Docstoc

IRP compiled 2004

Document Sample
IRP compiled 2004 Powered By Docstoc
					Contents
                                                                      Page

FOREWORD                                                                  2

ACKNOWLEDGEMENTS                                                          3

NIRP2 Reference Case Report                                               4

NIRP2 Stage 2: Risk and Sensitivity Analyses Report                      37

NIRP2 Stage 2 Appendix A:
Summary of Comments on NIRP2 Reference Case                              65

NIRP2 Stage 2 Appendix D:
Summary of ARC Comments on NIRP2 Stage 2 report                          76



Information provided on CD ROM                              inside back cover


NIRP2 Reference Case Appendix 1: Supply-side data,

NIRP2 Reference Case Appendix 2: Results,

NIRP2 Reference Case Appendix 3: Background supply-side data.

NIRP2 Stage 2 Appendix B: Summary of Plans

NIRP2 Stage 2 Appendix C: NIRP2 Stage 2 Database
    Foreword
    Background

    In the light of the Energy Policy White Paper for South Africa and the recent government initiatives for
    restructuring of the ESI, the NER introduced the development of a National Integrated Resource Plan (NIRP)
    as an independent information source to stakeholders and decision-makers for insuring security of the supply.
    The first National IRP (NIRP1) was completed and published in March 2002.

    At the beginning of the year 2003, the NER established the NIRP Advisory and Review Committee (ARC) with
    the function to provide wide stakeholders' guidance and contribution to the NIRP development process.

    Unlike its predecessor the NIRP 2003/4 (NIRP2) relies on average international data (modified to reflect South
    African labor and exchange rates) for the cost and performance of new generation plants. Other important
    change from NIRP1 is the inclusion of sensitivity analysis and scenarios to address risk factors and
    uncertainties. Further the NIRP2 takes into account the transmission integration costs and losses associated
    with the location of the new generation plants.

    NIRP2 Studies

    The NIRP2 has been generated under the guidance of the NER NIRP Advisory and Review Committee (ARC)
    by a NIRP team comprising Eskom Resources and Strategy Group, Energy Research Institute of UCT and the
    NER. The work carried out for NIRP2 is divided into two stages:
2

    & Stage 1: Development of a reference case
    & Stage 2: Development of risk and sensitivity analyses

    The NIRP2 reference case was completed and published on the NER Web site on 3 March 2004 while NIRP2
    Stage 2: Risk and Sensitivity Analysis - on 22 September 2004.

    This publication consists of NIRP2 Stage 1 and 2 reports. Some of the Appendices associated with the main
    reports are provided in an electronic format (CD).



    The NIRP 2 Reference Case publication contains the following parts:

    &   NIRP2 Reference Case Report
    &   Appendix 1: Supply-side data
    &   Appendix 2: Results
    &   Appendix 3: Background supply-side data.

    The NIRP 2 Stage 2: Risk and Sensitivity analysis includes the following five parts:

    &   NIRP2 Stage 2 Report,
    &   Appendix A: Summary of comments on NIRP2 Reference Case
    &   Appendix B: Summary of Plans
    &   Appendix C: NIRP2 Stage 2 Data Base Summary
    &   Appendix D: Summary of ARC Comments on NIRP2 Stage 2 report.
Acknowledgements
The NER acknowledge the guidance, contribution and
valuable suggestions of:

NER NIRP Advisory and Review Committee (ARC):         Project Team:

 Chairman: Prof A Eberhard, NER Board                    Dr Bianka Belinska, NER
 Members: Smunda Mokoena, CEO NER                        Steve McFadzean, Eskom ISEP
           Naresh Singh, EM NER                          Johan Prinsloo, Eskom ISEP
           Andre Otto, DME                               Zaheer Khan, Eskom ISEP
           Dr Elsa Du Toit, DME                          Andrew Etzinger, Eskom ISEP
           Robert Maake, DME                             Mavo Solomon, Eskom ISEP
           Tseliso Magubela, DME                         Moonlight Mbatha, Eskom ISEP
           Chris Gadsden, NT                             Mark Howells, ERI UCT
           Justice Mavhungu, DPE                         Glen Heinrich, ERI UCT
           Arnot Hepburn, EIUG                           Thomas Alfstadt, ERI UCT
           Piet van Staden, EIUG                         Andrew Kenny, ERI UCT
           Dick Kruger, Chamber of Mines                 Alison Hughes, ERI UCT
           Danny Vengedasamy, SACOB                      Rizia Buckas, NER
           Manfred Kuster, AMEU                          Willie Boeije, NER
           Mandla Tshabalala, AMEF                       Lambert du Plessis, NER
           Gerhard Loedolff, Eskom Generation
           Lynette Vajeth, Eskom Generation           WEB Publisher:                         3
           June Willis, Eskom Transmission
           Michael Barry, Eskom Transmission             Correy Sutherland, NER
           Segomoco Scheppers, Eskom Transmission
           Erica Johnson, Eskom Transmission          NIRP Contact Person:
           Jean Louis Pabot, Eskom KSACS
           Hermann FW Oelsner, Darling Windfarm IPP      Dr Bianka Belinska, NER
           Robert Siemers, Kelvin IPP                    Tel: 012-4014650
           Donald Bennett, Kelvin IPP                    Fax: 012-4014687
           Prof K Bennett, ERI UCT                       Email: bianka.belinska@ner.org.za

ARC Management Officer:

            Dr Bianka Belinska, NER
    National Integrated Resource Plan 2




              Reference Case

                       COMPILED BY
4
            ISEP Eskom (Resources and Strategy)

                            AND

                 Energy Research Institute
                 University Of Cape Town

                            AND

              The National Electricity Regulator




                      27 February 2004
                                                                                        Reference Case




Executive Summary
INTRODUCTION

In the light of the Energy Policy White Paper for South Africa and the recent government initiatives for
restructuring of the ESI, the NER introduced the development of a National Integrated Resource Plan (NIRP)
as an independent information source to stakeholders and decision-makers for insuring security of the supply.
The first National IRP (NIRP1) was completed and published in March 2002.

At the beginning of the year 2003, the NER established an IRP Advisory and Review Committee (ARC) to
provide wide stakeholders' contribution to the NIRP process. The main functions of the ARC are to approve the
primary assumptions (technical, economical, environmental and social), evaluate the supply- and demand
options for inclusion in the plan and oversee the development process.

This NIRP is a revision of the first NIRP issued in March 2002 and published on the NER Web site. For the
purposes of these analyses the first NIRP is referred as NIRP1 and the current study as NIRP2.

The NIRP2 has been generated under the guidance and approval of the NER NIRP Advisory and Review
Committee (ARC) by a NIRP team comprising Eskom Resources and Strategy Group, Energy Research
Institute of UCT and the NER.

Unlike its predecessor the NIRP2 relies on average international data (modified to reflect South African labor
and exchange rates) for the cost and performance of new generation plants.

Other important change from NIRP1 is the inclusion of sensitivity analysis and scenarios to address risk
                                                                                                                 5
factors and uncertainties such as performance of existing generation plants (Eskom and non-Eskom),
sustainability and delivery of Demand-side Management (DSM) options (including Interruptible load supplies
(ILS)) and changes in the electricity demand load shape. Further the NIRP2 takes into account Transmission
integration costs and credit for regional location of new capacity not included previously.

The growth in demand for electricity over the next twenty-year planning horizon is consistent with that
predicted in NIRP1 and follows a moderate forecast with moderate weather conditions.

OBJECTIVE AND PRIMARY ASSUMPTIONS

The objective of the National Integrated Resource Plan (NIRP) is to determine the least cost supply options to
the country, provide information to market participants on opportunities for investment in new power stations
and evaluate the security of the supply.

The NIRP2 reference case is based on the following primary assumptions approved by the ARC:

& The net discount rate (before tax) agreed for the studies is 10 % internal to South Africa;
& Options are compared on the basis of 1 January 2003 prices. Foreign capital is converted to South African
  Rand at the exchange rate of R9/$ reflective of a long-term planning approach;
& Plant availabilities for new plants are based on the World Energy Council (WEC) best quartile results for
  2002. For existing plants the studies use the current targets in Eskom adjusted independently for each
  individual station to give a weighted average for base-load capacity of 88% EAF; (7% PCLF: 3% UCLF with a
  provision of 2% for OCLF to cater for risk).
& The planning horizon for the study is 20 years from 2003 to 2022;
& Moderate electricity consumption and demand growth;
& Low DSM penetration;
& Inflation for all fuels and new technologies was at South African PPI.
    National Integrated Resource Plan 2




    CAPACITY OUTLOOK FOR NIRP2 REFERENCE PLAN

    The capacity outlook for the NIRP Reference plan developed in Stage 1 is illustrated graphically in Figure 1.


                                                      Reference capacity plan (10% Reserve margin) 2004 to 2022

                      58000   Greenfield PBMR (Base) - Earliest end 2013

                      56000   Greenfield PF (Base) - Earliest end 2013

                      54000   Greenfield Pumped Storage - Earliest 2013

                              FBC (Base) - Earliest end 2009
                      52000
                              CCGT (Base) - Earliest end 2008
                      50000
                              Komati PF - Earliest 2010
                      48000   OCGT - Earliest 2008
      Capacity (MW)




                      46000   Grootvlei PF - 2007

                      44000   Camden


                      42000

                      40000

                      38000

                      36000
                                                       Eskom Existing Capacity with Decommissioning       Non-Eskom Existing Capacity with Decommissioning
                      34000                            Imports-Cahora Bassa Hydro                         Simunye Eskom mothballed plants
                                                       Pumped Storage capacity                            Peaking capacity
                      32000                            Base load capacity                                 Peak Demand before DSM
                                                        Peak Demand after DSM                             Required capacity
                      30000
                          2004    2005    2006      2007   2008   2009     2010   2011   2012   2013   2014   2015   2016   2017   2018    2019   2020       2021   2022



6
    Figure 1: Capacity Outlook for NIRP2 Reference Plan




    CONCLUSIONS

    The main conclusions drawn from the NIRP2 reference case study could be summarised as follows:

    1) Options for diversification are insufficient to meet all of the forecast demand for electricity over the next 20-
       year planning horizon. Coal-fired options are still required for expansion during this period. For
       environmental benefit it is imperative to continue with efforts to reduce the costs of implementing clean
       coal technologies and improve the efficiency of coal-fired plants;
    2) Base load plants are required for commercial operation from 2010. Base load options competing, include;
       Pulverised Fuel Coal-fired (PF); Fluidised Bed Combustion (FBC); Combined Cycle Gas Turbine (CCGT).
       Given the cost and performance data used in the plan these options are broadly comparable, at 10% net
       discount rate, if regional siting and transmission benefits are included;
    3) At the current assumed cost of capital (10% net discount rate) and after returning the Eskom mothballed
       plant to service, fluidised bed combustion technologies are South Africa's most economic option, followed
       by investment in coal-fired plant. This in turn is followed by importing gas / LNG for CCGT plant in the
       Cape;
    4) It will be difficult to justify diversification on an economic basis, unless penalties for not doing so are
       included in future analyses. As the cost for diversification is becoming increasingly more expensive these
       penalties (or opportunities for emissions trading) will need to be substantial to offset the economic benefits
       of remaining with coal;
    5) The NIRP plans are based on attainment and sustainability of the DSM targets , power plants availability,
       imports and interruptible loads;
    6) The NIRP plans indicate that 920 MW OCGT peak load plants must begin commissioning from 2008;
                                                                                        Reference Case




7) Maintaining a higher reserve margin of 15% over the planning period will require acceleration of the RTS of
    the mothballed plants and coal-fired options together with commissioning of additional base load capacity
    (CCGT);
8) Diversified options are new technologies to South Africa. If these options for whatever reason are not able
    to be implemented it will mean a return to a dependency on new pulverised coal-fired plants earlier than
    shown in these plans;
9) There are supply options that have not been considered such as co-generation in industry, converting
    OCGT to CCGT, adding units onto existing power stations and new imports resulting from the
    development of Electricity Supply in the Southern African region;
10) Should interruptible supply and / or OCGT capacity not be implemented this will significantly advance new
    base-load capacity;




                                                                                                                 7
    National Integrated Resource Plan 2




    Table of Contents
    EXECUTIVE SUMMARY                                                             5

    TABLE OF CONTENTS                                                             8

    LIST OF TABLES                                                                9

    LIST OF FIGURES                                                               9

    ABBREVIATIONS                                                                 10

    1.INTRODUCTION - THE INTEGRATED RESOURCE PLANNING PROCESS                     11

    2.STRATEGIC FRAMEWORK AND PRIMARY ASSUMPTIONS                                 11

     2.1 STRATEGIC POSITION (DEFINED BY THE ADVISORY AND REVIEW COMMITTEE (ARC)
         OF THE NER)                                                              12
     2.2 PRIMARY ASSUMPTIONS                                                      12
     2.3 RISKS AND UNCERTAINTIES                                                  12
     2.4 OPTIMISATION PARAMETERS                                                  13
     2.5 CRITERIA FOR INCLUSION OF SUPPLY-SIDE AND DEMAND-SIDE OPTIONS            13

    3.NATIONAL ELECTRICITY FORECAST                                               13
8
      3.1 ECONOMIC GROWTH                                                         16
      3.2 LARGE INDUSTRIAL PROJECTS                                               16
      3.3 ELECTRIFICATION                                                         16
      3.4 ELECTRICITY INTENSITY                                                   17
      3.5 NATURAL GAS                                                             17
      3.6 FOREIGN FORECAST COMPONENT                                              17
      3.7 DEMAND PROFILES                                                         18

    4.DEMAND-SIDE OPTIONS (DSM)                                                   18
      4.1 THE DEMAND SIDE PLANNING BASIS FOR NIRP2                                20
      4.2 INTRODUCTION TO THE DEMAND-SIDE SCREENING RESULTS                       20
          4.2.1 Residential Energy Efficiency - REE                               21
          4.2.2 Industrial, Mining and Commercial Energy Efficiency               22
          4.2.3 Residential Load Management (RLM)                                 23
          4.2.4 Industrial and Mining Load Management (IMLM)                      23
      4.3 UNCERTAINTY AND RISK ASSOCIATED WITH DSM                                23
      4.4 PRIORITIES FOR THE DEVELOPMENT OF DEMAND-SIDE RESOURCES                 24

    5.SUPPLY-SIDE OPTIONS                                                         25
      5.1 ESKOM SYSTEM - EXISTING AND COMMITTED CAPACITY                          25
      5.2 NON-ESKOM SYSTEM - EXISTING CAPACITY                                    25
      5.3 RETURN TO SERVICE OF ESKOM MOTHBALLED PLANT (SIMUNYE)                   25
                                                                                              Reference Case




  5.4 NEW SUPPLY-SIDE OPTIONS                                                                          26
      5.4.1 New Pulverised Fuel (PF) Coal-Fired Stations                                               26
      5.4.2 New Gas-Fired Plant                                                                        26
      5.4.3 New Pumped Storage Schemes                                                                 27
      5.4.4 Greenfield Fluidised Bed Combustion                                                        27
      5.4.5 Conventional Nuclear (Advanced Light Water Reactor (ALWR))                                 29
      5.4.6 Research projects/programs                                                                 29
      5.4.7 Imported Hydro                                                                             29
  5.5 SCREENING CURVES                                                                                 29
  5.6 OTHER: ENVIRONMENTAL, EXTERNALITIES, TRANSMISSION EXPANSION                                      31

6.INTEGRATION AND SENSITIVITY ANALYSIS                                                                 31
  6.1 REFERENCE PLAN                                                                                   32
  6.2 ALTERNATIVE PLAN 1 TO REFERENCE PLAN                                                             33
  6.3 ALTERNATIVE PLAN 2 (SENSITIVITY STUDY)                                                           33
  6.4 ALTERNATIVE PLAN 3 (OPTIMAL RESERVE MARGIN)                                                      34

7.SYSTEM ANNUAL AVERAGE LONG RUN MARGINAL COST                                                         35

8.CONCLUSIONS                                                                                          36



LIST OF TABLES

Table 1: Average demand growth intervals                                                               14
Table 2: DSM aggregate megawatts displaced                                                             19
                                                                                                               9
Table 3: Residential Energy Efficiency                                                                 21
Table 4: Commercial energy efficiency                                                                  22
Table 5: Industrial and Mining Energy Efficiency                                                       22
Table 6: Residential load management programmes                                                        23
Table 7: Industrial and mining load management                                                         23
Table 8: Summary of cost and performance data of new supply-side options                               28



LIST OF FIGURES

Figure 1: Electricity sales forecast range - National plus foreign                                     14
Figure 2: Increase in annual peak demand and system losses (national plus foreign)                     15
Figure 3: Eskom track record                                                                           16
Figure 4: RSA electricity intensity                                                                    17
Figure 5: Typical hourly peak demand summer profile                                                    18
Figure 6: Typical hourly peak demand winter profile                                                    18
Figure 7: Life cycle levelised costs to build and operate base load plants                             30
Figure 8: Life cycle levelised costs to build and operate peaking plants                               30
Figure 9: Reference plan (10% Reserve Margin)                                                          32
Figure 10: Alternative 1 to reference plan (15% Reserve Margin)                                        33
Figure 11: Alternative 2 - sensitivity to reference plan (excludes interruptible supply options)       34
Figure 12: Annual average long run marginal costs of plans                                             35

APPENDIX 1: SUPPLY SIDE MODELLING DATA SUMMARY
APPENDIX 2: RESULTS
APPENDIX 3: BACKGROUND SUPPLY SIDE DATA
     National Integrated Resource Plan 2




     Abbreviations
     ALWR       Advanced Light Water Reactor
     ARC        Advisory Review Committee
     CEE        Commercial Energy Efficiency
     CCGT       Combined Cycle Gas Turbine
     CF         Coal Fired
     CF (DSM)   Capacity Factor for DSM
     CV         Calorific Value
     CUE        Cost of Unserved Energy
     DEAT       Department of Environmental Affairs and Tourism
     DME        Department of Minerals and Energy
     DSM        Demand Side Management
     DWAF       Department of Water Affairs and Forestry
     EAF        Energy Availability Factor
     EIA        Energy Information Administration
     EPC        Engineering, Procurement and Construction
     ESI        Electricity Supply Industry
     FBC        Fluidised Bed Combustion
     FOR        Forced Outage Rate
     GT         Gas Turbine
     HELM       Hourly Electricity Load Model
     HHV        High Heating Value
     HTF        Heat Transfer Fluid
10
     ICLM       Industrial and Commercial Load Management
     IMEE       Industrial and Mining Energy Efficiency
     IMLM       Industrial and Mining Load Management
     IEA        International Energy Agency
     IRP        Integrated Resource Plan
     ISEP       Integrated Strategic Electricity Planning
     LF         Load Factor
     LOLE       Loss of load expectation
     LOLP       Loss of load probability
     LNG        Liquefied Natural Gas
     LPG        Liquefied Petroleum Gas
     MCR        Maximum Continuous Rating
     MEUL       Minimum Energy Utilization Level
     NDR        Net Discount Rate
     MOU        Memorandum of Understanding
     NER        National Electricity Regulator
     O&M        Operation and Maintenance
     OCGT       Open Cycle Gas Turbine
     OCLF       Other Capability Loss Factor
     PCLF       Planned Capability Loss Factor
     PBMR       Pebble Bed Modular Reactor
     PF         Pulverised Fuel
     POR        Planned Outage Rate                               RLM    Residential Load Management
     PPI        Producer Price Index                              ROD    Record of Decision
     PV         Present Value                                     UE     Unserved energy
     PWR        Pressurised Water Reactor                         UCLF   Unplanned Capability Loss Factor
     REE        Residential Energy Efficiency                     WEC    World Energy Council
                                                                                           Reference Case




1.INTRODUCTION THE INTEGRATED RESOURCE PLANNING PROCESS

The prime objective of the National Integrated Resources Plan (NIRP) is to provide a long-term least-cost
resource plan for meeting the electricity demand consistent with the reliability of the electricity supply,
environmental, social and economic policies. The NIRP also serves as an information tool for potential project
developers and decision-makers.

The NIRP provides an assessment of the system adequacy and also addresses other public policies such as
environmental impacts and Demand Side Management (DSM).

The NIRP also takes into account the “The New Partnership for African Development (NEPAD)” by
incorporating committed contracts for imports and exports to South Africa from neighbouring States.

This is the second NIRP carried out under the auspices of the NER. The first NIRP was carried out during 2001-
2002 and published on the NER Web site. For the purposes of these analyses the first NIRP is referred as
NIRP1 and the current study as NIRP2. This analysis follows the steps already defined in the NIRP 1 as
follows:

& Develop the primary assumptions and selection criteria;
& Produce an electricity consumption and demand forecasts;
& Investigate a full array of demand- and supply-side options and identify those that meet the strategic
  selection criteria for inclusion in the plan;
& Determine an optimal combination of demand- and supply-side options from those selected for inclusion in
  the resource plan;
& Evaluate the risk factors associated with uncertainties such as load growth, plant availability, weather
  conditions, DSM penetration level, level of interruptible loads etc;
                                                                                                                    11
& Analyse the environmental, external and financial consequences;
& Select a preferred plan.

The IRP studies carried out in this process use several computer software models. A major model is the RP
Workstation, which consists of a suite of computer software programs, developed by the American Electricity
Power Research Institute (EPRI) and specifically the optimisation module, Electric Generation Expansion
Analysis System (EGEAS).

Due to time constraints, the work carried out for this NIRP is divided into two stages as agreed within the terms
of the National Electricity Regulator (NER) Advisory and Review Committee (ARC). These two stages consist:

& Stage 1: Determine a reference plan for analysis and as basis for recommendation on capacity planning
  issues in the short-term;
& Stage 2: Develop sensitivity studies and scenarios in order to address targeted risk factors and
  uncertainties.

The first stage is intended for completion by Feb 2004, which constitutes this report, and the second for
completion by April 2004.



2.STRATEGIC FRAMEWORK AND PRIMARY ASSUMPTIONS

This serves as the strategic basis for the planning assumptions for the NIRP. The NIRP2 was based on the
planning assumptions approved by the ARC. These studies are based on January 2003 as its primary
reference date for cost parameters and for the start of the twenty- year planning horizon. The NIRP is a long-
term planning process and it is undertaken on an annual basis. The assumptions defined in this document
need to be considered in that context.
     National Integrated Resource Plan 2




     2.1 Strategic position (defined by the Advisory and Review Committee (ARC) of the NER)

     The following key points are worth highlighting:

     & The aim of the modelling is to determine the long-term least-cost electricity supply options to the country,
       independent of ESI structure and subject to the primary assumptions and constraints;
     & The NIRP includes the electricity market within and external to South Africa (imports to and exports from
       South Africa);
     & The NIRP may be used as a basis for identifying investment opportunities for suppliers in the ESI;
     & The NIRP's objective is to optimise the supply-side and demand-side mix to keep the price of electricity to the
       consumers as low as possible.

     2.2 Primary Assumptions

     The NIRP2 reference case is based on the following primary assumptions approved by the ARC:

     & The net discount rate (before tax) agreed for the study is 10 % internal to South Africa;
     & Options are compared on the basis of 1 January 2003 prices. Foreign capital is converted to South African
       rands at the exchange rate of R9/$ reflective of a long-term planning approach;
     & Plant availabilities for new plants are based on the World Energy Council (WEC) best quartile results for
       2002. For existing plants the studies use the current targets in Eskom adjusted independently for each
       individual station to give a weighted average for base-load capacity of 88% EAF; (7% PCLF: 3% UCLF with a
       provision of 2% for OCLF to cater for risk).
     & The planning horizon for the study is 20 years, from 2003 until 2022;
     & National moderate electricity consumption and demand growth;
     & Low DSM penetration;
12
     & Inflation for O&M and fuel resources at South African PPI unless stated otherwise;
     & New technologies costs are adjusted for Transmission integration or where applicable the avoided
       Transmission costs (and losses) according to regional site selection;
     & The NIRP2 does not make any assumptions on the ownership of the plants.



