This content is being transitioned into our department's parent site - following Feb 2020 Machinery of Government changes. In the meantime, you can access climate change content:

Can't find something? Tell us using the website feedback form

100 per cent renewable study - modelling outcomes

Australian Energy Market Operator, 2013

Executive briefing

1. Introduction

On 10 July 2011, the Australian Government announced its Clean Energy Future Plan.

As one initiative under that plan, the Department of Climate Change and Energy Efficiency (DCCEE) commissioned AEMO to undertake a study which explores two future scenarios featuring a National Electricity Market (NEM) fuelled entirely by renewable resources.[2] DCCEE specified a number of core assumptions on which AEMO was asked to base its study.

This document is an Executive Briefing for the full report of the study undertaken by AEMO.

Any study of future energy supply-particularly one based on solely renewable energy-must consider how existing technologies will develop, and how new technologies will mature and become commercially available. Some commercially available renewables have limited scope for future development; others are still emerging, but may have commercial potential.

This study considers two scenarios with differing views about how quickly renewable technologies will develop over time. Accordingly, power systems with differing configurations are expected to emerge in each scenario.

The modelling undertaken presents results for four selected cases, two scenarios at two years, 2030 and 2050. The first scenario is based on rapid technology transformation and moderate economic growth while the second scenario is based on moderate technology transformation and high economic growth. The modelling includes the generation mix, transmission requirements, and hypothetical costs for each.

Given its exploratory nature, this study should be regarded as a further contribution to the broader understanding of renewable energy. The findings are tightly linked to the underlying assumptions and the constraints within which the study was carried out. Any changes to the inputs, assumptions and underlying sensitivities would result in considerably different outcomes.

  1. The results indicate that a 100 per cent renewable system is likely to require much higher capacity reserves than a conventional power system. It is anticipated that generation with a nameplate capacity of over twice the maximum customer demand could be required. This results from the prevalence of intermittent technologies such as photovoltaic (PV), wind and wave, which operate at lower capacity factors than other technologies less dominant in the forecast generation mix.
  2. The modelling suggests that considerable bioenergy could be required in all four cases modelled, however this may present some challenges. Much of the included biomass has competing uses, and this study assumes that this resource can be managed to provide the energy required. In addition, while CSIRO believe that biomass is a feasible renewable fuel[3], expert opinion on this issue is divided.[4],[5]
  3. The costs presented are hypothetical; they are based on technology costs projected well into the future, and do not consider transitional factors to arrive at the anticipated cost reductions. Under the assumptions modelled, and recognising the limitations of the modelling, the hypothetical cost of a 100 per cent renewable power system is estimated to be at least $219 to $332 billion, depending on scenario. In practice, the final figure would be higher, as transition to a renewable power system would occur gradually, with the system being constructed progressively. It would not be entirely built using costs which assume the full learning technology curves, but at the costs applicable at the time.

It is important to note that the cost estimates provided in this study do not include any analysis of costs associated with the following:

  1. Land acquisition requirements. The processes for the acquisition of up to 5,000 square kilometres of land could prove challenging and expensive.
  2. Distribution network augmentation. The growth in rooftop PV and demand side participation (DSP) would require upgrades to the existing distribution networks.
  3. Stranded assets. While this study has not considered the transition path, there are likely to be stranded assets both in generation and transmission as a result of the move to a 100 per cent renewable future.

Costs for each of these elements are likely to be significant.

This report is not to be considered as AEMO's view of a likely future, nor does it express AEMO's opinion of the viability of achieving 100 per cent renewable electricity supply.

2. Assumptions and limitations

The assumptions below are fundamental to the study outcomes. Some are drawn from the scope of works published by the DCCEE in July 2012[6]; others result from AEMO's initial investigations, and were published and discussed with stakeholders in September 2012.

Assumptions given in the scope of works are as follows:

