand forecast, shown below in Figure 2 and Table 2. The moderate growth forecast provides the base case, which has an average growth rate of 2.16%. The high demand forecast has an average annual growth rate of 2.8%. The South African actual energy demand for the 2016 calendar year was 243.94 TWh, with a growth rate of -0.3% relative to 2015.

Energy Demand Forecast 275 3.50

270 3.00 2.50 265 2.00 260 1.50 255 1.00 ENERGY(TWH) 250 0.50

245 0.00 ENERGY GROWTH RATE (%) 240 -0.50 2015 2016 2017 2018 2019 2020 2021 CALENDAR YEARS Moderate Growth Rate High Growth Rate Moderate High April 2016 MTSAO Forecast

Figure 2: Energy demand forecast

The moderate growth forecast takes account of current economic conditions, and forecasts increase in demand as a result of Eskom’s drive to increase sales both locally and across the border.

January 2016 Moderate growth High MTSAO 2017 246 500 246 500 248 607 2018 250 100 250 200 251 730 2019 256 100 256 500 252 532 2020 263 300 264 200 253 416 2021 269 300 272 100 252 826 Table 2: Energy demand forecast comparison in GWh

The South African electric power system was characterised by load shedding from as far back as 2008 up to 2015. This was due to a higher plant failure rate amid slippages in commissioning new committed capacity, which reduced available generation supply to meet demand. Figure 3 below shows that actual demand has remained practically the same since 2008 due to system constraints, load shedding, slow economic growth, higher prices, and energy efficiency measures that have been implemented by the industry since 2008.

Historical Energy vs Energy Demand Forecast 300 0.08

250 0.06

200 0.04

150 0.02 Energy[TWh]

100 0 Growth [%] Rate

50 -0.02 AAR (1990-2007): 3.09% AAR (2008-2016): -0.49%

0 -0.04

1986 2004 1980 1982 1984 1988 1990 1992 1994 1996 1998 2000 2002 2006 2008 2010 2012 2014 2016 2018 2020 Calendar Years

Actual Energy Growth Rate Moderate Growth Rate High Growth Rate Actual Energy Moderate Growth High April 2016 MTSAO Forecast

Figure 3: Comparison of historical and forecast demand

4.2 Existing and committed supply resources Generation resources and demand-side initiatives are both used to meet the forecast demand. The capacities of the generation resources are, furthermore, grouped in terms of Eskom installed capacity, new build commissioning dates, non-Eskom capacity without the Renewable Energy Independent Power Producer Programme (REIPPP), and REIPPP capacity.

4.2.1 Eskom installed capacity Total Eskom installed capacity consists of coal, nuclear, pumped storage, diesel, hydro, and wind. For the purposes of the MTSAO, a conservative view was taken that coal-fired power stations would reach the end of their economic live after 50 in line with the IRP2010. Table 3 depicts the Eskom installed capacity over the study horizon, but excludes Medupi and Kusile.

2017 2018 2019 2020 2021 Coal 35 795 35 795 35 795 35 238 34 307 Nuclear 1 860 1 860 1 860 1 860 1 860 Pumped storage 2724 2 724 2 724 2 724 2 724 Diesel 2 409 2 409 2 409 2 409 2 409 Hydro 600 600 600 600 600 Wind 100 100 100 100 100 43 488 43 488 43 488 42 931 42 000

Table 3: Eskom installed capacity

4.2.2 New build commercial operation dates The official commercial operation dates (CoDs) for Medupi and Kusile are a blend of P50 dates for the first three units and P80 for the last three units. The study also used the expected CoDs, based on the current construction performance of the New Build Programme, as a scenario.

The current construction performance of new committed generated capacity has resulted in full commercial operation of Ingula, earlier than assumed in the previous MTSAO study, commercial operation of Medupi’s Unit 5, and earlier synchronisation of Kusile’s first unit, further increasing available supply.