     2.3 Risks and Uncertainties

     This Report does not address risk rigorously but rather addresses it through imposing a minimum reserve
     margin on the plan of 10%. Imposing a deterministic reliability index (reserve margin of 10%) as a constraint
     does not directly reflect specific risk factors such as FOR, generation mix and unit size. However, it does
     provide a reasonable estimate of reliability performance when other parameters remain constant over the
     planning period.

     Due to time constraints, it was agreed at the ARC as a first pass, to develop a reference plan using a 10%
     reserve margin constraint as proxy for risk. The 10% reserve margin is reflective of current inter utility
     agreements in the Southern African Power Pool (SAPP). In addition, a second plan is developed based on a
     15% reserve margin constraint reflective of international practise.

     There are a large number of risks confronting the ESI in the future. These risks are both short-term and long-
     term. For example in terms of the load forecast a short-term risk could consist of an unexpected cold weather
     snap, whereas a long-term risk could be unexpected sustained increase in demand for electricity. Some of
     these risks include:

     & Plant failure leading to longer than expected plant outage;
     & Unavailability of municipal / Eskom / imported generating capacity;
     & Degree of market penetration of DSM and maintaining current level of interruptible loads;
                                                                                                Reference Case




&   Unexpected decrease / increase, spurious or sustained, of electricity demand;
&   Changes to the load shape associated with the forecast electricity demand;
&   Unexpected decommissioning / de-rating of existing generating capacity
&   Uncertain and prolonged lead times for building new plant;
&   Project slippage
&   Inclusion of co-generation options;
&   Embargoes on nuclear energy;
&   Shortage of skills to maintain and grow the system;
&   Other energy forms displacing electricity in the energy market;
&   Revolutionary technologies coming on the scene and stranding existing assets;
&   Internalisation of externalities, such as Introduction of a carbon tax and environmental levy;
&   Plant life expectations not met;
&   Retail choice;
&   Deterioration in credit rating, exchange rates etc. resulting in a higher cost of capital;
&   Electricity supply and sales contracts (import and export contracts) being reneged upon;
&   Effect of AIDS on the electricity market.
&   Drought and Floods

2.4 Optimisation parameters

The basis for the optimisation of this NIRP is the least cost of electricity for the supply life cycle. This takes into
consideration the cost of un-served energy (CUE) to the consumer. For this NIRP2, the CUE is assumed to be
R20 666/MWh (Eskom 2003).

This CUE was derived from a customer survey of the market sectors of the electricity supply industry (ESI).
Because of an expected marketing drive for DSM, mainly in the residential sector, but also in the commercial
                                                                                                                          13
and industrial sectors, the high CUE associated with the industrial sector was chosen as representative. This is
because DSM initiatives that could change the load profile may become exhausted over the 20-year planning
horizon. Previous NIRP1 used CUE of R 19000/MWh associated with the industrial sector.

2.5 Criteria for inclusion of Supply-side and Demand-side options

The criteria listed below have been approved by the NER's Advisory and Review Committee (ARC) and are
intended to give guidance in determining whether an option is formally included in the base case.
Technologies that are not included in the base case will be evaluated in sensitivity studies. For technologies to
be included in the reference case they should:

&   Be technologically feasible;
&   Be economically viable;
&   Have adequate accuracy of costs;
&   Be under national control either via equity participation, ownership or secure contracts;
&   Be socially, politically and environmentally acceptable;
&   Be dispatchable;
&   In line with World Bank emission standards.

3.NATIONAL ELECTRICITY FORECAST

The load forecast is the foundation upon which resources planning is based. It is an endeavour to forecast the
most likely futures based on selected long-term Southern African economic forecasts. The forecast takes as its
starting point the current position in the electricity supply market but in projecting the future, it excludes any
further demand-side interventions.

A detailed sectoral approach has been used over the last ten years to develop the long-term energy (GWh)
     National Integrated Resource Plan 2




     forecast. This method has been improved and refined over time. Using a sectoral approach, which considers
     about 110 sectors or major customers individually, is one of the ways to lower the forecast risk. The forecast
     has also been updated on average more than once per year during the past ten years.

     The national plus foreign electricity forecast used in these studies is based on an average annual economic
     growth rate of 2.8%, over the planning horizon and moderate temperatures throughout the year. The foreign
     portion includes normal sales to traditional neighbouring states plus the Scorpion zinc project in Namibia and
     the aluminium smelters Mozal 1 & 2 in Mozambique.

     To address the major inherent uncertainty in the environment, a cone of uncertainty approach is used to
     develop low, moderate and high-energy forecasts. The electricity database consists of the best quality of data
     for all the 110 individual sectors, going back as far as 1980. The load forecast has been developed by Eskom
     together with contributions from major sectors/customers and wide consultations inside and outside Eskom.

     The National forecast range is illustrated in Figure 1 below.


                                        Electricity sales forecasts - national plus foreign
                     400000
                                 High
                                 Moderate
                     350000      Low



                     300000
               GWh




14
                     250000



                     200000



                     150000
                          2000   2002   2004   2006   2008   2010    2012   2014   2016   2018    2020   2022



     Figure 1: Electricity sales forecast range - National plus foreign

     The average growth in energy in intervals over a five-year period within the planning horizon is given in Table 1:

     Table 1: Average demand growth intervals

                                        High (%)                       Moderate (%)                  Low (%)
         2003 - 2008                      4.3                           3.2                            1.6
         2008 - 2013                      3.1                           2.3                            1.0
         2013 - 2018                      2.6                           1.8                            0.9
         2018 - 2022                      2.5                           1.8                            0.8

     The electricity growth in the early years until 2006 is high (4% pa) due in part to cater for major expansions in
     platinum mining and high demand for ferrochrome as experienced in 2002 and 2003 and one further 850MW
     aluminium smelter at Coega. However this high growth is not sustained after 2006. Eskom sales growth has
     been high in 2002 and 2003 but the average sales growth is only 2.2% per annum over the last six years. The
     long-term forecast allows for a growth in fixed investment but there are currently no new major projects, which
     have been officially approved, other than the platinum mine expansions.
                                                                                                          Reference Case




An hourly electricity load model (HELM) is used to develop the maximum demand forecast where the annual
energy forecast has been converted to an annual maximum demand forecast using updated sectoral
customer usage profiles. It is expected that system energy utilisation with respect to annual peak demand will
deteriorate slightly over time and the system annual load factor worsen from a current average of 74% to about
73% by 2017.

Of specific importance is the impact of system losses on the forecast demand. The Eskom system losses have
been increasing over the last number of years from 5% in 1990 to 7.7% in 2002. The system losses for the
National forecast are estimated to be currently 9% increasing to in excess of 10% over the planning horizon.

Details of the moderate annual forecast (energy and demand) are shown in the Tables given in Appendix 2 to
this report. The demand forecast is illustrated in Figure 2: Increase in annual peak demand and system losses
(national plus foreign) below.



                             Increase in forecast annual peak demand (National + Foreign) and losses

                12.0%




                10.0%



                8.0%
                                                     National + Foreign annual
                                                     system losses (%)
 (%) Increase




                                                     Annual Increase in peak
                6.0%                                                                         Forecast Annual Peak Demand in
                                                     Demand (National + Foreign) (%)
                                                                                             2022 = 53256MW                    15
                                                                                             I e. Growth from 2004 to 2022 =
                4.0%                                                                         19611MW



                           Forecast Annual
                2.0%       Peak Demand in
                           2004 = 33645MW


                0.0%
                        2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022


Figure 2: Increase in annual peak demand and system losses (national plus foreign)



The short-term accuracy of the forecast and the track record of the long-term forecast are being monitored on
an annual basis. The short-term variance is around the zero line and is not always positive or always negative.

Figure 3 below tracks Eskom's (as opposed to national plus foreign) forecast predictions over the last ten years
within a cone of 1.5% to 4% growth per annum. Actual sales are shown in the thick black line.

Over the last few years these results show that the actual values achieved were slightly lower than the forecast
values but in the last two years actual values were slightly higher. It is also important to note that the forecast
value for the year 2007, as seen at this stage is very close to the value forecasted for the same year in 1996.
     National Integrated Resource Plan 2




                              Eskom long term sales forecast track record
            370000
                           1994
                           1995
                           1996
                           1997
            320000         1998
                           1999
                           2000
                           2001
                           2002
            270000
      GWh




                           2003
                           Actual
                           High 4%
                           Low 1.5%

            220000



            170000



            120000
                 1990                 1995             2000              2005               2010               2015

16
     Figure 3: Eskom track record

     The following is a brief summary of the main assumptions of the long-term forecast (national plus foreign).

     3.1 Economic Growth

     Long term economic growth rate for the moderate forecast is 2.8% average annual growth over the planning
     horizon. This economic growth rate takes into account the impact of HIV / AIDS.

     3.2 Large Industrial Projects

     New large industrial projects, such as the proposed Pechiney aluminium smelter, Billiton Hillside aluminium
     smelter expansions, Columbus Stainless Steel expansions, Sasol new projects and expansions and Mozal 2,
     have been included in the forecast. Further expansions to ferrochrome plants are also provided for. Current
     platinum mining expansions have been included, with provision for limited expansions after 2007/8. It was
     assumed that new gold mining projects would replace old ones. Gold production in South Africa is assumed to
     continue to decrease slightly over the long term. With regard to the longer term, allowance had also been made
     in the energy forecast for unknown new projects, based on past experience. A list is regularly revised for all
     possible new electricity intensive projects, including projects with a very low probability of occurrence.

     3.3 Electrification

     The forecast makes provision for the electrification of homes by Eskom and the municipalities. Electrification
     carried out by Eskom forms the bulk of its prepaid sales category. The number of connections obtained from
     planners and an estimated consumption per connection are used as a guide. It is assumed that the
     electrification rate will level off over time. The prepaid category only formed about 2.5% of total 2002 electricity
     sales in the country.
                                                                                            Reference Case




3.4 Electricity Intensity

The growth in the long term forecast is high in the near term due to major expansions of some electricity
intensive industries. However, the growth is expected to reduce in the latter period of the planning horizon due
to the fact that the electricity intensity of the South African economy is expected to decrease over time. This is
characteristic of an economy, which is moving towards a more service based structure as is the case in South
Africa and implies that the GDP growth rate in future will be consistently higher on average than the electricity
growth rate. This has already been experienced in South Africa during the four-year period 1998 to 2001.

The electricity intensity of the South African economy is shown in Figure 4. The current major expansions of
electricity intensive industries are expected to decline and platinum mining is not expected to continue to
maintain the current high growth rates after 2007/8. With more beneficiation of minerals, less energy will be
required per unit of production. Once there is no more excess generating capacity, electricity prices will have to
increase in real terms. Although it is assumed that this increase will not be extraordinary, it is expected that it
will marginally affect electricity consumption.



                                            RSA electricity intensity
                    0.35


                    0.30


                    0.25
       kWh / rand




                    0.20
                                                                                                                      17
                    0.15


                    0.10


                    0.05


                    0.00
                       1950   1955   1960   1965   1970   1975   1980   1985    1990    1995    2000



Figure 4: RSA electricity intensity

3.5 Natural Gas

Allowance is made for the loss in electricity sales due to consumers switching from the use of electricity to the
use of natural gas.

3.6 Foreign Forecast component

This part of the forecast includes sales that are currently in place - mostly under contract. Also included are the
Mozal aluminium smelter in Mozambique and the Skorpion zinc project in Namibia. Provision had also been
made for the Corridor heavy mineral sands project in Mozambique over the longer term.
     National Integrated Resource Plan 2




     3.7 Demand Profiles

     Figure 5 shows a typical demand profile for the total Eskom system for a summer week.


                            Eskom integrated system typical summer week hourly demand profile
                                                   January/February
                    27000
                                                                                                 2002
                                                                                                 2001
                    25000                                                                        2000


                    23000
               MW




                    21000


                    19000


                    17000


                    15000
                              Mon       Tue       Wed        Thu         Fri        Sat          Sun



     Figure 5: Typical hourly peak demand summer profile

     Figure 6 and shows a typical demand profile for the total Eskom system for a winter week.
18
                             Eskom integrated system typical winter week hourly peak demand
                                                          July
                    33000
                                                                                                  2002
                    31000                                                                         2001
                                                                                                  2000
                    29000

                    27000
               MW




                    25000

                    23000

                    21000

                    19000

                    17000

                    15000
                              Mon       Tue       Wed         Thu        Fri        Sat          Sun


     Figure 6: Typical hourly peak demand winter profile



     4.DEMAND-SIDE OPTIONS (DSM)

     This NIRP2 assumes a significantly (approximately 50%) lower penetration of DSM programmes than the
     previously in NIRP1. Previous estimates were based on desktop studies. Pilots and experience in the field
     now indicate that a more conservative approach should be followed in the estimates for DSM programmes
     both in terms of capacity and cost. Five DSM programmes are targeted:
                                                                                           Reference Case




&   Residential Energy Efficiency (REE)
&   Commercial Energy Efficiency (CEE)
&   Industrial and Mining Efficiency (IMEE)
&   Residential Load Management (RLM)
&   Industrial and Mining Load Management (IMLM)

Each programme has been modelled on a basis that capital will be expended over a nine-year period to ensure
increasing annual displacements of aggregate Megawatts will be deducted from the load associated with a
specified energy usage profile as indicated in the Table 2.

Table 2: DSM aggregate megawatts displaced

     Programme                                Annual MW Displacements         Annual Energy Displaced GWh

    Residential Energy Efficiency                          32                                129
    Commercial Energy Efficiency                           14                                 69
    Industrial and Mining Efficiency                       16                                101
    Residential Load Management                            49
    Industrial and Mining Load Management                  41

Each aggregate annual displacement is maintained over a period of twenty years to ensure no deterioration of
displacement takes place. (I.e. maintenance will continue for a further 11 years beyond the nine-year
investment period).

Interruptible supply options are modelled separately to the above DSM programmes. Eskom has several
interruptible supply agreements with key customers. These agreements are severely energy constrained and
                                                                                                                    19
their capacity impact in the long-term has been reduced. There have been several comments by various ARC
members to consider these interruptible supply agreements as emergency capacity only. This issue will be
addressed in the next round of the NIRP in terms of the uncertainty inherent in this assumption.

The DSM team of Eskom in their endeavours to continually improve decision-making on the role of DSM in
resources planning have taken several actions one of which is to improve information on DSM before its
inclusion.

In terms of residential DSM, Eskom has good statistics on device populations and unit gains in the residential
markets. The homogeneity in residential markets provides confidence in DSM planning data, because end-
use devices are generally characterised within narrow ranges of diversity.

The need to upgrade confidence for this NIRP was in the commercial and industrial market segments,
because previously there was no sound strategy to deal with the diversity of information. Fortunately since the
previous National Integrated Resources Plan, NIRP1, the DSM team has had access to a document on market
assessment for electric motors prepared for the Department of Energy (DOE) in the USA. This document
clearly spells out the opportunities for energy efficiency in the majority of the commercial and industrial end-
use sectors. Electric motors make up nearly 50% of Eskom's total sales, or more than 65% of the sales to
industrial and commercial clients.

A major task in preparation of DSM data for NIRP2 has been the mapping of the DOE information on DSM
opportunities in the USA into the South African context. This has now been done, and together with other desk
research it has been revealed that the situation in SA is very different to the USA. South Africa needs to devise
its own strategy to contend with issues that do not exist in the case of the USA.

There has been a major focus on the industrial and commercial sectors, providing a better insight into the CEE,
IMEE and IMLM strategies to be adopted in the future. A further improvement in this NIRP is in the provision of
     National Integrated Resource Plan 2




     an aggregated summary for the ICEE and ICLM programmes compared with the previous industrial and
     commercial components.

     The NIRP1 placed more emphasis on energy efficiency as opposed to load management programmes. Better
     experience has now shown this emphasis to be misplaced, in that energy efficiency programmes are difficult to
     implement and to achieve performance targets due to the many barriers involved, and are turning out more
     costly than previously estimated.

     4.1 The Demand Side Planning Basis for NIRP2

     The basis for end-use load forecasting adopted for DSM planning in this NIRP is the “moderate forecast” as
     detailed above. The amount of DSM resources to be developed over a planning period is very sensitive to the
     load forecast. For sensitivity studies there is also the facility to use a “high load growth” and a “low load growth”
     forecast. A high growth forecast would imply higher targets for DSM, and a low load forecast would suggest
     lower targets than required for the moderate forecast. This issue will also be addressed in the next round of the
     NIRP in terms of the uncertainty in DSM penetration.

     4.2 Introduction to the Demand-Side Screening Results

     The following summarises the DSM resources submitted for inclusion in the NIRP2. Resources that are not
     currently included may qualify for inclusion in future version if further investigation provides the information
     needed to obtain positive screening results. The screening results for energy efficiency options are given in
     Table 1 to 3 below. For comparison, separate columns give the NIRP1 and NIRP2 assumed market
     penetration.

     In Tables 6 and 7, pertaining to the load management resources, there are high and low capacity factor (CF)
20
     resource capacities, given in MW, and the balance, which is the difference between the two.

     Capacity factor is defined as the percentage of time that the capacity of the load management resource can be
     fully utilised. “Hi CF MW” means that control algorithms are utilised to achieve a high capacity factor. There is a
     maximum percentage of time that the residential load management (RLM) resource can be utilised. In the
     maximum percentage mode the system peak MW can only be reduced by the amount reflected in the “Hi CF
     MW” column.

     The system peak can be further reduced by the amount in the “Balance of MW” column, when operated in the
     low capacity mode. The “Lo CF MW” column indicates the total reduction of the system peak if the entire load
     management resource is only operated in low capacity factor mode.
                                                                                                  Reference Case




4.2.1 Residential Energy Efficiency - REE

Table 3 gives the results of screening Residential Energy Efficiency.

Table 3: Residential Energy Efficiency

  OPTION                                        No of units                                         Assumed Market
                                                                                                       Penetration
                                                MW/a            GWh/a        IN NIRP2           NIRP1     NIRP2
  INTEGRAL CFL'S                                25.20           91.97          YES                33%       20%
  HOT WATER SYS. EFFICIENCY                     2.93            10.13          YES                28%       10%
  LOW FLOW SHOWERHEADS                          2.38             8.22          YES                34%       10%
  HOTWATER CONSERVATION                         0.47             1.63          YES                30%       10%
  COOKING AWARENESS                             1.34             3.44          YES                30%       10%
  EFFICIENT COOL STORAGE                                                        NO                30%       10%
  THERMAL EFFICIENCY                                                            NO                30%       10%
  DEMARKETING TO GAS                                                            NO                30%       15%
  MODULAR CFL'S                                                                 NO                33%       20%
  Total                                         32.39           129.0

The REE resources submitted for NIRP are predominated by integral CFL's.

The energy efficiency measures related to hot water systems have also been included but it was decided not to
include efficient cool storage appliances, thermally efficient building practices and de-marketing to gas. Also
included are integral CFL's and energy efficiency measures on storage water heaters. The integral version of
the CFL costs less to sustain this technology over the planning period. The major problem associated with
                                                                                                                             21
targeting integral CFL's for the efficient residential lighting initiatives is the risk of sustainability (i.e. conversion
back to inefficient incandescent light bulbs in the future).

The strategy with respect to electrical water heating is to encourage storage water heating. Indications are that
residential load management is a lower cost resource than building peaking plant. Instant water heaters have
not found favour, because they do not lend themselves to load shifting. For storage water heating to hold its
ground in SA it is important to demonstrate that these systems can be energy efficient, and that the per capita
water consumption can be contained, especially in the event of water shortages at a future time. For water-
heater system efficiency measures like hot water conservation, low flow showerheads and thermal insulation
are included.

Other DSM options that have been included in residential energy efficiency strategies in other countries have
yet to be developed to a stage where they are mature enough to include in the REE strategy with some
confidence. For example energy-efficient fridges and freezers will only be included in the strategy at a later
stage when there is more clarification on appliance labelling policy. Information is required on how appliance
labelling would help to create demand and to make the program economically more attractive. The strategy on
fridges and freezers is therefore to closely monitor the DME initiatives regarding appliance labelling.

Residential space-heating loads are not included but should be targeted as a demand-side measure, because
they are one of the main reasons for poor utilization of supply capacity, and therefore not very economical to
supply. Previous work by the DSM team has indicated that gas heating is likely to be more expensive to the
consumer, yet electrical space heating is losing market share to gas space heating. With rising oil prices, and
therefore LPG prices, it is uncertain, whether there will be a significant shift to gas for space heating. Better
information is needed on how thermal efficiency measures impact on space heating demand before deciding
on a relevant strategy. In particular, there is a need to know what the most marketable measures are to de-
sensitise system space-heating demand to weather-related events like falling temperatures and cloud cover
during winter months.
     National Integrated Resource Plan 2




     4.2.2 Industrial, Mining and Commercial Energy Efficiency

     Table 4 gives the results of screening CEE programmes. For CEE, the assumptions on market penetration
     remain more or less the same as with the NIRP1

     Table 4: Commercial energy efficiency

       OPTION                                  No of units                                     Assumed Market
                                                                                                 Penetration
                                                 MW/a          GWh/a        IN NIRP2         NIRP1      NIRP2
       SUPERVISION                               2.05          10.60          YES              14.1%       15%
       FANS/PUMPS                                0.64           4.68          YES              26.0%       26.0%
       COMP AIR
       LIGHTING                                  9.53          43.54          YES              27.0%       27.0%
       VSD's                                     0.26           1.90          YES               3.0%       3.0%
       REPLACE                                   0.68           3.50          YES              15.0%       10.0%
       UPGRADE                                   0.66           3.40          YES              13.0%       10.0%
       DOWNSIZE                                                                NO
       COOLING SYSTEM EFF                                                      NO
       HEAT RECOVERY                                                           NO
       INSULATION                                                              NO
       Total                                     13.82         69.38




     Table 5 gives the results of screening of Industrial and Mining energy efficiency programmes
22

     Table 5: Industrial and Mining Energy Efficiency

       OPTION                                    No of units                                     Assumed Market
                                                                                                   Penetration
                                                 MW/a          GWh/a        IN NIRP          NIRP1      NIRP2
       SUPERVISION                               2.24          15.33          YES              14.1%        2.0%
       FANS/PUMPS                                1.31           9.26          YES              26.0%        5.0%
       COMP AIR                                  2.26          16.54                                        5.0%
       LIGHTING                                  3.77          23.44          YES              27.0%      10.0%
       VSD's                                     4.40          29.3           YES              3.0%         3.0%
       REPLACE                                   1.13           7.72          YES              15.0%        2.0%
       UPGRADE                                   1.06           7.02          YES              13.0%        2.0%
       DOWNSIZE                                                                NO
       COOLING SYSTEM EFF                                                      NO
       HEAT RECOVERY                                                           NO
       INSULATION                                                              NO
       Total                                     16.15         100.93



     These results suggest that most of the drive-power programs qualify for submission. The demand-side plan
     assumes that ESCO's will be enabled that specialize in the delivery of CEE and IMEE resources to the
     industrial market, and that they are competitive and operational in SA in the future. Under these assumptions it
     is possible that this target can be achieved within 10 years. This assumption assumes support from
     governance authorities in terms of policy.

     Additional energy efficiency resources would become available were commercial and industrial participants to
                                                                                             Reference Case




focus their attention on heating and cooling systems that derive their energy from an electricity supply. These
are indicated in the demand-side plan as a requirement for further investigation.