  • The scope of works acknowledges the inherently uncertain nature of this study given the uncertainty around technologies that could emerge in the intervening 40 years, the cost of those technologies, and the potential for regulatory change in that timeframe.
  • The modelling data was taken from the 2012 Australian Energy Technology Assessment (AETA 2012) produced by the Australian Government Bureau of Resources and Energy Economics (BREE). CSIRO and ROAM Consulting were commissioned to provide other key data, including projected resource availability and technology development rates.
  • The study limits consideration to the electricity sector, and does not include the associated social, political and economic changes likely to arise from the scenarios modelled.
  • The transition path from the current power system to the modelled 100 per cent renewable power systems is not considered. The hypothetical capital costs assume building all the new generation and transmission infrastructure at the estimated 2030 or 2050 costs. This means that the full advantage of anticipated technology cost reductions and performance improvements are included.
  • Distribution system costs are not included in this study. This does not imply that distribution systems would be unaffected.
  • No allowance has been made for land acquisition costs, the costs of stranded assets or possible research and development expenditure needed to drive the forecast cost reductions.
  • The presented capital costs are lump sum estimates and exclude all financing costs. The impact on wholesale prices does include an assumed weighted average cost of capital but excludes interest during construction.
  • The scope explicitly excludes any consideration being given to nuclear, gas, coal, and carbon capture and storage generation.
  • The study does not consider electricity supply outside of the NEM regions. This means Western Australia and the Northern Territory are excluded.

The following assumptions and limitations are also relevant when considering the modelling results. Some are drawn from AEMO's Modelling Assumptions and Input Report released in September 2012.[7]

  • While AEMO sourced the best cost estimates currently available for renewable generation technologies under Australian conditions[8], these estimates are likely to change over time as the technologies evolve.
  • No consideration is given to costs of government policies that may be needed to drive the transition to the modelled 100 per cent renewable power systems.
  • Other than an anticipated uptake of electric vehicles (EVs), no other fuel shifting from gas or petrol is considered.
  • The demand assumptions used in this report are based on AEMO's 2012 National Electricity Forecasting Report (NEFR) with revisions to fit with the 100 per cent renewables scenarios extended to 2030 and 2050 using a regression model.
  • The costs of developing the demand side participation (DSP), energy efficiency measures, and EV infrastructure assumed in the modelling have not been considered.

3. Approach

The scope of works published by DCCEE in July 2012 requested AEMO to explore optimised combinations of renewable electricity generation sources, and associated transmission infrastructure and energy storage systems under two scenarios:

  • Scenario 1: Rapid transformation and moderate growth-this scenario assumes strong progress on lowering technology costs, improving demand side participation (DSP), and a conservative average demand growth outlook in the lead up to the year being modelled.
  • Scenario 2: Moderate transformation and high growth-this scenario assumes current trends in lowering technology costs, moderate DSP, and robust economic growth in the lead up to the year being modelled.

Each scenario was modelled under two timeframes: 2030 and 2050. This resulted in a total of four cases being modelled: Scenario 1 (2030), Scenario 1 (2050), Scenario 2 (2030), and Scenario 2 (2050). Under the scope of works, AEMO was required to prepare a report containing the following:

  • Scenario inputs for a 100 per cent renewable electricity supply at 2030 and 2050.
  • Generation plant and major transmission networks required to support each case.
  • Hypothetical capital cost requirements for each case based in today's dollars.

In line with the published scope of works, AEMO undertook the following key steps:

  1. Resource investigation
    AEMO engaged expert consultants to estimate the potential quantity and quality of a range of renewable energy resources that would be accessible by 2030 and 2050 for use in electricity generation or energy storage technologies in selected NEM locations.
  2. Scenario input development
    Based on the resource investigation, AEMO developed modelling inputs consisting of the availability of various generation and storage technologies and their projected capital and operating costs in 2030 and 2050.

The Bureau of Resources and Energy Economics (BREE) Australian Energy Technology Assessment 2012 (AETA 2012), which estimates the generation costs for a range of technologies to 2050, was taken as a starting point for the costs. The AETA estimates were augmented with further inputs on the future costs of some technologies provided by CSIRO and ROAM Consulting.

Using the 2012 NEFR as a starting point, AEMO also developed specific annual electricity consumption projections for each of the four cases to suit the scope of works.

Steps 1 and 2 were documented in the Modelling Assumptions and Input Report released in September 2012.[9]

  1. Modelling
    Using information from the steps above, AEMO undertook modelling to determine the following:
    • The generation and energy storage combination most suited to each case that met the reliability standard at least cost.
    • The likely scale of transmission network augmentation required under each case.
    • The hypothetical capital costs for each case, including indicative estimates of energy price outcomes for consumers.

4. Modelling method

4.1 Modelling overview

In line with the scope of works, AEMO used least-cost modelling to determine an optimal combination of generation, storage and transmission investments to match the forecast customer demand for each case. The modelling also factored in a requirement to meet the current reliability standard in the NEM.

For each case, two different modelling tools were used:

  • A probabilistic generation expansion model.
  • An hourly time-sequential model for the year being studied.