April 2016 MTSAO Station unit Official CoD Expected CoD CoD MEDUPI Unit 6 Commercial Commercial Commercial Unit 5 Commercial Commercial 2018-March

Unit 4 2017-Dec 2017-Dec 2018-July

Unit 3 2019-Jun 2019-Jun 2019 Jun

Unit 2 2019-Dec 2019-Dec 2019 Dec

Unit 1 2020-May 2020-May 2020-May Table 4: Medupi Power Station unit commercial operation dates

Table 4 above and Table 5 below show that, relative to the April 2016 MTSAO CoD submission, Medupi Unit 5 added additional capacity to the grid earlier than projected in April 2016, and the commercial operation of Kusile Unit 1 was brought forward.

April 2016 MTSAO Station unit Official CoD Expected CoD CoD KUSILE Unit 1 2018-Apr 2017-Sept 2018-July

Unit 2 2019-Apr 2019-Apr 2019-July

Unit 3 2020-May 2020-May 2020-Aug

Unit 4 2021-Mar 2021-Mar 2021-Mar Unit 5 2021-Nov 2021-Nov 2021-Nov Unit 6 2022-Sep 2022-Sep 2022 Sep Table 5: Kusile Power Station unit commercial operation dates

4.2.3 Non-Eskom capacity without the REIPPP Table 6 depicts the non-Eskom and cross-border import capacities assumed in the study and is based on the latest Eskom and NERSA information.

2017 2018 2019 2020 2021 Kelvin 160 160 0 0 0 Sasol Infrachem coal 125 125 0 0 0 Sasol Synfuel coal 600 600 600 600 600 DoE Peaker 1 005 1 005 1 005 1 005 1 005 Other gas 140 140 140 0 0 Sasol Infrachem gas 175 175 175 175 175 Sasol Synfuel gas 250 250 250 250 250 Cahora Bassa 1 548 1 548 1 548 1 548 1 548 Mondi 144 144 144 144 144 Other cogen 140 140 140 140 140 Other hydro 12 12 12 12 12 Other wind 5 5 5 5 5 Sappi Ngodwana 174 174 174 174 174 Steenbras 180 180 180 180 180 Colley Wobbles 65 65 65 65 65 Table 6: Non-Eskom supply sources, including imports (MW)

4.2.4 REIPPP Table 7: REIPPP committed capacity

Table 7 below shows the installed capacity that is considered committed in this MTSAO for the REIPPP. The REIPPP capacities considered are from Bid Windows 1 up to 3.5, excluding the 100 MW of CSP not yet signed for Bid Window 3.5.

Installed capacity (MW) 2017 2018 2019 Wind 1 470 1 982 1 982 PV 1 474 1 474 1 474 CSP 200 300 400 Landfill gas 11 13 13 Small hydro 14 14 14 Biomass 0 17 17 3 169 3 800 3 900 Table 7: REIPPP committed capacity

4.2.5 Demand-side management Only 110 MW of DSM was considered in this study, 48 MW of which is allocated to the residential savings programme, while 62 MW is allocated to the commercial and industrial savings programme.

4.3 Eskom plant performance Since the 2008 electricity supply crisis, Eskom had been able to meet electricity demand through delaying maintenance of the generation fleet. That led to the deterioration in performance of the aging fleet, which exacerbated the past crisis, but also had a longer-term impact on the effectiveness of the fleet to meet future demand. Consequently, Eskom put in place strategies to arrest the decline in performance and return the average energy availability factor (EAF) of the current fleet to 80%. The assumed plant performance for the study builds on the “80:10:10” strategy. This results in an 80% EAF by FY2020 for the Eskom fleet, which is then maintained beyond the study period of this MTSAO. The EAF for the year ending 31 March 2017 was 77.3%. The plant performance that was considered in this study is depicted in Figure 4 and shows an EAF of 77.99% in calendar year 2018.

Plant Performance: July 2017 MTSAO

12 80.5

80 11.5 79.5

11 79

78.5 10.5

78 (%) EAF

10 77.5

Planned and Unplanned Outages (%) Outages Unplanned and Planned 9.5 76.5

9 76 2017 2018 2019 2020 2021 Calendar Years Unplanned Outages (%) Planned Outages (%) EAF(%)

Figure 4: Plant performance for July 2017 MTSAO

Figure 5 below compares the EAF in this MTSAO to the EAF assumed in the April 2016 MTSAO and shows an improvement of about 4% in 2017.