The most viable area for industrial and commercial energy efficiency is motor supervision and control.
Maintenance and supervision of compressed air or gas compressor systems is another area for potentially
very cost-effective energy efficiency measures. Companies that employ the practice of re-winding motors
rated lower than 45kW may require advice on the economics of re-winding versus replacing. In some cases it
may even pay to upgrade motors to premium efficiency motors.

It is most important to transform the market for new construction to energy efficient lighting, especially in
commercial buildings. Lighting controls should be installed to avoid unnecessary lighting usage. Retrofits
typically do not offer very short paybacks, but are in the range where respectable rates of return can be
guaranteed. Lighting retrofits that would also reduce air-conditioning load should be targeted first. It is a matter
of major concern that the majority of industrial and commercial lighting sales into the new construction market
still conform to old and outdated standards for energy efficiency.

Every effort should be made to apply best practices in the design of systems employing fluid or airflow. In
particular excessive throttling of flow should give way to variable speed drives or staged switching of drives,
and pipes and ducts should be sized to minimise total resource costs. Every effort should be made to promote
the adherence to best practices on systems with pumps, fans and compressors. CEE and IMEE retrofits
should initially target systems of large drives, i.e. 200 kW and larger.

4.2.3 Residential Load Management (RLM)

In the case of RLM, the screening criteria have led the DSM Team to firm conclusions on the RLM strategy. Up
until now too many conflicting DSM strategies were considered. Table 6 gives the results of screening the
                                                                                                                       23
various RLM strategies and options.

Table 6: Residential load management programmes

                                       SUMMARY OF RLM SCREENING RESULTS
    Options                            Hi CF MW/a      Balance of MW/a                       Lo CF MW/a
    Ripple Control                         49               49                                   98

4.2.4 Industrial and Mining Load Management (IMLM)

The pre-integration study data shows that for IMLM there is a potential 600 MW of controllable load that can be
developed, assuming the resource is operated at high capacity factor.

A summary of the first-round screening results for ICLM is given in Table 7

Table 7: Industrial and mining load management

    Options                            Hi CF MW/a              Balance of MW/a               Lo CF MW/a
    INDUSTRIAL/ MINING LOAD                40.9                     40.9                         82

Under system emergencies the resource can be operated at low capacity factor. The low capacity factor
figures assume that the resource is required for 2 hours per day.

4.3 Uncertainty and Risk Associated with DSM

The DSM team has had to come to grips with the reality that it is developing DSM plans for a fast-changing
energy market. Many of the decisions about the future structures of the Electricity Supply Industry (ESI) in
     National Integrated Resource Plan 2




     South Africa, which have yet to be made, may impact on DSM. Utilities will have no control over the way the
     industry will develop in future.

     To deal with uncertainty about the future, the DSM team has to make certain assumptions regarding the
     conditions for sustainable demand-side resources to exist. The DSM programmes assume that the
     governance authorities in consensus will set up the specified conditions for a DSM resource market with key
     stakeholders and energy consumers.

     Risk is closely linked to a lack of dependable information. The Demand-side plans make many assumptions on
     important factors that affect market penetration. The DSM assumptions and resource plans are based on
     targets that are believed to be realistic. The demand-side plan is exposed to risk if any of the assumptions are
     proven to be unrealistic. It is prudent to firm up on much needed information about the market potential for
     demand-side resources.

     4.4 Priorities for the Development of Demand-Side Resources

     These demand-side plans are based on the assumption that the SA market will be willing and able to transform
     towards a preference for the development of demand-side resources. Historically the preference has been to
     develop supply-side resources, so the SA market is currently not richly endowed with the resources and skills
     required to deliver demand-side resources on a large scale. A prime issue of concern is the fact that DSM is
     often seen from a utility perspective as a loss leader in that their core business is in selling electricity. This is
     especially relevant to energy efficiency programmes and it is imperative that policy and legislation be adopted
     in future to address this problem. Another major problem is in sustainability. For example in the Table View load
     management pilot study, previously about 8-10MW could be achieved through geyser control. Recent tests
     show only 4MW can currently be switched off. Customers are bypassing the system because they see no
     personal benefit as no time-of-use tariff is in place for individual residential customers.
24
                                                                                               Reference Case




5.SUPPLY-SIDE OPTIONS

Technology options are compared on the basis of discounted cash flows (total capital & operating costs) over
the option lifetime using 10% net discount rate as dictated by the ARC of the NER. The levelised cost per unit
output of the option is obtained by dividing the present value by the total discounted lifetime generation.
Levelised costs are calculated as a function of plant load factors (screening curves) to illustrate the relation
between the levelised cost and plant load factor.

5.1 Eskom System - Existing and Committed Capacity

The current (2003) total net base load capacity in operation of the Eskom system is given in Appendix 1 to this
report and is 33 871 MWe. The total net peaking capacity in operation is 2319 MWe. This includes Acacia and
Port Rex gas turbines, Drakensberg and Palmiet pumped storage, Gariep and Vanderkloof hydro plants.

Eskom has a long-term contract to purchase power from Cahora Bassa hydro plant in Mozambique. Available
net capacity from Cahora Bassa is at present 912 MWe (after losses). A further 369 MWe (after losses) is
committed from 2004. Eskom is importing about 100 MWe of peaking capacity from Zesco (Zambia) via
Zimbabwe and about 110 MWe from the DRC. These latter are excluded from the study as Eskom has been a
net exporter of similar proportions to neighbouring states. (I.e. as these imports and exports are effectively
equivalent the exports are not included in the load forecast nor are the imports included in the capacity plan).

5.2 Non-Eskom System - Existing Capacity

The following options are considered in the plan:

Appendix 1 also gives the net capacity of non-Eskom generating plant. These have been aggregated into
                                                                                                                         25
blocks of similar capacity units for modelling purposes as follows:

&   Munic 1: consisting 12x60 MW sent out coal-fired capacity
&   Munic 2: consisting 21x30 MW sent out coal-fired capacity
&   Sasol: consisting 12x60 MW sent out coal-fired capacity
&   Steenbras pumped storage plant: 3x60 MW sent out
&   Mini Hydro: 1x65 MW sent out
&   Minic OCGT: 6x50 MW sent out

The details of these plant and capacities are given in Appendix 1.

In addition to the Eskom and non-Eskom supply capacity there is a total 1510 MW interruptible supply capacity
included in the plan. The interruptible load accepted for inclusion in the plan, although based on current
contracts, does not necessarily represent the existing contracts. It is assumed that some level of interruptible
capacity, either in the form of contracts for reserve or as a demand-side bidding option, will exist in the future. In
total the existing capacity assumed on the National system in 2003 amounts to 41378 MWe sent out.

5.3 Return to Service of Eskom Mothballed Plant (Simunye)

These are the stations that were put into storage during the period of high excess capacity on the Eskom
system: Camden (1520 MW), Grootvlei (1130 MW) and Komati (909 MW). These capacities are on a sent out
basis. These stations all use PF coal. When refurbished, Camden is anticipated to run as base load, whereas
Grootvlei and Komati would most probably be two-shifted i.e. operating at an annual load factor of less than
30%.

Results of a pre-feasibility study initiated and carried out by Eskom indicate that repowering some of Eskom's
old plant with Fluidised Bed Combustion (FBC) technology (such as at Komati) could be an option in
     National Integrated Resource Plan 2




     comparison with other future base load plant. The remaining half of Komati (456 MW) could be re-powered
     with FBC technology to operate as base load.

     5.4 New Supply-Side Options

     Table 8 below summarises the data for all the new supply-side technologies considered in the plan. Detailed
     description of the cost and performance of the considered supply-side options is shown in Appendix 3. The
     ARC, applying the approved selection criteria, determined the inclusion of the supply side technologies in the
     plan. The capital costs are averages of the international data evaluated. Operating and Maintenance costs
     shown are averages from the international literature, after they have been adjusted for South African labour
     conditions. Appendix 3 contains additional information about the supply-side technologies, including costs,
     performance parameters, lead times and data statistics.

     A transmission benefit has been given to stations built in the Cape close to the load centres. This is to account
     for the line losses that occur when transmitting electricity from an equivalent station in the Highveld and for
     strengthening the grid along transmission lines.

     PBMR data is given for both the initial multi-module and, after several multi-modules have been deployed, a
     cheaper multi-module (costs are likely to decrease, benefiting from technology learning).

     Adding units to existing coal plants and co-generation options have not been considered in the plan.

     5.4.1 New Pulverised Fuel (PF) Coal-Fired Stations

     In the absence of more cost competitive options, coal-fired plants are likely to remain the backbone of the
     electricity supply industry in South Africa for a long time. Their proven design, the vast operating experience
26
     gained locally, their wide availability, reliability and relatively low cost make them an attractive option for base
                                                                                              1
     load generation. This is often the case in other parts of the world, such as China , where low-cost coal is
     available. The future stations considered consist of 6 units, each of 700 MW installed capacity, with dry-
     cooling (or hybrid) systems. Coal stations can operate with or without flue gas desulphurisation (FGD) to
     reduce emissions but only the option with FGD was considered for the plan because of the decision to adhere
     to World Bank emission standards. Detailed information on the cost and performance of the PF plants is
     shown in Appendix 3.4

     5.4.2 New Gas-Fired Plant

     Potentially Namibia, Mozambique or Angola could supply Gas. New gas-fired plants considered are open
     cycle gas turbines (OCGT) and combined cycle gas turbines (CCGT). Open cycle gas turbines, by virtue of
     their low efficiencies and resultant high cost of fuel when operating at low load factors are considered for
     peaking options only. Combined cycle gas turbines with much improved efficiency levels are regarded as non-
     peaking technologies due to the likely constraints on gas contracts (take or pay) although they could
     technically follow load.

     The OCGT stations considered have two 120 MW units. The CCGT stations considered have 5 x 400 MW
     units.

     If OCGT turbines were built, and in the future a gas network was developed, it is likely that OCGT plants would
     be converted to CCGT plants, which would increase the capacity and efficiency of the stations. This has not
     been taken into account in the plan.



     1 Future Implications of China’s Energy-Technology Choices, July 2001
                                                                                           Reference Case




Fuels considered for OCGT plants are kerosene, LPG, local syngas and LNG. Fuels considered for CCGT are
pipeline gas and LNG. Fuel prices for the gas options are given for all the fuels considered for each technology.

Detailed information on the cost and performance of the OCGT and CCGT plants is shown in Appendix 3.2 and
3.3 respectively.

5.4.3 New Pumped Storage Schemes

Two pumped-storage schemes have been considered in the plan. Braamhoek (which has already received a
record of decision (ROD)) and a generic scheme which would follow Braamhoek. Braamhoek consists of four
333 MW units, while the generic option consists of three 333 MW units. It must be noted that the costing of
pumped storage schemes is site specific. Additional details on the pumped storage plants are provided in
Appendix 3.8.

5.4.4 Greenfield Fluidised Bed Combustion

The fluidised bed combustion (FBC) stations considered have an installed capacity of 500 MW. They are
comprised of two 250 MW boilers and a 500 MW turbine, and have a 35 year lifetime. The price of duff coal will
have to be negotiated between the station and the mines. Additional information on the FBC is provided in
Appendix 3.5.




                                                                                                                    27
                                                                           28
Table 8: Summary of cost and performance data of new supply-side options
                                                                                National Integrated Resource Plan 2
                                                                                                Reference Case




5.4.5 Conventional Nuclear (Advanced Light Water Reactor (ALWR))

The new nuclear plant considered is an advanced light water reactor. It consists of two 900 MW units and will
be built at the Koeberg (existing nuclear) site near Cape Town.

The ALWRs are based on existing light water reactors but are designed for simplicity and ease of construction
and maintenance. They have a considerable degree of inherent safety (safety built in to the design rather than
reliant on active safety mechanisms). Examples of such reactors are the Westinghouse AP1000 and the
General Electric ABWR. There are currently no ALWR plants in operation. Additional information on the ALWR
is provided in Appendix 3.6.

5.4.6 Research projects/programs

In addition to the “mainstream” supply-side options listed above, the following are amongst the technologies
that are being researched and are considered in the screening curve analysis:

5.4.6.1 Wind energy
There are several areas in South Africa, particularly in the coastal regions, which have been identified as
having good potential for wind power. The proposed wind farm consists of twenty (1 MW) wind turbines.
There is little detailed data available for specific sites in South Africa. Additional details on wind turbines are
provided in Appendix 3.9.

5.4.6.2 Solar thermal
The Solar Thermal power station would be built near Upington in the Northern Cape. It would have three 110
MW units and the capacity for storage. It is therefore regarded as a dispatchable station in the plan. Additional
details on solar thermal power stations are provided in Appendix 3.10.
                                                                                                                          29

5.4.6.3 PBMR
If the necessary approvals are given and a commercial decision is taken to build one, the first full-sized
demonstration unit should be on line by 2008. If successful, further units will be built with the aim of producing
a power station consisting of eight units, each unit with an installed capacity of 170 MW. Additional details on
the PBMR plant are provided in Appendix 3.7

5.4.7 Imported Hydro

There is a large potential for South Africa to import hydro-electricity from the rest of Africa. A promising site is at
Mepanda Uncua in Mozambique about 60 km downstream of the existing Cahora Bassa Power Station on the
Zambezi River. In the first stage of this project the installed capacity at the HV terminals would be 1300 MW.

The firm power capacity at 95% availability is 827 MWe. Additional details on the imported hydro option are
provided in Appendix 3.11.

5.5 Screening Curves

To evaluate the cost of generation from new options, a screening curve analysis was undertaken. A screening
curve evaluates the cost of generating electricity at different average levels of production or load factors over
the life of the plant. This gives an indication of the economics of running the plant at these load factors.

The following screening curves have been drawn from the levelised capital, O&M and fuel costs for the various
technologies at different load factors. The analysis has been split into two sections: non-peaking stations
(plants expected to run at high load factors) and peaking stations (plants expected to run at low load factors).

Non-peaking plants operate as base load when their operating (including fuel) costs are lower than peaking
     National Integrated Resource Plan 2




     stations. There are some exceptions where stations with high operating costs are included as must run plants
     by virtue of fixed price contracts or technical constraints. CCGT options are examples whereby the gas
     contracts are placed on the basis of firm commitments to take minimum quantities by virtue of purchasing
     power agreements. These plants can be considered as inflexible options to system dispatch and construed as
     must run options.

     Peaking stations run at low load factors by virtue of their high operating costs generally have lower capital and
     fixed operating expenses than non-peaking stations. The ranges reported in the screening curves reflect the
     deviation in the data collected from international literature.

     Figure 7 illustrates the life cycle costs to build and operate the base load plants analysed in this report.

                        1500.00



                        1300.00



                        1100.00



                         900.00
                R/MWh




                                                                                                               Solar Thermal
                                                                 Conventional Nuclear
                         700.00
                                                                                         Imported Hydro
                                        CF with FGD
                                                                           PBMR
                         500.00
                                                                                                                               CCGT LNG
                                                                                                                                    CCGT Pipe

                         300.00

                                                Greenfield FBC
30                       100.00
                                                                                                                                                1
                                   30                 40              50                 60               70              80              90

                                                                      Load Factor (%)


     Figure 7: Life cycle levelised costs to build and operate base load plants

     Figure 8 illustrates the life cycle costs to build and operate the peaking plants analysed in this report.

                            4000



                            3500

                                                           Gas Turbine (LNG)
                            3000

                                                            Gas Turbine (Local Syngas)
                            2500
                    R/MWh




                            2000


                            1500

                                                                                    Wind (Non-dispatchable)
                            1000
                                                                                                               Pumped Storage (Generic)

                            500

                                                                                                               Pumped Storage (Braamhoek)
                              0
                                   0                  5               10                 15               20              25              30

                                                                            Load Factor (%)

     Figure 8: Life cycle levelised costs to build and operate peaking plants
                                                                                              Reference Case




The open cycle gas turbine levelised cost curves cross the Braamhoek Pumped Storage scheme cost curve
near 6% load factor. They cross the Generic Pumped Storage Scheme near 10% load factor. (Note: Wind
turbines are considered non-dispatchable options operating at low load factor due to paucity of wind resources
in South Africa).

Although some of the technologies overlap in the screening curves shown above, there is a significant
statistical difference between the capital costs of the various technologies at a 95 % confidence interval (see
Appendix 3). The screening curves show that, once capital costs are combined with O&M and fuel costs, the
differences between the total levelised costs of certain technologies are not statistically significant (they
overlap) at certain load factors.

5.6 Other: Environmental, Externalities, Transmission Expansion

Certain sites and technologies will reduce the need to strengthen the electricity grid as they will be built close to
load centres that are currently far from existing generation. This is taken into account in the levelised cost
curves and is described further in the detailed descriptions of the technologies.

All the new coal fired generating options considered meet World Bank emissions standards. Conventional coal
stations will be equipped with flue gas desulphurisation. Due to the limited water resources all new coal
stations are dry cooled.



6.INTEGRATION AND SENSITIVITY ANALYSIS

Resources Planning has to deal with multiple conflicting objectives, a broad range of options and pervasive
uncertainty. In this context it has to do with dominance and finding plans representing reasonable trade off
                                                                                                                        31
amongst various conflicting objectives. If, for example, the objective were to minimise cost then those plans
with higher cost would be considered inferior to those with lower cost. Lower cost plans would therefore
dominate the higher cost plans.

The focus of the traditional resources planning approach is to provide a robust (and flexible) plan determined
on the basis of an analysis of risk from either under or over investment within a range of forecast demand for
electricity over a specified planning horizon. (A plan that is 100% robust is one which lies in the decision set for
all futures and would lead to no regret). Flexibility is of equal importance to robustness, if not more important in
the uncertain planning environment. Flexibility can be defined as the ability to modify a resource plan without
significantly degrading system reliability and economics, in response to actual data that have deviated from
the forecast data.

The objective is therefore to provide long-term strategic projections of supply-side and demand-side options
added or deducted from the system over a specified planning horizon taking into account investment risk and
option lead time under a range of different planning scenarios.

The risks associated with this planning process have been outlined previously but predominantly focus on
those uncertainties, which will influence decisions on:

& The timing for new supply-side and demand-side initiatives
& The plant mix
     National Integrated Resource Plan 2




     Due to time constraints, this round of the NIRP focuses on the development of a reference plan only. Elements
     of uncertainty and risk are taken into account by imposing constraints on plant operating reserve. In terms of
     this process two plans have been developed for this round of the NIRP, namely:

     & Reference Plan - constrained to 10% reserve margin
     & Alternative Plan 1 - constrained to 15% reserve margin
     & Alternative Plan 2 - sensitivity study conducted on the reference plan in order to determine the impact of
       excluding Interruptible supply agreements from the plan
     & Alternative Plan 3 (optimum reserve margin) - non-constrained plan providing as an output the optimum RM
       based on trade-off between the cost of reliability of supply and the cost of un-served energy to the
       consumer.

     The details of these plans are given in Appendix 2 to this report.

     6.1 Reference Plan

     The reference plan has a minimum constraint of 10% on the reserve margin as proxy to take account of the risk
     elements detailed above. In addition to the constraint on the reserve margin, two further constraints are
     imposed.

     The first is in terms of the extent of un-served energy that would be allowed. In this case the un-served energy
     is constrained to a maximum of 0.011% of total energy demand in any specific year. This is reflective of
     historical values utilised in Eskom in previous years in order to ensure a loss of load expectation (LOLE) of 22
     hours is not exceeded in any given year.

     The second is in terms of the number and capacities of Open Cycle Gas turbines accommodated in the plan
32
     because of perceived limitations in the gas supply and the possibility of these options becoming stranded
     assets in the event more interruptible supply options are implemented in the future. It was decided in this
     instance to consider a maximum number of 10 stations limited to 2x120 MW each could be built in the twenty-
     year planning horizon.

     The capacity outlook for this plan is illustrated in Figure 9 below and the detailed results are given in Appendix
     2 to this report.


                     Mothballed             Coal-Fired        FBC         GT                Pumped Storage                        Demand-side Options

     YR      Cam (PF) Gr'tvlei    Kom      PF (1)   PF (2)     Green-    OCGT      PS (A)     PS (B)   PS (C)   PS (D)     CEE      IMEE     IMLM      REE      RLM       Syst Res
                       (PF)       (PF)                       field FBC


             Committed                                                            Committed                              CommittedCommittedCommitted CommittedCommitted
     2003                                                                Decide                                             14       16       41        32       49         24%
     2004                                                     Decide                          Decide                        14       16       41        32       49         21%
     2005      380       Decide                                                                        Decide               14       16       41        32       49         18%
     2006      380                Decide                                                                                    14       16       41        32       49         15%
     2007      380                         Decide                                                               Decide      14       16       41        32       49         14%
     2008      380                                  Decide                240                                               14       16       41        32       49         12%
     2009                 377                                             480                                               14       16       41        32       49         12%
     2010                 377      303                                    480                                               14       16       41        32       49         12%
     2011                 377      303                                    480                                               14       16       41        32       49         10%
     2012                          303                         466        480       333                                                                                     11%
     2013                                                                 240       999                                                                                     11%
     2014                                                      466                             666                                                                          11%
     2015                                                                                      333      333                                                                 10%
     2016                                                      466                                      666                                                                 11%
     2017                                                      932                                               333                                                        11%
     2018                                                      466                                               333                                                        11%
     2019                                   642                                                                                                                             10%
     2020                                   642     1284                                                                                                                    12%
     2021                                  1284     1284                                                                                                                    12%
     2022                                  1284     1284                                                                                                                    11%
     TOTAL     1520      1130      909     3852     3852       2796      2400       1332       999      999      666       124       145      368      292       441




     Figure 9: Reference plan (10% Reserve Margin)
                                                                                                                                                        Reference Case




These results indicate a commitment to ensuring the immediate return to service of Camden power station in
the suite of Eskom Simunye mothballed plants. In addition an immediate decision is required to build at least
720 MW of OCGT plant for commissioning and commercial service by 2009. Furthermore, activities
(monitoring and evaluation) need to be implemented immediately to ensure Braamhoek pumped storage
remains on track for commercial service in 2012, and also that the DSM programmes achieve their targets over
the ensuing years.

6.2 Alternative Plan 1 to Reference Plan

There is a concern that carrying a 10% reserve margin on the moderate national forecast is too conservative in
the light that most international utilities are reluctant to carry a RM below 15%. An additional plan has been
developed which is constrained to a reserve margin of 15%. This plan is detailed in Figure 10 below.



             Mothballed               Coal-Fired FBC                      GAS                    Pumped Storage                              Demand-side Options

 YR     Cam (PF) Gr'tvle     Kom      PF (1)   PF (2)   Green-   CCGT CCGT (2)     OCGT      PS (A)     PS (B) PS (C)     PS (D)     CEE       IMEE     IMLM      REE       RLM     Syst
                 i (PF)      (PF)                        field    (1)                                                                                                               Res
                                                         FBC

        Committed                                                         Decide            Committed                              Committed Committed Committed Committed Committed
 2003               Decide                              Decide   Decide            Decide                                             14        16        41        32        49     24%
 2004                        Decide   Decide                                                            Decide                        14        16        41        32        49     21%
 2005      380                                                                                                                        14        16        41       32        49     18%
 2006      380                                                                                                   Decide               14        16        41       32        49     15%
 2007      380       188                                                                                                  Decide      14        16        41       32        49     14%
 2008      380       377      202              Decide                               480                                               14        16        41       32        49     15%
 2009                377      303                                                   480                                               14        16        41       32        49     15%
 2010                         202                        466               387      240                                               14        16        41       32        49     15%
 2011                                                    466      774                                                                 14        16        41       32        49     14%
 2012                188                                         1161                          333                                                                                  15%
 2013                         202                                                              999                                                                                  15%
 2014                                                    932                                             333                                                                        15%
 2015                                                                                                    666                                                                        14%
 2016                                  642                                                                        333                                                               15%
 2017
 2018                                  642
                                                         466
                                                         466
                                                                                                                  666      333                                                      15%
                                                                                                                                                                                    15%
                                                                                                                                                                                           33
 2019                                                                                                                      333                                                      13%
 2020                                          1284                                                                        333                                                      15%
 2021                                 1284     1284                                                                                                                                 15%
 2022                                 1284     1284                        387                                                                                                      14%
TOTAL     1520      1130      909     3852     3852     2796     1935      774     1200       1332       999      999      999        124      145       368       292      441




Figure 10: Alternative Plan 1 to reference plan (15% Reserve Margin)

Alternative 1 requires additional capacity of 1842 MW over the planning horizon, at the additional cost of 15
552 million Rand, in order to maintain the higher RM of 15%. The plan also requires acceleration of the return to
service of the mothballed plants and base load PF plant.