The mathematical modelling results were reviewed from an operational perspective (to check that the resulting power system could be securely managed) and from a transmission network perspective (to estimate the transmission capability required to transport generation to load centres).

This process was repeated several times for both modelling tools, to take into account operational and transmission review feedback. After several iterations, the modelling for each case produced an optimised generation mix and transmission network which satisfied the operational and transmission assessments. The process is shown pictorially in Figure 1.

Figure 1: Methodology process overview

4.2 Probabilistic modelling

The probabilistic modelling used Monte Carlo methods to simulate 5000 random days for all four cases. Each random day contained hourly profiles of each renewable energy resource by location as well as customer demand and observed the historical correlations between each renewable resource, and between renewable resources and demand. The model simulated the dispatch of generation, demand side participation and daily storages (such as at Concentrating Solar Thermal (CST) plants and pumped hydro) to meet the customer demand at least cost across each of the random days.

For each case, the model was used to find the lowest cost mix of generation and storage that met the current reliability standard.

4.3 Time-sequential modelling

Time-sequential methods were used to compare the hourly demand calculated for 2030 and 2050 with the actual renewable resource data for each renewable technology using a typical year's climate data. This method addressed the following:

  • Capacity sufficiency (the ability to meet maximum demand with the available renewable resources).
  • Energy sufficiency (the ability to manage demand during sustained periods of time when generation from intermittent sources is low).

The time-sequential modelling was also used to calculate the power flows across the transmission system, which was then reviewed in the transmission assessment. Finally, the time-sequential modelling was used to evaluate technological issues such as generator ramp rates, share of non-synchronous generation and other metrics identified in the operational assessment.

4.4 Operational and transmission assessments

The modelling assumed that the existing transmission system was available in all four cases. The transmission assessment investigated what additional transmission assets would be required to transport the modelled generation production to load centres at the lowest overall cost. This investigation explored both new transmission lines as well as upgrades to the existing transmission system.

The operational assessment considered a range of technical issues including frequency control and system inertia. Operational assessments also aimed to identify any generation mix adjustments likely to be required for system security purposes.

5. Key inputs

5.1 Electricity demand projections

A fundamental input to power system modelling is a forecast of customer demand. However, forecasting 20 and 40 years into the future is inherently difficult and there are many social, economic and technological changes which could invalidate these forecasts.

Table 1 shows the forecasts for annual energy consumption and diversified[10] maximum demand used in the modelling.

Table 1: Electricity demand projections

2011 (for comparison)

Scenario 1 2030

Scenario 1 2050

Scenario 2 2030

Scenario 2 2050

Annual energy (TWh) not accounting for PV and EV






Annual energy (TWh) Rooftop PV generation






Annual energy (TWh) EV






Annual energy (TWh) accounting for rooftop PV and EV






Maximum demand (GW) not accounting for rooftop PV, EV, and DSP*






Maximum demand (GW) accounting for rooftop PV, EV, and DSP*






* Most probable, or 50% POE

The photovoltaic (PV) figures are based on AEMO's 2012 Rooftop PV Information Paper.[12] The electric vehicle (EV) figures were modelled for this report based on the EV modelling used in AEMO's 2011 Electricity Statement of Opportunities.[13]

5.1.1 Trends affecting demand: energy efficiency, rooftop PV, and demand side participation

These projections demonstrate relatively low growth in demand, reflecting a less energy-intensive future which is primarily driven by energy efficiency, rooftop PV and demand side participation (DSP).

This is particularly evident in Scenario 1, which assumes rapid transformation of renewable technologies, and where PV, energy efficiency and DSP more than counter any demand increases caused by expected EV uptake.

In all four cases, anticipated rooftop PV generation growth is high enough to contribute to the NEM becoming winter peaking - a major change from the situation today.

Expected DSP increases result from appropriate incentives being implemented to enable consumers to alter the quantity and timing of their energy consumption to reduce costs. This drives a shift in consumption patterns that responds to market needs and takes advantage of high renewable generation availability (usually when PV is peaking) to reduce energy spills.

Scenario 1 assumes up to 10% of demand is available for DSP and Scenario 2 assumes up to 5%. For each case modelled, half of the DSP is assumed to be curtailable load (that is, demand which can be reduced at a given cost) and half is modelled as 'movable demand' which can be consumed at an alternative time that day.

Both components of DSP represent voluntary customer behaviour. These are separate to unserved energy (USE), which is involuntary curtailment of customer demand. The reliability standard discussed in Section 5.2 refers to USE only, not DSP.