82 EAF COMPARISON

EAF (%) EAF 74

68 2017 2018 2019 2020 2021

April 2016 MTSAO July 2017 MTSAO

Figure 5: April 2016 and July 2017 MTSAO, % EAF Comparison

5. Modelling scenarios The following scenarios were, therefore, considered in the MTSAO July 2017, as shown in Figure 6 below. All the scenarios were based on a 50-year life of plant (LOP) for coal power stations and the Eskom plant performance discussed in Figure 5 above. Two demand forecasts were tested, a moderate growth demand forecast and a high growth demand forecast, as depicted in Figure 2 above. For each forecast, two different CoD scenarios were tested for the new build commercial operation dates, as shown in Table 4 and Table 5 above.

Existing fleet Demand forecast New build CoD

Official CoD Moderate growth

Expected CoD

50-year LOP

Official CoD

High

Expected CoD Figure 6: Scenarios Considered

The scenario with the moderate growth demand and official commercial operation dates is considered the base case. Scenarios not considered include the risks of inadequate coal stock levels at multiple power stations and drought in the Western Cape, which have been identified in the short term, since they have treatment plans and are managed.

6. Results and conclusion This MTSAO study has shown that the system is adequate in the medium term to meet demand from 2017 to 2021 in all the scenarios studied. This is similar to the October 2016 publication for the MTSAO. Lower-than-expected demand, improvements in the EAF of Eskom coal-generating sources, and the earlier commercial operation dates of Eskom’s new build have contributed to the improved adequacy since the April 2016 study. The system has excess capacity throughout the study horizon of this MTSAO. Figure 7 below indicates the excess capacity associated with the base case of this MTSAO study (moderate growth with official CoDs). The excess is based on the average growth of 2.16%, which is higher than what is currently being observed. The existing fleet has no mid-merit-type generators, only baseload and peaking. Therefore, the excess is made up mainly of baseload – therefore, the Eskom coal-fired plant.

Forecast surplus capacity 5000

4000

3000

2000

Capacity[MW] 1000

0 2017 2018 2019 2020 2021 Year

Figure 7: Forecasted excess capacity (MW) from 2017 to 2021

Furthermore, it can be concluded that extending the life of plant beyond 50 years, further reduction in demand, and additional IPPs beyond Bid Window 3.5 will increase the excess capacity beyond what is indicated in the base case. The October 2017 publication will be more comprehensive and will look at additional scenarios, which may include quantifying the magnitude of excess capacity due to additional IPPs (beyond Bid Window 3.5), lower demand, and longer life of plant for the Eskom coal-fired stations. The October 2017 study will also seek to determine the “meetable” country demand, which is the demand growth that can be adequately supplied by the country supply system, demand-side management, and demand response measures.

7. Glossary and abbreviations “Adequacy” relates to the existence of sufficient facilities in the system to satisfy the customer load demand or system operational constraints. Adequacy is, therefore, associated with static conditions, which do not include system dynamic and transient disturbances. “Adequacy metrics” are the output parameters analysed to determine “adequacy”. “Baseload” represents plant capable of generating all day. “AGR” means annual growth rate. “CoD” means commercial operation date. “CSP” means concentrated solar power. “EAF” means energy availability factor and reflects a unit, plant, or industry’s availability to produce energy. The energy availability factor is the ratio of available energy over the nominal energy (sent-out energy capability) and refers to the energy that could have been produced at available capacity for the reference period over the nominal energy for the same period. “Non-Eskom capacity” means generation capacity from external sources. “IPP” means independent power producer. “Load factor” reflects the ratio of the actual generated energy against the nominal energy (sent-out energy capability) and, thus, represents the extent to which the installed capacity is utilised. The calculation method of this measure is similar to the term “capacity factor”, which is used in certain electricity generating references. “LOP” means life of plant. “Mid-merit” represents plant typically generating from before the morning peak demand to after the evening peak demand. “NERSA” means National Energy Regulator of South Africa. “P50” means a probability of 50% of meeting the target. “P80” means a probability of 80% of meeting the target. “Peaking” represents plant generating only during the peak demand or emergency hours. “PV” means solar photovoltaic. “REIPPP(s)” means Renewable Energy Independent Power Producer Programme(s). “SO” refers to the System Operator who is responsible for dispatch of power.