6.3 Alternative Plan 2 (Sensitivity Study)

A prime concern of ARC members is the uncertainty surrounding the sustainability of interruptible supply
options. In order to determine the impact of interruptible supply options, a further plan (Alternative Plan 2) was
developed based on the primary assumptions and constraints used to develop the reference plan but
excluding all interruptible supply options from the plan. This plan is detailed in Figure 11 below.

This plan requires significant acceleration of the return to service programme of the mothballed plants in the
near term. In this, the plan is at risk, because the return to service programme of the Simunye plant, even with
the best endeavours, might be unable to be accelerated sufficiently to meet the demand in the near term
(2006).

There is a PV increase in costs of this plan when compared to the reference plan of 2 997 Million Rand. The
higher cost of this plan is largely due to the fact that there is a penalty paid in interrupting customers in the early
years at a high cost of un-served energy of 20 660 R/MWh.
     National Integrated Resource Plan 2




                      Mothballed                Coal-Fired        FBC        Gas                Pumped Storage                         Demand-side Options

      YR     Cam (PF)    Gr'tvlei   Kom (PF)   PF (1)   PF (2)     Green-    OCGT      PS (A)     PS (B)   PS (C)   PS (D)     CEE       IMEE     IMLM      REE       RLM     Syst
                          (PF)                                   field FBC                                                                                                    Res


             Committed                                                                Committed                              Committed Committed Committed Committed Committed
      2003               Decide      Decide                       Decide     Decide                                             14        16        41        32        49     19%
      2004                                                                                        Decide   Decide               14        16        41        32        49     16%
      2005     380                                                                                                              14        16        41       32        49     13%
      2006     380                             Decide                                                               Decide      14        16        41       32        49     10%
      2007     380         377        101                                                                                       14        16        41       32        49     10%
      2008     380                    101               Decide                480                                               14        16        41       32        49     11%
      2009                 377        202                                     480                                               14        16        41       32        49     11%
      2010                 377        303                                     480                                               14        16        41       32        49     12%
      2011                            202                          466        480                                               14        16        41       32        49     11%
      2012                                                         466        480       333                                                                                   11%
      2013                                                                              999                                                                                   11%
      2014                                                                                         999      333                                                               11%
      2015                                                                                                  333                                                               10%
      2016                                                         466                                      333      333                                                      10%
      2017                                                         932                                               333                                                      11%
      2018                                      642                                                                  333                                                      11%
      2019                                                         466                                                                                                        10%
      2020                                      642     1284                                                                                                                  11%
      2021                                     1284     1284                                                                                                                  11%
      2022                                     1284     1284                                                                                                                  11%
     TOTAL     1520       1130        909      3852     3852       2796      2400       1332       999      999      999        124      145       368       292       441




     Figure 11: Alternative Plan 2 - sensitivity to reference plan (excludes interruptible supply options)

     6.4 Alternative Plan 3 (Optimal reserve margin)

     The issue of determining an optimal reserve margin for electricity generation in South Africa has not been
     addressed formally in any forum in the country to date. In optimal integrated resources planning studies, the
     calculation of the reserve margin is an outcome, based on the primary assumptions underpinning the
     development of the plan.
34
     It is therefore expedient to determine a reserve margin based on the outcome of the optimisation process, but
     taking into account specified risk(s) associated with key uncertainties in the primary planning assumptions.
     Such a plan will result in a different mix and timing of new plant options compared to one developed to meet a
     deterministic reliability index.

     The development of such a plan is intended as the outcome of the next round of studies aimed at producing a
     recommended set of strategies (plan) for NIRP2 based on an analysis of risk factors associated with the
     primary assumptions.

     As a starting point to this exercise, an optimal plan has been developed, based on the primary assumptions
     listed above without imposing any constraints and in particular on the reserve margin, un-served energy and
     number of peaking (OCGT) plants.

     This plan is detailed in the Appendix 2 to this report. It is intended that this plan will serve as the basis for the
     next round of studies where several different plans will be developed on the basis of meeting a number of risk
     scenarios associated with the uncertainties surrounding several of the key assumptions. In each of the plans,
     the reserve margin will be an outcome as a function of the risk associated with specified uncertainties.

     It is important to state that given certainty in the current primary assumptions, the reserve margin required
     would be of the order 4% in the long-term. This plan would however not be sufficiently flexible to meet any of
     the short or long-term risks detailed above.
                                                                                                                                             Reference Case




7.SYSTEM ANNUAL AVERAGE LONG RUN MARGINAL COST

In this report, the terms Long Run Marginal Cost (LRMC) and Long Run Incremental Cost (LRIC) are used
interchangeably. Theoretically the LRMC refers to the incremental cost of providing one additional unit of
energy to supply the increase in demand; whereas the LRIC refers to the incremental cost of providing an
additional increment of energy.

The LRMC curve is calculated by determining the difference in costs between two generation expansion plans
developed over the twenty-year planning horizon:

(1) Base case: to meet the forecast demand in electricity and,
(2) Marginal case: the forecast demand increased by an increment (500 MW in this case)

The annual difference in optimal cost for building and operating plant to meet these two plans is divided by the
difference in energy associated with each of these annual forecast demands. (I.e. the base demand and the
demand increased by the increment of 500 MW at the system annual average load factor).

The LRMC has been developed using this methodology, for the reference plan and each of the alternative
plans to the reference plan. The detailed results are given in Appendix 2 attached to this report and illustrated
graphically in Figure 12.


                                  System Marginal Cost of NIRP Ref Plan at 2003 prices (10% NDR)
                   280.0

                   260.0

                   240.0
                                                                                                                                                                            35
                   220.0

                   200.0
   LRMC in R/MWh




                   180.0

                   160.0

                   140.0                                                                            Alt Plan 2 (Ref excluding interruptible loads , (R/MWh)

                   120.0
                                                                                                    Alt Plan 1 (15% RM) (R/MWh)
                   100.0

                    80.0                                                                            Ref Plan (R/MWh)

                    60.0

                    40.0

                    20.0

                     0.0
                           2003   2004   2005   2006   2007   2008   2009   2010   2011   2012   2013   2014   2015   2016    2017   2018    2019   2020      2021   2022

                                                                                          Years

Figure 12: Annual average long run marginal costs of plans

Of particular significance in these results are the high values of the LRMC in the early years associated with the
Alternative 2 excluding interruptible supply options. This is because the Simunye mothballed plant cannot be
returned to service in time to avoid interrupting customers at a high cost to the economy.
     National Integrated Resource Plan 2




     8.CONCLUSIONS

     The main conclusions drawn from the NIRP2 reference case study could be summarised as follows:

     1) Options for diversification are insufficient to meet all of the forecast demand for electricity over the next 20-
         year planning horizon. Coal-fired options are still required for expansion during this period. For
         environmental benefit it is imperative to continue with efforts to reduce the costs of implementing clean
         coal technologies and improve the efficiency of coal-fired plants;
     2) Base load plants are required for commercial operation from 2010. Base load options competing, include;
         Pulverised Fuel Coal-fired (PF); Fluidised Bed Combustion (FBC); Combined Cycle Gas Turbine (CCGT).
         Given the cost and performance data used in the plan these options are broadly comparable, at 10% net
         discount rate, if regional siting and transmission benefits are included.
     3) At the current assumed cost of capital (10% net discount rate) and after returning the Eskom mothballed
         plant to service, fluidised bed combustion technologies are South Africa's most economic option, followed
         by investment in coal-fired plant. This in turn is followed by importing gas / LNG for CCGT plant in the
         Cape. The studies also show that after taking the avoided cost of transmission into account (including
         losses) that is not sufficient to make the CCGT plant competitive with coal;
     4) It will be difficult to justify diversification on an economic basis, unless penalties for not doing so are
         included in future analyses. As the cost for diversification is becoming increasingly more expensive these
         penalties (or opportunities for emissions trading) will need to be substantial to offset the economic benefits
         of remaining with coal;
     5) The NIRP plans are based on attainment and sustainability of the DSM targets , power plants availability,
         imports and interruptible loads;
     6) The NIRP plans indicate that 920 MW OCGT peak load plants must begin commissioning from 2008;
     7) Maintaining a higher reserve margin of 15% over the planning period will require acceleration of the RTS of
         the mothballed plants and coal-fired options together with commissioning of additional base load capacity
36
         (CCGT);
     8) Diversified options are new technologies to South Africa. If these options for whatever reason are not able
         to be implemented it will mean a return to a dependency on new pulverised coal-fired plants earlier than
         shown in these plans;
     9) There are supply options that have not been considered such as co-generation in industry, converting
         OCGT to CCGT, adding units onto existing power stations and new imports resulting from the
         development of Electricity Supply in the Southern African region;
     10) Should interruptible supply and / or OCGT capacity not be implemented this will significantly advance
         new base-load capacity;
                                          Risk & Sensitivity Analysis




National Integrated Resource Plan 2




                   Stage 2

    Risk & Sensitivity Analysis
                                                                        37




                    COMPILED BY



          The National Electricity Regulator



        ISEP Eskom (Resources and Strategy)



             Energy Research Institute

              University Of Cape Town
     National Integrated Resource Plan 2




                                    NIRP 2 STAGE 2:
                            RISK AND UNCERTAINTY ANALYSIS

                                        TABLE OF CONTENTS

     1   INTRODUCTION                                                            40


     2   SCOPE OF STUDY                                                          40
         2.1 APPROPRIATE RESERVE MARGIN                                          40
         2.2 IDENTIFIED RISKS AND UNCERTAINTIES                                  41
         2.3 IMPACT OF THE SELECTED RISK PORTFOLIOS                              42


     3   MODELED UNCERTAINTIES                                                   43
         3.1 FORECAST                                                            43
            3.1.1   Background                                                   43
            3.1.2   Compilation of the energy component of the higher forecast   43
            3.1.3   Compilation of the demand component of the higher forecast   44
            3.1.4   Forecast Data                                                46
         3.2 AVAILABILITY OF PLANT                                               47
         3.3 SUSTAINABILITY OF EXISTING INTERRUPTIBLE LOADS                      48
38
         3.4 SUSTAINABILITY OF EXISTING CAPACITY                                 48


     4   SCENARIO DEVELOPMENT                                                    49


     5   RANKING OF PLANS/CANDIDATE PLANS                                        51
         5.1 METHODOLOGY                                                         51
         5.2 TOTAL PROBABILITY-WEIGHTED COST                                     52
         5.3 RELIABILITY ASSESSMENT                                              53
         5.4 EMISSIONS AND DIVERSIFICATION ASSESSMENT                            54


     6   PREFERRED RESOURCE PLAN                                                 56
         6.1 PREFERRED PLAN 14                                                   56


     7   SENSITIVITY ANALYSIS                                                    57
         7.1 COMPARISON OF DATA CHANGES TO NIRP2 REFERENCE PLAN                  57
         7.2 THE IMPACT OF NET DISCOUNT RATE ON OPTIMUM CHOICE OF TECHNOLOGY     59
         7.3 RENEWABLE ENERGY TECHNOLOGIES                                       62


     8   CONCLUSIONS                                                             64


     9   APPENDICES                                                              64
                                                                         Risk & Sensitivity Analysis




LIST OF FIGURES

Figure 1.   Impact of Modelled Uncertainties                                                  43
Figure 2.   Energy Forecasts                                                                  44
Figure 3.   Maximum Demand Forecasts                                                          45
Figure 4.   The proportion of consumption increases over the reference case.                  47
Figure 5.   Variability of total cost.                                                        53
Figure 6.   Reserve Margin under own future.                                                  53
Figure 7.   Reserve Margin under A1 future.                                                   54
Figure 8.   Ranking of Plans                                                                  56
Figure 9.   Capex Costs for Base Options                                                      58
Figure 10.  O&M Costs for NIRP 2 Reference Plan and NIRP 2 Risk & Sensitivity Plan            58
Figure 11.  Fuel Costs for NIRP 2 Reference Plan and NIRP 2 Risk & Sensitivity Plan           59
Figure 12.  Levelised costs of base-load plants versus Net Discount Rate for the
            NIRP 2 Reference Plan                                                             61
Figure 13. Levelised costs of base-load plants versus Net Discount Rate for the
            NIRP 2 Risk & Sensitivity analysis                                                61
Figure 14. Preferred Plan 14 with Renewables                                                  63
Figure 15. Reduction in environment emissions in preferred Plan 14 with renewables.           63


LIST OF TABLES

Table 1.     Forecasting Data                                                                 46
Table 2.     Electricity demand in the high growth scenario (GWh)                             47
Table 3.     Availabilities of new plant.                                                     48
Table 4.     Probabilities associated with primary assumptions.                               49
                                                                                                       39
Table 5.     Scenarios and their probabilities.                                               50
Table 6.     PWC Ranking of the Plans                                                         52
Table 7.     Comparison of Candidate Plans                                                    55
Table 8.     Preferred Plan 14.                                                               57
Table 9.     Levelised costs of candidates for selection in the NIRP2 Reference Plan          60
Table 10.    Levelised costs of candidates for selection for the NIRP2 Risk and Sensitivity
             analysis                                                                         60
Table 11.    Additional Cost of Renewable Technologies                                        62
     National Integrated Resource Plan 2




     1.Introduction

     This Report is an update of the NIRP2 Reference Plan. This work was commissioned following comments
     received from both the Public and the NER Advisory Review Committee (ARC) after the publication of the
     NIRP2 Reference Plan.

     This Report should be read in conjunction with that of the NIRP2 Reference Plan since it uses as basis the
     primary planning assumptions (including the data base) of the reference plan unless stated otherwise. Where
     data revisions have been made to the reference data, these are explained in the Report. The list of
     abbreviations and bibliography of the NIRP2 Reference Plan apply to this Report.

     The objective of this study is to provide a robust plan for NIRP2 based on an appropriate net reserve margin
     above the expected annual maximum demand for electricity assumed in the Reference plan. This reserve
     margin is calculated as an outcome based on an analysis of specified probability weighted risk portfolios.

     A complete listing of comments from the Advisory Review Committee (ARC) of the NER plus comments from
     individual members of the Public concerning the NIRP2 Reference Plan are contained in Appendix A attached
     to this Report. The most important aspects leading to the need to carry out a risk and sensitivity analysis are
     summarized below:



     & The use of deterministic reserve capacity margin as a substitute for risks, without explicitly identifying the
       mitigated risks associated with the planning uncertainties (plant availability, load forecast deviations, fuel
       availability, interruptible supplies etc), is not ideal planning method;

     & The appropriate reserve margin should be an outcome of a risk and uncertainty analyses, characteristics of
40
       the existing and future plants and costs. The reserve margin should be an outcome of the modeling process
       based on sets of defined assumptions for different scenarios;

     & The impact of the discount rate on the optimum choice of technology should be considered in the analyses;

     & Considerations regarding resource alternatives under high growth scenarios should be included;

     & The costs of some generation sources such as FBC and imported hydro, as well as the cost of the coal are
       underestimated;

     & The Inclusion of Government renewable energy targets.

     2.Scope of study

     The major concern expressed by the ARC and members of the Public were expressed in the many comments
     regarding an appropriate reserve margin to cater for risk. This report therefore focuses on developing a robust
     plan with an appropriate reserve margin accounting for risk. It also takes into account as far as possible
     attributes other than least cost such as environmental emissions, flexibility and a diversified portfolio of
     options.

     2.1 Appropriate Reserve Margin

     The reference plan used two levels of deterministic net reserve margin (10% and 15%) to take account of risk.
     The use of a deterministic net reserve margin to address risk, without explicitly identifying and quantifying the
     risks associated with the primary planning uncertainties (e.g. plant availability, load forecast deviations, fuel
     availability, interruptible supplies etc), is not an ideal planning method.
                                                                                 Risk & Sensitivity Analysis




An appropriate net reserve margin should consist of a plant mix which matches the risk impacting the capacity
shortage; E.g. If Cahora Bassa units became unavailable, base-load capacity would be required to meet the
shortfall. If interruptible capacity became unavailable, peaking plant would be required.

Following discussions at the Advisory Review Committee and resulting from the various comments received, it
was decided to adopt the following approach to determine a appropriate net reserve margin for stage2 of
NIRP2.

& Develop scenarios and assign probabilities
& Develop plans for each scenario
& Carry out trade-off analyses (I.e. Test each plan under each scenario and calculate the total probability
  weighted cost (cost of supply plus cost of non-supply) of each plan multiplied by the probability of
  occurrence of the scenario
& Select the most robust (flexible) plan under agreed attributes
Evaluate the net reserve margin of this plan in respect of the expected forecast

2.2 Identified risks and uncertainties

Following the NIRP2 Reference plan, the following short-term and long-term risks and uncertainties were
identified for inclusion in the risk and uncertainty analysis.

&   Plant failure leading to longer than expected plant outage;
&   Unavailability of municipal / Eskom / imported generating capacity;
&   Degree of market penetration of DSM and maintaining current level of interruptible loads;
&   Unexpected decrease / increase, spurious or sustained, of electricity demand;
&   Changes to the load shape associated with the forecast electricity demand;
                                                                                                                    41
&   Unexpected decommissioning / de-rating of existing generating capacity
&   Uncertain and prolonged lead times for building new plant;
&   Project slippage
&   Inclusion of co-generation options;
&   Embargoes on nuclear energy;
&   Shortage of skills to maintain and grow the system;
&   Other energy forms displacing electricity in the energy market;
&   Revolutionary technologies coming on the scene and stranding existing assets;
&   Internalisation of externalities, such as introduction of a carbon tax and environmental levy;
&   Plant life expectations not met;
&   Deterioration in credit rating, exchange rates etc. resulting in a higher cost of capital;
&   Electricity supply and sales contracts (import and export contracts) being reneged upon;
&   Effect of AIDS on the electricity market.
&   Drought and Floods

Many of these risks and uncertainties are inter-dependent. By aggregating some risk events into independent
risk portfolios it is possible to provide a net reserve margin to cater for a number of risk events under the
umbrella of one independent risk portfolio. This can best be illustrated by the following examples:

Drought: The impact of drought on Hydro imports (Cahora Bassa) can be taken into account by reducing the
capacity of this option. The level of capacity reduction from Cahora Bassa can also be increased to cater for
reduced imports due to increasing capacity withdrawals to neighbouring states (notwithstanding current
agreements) in the future. Although strictly speaking these two events can be construed as separate events
each with their own probability of occurrence, for the purposes of this report the level of capacity reduction is
set to accommodate both with the same probability.

No account is taken of the impact of drought on existing Hydro capacity because the amount of such Hydro in
     National Integrated Resource Plan 2




     this country is very low. Drought also impacts both cooling water supplies of existing and new coal-fired plants.
     This is mitigated to some extent by the fact that all new inland base plants are considered to be dry-cooled (or in
     the least dry-cooled with wet assistance at high ambient temperatures). Those located at the coast use once
     through sea water cooling.

     Flood: This impacts the availability of fuel supplies to the coal-fired plants, mitigated to some extent by those
     stations with open cast mines (these are more likely to be impacted) carrying higher stock levels. From the
     point of view of imported Hydro, a substantial energy loss could be averted in times of flood, through
     implementing a SAPP Zambesi River Authority to ensure for example that where possible water is not let out of
     higher altitude dams (Kafua) whilst lower altitude dams (Cahora Bassa) are spilling.

     Impact of fuel cost / price: The South African Electricity Supply is heavily dependent on coal as feedstock. In
     the event that there was a major price increase in the cost of coal, the economy would be at risk. There are
     proponents to providing a diversified approach to building new generating options which may also be more
     favorable in reducing environmental and greenhouse gas emissions. This aspect is addressed in that three
     new base technologies other than conventional coal-fired plants are considered: Fluidised Bed Combustion
     (FBC), Nuclear Pebble Bed Modular Reactors, Gas Combined Cycle (CCGT) options.

     Following discussions with various experts in their specific fields, it was decided to address the following four
     risk portfolios, considered to be largely independent of each other and as having a major impact on the plant
     mix, timing and requirement for carrying reserve, namely:

     & Level of risk associated with the moderate national forecast being higher and particularly in the near-term to
       medium-term;
     & Level of risk associated with Eskom and non-Eskom generators being able to sustain high levels of plant
       availability once the system becomes stressed and the plants age;
42
     & Level of risk of Non-Eskom and Import capacity becoming unavailable or unreliable (i.e. being de-rated) and
       particularly in the near-term;
     & Availability of interruptible supply options continuing to serve as reserve capacity. Already these options are
       severely energy constrained and their capacity impact has been reduced.

     2.3 Impact of the selected risk portfolios

     Following discussions with respective experts the following levels of risk were identified for each of the risk
     portfolios:

     & National moderate Load forecast increased by 2570MW in 2007 to 5078MW in 2022 to account for higher
       GDP growth and colder weather (Base-load to Peaking)
     & Plant availability reduced from 88% to 85% due to poorer FOR (Base-load).
     & Contracted Interruptible Supply Excluded:
       > Mozal, Alusaf, Ferrochrome and other industries (Peaking)
     & Capacity Reductions:
       > Cahora Bassa capacity reduced from 1281MW by 250MW in 2004 and 500MW thereafter (Base-load)
       > Non-Eskom OCGT capacity (180MW of 270MW) assumed unavailable over the planning horizon
         (Peaking)
       > Non-Eskom coal-fired capacity (600MW of 1920MW) assumed unavailable over the planning horizon
         (Base-load)
                                                                                                                              Risk & Sensitivity Analysis




The impact of these levels of uncertainty over the planning horizon is illustrated graphically in Figure 1 below.

                                                                Impact of Modelled Uncertainties
                       10000


                       9000


                       8000


                       7000
       Capacity (MW)




                       6000
                                                                                                  Capacity impact of excluding Import and Munic capacity (MW)
                       5000
                                                                                                  Capacity Impact of excluding DSM Interruptible load (MW)

                       4000                                                                       Capacity impact of increased load forecast (MW)

                                                                                                  Capacity impact of reduced plant availability (89% to 85%) (MW)
                       3000


                       2000


                       1000


                           0
                           2003   2004   2005   2006   2007   2008   2009   2010   2011   2012   2013   2014   2015    2016    2017   2018    2019    2020   2021   2022
                                                                                            Years

Figure 1: Impact of Modelled Uncertainties

3.Modelled Uncertainties

These risk portfolios were generated on the basis of the comments made on the primary assumptions by the
ARC and in consultations with various experts. Following is a description of the levels or risk associated with
                                                                                                                                                                           43
the primary assumptions as detailed in the NIRP2 Reference Plan.