Figure 2 shows a sample forecast demand profile from the study, and demonstrates how DSP results in demand shifting from evening to midday.

Figure 2: A sample forecast demand profile demonstrating load shape changes

5.2 Reliability standard

In accordance with the scope of works, AEMO used the current NEM reliability standard as the benchmark for this study. The current standard is that the maximum USE is less than 0.002% of annual energy consumption.

5.3 Energy resources and location

AEMO's consultants investigated a range of historical weather and spatial data to develop estimates of the energy resource available from each technology at each of the 43 NEM areas selected for this study. Consideration was given to issues that might limit access to these resources, such as competing land uses, topography and population density.

The consultants then used these energy resource estimates to calculate the maximum installable generation capacity at each location.

A summary of this work is shown in Table 2. More information on each technology considered under the types listed below is provided in the full report.

Table 2: Total resource by technology


Maximum installable generation capacity (GW)

Maximum recoverable electricity


Wind - onshore (greater than 35% capacity factor)



Wind - offshore (greater than 50% capacity factor)



Solar - Concentrating Solar Thermal (CST)/Photovoltaic (PV)

18,500 / 24,100

41,600 / 71,700

Geothermal (EGS or hot rocks)



Geothermal (HSA or sedimentary aquifer)













25,700 / 31,300

86,800 / 116,900

Current generation in the NEM (all technologies)



AEMO chose a subset of NEM locations to include in the modelling. Where possible, this subset provides the following:

  • The best resource for each technology in terms of energy production capacity factor and minimal seasonal variation.
  • Reasonable spread across the entire NEM, to minimise fluctuations due to local weather conditions.
  • Other geographical advantages, such as siting generation reasonably close to the transmission system and major load centres.

As a result, the modelling used renewable technologies distributed over a wide area. The general location used for each technology is indicated by shaded circles on Figure 3. For simplicity, the size of each shaded circle has been kept small, but each represents deployment of technologies distributed over a much larger area, including neighbouring locations with equally good energy resources.

Figure 3: Subset of selected technology locations

5.4 Transmission network

Many of the renewable resources within the subset identified are in locations that are remote from the current transmission system. Based on these locations, high-level transmission options were developed for use in this study. These included potential upgrades to the existing transmission system.

Figure 4: High-level transmission options

The red arrows show existing cross-border interconnectors that connect the major load centres (grey dots) in the NEM. These interconnectors (with the required reinforcements defined for each case) will distribute renewable supply between the major load centres.

Significant amounts of generation will be connected to the transmission system (yellow arrows) from more remote renewable generation clusters (blue dots). An estimate of local demand in these generation clusters was used when determining new transmission requirements.

5.5 Cost inputs and assumptions

5.5.1 Generation technology costs Capital costs

As new technologies emerge and mature, their costs generally follow a curve such as the generic learning curve, or Grubb curve, shown in Figure 5.[14] The renewable energy resources considered in this study are currently at varying stages of maturity and are likely to differ in terms of performance improvements and cost reductions in the coming decades.

Figure 5: Generic Grubb curve showing typical technology cost cycle

The generation technology costs used in this study account for expected cost improvements by 2030 and 2050 using outputs from CSIRO's Global and Local Learning Model (GALLM).

While technology costs are expected to fall over time, resource costs will generally increase as higher quality resources and more favourable sites close to the transmission system diminish, leaving lesser resources available. This leads to longer-term stabilisation of generation costs for some technologies.

Table 3 lists the projected capital costs for each generation technology (using real 2012 dollars per kilowatt of electricity generation capacity) taken from the AETA 2012 or provided by AEMO's consultants, and includes the AETA 2012 capital costs for comparison.[15]

While capital costs are only one part of the picture, they do illustrate the effects of expected learning curves for each case.

Table 3: Hypothetical generation capital costs

Electricity generation technology

2012 (NSW)


Scenario 1 2030


Scenario 1 2050


Scenario 2 2030


Scenario 2 2050


Wind - onshore






Wind - offshore






CST - central receiver with energy storagea






PV - rooftop, non-tracking






PV - utility scale, single axis tracking






Geothermal (EGS)

Technology not available





Geothermal (HSA)

Technology not available











Biogas-fuelled OCGTsb







Technology not available





Pumped hydroc






a The CST plant used in the modelling had a higher solar multiplier and larger storage than assumed in the 2012 AETA, so these costs are about 23% higher than reported in that study.

b Similar costs apply to biogas-fuelled open cycle gas turbines (OCGTs) in all scenarios as this is considered a mature technology.

c Pumped hydro costs were not covered in the AETA 2102, so these were based on costs provided by ROAM Consulting.[16]

The costs in Table 3 show the modelled development of several technologies which are not commercially available yet, including geothermal and wave. Costs of more mature technologies (such as on-shore wind, biomass and hydro) are expected to reduce at a lower rate than developing technologies such as CST.