3.1 Forecast

3.1.1 Background

The NIRP2 reference plan is based on a moderate demand for electricity developed by Eskom. This forecast
used moderate GDP growth and moderate weather conditions. It was recommended by the ARC to also
develop a forecast of accelerated demand growth for the risk and uncertainty analysis, reflecting colder
weather conditions and higher GDP growth (especially in the near-term). It was not considered necessary by
the ARC to use a lower forecast than the moderate, the moderate being considered to be already reflective of a
lower than expected scenario. It was agreed that the Energy Research Centre (ERC) would develop the high
growth forecast independent of Eskom. Eskom resources would be used to transform the energy forecast into
hourly annual demand values according to Eskom sector profiles.

3.1.2 Compilation of the energy component of the higher forecast

The higher energy forecast in demand for electricity was developed by ERC based on energy forecasting
techniques incorporated in the LEAP model (SEIB 2002; Howells et al 2002; and Hughes et al, 2003), and the
Integrated Resource Planning manual of the United Nations Environment Program (UNEP 1998). It considers
a relatively detailed breakdown of electricity consuming sectors. For example, the mining sector was divided
into nine different sub-sectors, as was the industrial sector. Added to this was the commercial, transport,
agricultural and residential sectors. A relationship was developed between the outputs of these sectors and the
future Gross Domestic Product (GDP) (with the exception of the residential sector) as a basis for future
electricity demand. The relationship between GDP and sector output is time dependant with the proportional
contribution of each sector changing over time. The sector outputs considered was either value added, or
physical output.
     National Integrated Resource Plan 2




     In line with the assumptions approved by ARC, the ERC developed a high forecast which used higher than
     expected growth in the near- to medium-term in specific sectors. The growth associated with this forecast is
     higher than that of the moderate forecast of the NIRP2 Reference plan by an estimated half a percentage point
     in the short to medium term.

     When the results of the ERC energy forecast are compared to that of the Eskom model they will differ in certain
     respects. A forecast with a higher growth rate will automatically imply a different sectoral energy mix.
     Furthermore, the energy model used by the ERC makes use of sectoral data of a more aggregated nature than
     the Eskom sectoral model.

     A comparison of the energy forecasts for the previous NIRP1, NIRP2 reference case and this analysis is
     shown in the Figure 2 below.

                                                                               Energy sent out
                                            400000
             Annual energy sent out [GWh]




                                            350000


                                            300000


                                            250000

                                                                              NIRP1 Moderate growth and moderate weather
44                                          200000                            NIRP2 Moderate growth and moderate weather
                                                                              NIRP2 High growth and cold weather

                                            150000
                                                 01

                                                       03

                                                             05

                                                                   07




                                                                                11

                                                                                      13



                                                                                                  17

                                                                                                        19

                                                                                                              21

                                                                                                                    23
                                                                         09




                                                                                            15




                                                                                                                           25
                                                            20

                                                                  20




                                                                                                                   20
                                               20

                                                      20




                                                                        20

                                                                              20

                                                                                     20

                                                                                           20

                                                                                                 20

                                                                                                       20

                                                                                                             20




                                                                                                                         20




                                                                                     Year
     Figure 2:. Energy Forecasts

     3.1.3 Compilation of the demand component of the higher forecast

     The energy forecast is transformed into an hourly demand forecast (instantaneous demands averaged over
     each hour) with the HELM (Hourly Electric Load Model) computer model. The model uses:

     & Detailed sectoral load profiles developed by Eskom
     & Energy sales forecast and losses as inputs
     & Temperature as an input where the impact is determined through pre-determined formulae (E.g. for
       extremely cold weather).
     & Output is hourly maximum demands in each year over the whole planning horizon

     The following assumptions were used in developing this demand forecast:

     & The ERC energy sales forecast as described above;
     & Updated sectoral demand profiles, which were obtained from recent load research carried out by Eskom
       consultants in 2003;
     & The system load factor was constrained to deteriorate within a narrowband (from 72% in 2004 to 70% in
       2022).
                                                                                                  Risk & Sensitivity Analysis




& Based on 1996 weather conditions which is recorded as the coldest weather Nationally over the last 12 years
& No new demand side management initiatives are included in the forecast from 2003 (I.e. from the beginning
  of the planning horizon)

It is pertinent to note that the system losses as calculated by the ERC and used in this analysis are marginally
different to those calculated by Eskom. The annual maximum demands for the various NIRP studies are
compared in Figure 3below.

                                                           Maximum demand
                               65000

                               60000
         Maximum demand [MW]




                               55000

                               50000

                               45000

                               40000
                                                                     NIRP1 Moderate growth with moderate weather
                               35000
                                                                     NIRP2 Moderate growth with moderate weather
                               30000
                                                                     NIRP2 High growth with cold weather
                               25000
                                                                         13
                                   01


                                         03


                                               05


                                                     07


                                                           09


                                                                 11




                                                                                15


                                                                                      17


                                                                                             19


                                                                                                    21


                                                                                                            23


                                                                                                                  25
                                                    20


                                                          20


                                                                20


                                                                       20
                                 20


                                        20


                                              20




                                                                              20


                                                                                     20


                                                                                           20


                                                                                                   20


                                                                                                           20


                                                                                                                 20
                                                                         Year                                                   45

Figure 3:. Maximum Demand Forecasts

3.1.3.1.1 Weather

The 1996 National weather pattern was used to estimate weather impact because statistically it was found to
be the coldest over the last twelve years. The impact of this colder weather is to increase the annual peak
demand of the moderate forecast by approximately 3% in 1996 and extrapolated for the other years in the
planning horizon.
     National Integrated Resource Plan 2




     3.1.4 Forecast Data

     The relevant data supporting the Figures illustrated above are given in the Table 1 below.



                        NIRP1 (Moderate)                              NIRP2 (Moderate)                                NIRP2 (High)
                 Moderate Growth and moderate                  Moderate Growth and moderate                  Higher Growth and cold weather
               weather (Forecast energy and losses           weather (Forecast energy and losses               (Forecast energy and losses
                     determined by Eskom)                          determined by Eskom)                            determined by ERC)

      Year     NIRP1       MW       Sales   Losses Annual    NIRP2      MW       Sales   Losses   Annual   NIRP2    MW      Sales    Losses   Annual
               Moderate Increase    GWh       %    Energy    Moderate Increase   GWh       %      Energy    High Increase   GWh        %      Energy
                           per                     (GWh)     Demand per                           (GWh)    Demand   per                       (GWh)
                           Year                               (MW)      Year                                (MW)    Year
                          (MW)                                         (MW)                                        (MW)



      1999                         177689
      2000                         181908
      2001     32316               185171    9.88   205476
      2002     33283      967      188747   10.83   211665
      2003     34499     1216      194240   11.63   219794   33645               197189 9.15 217044        34624            197401     9.21   217426
      2004     35684     1185      202858   10.81   227439   34914     1269      205159 9.15 225824        36817    1274    209722     9.33   231303
      2005     36784     1100      212559    9.19   234066   36146     1232      212415 9.15 233805        38394    1238    218274     9.45   241054
      2006     37867     1083      220649    8.24   240457   37632     1486      221593 9.27 244242        39931    1466    227001     9.57   251024
      2007     38811      944      225444    8.22   245623   38528     896       226305 9.27 249423        41098     902    232602     9.69   257560
      2008     39798      987      230669    8.15   251128   39440     912       231064 9.26 254657        42166     919    237671     9.81   263523
      2009     40784      986      235125    8.51   256984   40377     937       236245 9.33 260559        43387     930    243686     9.93   270552
      2010     41835     1051      239711    8.84   262952   41389     1012      241799 9.33 266681        44574    1018    249547    10.05   277429
      2011     42864     1029      244139    9.26   269068   42342     953       247295 9.34 272762        45777     961    255450    10.17   284370
      2012     43894     1030      248886    9.53   275118   43389     1047      253220 9.41 279513        47091    1039    262064    10.29   292123
      2013     44893      999      253810    9.64   280896   44343     954       258581 9.41 285448        48275     962    267718    10.42   298859
      2014     45922     1029      258683    9.83   286883   45271     928       263829 9.42 291258        49423     936    273173    10.54   305358
      2015     46971     1049      263584   10.04   292993   46198     927       268355 9.55 296678        50520     903    278263    10.66   311465
46    2016     47980     1009      268586   10.22   299148   47034     836       273055 9.56 301910        51597     844    283176    10.78   317391
      2017     49068     1088      273812   10.37   305500   47939     905       277825 9.56 307194        52692     913    288139    10.90   323388
      2018     50199     1131      278976   10.64   312198   48901     962       282731 9.63 312847        53839     954    293445    11.02   329788
      2019     51350     1151      284268   10.88   318978   49841     940       287716 9.63 318375        54976     948    298636    11.14   336075
      2020     52507     1157      289652   11.10   325802   50790     949       292725 9.63 323929        56118     958    303850    11.26   342405
      2021     53739     1232      295143   11.32   332817   52249     1459      297965 10.64 333445       57101    1072    308905    11.38   348572
      2022     54936     1197      300639   11.55   339879   53256     1007      303304 10.64 339407       58334    1013    314602    11.50   355483
      2023     56182     1246      306176   11.74   346911
      2024     57400     1218      311731   11.96   354066
      2025     58673     1273      317332   12.16   361269
     Average             1098                                          1032                                         1013



     Table 1: Forecasting Data



     3.1.4.1 Changes in higher forecast from reference case

     The difference between the NIRP2 Reference case forecast and this high growth forecast is due mostly to an
     increase in industrial demand. This is illustrated in a breakdown of differences in consumption for the year
     2013 as shown in Figure 4.
                                                                                  Risk & Sensitivity Analysis




                                                    Transport

                              Residential                           Mining




                        Agriculture



                       Commerce




                                                                  Industry



Figure 4: The proportion of consumption increases over the reference case.

A detailed breakdown in GWh of demand consumption for the high growth case is given in Table 2 below.

Table 2: Electricity demand in the high growth scenario (GWh)

  Sector/Year      2003     2004      2005     2006  2007       2008     2009     2010     2011     2012     2013
  Transport        3432     3569      3656     3752  3782       3807     3847     3886     3930     3987     4029
  Mining           31923    33204     34079    34987 35228      35326    35528    35678    35966    36398    36694
  Total Industry   91658    97999     101415   105855108896     111698   115107   118308   121313   124531   127563
  Commerce         18360    19537     20450    21449 22058      22633    23317    23968    24660    25457    26161
  Agriculture      6310     6728      7055     7408 7621        7819     8052     8267     8493     8755     8985
  Residential      32884    34866     36325    37431 38882      40202    41596    42944    44296    45650    46982    47
  Total domestic
  ex own
  generation       184568 195903 202981 210881 216468 221484 227446 233052 238659 244779 250414

3.2 Availability of plant

The plant availabilities of new plants for the NIRP2 Reference Plan are based on the World Energy Council
(WEC) best quartile results for 2002. For existing plants the NIRP2 Reference Plan uses the current targets in
Eskom adjusted independently for each individual station to give a weighted average for base-load capacity of
88% EAF; (7% PCLF: 3% UCLF with a provision of 2% for OCLF to cater for risk).

The above availability levels were considered too optimistic by the ARC. It was decided to use average results
rather than best quartile taken from the WEC Report 2001 as being reflective of a lower level of plant
availability. (Note: Later editions of WEC Reports do not accurately reflect global plant availability levels
because fewer and fewer utilities are providing input to the WEC database as competition in the Electricity
Supply Industry increases. Figures are now biased towards a few remaining large utility players still
contributing). For existing plants this study uses lower Eskom targets adjusted independently for each
individual station to give a weighted average for base-load capacity of 85% EAF; (7% PCLF: 6% UCLF with a
provision of 2% for OCLF).

A comparison of plant availability levels for this analysis as compared to the NIRP2 Reference Plan is given in
Table 3 below for all new plant options.
     National Integrated Resource Plan 2




     Table 3: Availabilities of new plant.


         Option              Expected           Low          Option              Expected            Low
         New Coal with FGD                                   New Gt Open Cycle (LNG)
         PORg                  7.20%               7.20%     PORg                 7.08%                 7.08%
         FORg                  5.17%               7.97%     FORg                 7.31%                13.90%
         Availability         88.00%              85.40%     Availability        86.13%                80.00%
         New Coal without FGD                                New Gt Open Cycle (Kerosene)
         PORg                  7.20%               7.20%     PORg                 7.08%                 7.08%
         FORg                  5.17%               7.97%     FORg                 7.31%                13.90%
         Availability         88.00%              85.40%     Availability        86.13%                80.00%
         New Pumped Storage                                  Renewables (CSP)
         PORg                  1.66%               1.66%     PORg                 3.85%
         FORg                  1.26%               4.82%     FORg                 1.98%
         Availability         97.10%              93.60%     Availability        94.25%
         New Fluidised Bed                                   Renewables (Wind)
         PORg                 10.20%               7.69%     PORg                 2.00%                 2.00%
         FORg                  4.57%               8.89%     FORg                 1.02%                 8.16%
         Availability         85.70%              84.10%     Availability        97.00%                90.00%
         CCGT Pipeline Excl Trans                            Public PBMR (1st MM) Excl Tran
         PORg                  8.73%               8.73%     PORg                 4.04%                 4.04%
         FORg                  6.46%               9.61%     FORg                 4.13%                 8.29%
         Availability         85.37%              82.50%     Availability        92.00%                88.00%
         CCGT LNG Exc Trans                                  Public PBMR (1st MM) Inc Tran
         PORg                  8.73%               8.73%     PORg                 4.04%                 4.04%
         FORg                  6.46%               9.61%     FORg                 4.13%                 8.29%
48       Availability         85.37%              82.50%     Availability        92.00%                88.00%
         CCGT LNG Inc Trans                                  Public PBMR Series No Trans
         PORg                  8.73%               8.73%     PORg                 4.04%                 4.04%
         FORg                  6.46%               9.61%     FORg                 2.04%                 6.21%
         Availability         85.37%              82.50%     Availability        94.00%                90.00%
         New Gt Open Cycle (Local Syngas)                    Nuclear
         PORg                  7.08%               7.08%     PORg                15.00%
         FORg                  7.31%              13.90%     FORg                 7.65%
         Availability         86.13%              80.00%     Availability        78.50%


     3.3 Sustainability of existing interruptible loads

     The current contracts entered between customers and Eskom concerning interruptible supply are severely
     constrained in terms of their usage requirements. Studies carried out in Eskom comparing these options with
     equivalent OCGT capacity showed that they should be de-rated to obtain equivalence. In anticipation of future
     Demand Market Participation or other Demand Response mechanisms becoming available, it was considered
     prudent by the ARC to consider the possibility that current interruptible supply contracts may be re-negotiated
     in the future. As such some plans should be developed on the basis of excluding them from the planning base
     in the risk analysis.

     3.4 Sustainability of existing capacity

     As stated previously, some risk events are aggregated into independent risk portfolios. The aggregated
     options in each portfolio are assigned the same probability of occurrence for purposes of this Report.

     The following capacity reductions are included as aggregated options in this analysis. The levels of capacity
     reductions are based on either ARC recommendations or expert opinion:
                                                                                   Risk & Sensitivity Analysis




& Cahora Bassa capacity is reduced by 250MW (from 1281MW) in 2004 and by 500MW thereafter. This is a
  base-load capacity reduction;
& Non-Eskom OCGT capacity is reduced by 180MW (from 270MW) and assumed unavailable over the
  planning horizon. This is a Peaking capacity reduction;
& Non-Eskom coal-fired capacity is reduced by 600MW (from 1920MW) and assumed unavailable over the
  planning horizon. This is a base-load capacity reduction. In addition, this latter option targets Municipal
  capacity destined for decommissioning from 2011.

4.Scenario Development

It was agreed by the ARC to carry out the risk and uncertainty analysis to develop an appropriate net reserve
margin using a Resources Planning approach. The focus of this approach is to provide a best (robust and
flexible) plan developed from an analysis of the risk portfolios and then compare the capacity of the selected
plan with the expected demand to establish an appropriate net reserve margin.

(Note: A plan that is 100% robust is one, which lies in the decision set for all futures and would lead to no regret
whereas flexibility is the ability to modify a resource plan without significantly degrading system reliability and
economics, in response to actual data that have deviated from the forecast data).

There are four risk portfolios defined each consisting two alternatives. This results in sixteen scenarios
incorporating the alternative risk events from each risk portfolio. Optimal plans are then developed for each of
the sixteen scenarios where the basis for the optimisation is the least cost of electricity for the supply life cycle.
This considers the least cost of each plan in terms of the cost of supply and the cost of non-supply (CUE) to the
consumer. For this NIRP2 as explained in the NIRP2 Reference Plan, the CUE is assumed to be
R20 666/MWh.
                                                                                                                         49
Probabilities were assigned to each of the risk portfolios using a Delphi approach in consultation with the
various experts in the field associated with developing the portfolios.

Table 4 below details the probabilities assigned to each of the risk portfolios.

Table 4: Probabilities associated with primary assumptions.



Uncertainties:                                                       Probability                         Probability
Load Forecast                    Expected (Moderate) (ED)               0.6         High (HD)               0.4
Plant outage                     Expected (EF)                          0.4         High (HF)               0.6
Capacity availability            Expected (EC)                          0.3         Low (LC)                0.7
DSM Interruptible Load           Expected (EI)                          0.4         Low (LI)                0.6




Table 5 below details the sixteen scenarios incorporating the alternative risk events from each risk portfolio
with the resultant probabilities assigned to each scenario.
     National Integrated Resource Plan 2




     Table 5: Scenarios and their probabilities.


          Futures                             Plan A-1 / Scenario 1     Plan A-2 / Scenario 2
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                           Expected (EF)               High (HF)
          Capacity availability                  Expected (EC)                Low (LC)
          DSM Interruptible Supply               Expected (EI)                Low (LI)
                           Probability               0.0288                    0.1008

          Futures                             Plan A-3 / Scenario 3     Plan A-4 / Scenario 4
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                             High (HF)               Expected (EF)
          Capacity availability                     Low (LC)               Expected (EC)
          DSM Interruptible Supply                  Low (LI)               Expected (EI)
                           Probability               0.1512                    0.0192

          Futures                             Plan A-5 / Scenario 5     Plan A-6 / Scenario 6
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                             High (HF)                 High (HF)
          Capacity availability                  Expected (EC)                Low (LC)
          DSM Interruptible Supply                  Low (LI)               Expected (EI)
                           Probability               0.0648                    0.0672

          Futures                             Plan A-7 / Scenario 7     Plan A-8 / Scenario 8
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                           Expected (EF)               High (HF)
          Capacity availability                     Low (LC)               Expected (EC)
50        DSM Interruptible Supply                  Low (LI)               Expected (EI)
                           Probability               0.1008                    0.0288

          Futures                             Plan A-9 / Scenario 9    Plan A-10 / Scenario 10
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                           Expected (EF)             Expected (EF)
          Capacity availability                  Expected (EC)             Expected (EC)
          DSM Interruptible Supply                  Low (LI)                  Low (LI)
                           Probability               0.0432                    0.0288

          Futures                            Plan A-11 / Scenario 11   Plan A-12 / Scenario 12
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                              High (HF)              Expected (EF)
          Capacity availability                  Expected (EC)                Low (LC)
          DSM Interruptible Supply                Expected (EI)             Expected (EI)
                           Probability               0.0432                    0.0448

          Futures                            Plan A-13 / Scenario 13   Plan A-14/ Scenario 14
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                              High (HF)              Expected (EF)
          Capacity availability                     Low (LC)                  Low (LC)
          DSM Interruptible Supply                Expected (EI)               Low (LI)
                           Probability               0.1008                    0.0672

          Futures                            Plan A-15 / Scenario 15   Plan A-16/ Scenario 16
          Load Forecast                          Expected (ED)               High (HD)
          Plant outage                           Expected (EF)               High (HF)
          Capacity availability                     Low (LC)               Expected (EC)
          DSM Interruptible Supply                Expected (EI)               Low (LI)
                           Probability               0.0672                    0.0432
                                                                                Risk & Sensitivity Analysis




It is necessary to state that after consultation with the various experts in their fields, that some of the expected
events (used as basis for developing the NIRP2 Reference Plan) were in fact less realistic than originally
thought. This resulted in some cases, higher probabilities being assigned to the alternative case than the
expected case.

5.Ranking of Plans/Candidate Plans

The assessment of the plans/scenarios includes the following steps:

&   Development of optimal resource plan for each of the 16 scenarios;
&   Testing each of the plans under the assumptions of the remaining scenarios;
&   Calculation of the total probability weighted cost (PWC) of each plan;
&   Ranking of the plans in terms of the attributes
&   Selection of the most robust plan

The output of the trade-off analyses would be the most robust optimum plan/strategy for capacity development
over the planning period.

The selected attributes (measures of goodness) for evaluation of the plans are:

&   Total probability weighted cost
&   Reliability
&   Emissions;
&   Diversity of fuel sources.

5.1 Methodology
                                                                                                                       51

The main criteria for measuring the performance of the plans is the total probability-weighted cost (PWC). The
probability-weighted cost is calculated as the sum of the total cost of the plan for each of the 16 scenarios
multiplied by the probability of occurrence of the corresponding scenario.

The 16 plans are ranked in ascending order (minimum cost) in terms of total probability-weighted cost (PWC).
The first three plans with the lowest probability-weighted total cost are then selected as candidates for the
preferred plan.

In order to test the sensitivity of the probability-weighted cost to the length of planning period, these analyses
were conducted for planning horizon of 8, 12, 16 and 20 years.

To test the sensitivity of the probability weighted cost to the probability assigned to each of the primary
uncertainties, i.e. the analyses were conducted for equal probability of each of the primary uncertainties (50%
probability of occurrence of each of the primary uncertainties), as well as for 60% probability of occurrence of
the expected uncertainties.

The second attribute considered is the reliability of the plans measured as the level of the reserve margin
provided by the plan over the planning horizon. Higher reserve margin will provide higher system adequacy
and reliability. The candidate plans are ranked in descending order according to the RM they provide.

The evaluation of the plans in terms of the emissions is conducted by comparing the total emission level (CO2,
SO2, NO2 and particulate emissions) of the three candidate plans over the planning horizon.

The fourth attribute for comparison is the diversity of the plans. The diversity of the plans is measured as the
ratio of non-coal and coal based generation resources.
     National Integrated Resource Plan 2




     5.2 Total Probability-Weighted Cost

     The ranking of the plans is conducted in terms of the attribute total probability weighted cost. The first three
     plans are selected as candidates for the most robust plan/strategy for implementation.

     The ranking of the 16 plans and their associated expectation cost over the whole planning period (20 years), 16
     years, 12 years and 8 years is illustrated in Table 6 below. While under the assumed probabilities of the
     primary uncertainties Plan 14 enjoys the least cost or most robust position, Plan 8 and Plan 16 follow closely its
     performance. The ranking of these three plans remains unchanged within a range of planning period from 12 to
     20 years.

     The ranking of the plans for the first 8 years indicates that the Plan 14 and Plan 08 maintain their position as the
     best performers while Plan 3 emerges as the second most robust plan for the first 8 years of the resource plan
     study period.

     Table 6: PWC Ranking of the Plans


     Ranking of Expectation over Ranking of Plans Expectation over Ranking of             Expectation    Ranking of    Expectation
     Plans over 8 8 years, mR    over 12 years    12 years, mR     Plans over 16          over 16 years, Plans over    over 20 years,
     years                                                         years                  mR             20 years      mR

                 Assumed probabilities
        Plan14        102 651             Plan14             145 335          Plan14          181 830        Plan14        212 601
        Plan03        102 790             Plan08             145 712          Plan08          182 547        Plan08        214 053
        Plan08        102 872             Plan16             146 098          Plan16          183 084        Plan16        214 546
                 50% Probabilities of primary uncertainties (Each plan has an equal probabilty of 6.25%)
        Plan14        102 267             Plan14             144 854          Plan14          181 364         Plan14       212 175
        Plan03        102 361             Plan08             145 250          Plan08          182 154         Plan08       213 722
        Plan08        102 379             Plan16             146 046          Plan12          183 077         Plan12       213 968
52               60% Probablities of expected primary uncertainties (Plan A1 has the highest probability of 12.96%)
        Plan10        100 341             Plan10             141 587          Plan12          177 074         Plan12       207 187
        Plan12        100 341             Plan12             141 633          Plan10          177 122         Plan10       207 203
        Plan05        100 456             Plan14             141 780          Plan14          177 241         Plan14       207 444



     The above results show that the ranking of Plan 14 is not sensitive to small changes in the assigned
     probabilities and the length of the planning period.

     The difference in the PWC of the first three plans is marginal and less than 1%.

     The actual variability of the plans PV cost under the modeled future conditions is illustrated in Figure 5 below.
     For a comparison the graph also includes plan A2 which is developed for the most pessimistic conditions
     (highest capacity requirements) and therefore carries the least variability of the costs under the remaining 15
     more favorable futures.

     The performance of the first three plans, Plan 14, 08 and 16 is further evaluated in terms of the attributes
     reliability, emissions and diversity.
                                                                                                                                        Risk & Sensitivity Analysis




                           310 000

                                                                                                                                                     Plan 2 NPV 20
                                                                                                                                                     Plan 3 NPV 20
                           290 000                                                                                                                   Plan 8 NPV 20
                                                                                                                                                     Plan 14 NPV 20
                                                                                                                                                     Plan 16 NPV 20

                           270 000
          Total cost, mR




                           250 000




                           230 000




                           210 000




                           190 000
                                 S01      S02      S03        S04    S05      S06      S07        S08      S09     S10         S11     S12     S13      S14        S15         S16

                                                                                                  Scenario
Figure 5: Variability of total cost.

5.3 Reliability Assessment

The reliability assessment is conducted in terms of the reserve margin carried by the plans under the future
scenario they were developed for and the reserve margin that the plans would carry had the expected
conditions (business as usual) prevailed over the planning period. The reserve margin under own future and
under the expected (A1) future are shown in Figure 6 and Figure 7, respectively.


                                                                                                                                                                                     53
                           25.00%




                           20.00%
                                                                                                                                                                         A3
                                                                                                                                                                         A14
                                                                                                                                                                         A16
                                                                                                                                                                         A8
                           15.00%
              RM, %




                           10.00%




                            5.00%




                            0.00%
                                 2003   2004    2005   2006   2007   2008   2009    2010   2011    2012   2013   2014   2015    2016   2017   2018   2019   2020    2021      2022

                                                                                                          Year
Figure 6: Reserve Margin under own future.
     National Integrated Resource Plan 2




                      25.00%
                                                                                                                                                           A3
                                                                                                                                                           A14
                                                                                                                                                           A16
                                                                                                                                                           A8
                      20.00%




                      15.00%
              RM, %




                      10.00%




                      5.00%




                      0.00%
                           2003   2004   2005   2006   2007   2008   2009   2010   2011   2012   2013   2014   2015   2016   2017   2018   2019   2020   2021   2022

                                                                                          Year


     Figure 7: Reserve Margin under A1 future.

     Plan 14 carries a higher than 10% reserve margin (RM) under its own future from 2008 to 2022 and 12% % RM
     under expected future (A1). In comparison Plan 16 carries a better RM under A1 future but under its own future
     scenario its RM is relatively low.

     5.4 Emissions and Diversification Assessment
54
     A summary of the comparison of the plans in terms of diversity of resources, emissions as well as costs and
     reliability parameters is provided in Table 7 overleaf.

     In terms of emissions the least-cost candidate plans have quite similar performance. This is a result of the large
     share of the coal-fired generation in all 16 plans. For the same reason the diversification of the plans is also low
     and varies between 9 and 12%. However, regardless of the close performance of the plans in terms of
     emissions, Plan 3 ranks marginally as the plan with the lowest emissions level, followed by the remaining three
     plans with almost equal pollution level.

     Higher diversification of resource is achieved in the alternative of the Plan 14, presented hereafter, which
     factors the Government renewable energy (RE) target of 10 000 GWh by 2012 in the capacity mix.
                                                                  Risk & Sensitivity Analysis




Table 7: Comparison of Candidate Plans



NIRP 2 STAGE 2 COMPARISON OF CANDIDATE PLANS
                     Plan                          A14           A8        A16          A3
                  Probability                     0.0672         0.0288     0.0432    0.1512
                   Demand                          High         High       High          E
                     FOR                             E          High       High        high
               Capacity Avail                      Low           E          E          Low
                    DSM IL                         Low           E         Low         Low
      RM(weighted average 2006-2022)              10.61%         6.85%      6.23%     8.90%
 RM(weighted average 2006-2022) under Ref
                   case A1                        14.53%         16.57%     17.70%   12.80%
LOLP (average over 20 yrs)                        0.00514     0.0054066   0.007379   0.00568
USE (over 20yrs), GWh                               919          940       1307        888

Probability weighted PV cost over 20yrs, mR        212 601      214 053    214 546    216 969
Probability weighted PV cost over 20yrs, %        100.00%      100.68%    100.91%    102.05%
PV System Cost, mR                                 209 169      221 189    230 200   202 638
PV of Capex, mR                                    61 009      69 261       70 534    58553

PV of Energy Prod (20 yrs) GWh                    2 455 643   2 545 521 2 545 268    2 455 613
Constant cost, R/KWh                               0.0852      0.0869    0.0904       0.0825
Constant pobability weighted cost, R/KWh           0.0866      0.0841    0.0843       0.0884

New capacity, MW
                       Mothballed plants, MW       3559        3559       3559        3559
                               First unit year     2005        2005       2005        2005       55
                            OCGT(peak), MW         2880        2880       3360        1920
                               First unit year     2 006       2 006      2 006       2006
                     Pump Storage (peak), MW       3330        3330       3330        3330
                               First unit year     2 012       2 012      2 012       2012
                       CCGT (base load), MW          -            -         -           -
                               First unit year       -            -         -           -
                           PF(base load), MW       9630          10914      11556     10272
                               First unit year     2016         2015      2015        2015
                        FBC (base load), MW        2330         2796      2330        1398
                               First unit year     2012         2010      2010        2010

                            Total capacity MW      21 729      23 479     24 135      20 479
                             PV of Capex, mR       61 009      69 261      70 534     58553

                                Coal based, %     86.75%       87.73%     86.08%     90.62%

                        Diversity (non-coal), %   13.25%       12.27%     13.92%      9.38%

CO2 Emissions (Mt)                                 5397         5340       5336       5180
SO2 Emissions (Mt)                                  47           47         46         44
NO2 Emissions (Mt)                                  20           20         20         19
Particulate Emissions (Mt)                          1            1          1          1
Water Usage (Mcub m)                               7232         7028       7034       6951
Total emissions, Mt                                5464         5407       5402       5243
     National Integrated Resource Plan 2




     6.Preferred Resource Plan

     The overall ranking of the plans is conducted in terms of the attributes PWC, RM own, diversity and emissions
     with a weighting of 40, 40, 10 and 10% respectively. Sensitivity analyses are performed with weightings of
     30/30/20/20 and 50/30/10/10. The results of the ranking of the plans are illustrated in Figure 8 below.


                1.100
                                                                                            Best plan =1
                                                                                                                                A14
                1.000                                                                                                           A3
                                                                                                                                A8
                                                                                                                                A16

                0.900




                0.800




                0.700




                0.600




                0.500




                0.400
                        RM own    Prob Weighted PV   Emmissions   Diversity   Weight 40/40/10/10 Weight 30/30/20/20 Weight 50/30/10/10



     Figure 8: Ranking of Plans
56
     Plan 14 remains the most robust plan, followed by Plan 3, plan 8 and Plan 16. Therefore in terms of the
     attributes Plan 14 should be selected as the optimum most robust plan/strategy for implementation. It has to be
     highlighted that the four candidate plans are very similar regarding the capacity mix and timing. All plans,
     except Plan 3 are developed for high growth scenario which is the planning uncertainty with the highest impact
     on the future capacity requirements, timing and capacity mix.

     The preferred Plan 14 is developed under high growth demand forecast, expected plant outage rate, low
     import and municipal capacity availability and no interruptible loads.

     6.1 Preferred Plan 14

     The capacity mix and timing of Plan 14 are shown in Table 8. Details about Plan 14 attributes are illustrated in
     Table 5.2.

     Similarly to the remaining 15 plans, Plan 14 is also based on RTS of all mothballed plants by the year 2011.

     In terms of peaking plants, the plan requires construction of 2880 MW of OCGTs starting from the year 2006 to
     meet the medium term requirements for peaking capacity. The peaking requirements in the second half of the
     planning period are satisfied by the commissioning of 3330 MW of pump-storage plants starting from the year
     2012.

     The first FBC base-load capacity is required in 2012 while the first PF coal-fired unit must be commissioned in
     2016.

     The plan provides certain degree of flexibility regarding modification of the timing of the required capacity.
     Depending on deviations of the actual demand from the forecasted, the decisions for resource commissioning
     could be accelerated or delayed.
                                                                                                                Risk & Sensitivity Analysis




Table 8: Preferred Plan 14.

                    Mothballed                         Coal-Fired              FBC         Gas           Pumped Storage             DSM

 YR      Cam (PF)      Gr'tvlei    Kom (PF)   PF (1)     PF (2)     PF (3)   Green-field   OCGT       PS (A)     PS (B)   PS (C)   CEE IMEE    Reserve on
                        (PF)                                                    FBC                                                IMLM REE     moderate
                                                                                                                                      RLM       forecast



        Committed                                                                                   Committed                      Committed
 2003                Committed    Committed                                                Decide                                    152          24%
 2004                                         Decide                                                             Decide              152          21%
 2005      380                                                                 Decide                                                152          18%
 2006      380                                                                              720                                      152          17%
 2007      570           188                                                                                                         152          16%
 2008      190           377         101                 Decide                             480                                      152          16%
 2009                    565         202                            Decide                  720                           Decide     152          17%
 2010                                303                                                    720                                      152          17%
 2011                                303                                                                                             152          13%
 2012                                                                           932                    333                                        13%
 2013                                                                           466                    999                                        13%
 2014                                                                           932                               999                             15%
 2015                                                                                                                                             13%
 2016                                         1284                                                                                                13%
 2017                                         1284                                                                                                14%
 2018                                         1284                                                                                                14%
 2019                                                                                                                      999                    14%
 2020                                                     1284                                                                                    15%
 2021                                                     1284      642                                                                           14%
 2022                                                     1284      1284                   240                                                    13%
TOTAL      1520         1130         909      3852        3852      1926        2330       2880        1332       999      999       1370




7.Sensitivity Analysis

The ARC requested the following sensitivities be investigated:
                                                                                                                                                            57
& The impact of the net discount rate on the optimum choice of technology;
& The impact of Renewable Energy Technologies on the preferred (robust) plan;

These sensitivity studies were considered necessary to test the robustness of the conclusions and
recommendations emanating from the NIRP2 Reference Plan, and their impact on important energy policy
issues.

Before discussing the sensitivity analysis it is necessary to highlight data changes made to the demand- and
supply-side data base of the NIRP2 Reference Plan in this study.

A summary of the complete data base of new plant capital, O&M, Fuel costs and performance parameters is
given in Appendix C of this Report

7.1 Comparison of Data changes to NIRP2 Reference Plan

No data changes were made to any existing plant options or new demand-side initiatives from the NIRP2
Reference Plan. Marginal changes were made to some specific new supply-side options. To highlight these
changes more effectively, the candidate plants are separated into Peaking and Base Load options.

These options are compared in terms of their levelised costs at their maximum production levels and
separated into O&M, Fuel and Capex components. It is important to note that the CCGT option in the Cape
using natural gas was excluded from the analysis in this Report. However this was replaced with a CCGT plant
(located in Namibia) using natural gas from the Kudu field.

The results in Figure 9, Figure 10 and Figure 11 below, show that the most significant changes were made to
the cost of coal. There is an increase in coal costs in the case of the PF plants and an increase in O&M cost for
new FBC options to account for an increased requirement for sorbent.
     National Integrated Resource Plan 2




                                    Comparison of Change in Cost at 10% NDR between the Reference NIRP2 Plan and the
                                                                 Risk & Sensitivity Plan


                                     Reference Plan Plant Capex
                       400.00
                                     Risk & Sensitivity Plant Capex

                       360.00


                       320.00
        CAPEX R /MWh




                       280.00


                       240.00


                       200.00


                       160.00


                       120.00


                        80.00


                        40.00


                         0.00
                                CF with FGD     Greenfield FBC   CCGT Incl Trans      CCGT Kudu         CCGT Incl Trans   Mepanda Uncua   PBMR (1st MM    PWR (Incl trans
                                                  with FGD          (LNG)          (Offshore nat gas)      (pipe)                           Incl trans)     benefits)




     Figure 9: Capex Costs for Base Options



58



                                    Comparison of Change in Cost at 10% NDR between the Reference NIRP2 Plan and the
                                                                 Risk & Sensitivity Plan

                                                                      Reference O & M Cost

                                                                      Risk & Sensitivity O & M Cost
                       60.00
        CAPEX R /MWh




                       40.00




                       20.00




                         0.00
                                CF with FGD     Greenfield FBC   CCGT Incl Trans      CCGT Kudu         CCGT Incl Trans   Mepanda Uncua   PBMR (1st MM    PWR (Incl trans
                                                  with FGD          (LNG)          (Offshore nat gas)      (pipe)                           Incl trans)     benefits)




     Figure 10: O&M Costs for NIRP 2 Reference Plan and NIRP 2 Risk & Sensitivity Plan
                                                                                                                         Risk & Sensitivity Analysis




                                Comparison of Change in Cost at 10% NDR between the Reference NIRP2 Plan and the
                                                             Risk & Sensitivity Plan


                                                                                                                           Reference Fuel Working Cost

                                                                                                                           Risk & Sensitivity Fuel Working Cost

                   200.00
    CAPEX R /MWh




                   150.00




                   100.00




                    50.00




                     0.00
                            CF with FGD   Greenfield FBC   CCGT Incl Trans      CCGT Kudu         CCGT Incl Trans   Mepanda Uncua   PBMR (1st MM    PWR (Incl trans
                                            with FGD          (LNG)          (Offshore nat gas)      (pipe)                           Incl trans)     benefits)




Figure 11: Fuel Costs for NIRP 2 Reference Plan and NIRP 2 Risk & Sensitivity Plan



7.2 The impact of net discount rate on optimum choice of technology
                                                                                                                                                                      59

The timing of building new plant options will not be impacted to any great extent by small changes to the net
discount rate. However changes to the net discount rate will impact on which base-load and peaking
technologies are selected in the plant mix.

Due to time constraints required to prepare sets of plans for different net discount rates, it was agreed by the
ARC to carry out this analysis, external to the computer modelling process, by studying the impact of different
net discount rates on the life-cycle levelised costs of building and operating the specific plant options.

The analysis is carried out on both the data as detailed in the NIRP2 Reference Plan database as well as for the
updated costs as applied in the database for this study. Where applicable, differences between the two sets of
data are highlighted.
     National Integrated Resource Plan 2




     Table 9 below shows the impact of changes in the net discount rate on the life-cycle levelised costs of all the
     options available to the NIRP2 Reference Plan at their maximum operating levels.

     Table 9: Levelised costs of candidates for selection in the NIRP2 Reference Plan


                                                                              Net Discount Rate
           NIRP 2 Reference Plan Options         Load Factor
                                                                 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0%
      PF with FGD                                    88%        137.4 160.1 185.6 214.0 244.7 277.7 312.7
      PF without FGD                                 88%        127.5 148.3 171.7 197.6 225.8 256.0 288.1
      Greenfield FBC without FGD                     86%        131.7 153.0 176.9 202.9 230.8 260.5 291.5
      Greenfield FBC with FGD                        86%        134.2 155.3 179.0 205.0 232.9 262.5 293.5
      CCGT Incl Trans (LNG)                          85%        310.7 318.7 327.6 337.1 347.2 357.8 368.7
      CCGT Excl Trans (LNG)                          85%        328.1 337.4 347.6 358.6 370.3 382.5 395.0
      Single cycle GT (kerosine)                     13%       1158.3 1199.8 1244.3 1291.3 1340.0 1389.8 1440.2
      Single cycle GT (LNG)                          13%        706.9 749.5 795.1 843.0 892.5 943.0 994.2
      Single cycle GT (local syngas)                 13%        663.3 706.3 752.3 800.6 850.5 901.4 952.9
      Single cycle GT (LPG)                          13%       1005.8 1047.7 1092.6 1139.9 1188.8 1238.9 1289.5
      CCGT Incl Trans (pipe)                         85%        219.5 227.7 236.7 246.4 256.5 267.2 278.1
      CCGT Excl Trans (pipe)                         85%        234.8 244.3 254.7 265.8 277.5 289.8 302.4
      PS (based on Braamhoek public data)            23%        185.2 225.3 275.2 335.4 406.6 489.5 585.4
      PS (generic)                                   20%        269.1 348.9 448.2 568.1 710.0 875.5 1066.8
      Mepanda Uncua                                  58%        242.3 303.9 375.4 456.5 547.4 647.8 758.0
      Wind                                           33%        256.0 292.6 331.9 373.8 417.9 463.9 511.4
      Concentrating Solar Power                      70%        336.6 409.2 486.3 566.0 646.9 728.0 808.3
      PBMR (1st MM incl. trans)                      86%        187.7 220.6 258.3 299.9 344.9 392.8 443.4
      PBMR (1st MM excl. trans)                      86%        206.6 244.0 286.7 334.0 385.0 439.4 496.7
      PBMR (Series MM excl. trans)                   88%        154.7 174.7 197.7 223.1 250.5 279.7 310.4
      PWR (inc trans benefits)                       79%        236.9 268.4 303.5 341.2 380.8 421.9 464.1
      PWR (exc trans benefits)                       79%        248.6 280.4 315.7 353.8 393.8 435.3 477.9

60
     Table 10 below shows the impact of changes in the net discount rate on the life-cycle levelised costs of all the
     options available in this study also at their maximum operating levels.

     Table 10: Levelised costs of candidates for selection for the NIRP2 Risk and Sensitivity Analysis

                                                                                    Net Discount Rate
      NIRP 2 Risk and Sensitivity Plan Options     Load Factor
                                                                   4.0%     6.0%    8.0%    10.0%    12.0%    14.0%    16.0%
      PF with FGD                                      88%        145.6    167.9    193.2    221.1    251.5    284.1    318.7
      PF without FGD                                   88%        136.0    156.4    179.6    205.1    233.0    262.8    294.5
      Greenfield FBC without FGD                       86%        134.8    156.1    179.9    206.0    233.9    263.6    294.7
      Greenfield FBC with FGD                          86%        150.2    171.8    195.9    222.1    250.2    279.9    311.1
      CCGT Incl Trans (LNG)                            85%        308.0    315.9    324.6    334.0    343.9    354.3    365.1
      CCGT Excl Trans (LNG)                            85%        320.4    329.5    339.5    350.3    361.7    373.6    385.9
      Single cycle GT (kerosine)                       13%       1158.3   1199.8   1244.3   1291.3   1340.0   1389.8   1440.2
      Single cycle GT (LNG)                            13%        706.9    749.5    795.1    843.0    892.5    943.0    994.2
      Single cycle GT (local syngas)                   13%        663.3    706.3    752.3    800.6    850.5    901.4    952.9
      Single cycle GT (LPG)                            13%       1005.8   1047.7   1092.6   1139.9   1188.8   1238.9   1289.5
      CCGT Kudu (Offshore nat gas)                     85%        228.8    238.4    249.0    260.4    272.5    285.1    298.2
      PS (based on Braamhoek public data)              23%        193.7    233.8    283.7    343.8    415.0    498.0    593.8
      PS (generic)                                     20%        277.6    357.4    456.7    576.6    718.5    884.0   1075.2
      Mepanda Uncua                                    58%        216.0    280.6    355.8    440.8    535.1    638.5    751.1
      Wind                                             33%        251.1    281.9    314.3    347.7    381.9    416.5    451.3
      Concentrating Solar Power                        70%        492.6    565.2    642.3    722.1    803.1    884.2    964.5
      PBMR (1st MM incl. trans)                        92%        168.6    199.3    233.7    271.0    310.4    351.6    394.2
      PBMR (1st MM excl. trans)                        92%        183.8    218.7    257.7    299.9    344.7    391.4    439.7
      PBMR (Series MM excl. trans)                     94%        124.3    143.0    163.9    186.7    210.8    235.9    261.8
      PWR (inc trans benefits)                         79%        239.3    270.8    305.8    343.6    383.2    424.3    466.4
      PWR (exc trans benefits)                         79%        255.4    287.1    322.5    360.6    400.6    442.1    484.6

     Due to the levelised cost being sensitive to the load factor applied, the following analyses focus on the
     candidate base-load technologies taken at their maximum operating levels.
                                                                                                                           Risk & Sensitivity Analysis




Selected plots for the various options based on type can be revealing. The base-load options with transmission
credits, referenced against a coal-fired plant are illustrated in Figure 12 and Figure 13 below for both the NIRP2
Reference Plan and this study. Options with a high capital component become less attractive with the increase
in net discount rate. The reverse is true at lower discount rates.



                                         Life cycle levelised cost (R/MWh) to build and operate new plant - NIRP2 Reference plan

                                 500.0
                                               PWR (inc trans benefits)
                                               CCGT Incl Trans (LNG)
                                 450.0
                                               PBMR (1st MM incl. trans)
                                               CCGT Incl Trans (pipe)
                                 400.0
                                               PF with FGD
                                               Greenfield FBC with FGD
                                 350.0
 Cost of Supply (R/MWh)




                                 300.0


                                 250.0


                                 200.0


                                 150.0


                                 100.0


                                  50.0


                                   0.0
                                             4.0%                 6.0%             8.0%           10.0%            12.0%     14.0%      16.0%

                                                                                          Real Discount Rate (%)

                                                                                                                                                         61
Figure 12: Levelised costs of base-load plants vs Net Discount Rate for the NIRP 2 Reference Plan




                                          Sensitivity of Levelised Lifecycle costs of Plant to Net Discount Rate - NIRP2 Risk & Sensitivity
                                                                                       Analysis

                                 500.0
                                                    PWR (inc trans benefits)
                                                    CCGT Incl Trans (LNG)
                                 450.0
                                                    PBMR (1st MM incl. trans)
                                                    CCGT Kudu (Offshore nat gas)
                                 400.0
                                                    PF with FGD
                                                    Greenfield FBC with FGD
        Cost of Supply (R/MWh)




                                 350.0


                                 300.0


                                 250.0


                                 200.0


                                 150.0


                                 100.0


                                  50.0


                                   0.0
                                             4.0%                  6.0%            8.0%           10.0%            12.0%     14.0%      16.0%

                                                                                          Real Discount Rate (%)
Figure 13: Levelised costs of base-load plants vs Net Discount Rate for the NIRP 2 Risk &
Sensitivity Analysis
     National Integrated Resource Plan 2




     7.3 Renewable Energy Technologies

     The levelised costs of renewable technologies (wind and concentrating solar power) are relatively high when
     compared to other equivalent supply options operating at similar levels of production. Because of this high cost
     in building renewable energy technologies, the computed economic least cost optimisation process excludes
     these options from the planning base.

     However, it is difficult to assess the externality benefits of these options in a purely economic comparison and it
     is stated Government policy to introduce some amount of renewable technologies in the Electricity Supply
     planning base.

     For purposes of this study, and to try and meet the aspirations of Government policy (10000GWh cumulative
     energy supplied by renewable energy sources (RES) wind, solar, biomass, small-scale hydro by 2013), two of
     the technologies; wind and concentrating solar power (CSP), are considered in this analysis only due to the
     lack of reliable data for other RES. It has to be underlined that the selected RES mix is not the least cost mix. A
     number of lower cost RES options have been identified by the White Paper on Renewable Energy (WPRE)
     and currently under investigation for implementation.

     The lead-time required to engineer, design and build these options is stated as 2 years for wind, and 3 years for
     CSP. This excludes EIA's, licensing, financing and other issues. For the purpose of this study and to be able to
     meet Government aspirations, wind turbines (20MW total) are brought into commercial service in 2007,
     expected to achieve a maximum load factor of at least 30% annually. CSP (300MW total), is anticipated to
     operate at a load factor of at least 70% (with battery assistance) are introduced from 2008.

     The additional cost of imposing renewable technologies on the system and the cumulative energy contribution
     can be seen in Table 11 below.
62

     Table 11: Additional Cost of Renewable Technologies

                             Capex (Rm)      O & M Costs      Total    PV Cost (Rm)       Total Energy SO
                                                 (Rm)        cost of        of                  (GWh)
                                                             Wind &    Renewables
                                                              Solar
                    Year     Wind    Solar   Wind    Solar    (Rm)     Total    Cum.      Annual    Cum.


                      2003
                      2004
                      2005
                      2006
                      2007     155               4               159      109       109        57       57
                      2008           10377       4     337     10717     6655      6763      1935     1992
                      2009                       4     337       341      192      6955      1935     3927
                      2010                       4     337       341      175      7130      1935     5862
                      2011                       4     337       341      159      7289      1935     7797
                      2012                       4     337       341      144      7434      1935     9732
                      2013                       4     337       341      131      7565      1935    11667
                      2014                       4     337       341      119      7684      1935    13602
                      2015                       4     337       341      109      7793      1935    15537
                      2016                       4     337       341       99      7892      1935    17472
                      2017                       4     337       341       90      7981      1935    19407
                      2018                       4     337       341       82      8063      1935    21342
                      2019                       4     337       341       74      8137      1935    23277
                      2020                       4     337       341       67      8204      1935    25212
                      2021                       4     337       341       61      8266      1935    27147
                      2022                       4     337       341       56      8321      1935    29082


     These results show the high cost of the solar contribution to the total additional cost of these two renewable
     technologies. The overall increase in cost to the system from implementing these options is estimated by
     carrying out a detailed computational analysis on the preferred plan 14.
                                                                                                                                              Risk & Sensitivity Analysis




The power supply profiles of these options will not always be consistent with the system demand profile. These
options are therefore modelled as non-dispatchable technologies. A power demand profile for the wind option
was supplied by ERC in consultation and assistance from the Darling Wind Farm IPP in the Western Cape, to
yield a load factor of 33%.

The profile for the CSP was assumed to be similar to that of the commercial and industrial energy efficiency
programmes yielding a load factor of 71%. These profiles were used to simulate the required load factor
expectations of these renewable options.

The results of the computational analysis show an increase in PV cost of R4773 Million cumulative over the
planning horizon compared to the preferred Plan 14 without renewables, as illustrated in Figure 14 below.


                     Mothballed                                Coal-Fired                      FBC           Gas                Pumped Storage                      Renewables                DSM

YR        Cam (PF)         Gr'tvlei     Kom (PF)     PF (1)          PF (2)      PF (3)      Green-field     OCGT            PS (A)      PS (B)       PS (C)        Conc.         Wind       CEE IMEE
                            (PF)                                                                FBC                                                                  Solar      Turbines     IMLM REE
                                                                                                                                                                    Panels                      RLM
                                                                                                                                                                    (CSP)


         Committed                                                                                                      Committed                                                            Committed
 2003                    Committed     Committed                                                             Decide                                                                            152
 2004                                                Decide                                                                              Decide                     Decide          Decide      152
 2005       380                                                                                Decide                                                                                           152
 2006       380                                                                                               720                                                                               152
 2007       570              188                                                                                                                                                      20        152
 2008       190              377          101                        Decide                                   240                                                    300                        152
 2009                        565          202                                                                 720                                                                               152
 2010                                     303                                    Decide                       720                                                                               152
 2011                                     303                                                                 240                                     Decide                                    152
 2012                                                                                           932                           333
 2013                                                                                                                         999
 2014                                                                                           932                                          999
 2015                                                                                           466
 2016                                                 1284
 2017                                                 1284
 2018                                                 1284                                                                                                                                               63
 2019
 2020                                                                1284
 2021                                                                1284                       466                                                    999
 2022                                                                1284            1284
 TOTAL      1520            1130          909         3852           3852            1284       2796          2640           1332            999       999           300              20       1370



Figure 14: Preferred Plan 14 with Renewables

The renewable options will reduce environmental emissions and water consumption of the preferred Plan14.
The extent to which the renewables reduce the environmental emissions is illustrated in Figure 15 below.

                                                                                                                                                     MTons { SO2, Particulates
             CO2 MTons                                                                                                                                  & NO2 Emmisions}

                   30                                                                                                                                                           0.3

                                      Cum CO2 Reduced Emissions
                                      Cum SO2 Reduced Emissions
                   25                                                                                                                                                           0.25
                                      Cum Reduced NO2 Emissions
                                      Cum Reduced Part Emissions
                   20                                                                                                                                                           0.2




                   15                                                                                                                                                           0.15




                   10                                                                                                                                                           0.1




                     5                                                                                                                                                          0.05




                     0                                                                                                                                                          0
                     2003     2004     2005   2006   2007     2008   2009     2010    2011   2012     2013   2014     2015    2016    2017    2018   2019    2020   2021     2022
                                                                                                Years

Figure 15: Reduction in environment emissions in preferred Plan 14 with renewables.
     National Integrated Resource Plan 2




     The cumulative total emission reduction of CO2 over the planning horizon is 28.2 Million Tons whereas the
     cumulative constant rand increase in the cost of the plan is R13547 Million. This represents a cost of R481/Ton
     of CO2 saved at 1 January 2003 prices.




     8.Conclusions

     The main conclusions of these studies in comparison with the NIRP2 reference are summarised as follows:

     1) The preferred plan has been developed on the basis of an evaluation of risk profiles and the reserve
         margin is an outcome of the optimisation process.
     2) The reserve margin of the preferred plan is higher than the 10% reserve margin used as a deterministic
         reliability criteria in the reference plan;
     3) The studies show that immediate decisions are required for peaking and base load plants from 2006 and
         2012 respectively;
     4) The base load plants competing for commercial operation from 2012 competing, include; Pulverised Fuel
         Coal-fired (PF); Fluidised Bed Combustion (FBC); Combined Cycle Gas Turbine (CCGT)
     5) In order to meet Government RES targets by 2013 immediate decisions are also required for renewable
         energy options;
     6) Options for diversification are still insufficient to meet all of the forecast demand for electricity over the next
         20-year planning horizon. Coal-fired options are still predominating the capacity during the 20-year
         planning horizon. For environmental benefit it is imperative to continue with efforts to reduce the costs of
         implementing clean coal technologies and improve the efficiency of coal-fired plants;
     7) At the current assumed cost of capital (10% net discount rate before tax) and after returning the Eskom
         mothballed plant to service, fluidised bed combustion technologies are South Africa's most economic
64
         option, followed by investment in coal-fired plant. This in turn is followed by importing gas / LNG for CCGT
         plant in the Cape;
     8) It will be difficult to justify diversification on an economic basis, unless penalties for not doing so are
         included in future analyses. As the cost for diversification is becoming increasingly more expensive these
         penalties (or opportunities for emissions trading) will need to be substantial to offset the economic benefits
         of remaining with coal;
     9) The NIRP2 Stage 2 plans are based on the accepted RTS program of the mothballed plants;
     10) The preferred plan indicates that 2880 MW OCGT peaking plants are required in the planning horizon in
         service from 2006;
     11) There are supply options that have not been considered such as co-generation in industry, converting
         OCGT to CCGT, adding units onto existing power stations and new imports resulting from the
         development of Electricity Supply in the Southern African region;
     12) The RES scenario is not fully reflective of the least cost RES mix currently investigated for implementation
         of the WPRE. This will be included in the next round of the NIRP update.
     13) Should interruptible supply and / or OCGT capacity not be implemented this will significantly advance new
         base-load capacity;




     9.Appendices

     Appendix A: Summary of comments on reference case report

     Appendix B: Detailed Plans

     Appendix C: Data Base
                              ARC and Public Comments on NIRP2 Reference Case




Appendix A
ARC AND PUBLIC COMMENTS ON THE NIRP2 REFERENCE CASE


I.  ESKOM SYSTEM OPERATION COMMENTS                                       66
II. ESKOM KSACS COMMENT                                                   67
III.EIUG COMMENT ON NIRP 2003/4 REFERENCE CASE VERSION 18 DECEMBER 2003   70
IV. ESKOM GENERATION COMMENTS ON NIRP 2003/4 REFERENCE CASE
    VERSION 18 DECEMBER 2003                                              72
V. COMMENTS FROM DARLING IPP                                              74
VI. COMMENTS FROM KELVIN IPP                                              75




                                                                                65
     National Integrated Resource Plan 2 - Appendix A



                                                             I
     I. ESKOM SYSTEM OPERATION COMMENTS

     Provided by Mr. Michael Barry

     1. The method to use pre-defined reserve margins and then derive the generation expansion plan is not
        ideal. As mentioned in the ARC meetings, I believe a set of data assumptions should be defined and used
        in the modeling. The reserve margin is then an outcome of the modeling process. Using this approach,
        one can relate each generation expansion plan to a set of assumptions, for example with x% reduction in
        generation performance, the expansion plan is advanced by y years. This set of assumptions is changed
        for the various sensitivity studies/scenarios.

     2. The NIRP Reference Plan is for the forecast moderate load growth. (The documentation makes mention of
        robust plans, no regret and flexibility, but from my understanding these are not specifically considered in
        deriving the reference plan). The Reference Plan then gives a conservative picture (building too late)
        when one considers the load growth uncertainty and the cost of supplied vs unsupplied electricty. The
        Reference Plan should, as part of the set of assumptions, include an additional plant margin for load
        growth uncertainty.

     3. The NIRP Reference Plan does not make provision for the inherent inertia in changing the "direction" of a
        "ship" the size of the electricity industry in South Africa. The country is used to many years of no significant
        generation capacity expansion. The Reference Plan should, as part of the set of assumptions, include an
        additional plant margin for this inertia.

     4. The Conclusions contain a number of "immediately implement" statements when referring to specific
66
        capacity additions. Is the NIRP saying a period of capacity shortages is forthcoming (and even more so
        considering the comments above)? If this is the case, should the NIRP not make some clear statement
        about the desirability of any capacity expansion or demand management in the next few years.

     5. From a strategic perspective, I believe a NIRP resulting in a reserve margin of anything less than 12% will
        harm the economy of South Africa. Potential international investors in the industrial arena might think
        "electricity interruptions" when seeing future reserve margins of less than 12%. Should the strategy not
        rather be to establish the capacity thereby stimulating highe growth?

     6. The reserve margin is calculated after the peak demand is reduced by the interruptible load. This implies a
        NIRP planning to interrupt customers. Is this the correct message?
                                         ARC and Public Comments on NIRP2 Reference Case



                                                      II
II.ESKOM KSACS COMMENT

Provided by Dr. Jean Pabot

With respect to SAPP, I did not find this word in your report, though you mention briefly an option for imported
hydro (Table 6). This option is not even discussed in the text (e.g. in Section 5). I assume you included the
transmission costs from the foreign plant to the Eskom network into your capital cost. However this option
appears not to have been considered in your simulations? If true, why?

I accept that there are no good recent reliable data for SAPP hydro plants, except maybe for Mepanda Uncua,
which may be the option mentioned inTable 6. If this is the case, it should be stated.

You might consider a scenario where both Mepanda Uncua and Inga 3 are built, e.g. in 2014, irrespective of
their costs, and study the impact thereof on your NIERP.

There is some confusing wording and apparently an inconsistency about imports-exports in Section 3.6 and
Section 5.1 (last 3 lines). Were exports included in the load forecast or not?

It would be useful to include a table showing the MW exports included in the load forecast. It might already be in
the appendices, but I did not see it.



With respect to the reserve margins used in the SAPP studies, these are:
                                                                                                                     67
We use:
& 7,6 % for hydro units,
& 10,6 % for thermal units,

There is a problem, even in Eskom, to justify a 10 % capacity reserve criterion to auditors, when many
utilities have reserve criteria ranging from 15 % to 25 %.

The minimum (or optimal) reserve margin is utility dependent. It depends on:

& The risks which the utiity wants to mitigate by adding reserve capacity, e.g risks associated with uncertainties
  on plant availability, load forecast, fuel availability, etc.
& The characteristics of the existing and future plants, e.g. unit size, plant availability, etc.
& The costs involved, e.g. plant cost, cost of un-served energy, etc.

Eskom is a SAPP member. A few years ago, the SAPP reduced its minimum installed capacity reserve
requirements (not the operating reserve requirements) to about 10 %, based on the results of a simple
optimization study. The methodology and assumptions are auditable, though auditors might disagree with all.

With hindsight, I now believe that one of the drought related assumptions is not valid, but this would have little
impact on thermal plant based utilities.

Additional minor comments on your report:

1. Interruptible supplies:

& In section 5.2, the report states: there is a total 1510 MW interruptible supply capacity included in the plan.
  Does this mean that you have 1510 MW of interruptible load/supplies assumed to be equivalent to 1510 MW
     National Integrated Resource Plan 2 - Appendix A




       of supply capacity? or is this 2500 MW or 3000 MW of interruptible load/supplies assumed to be equivalent
       to 1510 MW of supply capacity? Table 1 in Appendix 1 mentions it is derated capacity. This should be
       explained in the text.

     & Table 2 in Appendix 2 shows that the existing interruptible load (IL) decreases from 1510 MW to 500 MW in
       2012 and 384 MW and that it is not replaced by new IL. This does not make much sense.

     & IL is a very cost effective way to provide reserve to a system, and if Eskom's IL contracts come to an end,
       Eskom would surely seek new IL contracts.

     2. LRMC

     Table 12 in Appendix 12 shows that the LRMC does not vary significantly with the reserve margin (i.e. with the
     alternative plan) beyond 2015.

     With the methodology presented in Section 7, the LRMC should depend on
     the amount of plant to be built to meet a 500 MW load increase, which depends on the reserve margin criterion.

     3. Wind generator

     Section 9, in Appendix 3, should indicate the fuel/wind availability: when is it available during the day and with
     what probability?

     What capacity credit do you give the wind generator for reserve margin calculations?

     4. Imported hydro
68

     & The capital cost of imported hydro differs in Table 6 of the main report ( Rm 17044) and in Table 3 of Appendix
       1 (Rm 19366).

     & The life of the imported hydro plant is given as 30 years, much lower that the life of a pumped storage plant or
       a PBMR (40 years). This is not realistic. A dam can last 100 years, and a hydro generating unit 40 or 50 years.




     NIRP2 Reference Case Handling of Risks

     I noticed in your executive summary (Section 6) and in the main report (section 2.3) the way you handle risk:

     "The reference plan uses a constraint of 10% on the reserve margin. In addition the un-served energy is
     constrained to a maximum of 0.11% of total energy demand in any specific year.

     In order to address some risks to some extent and to determine reserve margin exposure to lead time, four
     alternative plans to the NIRP Reference plan were developed for increasing levels of constraint on the reserve
     margin
     As explained previously this Report does not address risk rigorously in this round but rather addresses it
     through imposing a minimum reserve margin on the plan of 10%."

     & Where did you get the 10% on the reserve margin criterion from and the 0.11 % UE criterion from?

     & Why do you need these criteria for planning? Theoretically, your optimization method should yield a plan,
       from which you could derive the planned reserve margin and UE.
                                          ARC and Public Comments on NIRP2 Reference Case




& Why do you need to have plans for various reserve margins, from 10 % to 17 %. This is not a random choice.
  Which on is the best and why?

& The above points should be explained in the report.

& SAPP uses a reserve margin criterion because this was the only way acceptable to all members to agree on
  what capacity each member should provide. However the reserve margin criterion itself was selected with a
  specific auditable process (see my previous message and attached paper) and was not arbitrary (though
  debatable).

I just have a few additional theoretical comments. You could find the same comments and additional /better
ones in the numerous articles available in the literature on this topic.

& In a generation planning problem, the planner has to handle risk (associated with uncertainties), i.e. he has
  to have a base/reference plan (e.g. over 10 years) and to define risk handling strategies to deviate from the
  plan if uncertainties are realized in a way or another (e.g. if the load forecast increases or decreases, if
  nuclear energy is phased out or encouraged, or if gas becomes very cheap or expensive, or if wind energy
  becomes a must have option).

& Long ago (20 or 30 or 50 years ago), the planners identified their planning risks, determined a reserve margin
  criterion (e.g peak load + 20 %), planned accordingly, and would revise their plans regularly as uncertainties
  become realized, and new ones develop.

The problem is that a reserve margin criterion becomes invalid if the load shape changes, or the
reliability changes, or uncertainties change.
                                                                                                                       69
& Later (25-35 years ago) the planners moved to a reliability criterion (e.g. 1 days in 10 years loss of load, or
  LOLE = 20 hours per year, or UE = .1 % or any similar one).

& This is better than a reserve margin criterion, but how do you select the reliability criterion?

& Later (20-30 years ago) the planners moved to a minimum cost to society approach (using the cost of un-
  served energy)

& Eskom generation planners moved from a reserve margin criterion (long ago, before my time) to a reliability
  criterion (10 % LOLP at the peak of the year) around 1979, but the reliability criterion was derived from the
  minimum cost to society approach (using a 0.5 R/kWh cost of un-served energy).

& Later around 1995 Eskom generation planners moved to a pure minimum cost to society approach, mainly
  because of the availability of more modern computing tools.

& I am not involved in Eskom ISEP, and I have not seen the plans ISEP9 and/or 9A. I understand that the
  Eskom ISEP planners are now also using a reserve margin criterion in addition to the minimum cost to
  society approach.

& I saw various Eskom documents where various people compared the reserve margin in Eskom with the
  reserve margin in other utilities. This is not very meaningful, unless you know well what risks each utility is
  exposed to (i.e. their associated uncertainties), and which risks are mitigated by a reserve margin (i.e.
  building reserve plant) and which risks are mitigated by additional strategies (e.g. plant mix, flexibility, etc),
  strategies which might be cheaper than building additional reserve plant.

& The problem with handling risk in planning, is that it changes from year to year in future (uncertainties
  increase as you look further in the future), therefore the reserve margin criterion should change every year in
     National Integrated Resource Plan 2 - Appendix A




       your planning process (e.g. 8 % in 2004, 9 % in 2005, 10 % in 2006, etc...) but there is a limit in future, where it
       is not meaningful to plan to build additional reserve, since there would be time to change the plans (e.g. in
       SAPP, the reserve margin criterion is optimized to mitigate risk over 4 years, e.g. if one notices a load
       forecast increase in year 1, the utility can do something about it in year 4, e.g. install a GT or contract for
       interruptible load).




     III. EIUG COMMENT ON NIRP 2003/4 REFERENCE CASE VERSION 18 DECEMBER 2003

     Provided by Mr Arnot Hepburn, EIUG Sectretariat

     Impressions

     1. The Eskom / ERI-UCT / NER team are to be congratulated on developing the NIRP document from what
        was previously a high level comparison of planning scenario options to a study report making specific
        project recommendations.

     2. EIUG Members welcome the publication of a significantly more comprehensive NIRP than previously
        issued which provides information that is essential to the business planning of energy intensive users.
70

     3. Although the NER was under the impression that they had given 4 weeks for comment the fact that the
        document was released just before the Christmas recess meant that most recipients were afforded less
        than 2 weeks to review and comment on this critical document. This has negatively impacted on the
        review and comments received.

     General Comments

     4. It is important to note that the optimum choice of technology is very dependent on the selected discount
        rate, with higher discount rates favouring less capital-intensive plant with shorter lead times. This mirrors
        the international trend towards more flexible technology options and away from traditional coal-fired and
        hydro stations.

     5. Due to the high regulatory uncertainty, independent power producers may well require rates of return
        higher than 10%, thereby affecting the appropriate technology choice. This effect should be taken
        into consideration.

     6. It is noted that although a range of demand projections are presented, the alternatives are only considered
        against a single demand forecast. There needs to be consideration of the course of action under the
        high demand scenario at least, or at the very least a qualitative description of the response to that
        scenario.

     7. A load duration curve and a forecast for its evolution over time should be included, so as to provide the
        minimum required information for attracting independent power producers into the ESI in the medium to
        long term.

     8. The full impact of the appropriate reserve margin should be considered - 10% appears low by
                                                         ARC and Public Comments on NIRP2 Reference Case




                                                                                               1
     international norms, which are around 20% for long-term planning purposes . While the
     appropriate reserve margin differs from market to market, and should be a result of a thorough
     LOLP and VOLL analysis, the specific characteristics of the South African system, i.e. a high
     degree of uncertainty on demand forecasts and the inflexible and long lead-time technologies
     used in SA (coal PF, pump storage) would indicate the need for a higher rather than lower reserve
     margin.

9. The EIUG have stated at IRP ARC meetings and would again like to place on record that we do not
   agree with interruptible load being considered as 'firm capacity' as this capacity can only be used in
   an emergency and is only 'firm' for a contractually limited time period each week.

10. We are surprised to find that media sources, such as the Engineering News, serve as references for cost
    estimations (Appendix 3 footnote 64, p25). This creates doubt over the reliability of this information and
    should be avoided in future.

11. The report has been enhanced by the addition of the appendices however there is very limited reference to
    specific sections in the appendices within the text of the report negating the effective transfer of
    information to the reader. This should be addressed in editing the report.

12. The opening paragraph of the report mentions taking into account committed contracts for imports and
    exports to South Africa from neighbouring states which is tie-line trading. The NEPAD initiative would
    require South Africa to be the anchor customer for bulk import of new hydropower as envisaged from the
    so-called Western Corridor (DRCANSA Interconnection). Lack of comment on the impact of such
    regional initiates on the NIRP appears to be an omission.

Specific Comments                                                                                                 71

13. In Table 6 of the main report (p18) entitled 'Summary Table of New Supply Side Options Data', the
    efficiency ascribed to the CCGT's appears low. A figure of 50% is routinely quoted in international papers.

14. In Table 8 of Appendix 1, no fuel costs have been allocated to the pumped storage units i.e. off-peak
    electricity purchases for pumping. It is important to know whether, in the evaluation of the viability of
    pumped storage units, the cost of the additional coal that has to be burnt for the pumping cycle was
    included as a variable pumped storage cost. Table 8 implies that this has not been done and if so, should
    be rectified.

15. Table 2 of Appendix 2 (p3) shows that 2,400 MW of new build SCGT is to be fired on local syngas. The
    assumption that as much as 2,400 MW of local syngas will actually be available seems quite optimistic.
    The other SCGT fuel source is Kerosene as it is currently cheaper than diesel, although kerosene is a
    government subsidised fuel and it is not certain if it would be appropriate to subsidise a peak power
    station's fuel source.

16. In paragraph 5 of Appendix 3 the Fluidised Bed Coal Fired Plant is discussed. The lower firing
    temperatures in a fluidised bed boiler certainly limits the formations of NOX's. However the SOX's are
    dependent on the coal characteristics and presumably limestone will be required. It appears as if
    limestone costs are not included in the variable costs of this technology. In addition, the coal
    costs appear to be low and should be confirmed with the owners of the reserves.

17. The efficiency range quoted (in note 51) of 35.2 to 42.1 presumably refers to pressurised and other
    advanced fluidised bed technologies. We assume that the technology in the NIRP is basic fluidised
    combustion. For this technology an efficiency of 37.36% seems to be very optimistic.

1 National Grid Company of the UK “Seven Year Statement” www.nationalgrid.com
     National Integrated Resource Plan 2 - Appendix A




     18. Points 4 to 7 in the Conclusions of the Executive Summary and the Report are welcomed as they are clear
         recommendations to implement specific projects but these should all have the required commissioning
         date (or first to last unit commissioning dates) as has been done with point 5 to stress the need.

     19. The Conclusions need to make a clear statement regarding why the import of bulk hydropower is
         not considered in the report.

     20. The Conclusions need to make a clear statement as to whether the studies identified any
         inadequacies or constraints in the transmission system that could impact on the quality of supply.

     21. Question - are the amounts of dispatched imbedded generation (municipal and industrial)
         included in the load forecast (fig 1 page 5) to give the total consumption since imbedded
         generation is included in the supply side availability figures.

     22. Executive summary text editing:
          Page I para 1 line 4 : ensuring not insuring
          Page IV section 5 line 2 : as dictated by the ARC of the NER
          Page VII fig 5 title : NIRP reference plan and alternatives at
          Page VII conclusion 1 last sentence : (including losses) it is not sufficient.
          Report text editing :
          Page 14 last line : time-of use tariff is available.




                                                            IV
72

     IV. ESKOM GENERATION COMMENTS ON NIRP 2003/4 REFERENCE CASE VERSION 18 DECEMBER
     2003

     Provided by Mr. Gerhard Loedolff

     1) Is this a report by Eskom, ERI and NER, or is it a report by Eskom & ERI for NER, or is it a report by NER,
        drafted by Eskom and ERI. Depending on the answer, the text needs to be reflective of the position the
        report is written from.

     2) In the Introduction, it is referred in par 1 that the study is conducted to provide information on the
        economics of new electrical generation. Is this really so, or is objective more fold, and directed at
        identifying the appropriate level of new capacity additions to meet future supply; assess the
        appropriateness of various solutions and reflect on the economic outcome of such solutions?

     3) Also in the Introduction, a number of bullets is listed as steps defined and carried forward from NIRP1.
        However, on many of them the objective (to achieve what?) is not specified. Therefore, as an example,
        you are 'Selecting a preferred Plan' to achieve what?; or you are analyzing financial consequences of
        What?

     4) At the top of page 2, the report refers to the "NER agreeing to...." . Surely the NER appointed ARC to
        develop the planning assumptions.

     5) Especially on DSM, the report contains a huge amount of debate and policy and management suggestions
        to NER. To my knowledge this is outside of the scope of this report. I would have expected a good review of
        assumptions, with possible implications of deviation on the study outcomes. Refer pages 14- 16
                                          ARC and Public Comments on NIRP2 Reference Case




6) Paragraph 3.3: 2% to 3% represents a movement of 50%. can you not

7) close the gap somewhat?

8) Par 3.5: Surely such a statement needs some explanation?

9) Par 3.7: No indication is provided on the relevance of this information

10) Par 3.7.2: No indication as to why sales growth is assumed to be higher in earlier than later years. Also
    impact of TOU based WEPS rates not referred to.

11) Who is DOE?

12) Wording around Eskom's plant performance is again confusing. Suggest it reflects that it is based on
    90:7:3, with an average long term EAF of 88%, as also discussed at the last ARC meeting, and as used.
    (p27)

13) In a number of places in the report, restructuring of the text to bullet format will substantially ease reading.

14) Par 5.4: New Supply Side Options were recognized, BUT not included. Why?

15) R60/ton for coal to the 6-pack appears underestimated by around R20/ton, while overnight capital appears
    about R100/kW higher that latest used in ISEP

16) Suggest that the screening curves should focus only on proven technologies, and not on the speculative /
    developing ones in the main report. Can include an annex to give relative position to adopted supply side
                                                                                                                       73
    options

17) Section 6 deals with some issues around reliability of supply. Rather than opening a debate here, which
    could also be based on invalid assumptions like Cost of non supply, should the current regulatory standard
    w.r.t. QoS criteria as applied via licenses not rather serve as the basis?

18) The plan on p29 indicates OCGT's being decided upon ahead of Komati and Grootvlei. From the
    discussion at the ARC meeting this appears to be the result of artificially constraining the return of
    Grootvlei and Komati. Is this still so; If so, Please correct!
     National Integrated Resource Plan 2 - Appendix A



                                                              V
     V. COMMENTS FROM DARLING IPP

     Provided by Mr. Herman Oelsner

     I would like to put on record our disappointment that input given in our E-mail dated 2nd of December (copied
     below) has not found its way into the modelling.

     The committee could find itself soon in a very embarrassing situation since Renewable Energy Sources and
     Technologies have largely been ignored. This despite clear government (RE White Paper) policy and targets
     and numerous country-wide spread activities in this sector.

     From the Renewable Sector's view it appears to be "business as usual".

     In order to enable a more valuable contribution from the Renewable Energy Service sector, I would like to
     propose the co-option of two new additional members from RES to the committee which can assist me in
     effectively contributing to the process.



     Since there was only short time to look at the documentation you distributed, I had no opportunity to query
     certain figures in the modelling results.

     For example:

     The lead times for wind are much shorter and should be recorded as:
74
     a Concept Phase                     one year
     b Feasibility phase                 two years (Darling as first-off is a bad example/ see
                                         Klipheuwel EIA took less a year)
     c Investment and construction       one year maximum for 20 MW

     Extremely short lead times and the possibility of small decentralized plants which can integrate into existing
     national grid are the outstanding advantages of wind electricity generation.

     a.Life (yrs)                      25 (not 20)
     b.Maximum Capacity factor         35% (not 19.27)
     c. Minimum Capacity factor        20 %(not 4,85)
     d.Operational Mode is not in base load but generation takes place dominantly during peak and standard times.
     e.Energy Limit per station of 20 MW is 54 GWh/a not 33.5

     There are several abbreviations in the documentation, which I do not recognize. A key would be of great help.

     Renewable Energy generation, solar / wind does not appear at all in the NIRP Plans for Publication projecting
     plans until 2022.

     At last weeks World Wind Energy Conference Honourable Minister Ms Phumzile Mlambo-Ngcuka announced
     that the government will ratify in Cabinet on Wednesday the 3rd of December the Renewable Energy White
     Paper.

     The White Paper sets a 10-years target of 10 000 GWh renewable energy contribution to final energy
     consumption by 2013, from wind, biomass, solar and small hydro. Since wind is arguably at present the most
     economic mode of electricity generation and a well developed technology, wind will play a dominant role in new
     generation plant in South Africa in the near future. Assuming that wind will contribute at least half for this target,
                                         ARC and Public Comments on NIRP2 Reference Case




would translate into 10 off 20 MW wind farms, which is very realistic and rather conservative.

The implementation progress will be evaluated mid-term in order to find out whether it will be necessary to
revise the White paper.

Based on the White Paper, the first draft renewable energy strategy is expected to be published in February
2004.

I respectfully submit that RE must be included in the NIRP Reference Plan.




                                                     VI
                                                                                                                    75


VI. COMMENTS FROM KELVIN IPP

Provided by Mr. Donald Bennett

I have been through the document and do not have any major comments about it in general. The only point
that I would like to make is that the reference to Kelvin Power Station is incorrect (Appendix 1 page 3 table 2).
Kelvin (A & B) is grouped with the “Munics 1 & 2”. Kelvin was in fact privatized in November 2001 and operates
as an IPP (although only serving the City of Johannesburg under a Long Term Agreement at this point in time).
Reference is also made in point 5.2 (Non-Eskom System Existing Capacity) page 15 of the main document to
the fact that Kelvin generation is included in the “Munic 1 & 2” blocks.
     National Integrated Resource Plan 2 - Appendix D




     Appendix D

     NIRP ARC COMMENTS ON THE NIRP2 STAGE 2 REPORT

     I.   ESKOM KSACS COMMENT                                           77
          Concept of expectation / Definition of the word “expected”    77
          Uncertainties considered (Section 4)                          77
          Range of values for uncertain parameters (Section 4)          77
          Definition of scenarios (futures) (Section 4)                 78
          Simulation of plans , and cost estimates (Sections 4 and 5)   79
          Renewable technologies (Section 7.3)                          80
          Conclusion                                                    80

     II. EIUG COMMENT ON NIRP2 STAGE 2 REPORT                           80

     III. DEPARTMENT OF MINERALS AND ENERGY COMMENTS
          ON RENEWABLE ENERGY SCENARIO                                  81




76
                                                NIRP ARC Comments on NIRP2 Stage 2 Report



                                                        I

I. ESKOM KSACS COMMENT                                                                           22 August, 2004




COMMENTS ON NIRP 2003/4 STAGE 2 REPORT

Personal comments by JL Pabot

The principle of the methodology used in the report to develop an optimal plan seems reasonable, but I have
serious doubts on the implementation techniques applied with this methodology to derive the recommended
plan, which in my opinion invalidates the results.

Therefore I cannot determine from the report if the recommended plan is truly the best plan.

Concept of expectation / Definition of the word “expected”

The word “expected” is mentioned at many places in the document, but It does not seem to have the usual
meaning, which is confusing for the reader.

For me, when a parameter (e.g. availability of plant) has a statistical distribution of values over a range (low,
median, high), the expected value is the probability weighted average over the range, usually close to the
median value.
                                                                                                                      77
In the whole report, the expected value of a parameter seems to always be at one end of the range, always the
least favourable (i.e. the best) end for capacity investment (the side requiring the least capacity installed). All
other values are worse, i.e. lead to higher capacity investment.

Uncertainties considered (Section 4)

The report addresses only uncertainties that have a direct impact on the load/capacity balance.

Other uncertainties that might have a significant impact on the plan, are not considered, e.g. fuel cost or fuel
availability (except for imported hydro).

Range of values for uncertain parameters (section 4)

As mentioned in comment 1 above, the range of values selected for the analysis for the uncertain parameters
is systematically biased towards the worst end of the range (i.e. the values requiring the highest capacity
installed). The expected value being the lowest value considered.

This obviously will lead to an optimal plan requiring more capacity than truly required.

Examples of parameters that could have more favourable values than expected, i.e. outside the range
considered in the study:

& Load growth forecast: Surely there is some probability that there might be an economic crisis due to some
  international event (high petrol price, financial crisis, war in the middle, crisis in a neighbouring country,
  impact of AIDS, etc) which might lead to lower energy consumption growth. Export sales to SAPP might also
  decrease if new stations are built in SAPP.
     National Integrated Resource Plan 2 - Appendix D




     & Imported energy: Surely there is some probability that Eskom could import more power from SAPP, e.g. from
       Inga 1-2 after refurbishment, from ZESCO in Zambia, when they have finished refurbishing their plant, from
       Kafue Gorge lower, from Kudu in Namibia, if the plant is built primarily for supplying Nampower, and if Eskom
       buys the balance of energy generated, etc…

     & Interruptible load (IL) and DSM: it is quite likely that in future Eskom may contract for more IL capacity,
       possibly with less usage constraints, and for more demand management programmes (DMP) or DSM
       programmes.

     The probabilities allocated to the parameters values on the worst (for MW to be built) end of the range seem
     excessive, e.g. 70 % for available capacity

     Definition of scenarios (futures) (Section 4)

     There is an implicit (not discussed in the text) assumption in the scenarios: the 4 uncertain parameters are of a
     random nature, and are independent.

     At least one parameter, the existing IL MW capacity, is not a random parameter. The IL contracts are to a large
     extent under Eskom's control, i.e. their continued existence cannot be used in a probabilistic analysis.

     The 4 parameters are not independent in the long term, and a few of them are also not independent in the short
     term.

     & IL MW capacity is not independent of the other parameters. Eskom is unlikely to cancel (or renegotiate)
       these contracts if there is a shortage of capacity to replace them (e.g. if there is some indication of a high load
       growth, or if imports are curtailed).
78

     & The same can be said of the municipal generation (it will not be retired if its capacity is needed) and of the
       plant availability (if low, Eskom will strive to improve it if there is a shortage of capacity, at least within a year or
       two).

     & The MW capacity of DSM programmes contracted by Eskom is not independent of the other parameters:
       Escom will invest more resources in developing these programmes if there is more need for them (e.g. for
       high load growth)

     & The MW capacity of imports is also not independent of the other parameters: Eskom will not give up part of its
       contractual MW allocation form Cahora Bassa at Songo if this capacity is needed in South Africa. Also
       Eskom will import more energy form SAPP if there is a need for it, mainly because Eskom can then offer a
       higher price.

     & The load forecast includes the energy exported to SAPP (about 2100 MW peak power in 2004). Most of
       Eskom's export contracts are expiring soon, part of the existing contract are non-firm (Eskom can decline to
       sell energy at 1 day notice). In future Eskom will reduce exports if there is a prospect of capacity shortage.

     Therefore the definition of the scenarios (by combining all the values of the uncertain parameters), and the
     derivation of their probabilities (by multiplication of the individual probabilities), is not appropriate for this
     expansion planning exercise.

     It would be more appropriate to define a set of 3 values for the uncertain parameters (low, medium, high),
     assess their interdependence, and derive by judgement a set of 4 or 5 realistic futures.

     With reference to my comment 1, it is very odd that the future (scenario 1) with all parameters having their
     expected values has a probability of 3 % and that all the other futures (with a combined probability of 97 %) are
                                                 NIRP ARC Comments on NIRP2 Stage 2 Report




all worse (in terms of capacity requirements), even much worse for many of them.

This reflects a bias in the study towards the worst cases, requiring higher installed capacity.

Simulation of plans , and cost estimates (sections 4 and 5)

The NIERP planners (the authors of the study) design 16 plans for the 16 futures. They then simulate each plan
for each future, over the study period (8 to 20 years) assuming that the plans remain as designed for the whole
duration of the study period.

This does not make sense: a plan designed for a low load growth forecast would not remain unchanged if a
higher load forecast would occur. Eskom would surely take some action within a year or two, e.g. order OCGTs
(lead time = 1 or 2 years), contract more IL, etc…The simulation should make some provision for the
adaptability of the plans to the future under which they are simulated.

The consequence of this methodology implementation technique is to derive very high costs for the plans
designed for favourable (i.e. good, as defined in Comment 1) futures when run under unfavourable (bad)
futures.

Example: The table below shows the PV cost of plans 1 (designed for future 1, the best) and 2 (designed for
future 2, the worst) when run under futures 1 and 2. It shows also the expected cost (probability weighted cost,
PWC) of the plans over the 16 scenarios. These costs were provided by the NIERP planners in a spreadsheet.



    Simulation              Scenario              Scenario              PWC
    20 years                future 1              future 2              cost
                                                                                                                      79

    Plan 1, cost Rm         182427                746958                403520
    Plan 2, cost Rm         215676                240722                222918



The high cost of plan 1 (low capacity installed) for future 2 is due to the high amount of unserved energy
calculated for that case, and to the high cost of unserved energy (20 R/kWh).

& The cost of unserved energy is unrealistic for that case. This is an outage cost, not a shortage cost. As
  mentioned extensively in the literature, outage costs are incurred by consumers when unexpected outages
  occur, and shortage costs are incurred when outages are expected (e.g. when there is a shortage of capacity
  over long periods, e.g. 6 months). Shortage costs are much lower that the outage costs. Obviously in the
  case of plan 1 run under future 2, outages would be expected, and a shortage cost should have been used
  for unserved energy, not an outage cost.

& The high amount of unserved energy of plan 1 for future 2 is totally unrealistic. Plan 1 would not remained
  unchanged over 20 years for future 2, nor for any future worst than future 1 (i.e. for all other futures) over 20
  years, nor over 8 years, nor over 4 years.

The impact of this invalid methodology implementation technique is to weigh all plans designed for the most
favourable futures with very high unserved energy costs, which leads to very high probability weighted costs
for these plans, and to their elimination on the basis of high PWC.

In effect the study is seriously biased towards plans designed for bad futres, i.e. with high capacity installed.

The study would be more valid if the plans were simulated as adaptable, i.e. if additional gas turbines could be
installed at short notice when needed, or if plant commissioning could be delayed at short notice (with
     National Integrated Resource Plan 2 - Appendix D




     additional costs) to cope with changing futures.

     Renewable technologies (Section 7.3)

     Plan 14 with 320 MW of renewable capacity installed over the nest 20 years only pays lip service to the
     government's desire to introduce renewable energy in the generation of electricity. I did not check the white
     paper on energy policy, so maybe the study is realistic in this respect.

     A more visionary scenario would consider a more commendable goal, e.g. 5 % or 10 % of all new installed
     capacity to be based on renewable energy.

     If in future South Africa moves to gas fired plants with expensive gas fuel (e.g. LNG) for electricity generation,
     wind power plants may become economical because of their associated fuel savings, even if they receive no
     capacity credit to meet the peak load:

     Conclusion

     The principle of the methodology used in the report to develop an optimal plan seems reasonable, but the
     planning techniques used in this study for the implementation of this methodology, are in my opinion invalid.

     In my opinion, the shortcomings identified in the above comments invalidate the results, i.e. I cannot determine
     from the report if the recommended plan 14 is truly the best.

     I also cannot determine if the recommendation to build 720 MW of Open cycle gas turbines in 2006 is
     reasonable.
80
     However since the DME representatives stated at a recent meting at the NER that Eskom and the DME had
     already agreed that these 720 MW of OCGT would be built in 2006, this capacity is in effect committed, and my
     opinion on this point is irrelevant.




                                                           II
     II.EIUG COMMENT ON NIRP2 STAGE 2 REPORT

     Provided by Mr Arnot Hepburn, EIUG Sectretariat

     EIUG members in their review of the NIRP2 Stage 2 report are in agreement that the comments submitted
     previously and the EIUG concerns raised again at the last NIRP ARC meeting are all that need be given for the
     current report. As stated by the EIUG at the last NIRP ARC meeting it is imperative that the release of the report
     should not be delayed in any way as the information is urgently needed by decision makers. It is appreciated
     that it will never be possible to achieve 100% projections but what is known must be published.

     The EIUG gives it's assurance that it will give what ever support possible to future NIRP studies and reports
     and submit to you the following comments for consideration in the preparation of NIRP3:

     Demand- and consumption growth forecasting

     1. There appears to be uncertainty and a lack of agreement on the actual figures in the baselines as used,
        and on the expected demand growth in future. For example, in table 1 in the NIRP HIGH column, the sales
                                                NIRP ARC Comments on NIRP2 Stage 2 Report




    in 2003 is shown as 197 401 GWh. From the Eskom annual report the actual Eskom sales was 196 980
    GWh. The Eskom figures excludes the energy sent out from non-Eskom generation, and therefore it
    appears that we are projecting future demand and consumption figures from a too low base. This may
    have a significant effect on the actual reserve situation in future.

    Further to the above, the 0,5% increase (on top of the NIRP 2 base case) in the demand growth from the
    ERC study to arrive at the high growth scenario, seems modest compared to the actual year on year
    growth of 5% over the last few years, even if we know of large new industrial loads that came on line
    recently. It the actual rate of growth does not taper off to 3% during the remaining months of 2004, we
    should consider another upward adjustment to the forecast next year.



2. In table 4, the probabilities assigned to the different primary assumptions seem debatable. A probability of
   0.4 is assigned to the high load forecast, while recent history and the current economic outlook seem to
   support a higher probability. Similarly, the 0.6 probability that plant outages will be higher implies that
   Eskom will not be able to maintain their equipment availability, against a history that they never dropped
   below 88%, even in a bad year like 2003. Given the fact that the impact of higher load growth is more
   severe than the rest (figure 1), these assumptions greatly influence the results in the 16 scenarios. If a 0.6
   probability is assigned to the high load growth assumption it will result in lower reserve margins and higher
   PWC in scenarios 2, 6, 8 and 16. If such a relative modest change in a primary assumption can lead to a
   scenario of zero reserve margin, the report should point that out, assign a probability to it and sensitize
   government to such a possible eventuality.

3. The report is silent on brown fields CF base load options, while it is common knowledge that these
   possibilities have shorter lead times and are cheaper than green fields options.
                                                                                                                       81
4. The inclusion of 300 MW CSP, at R10bn capital does not seem to be well researched. It is unclear why
   wind power cannot contribute more to the 10 000 GWh target of Government.




                                                       III

III. DEPARTMENT OF MINERALS AND ENERGY COMMENTS ON RENEWABLE ENERGY SCENARIO

Provided by Mr. Andre Otto

The levelised cost for wind is for all discount rates as indicated in table 9 lower than that for CSP wind, what was
then the decision rationale and calculation used to "choose" the significant 300 MW CSP vs only 20 MW wind?

The GEF South African Wind Energy Programme (GEF council approval of full size project in progress) aimed
at an installed wind farm capacity of up to 50 MW (5 MW Darling + 45 MW through SAWEP) by December
2009. 50 MW is regarded as the minimum installed wind farm capacity to attract investment and to start up a
local manufacturing and exporting capability which should lower further wind farm investment in SA.

The Eskom CSP demo plant (as far as I know) is based on 100 MW which still have to secure the necessary
financing and authorisation.

A macro-economic analysis (see attached) done (in cooperation with National Treasury) in motivating for
     National Integrated Resource Plan 2 - Appendix D




     Cabinet approval of the White Paper on Renewable Energy (RE WP) and forming the basis of the RE Strategy
     (in progress), indicates a least cost based approached to be followed in implementing the RE WP target of 10
     000 GWh RE contribution to energy consumption by 2013.

     The Renewable Energy Market Transformation project (REMT) which is approved and financed by the SA
     Government, GEF and PCF via the World Bank, investment will start in 2005, aimed at establishing RE
     capacity to produce the 1st 1000 GWh of the 10 000 GWh target by 2008. Pending the success, the project will
     be expanded to an additional 3000 GWh to be introduced the following 4 years. The REMT is based on the
     macro-economic least cost approached of introducing the RE WP. The following RE technologies and
     applications (see table below) have been identified through a DME and World Bank economic and financial
     due diligence of the REMT.



     Recommendation

     The NER NIRP be updated with the RE technologies and applications as identified through the DME and
     World Bank economic and financial due diligence of the REMT, including CSP and wind energy as indicated in
     the attached table.

     Please take note: The REMT envisaged to introduce 744 MW RE capacity over the 1st 8 years of the RE target
     period which is already in excess of 320 MW.
     Footnote in RE WP published in Government Gazette vol 466, 14 May 2004, No 26169:

         "It is estimated that if the cumulative renewable energy capacity by end of 2013 is
         approximately 1667 MW, then 10 000 GWh renewable energy would have been consumed over
         the 10 year period"
82

                                                        Dec 03 Jan 07         MW         Annual
                                                        to Jan to Dec                    load
                                                        07 GWh 11                        factor
                                                               GWh                       %
                       REMT                             Phase 1 Phase 2
                       Sugar      sugar mills spare            55        55         12        52
                                  capacity
                                  reduced process             109       109         24        52
                                  steam
                                  full scale cogen                      551        157        40
                       SWH                                    175     1,000        190        60
                       Pulp       Ngodwana                     65        65          8        90
                       &Paper
                                  Additional projects          20       170         21         90
                       Hydro      Identified projects         210       210         63         38
                                  Additional projects          75     1,000        300         38
                       LFG        Identified projects         240       240         34       80.8
                                  Additional projects          51       600         85       80.8
                       GEF
                       SAWEP
                       Wind                                             120         50        28
                       Eskom
                       CSP                                              613        100        70
                       Total                                 1000      4733        894


     Total 5733 GWh (57% of 10 000 GWh target)

     Balance 4267 GWh most probably taken up by biofuel
Notes




        83
84

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:47
posted:2/23/2011
language:English
pages:84