Key differences are also evident between the costs in each scenario, which is driven by the GALLM learning curves. In particular, Scenario 1 projects more aggressive reduction in PV generation costs than Scenario 2. Conversely, Scenario 2 projects faster reductions in wind generation costs than Scenario 1. These costs differences are reflected in the generation mix variations in the modelling results. Operations and maintenance costs

With the exception of bioenergy, the capital costs of the technologies are often the dominant cost factor in renewable generation as, once constructed, fuel cost for most technologies tends to be low or zero. Bioenergy requires fuel that is costly to collect and, in the case of biogas, costly to convert.

All generation plant requires maintenance, however, and for renewable technologies such as geothermal and wave, maintenance costs can be considerable.

The operations and maintenance costs (O&M) costs, both fixed and variable, are based on the mid-point estimates from the AETA 2012.

Scenario 2 uses the AETA 2012 assumption that O&M costs escalate at around 150% of Consumer Price Index (CPI), while Scenario 1 assumes that rapid technology transformation will drive real reductions in O&M costs. The bioenergy fuel costs are also taken from the AETA 2012.

The fixed and variable O&M costs for each case are detailed in the full report.

5.5.2 Energy storage technologies and costs

Maintaining system security requires supply and demand to be balanced at all times, and preserving this balance can be more challenging in a 100 per cent renewable power system.

Several key renewable energy sources are variable given they depend on weather conditions that vary on several time scales (minutes, daily, seasonally, annually). This means flexible supply and demand options will be required to achieve the balance traditionally provided by fossil fuel generators. Energy storage is likely to be central to providing this flexibility.

AEMO's consultants provided estimates on the availability and costs of five categories of large utility-scale energy storage technologies: batteries; biomass, as solid matter and as biogas; compressed air; hydro, including pumped hydro; and molten salt thermal energy storage associated with CST energy collection.

Based on this information, AEMO modelled a subset of storage technologies, selecting those that provide the required storage flexibility at least cost, and cover periods of high demand or low generation from other sources.

Existing pumped hydro in the NEM was assumed to remain, and the subset selected adds to this a mix of CST with molten salt, biogas (stored in the existing gas systems), biomass, and additional pumped hydro.

Given the chosen mix of generation from diverse sources across the NEM, investment in specific storage solutions such as batteries and compressed air did not emerge as being economic for large-scale deployment and were not included in the modelling.

5.5.3 Transmission costs

For each case modelled, the requisite new transmission lines and/or upgrades to existing transmission facilities were identified and costed. Transmission connection costs, to connect generators to the closest transmission node, are based on the 2012 NTNDP[17] connection cost estimates.

For technologies not covered by the NTNDP, AEMO developed specific cost estimates of the transmission lines, substations and easements required. The costs of transmission connection assets required for individual generators are included in the generation costs. Transmission cost estimates are included in the full report.

5.5.4 Other costs (not included)

The study does not include costs for any changes that might be required to local electricity sub-transmission or distribution networks, which could be significant.

In practice, the amount of PV generation forecast and the assumed DSP increases suggest that significant investment may be required to manage voltage swings, increased fault levels, to maintain power quality during more significant variations in flow, and to accommodate changes in flow direction.

The study does not include the costs associated with the acquisition of the land required for the renewable generation. This is likely to be significant and the actual acquisition may present some challenges.

The study also does not include the costs associated with any generation or transmission assets left stranded by the shift to the 100 per cent renewables.

6. Results

Each of the four cases modelled the optimal mix of generation, energy storage and associated transmission infrastructure which reliably met assumed customer demand.

6.1 Generation mix

6.1.1 Overview

Overall, the generation mix identified included a diverse range of resources and technologies.

The results show that no single renewable technology dominates, and that all versions of the 100 per cent renewable system would most likely require a mix of technologies. This mix is expected to be necessary to manage intermittent generation output variations and the technical issues related to maintaining reliability and system security.

The projected mix of generation technologies differs greatly across the four cases. This is primarily driven by generation costs, which vary by both scenario and the year being modelled given the learning curve variations for the technologies applicable in each case.

The optimised generation mix determined for each case is shown below in terms of installed technology capacity (Figure 6) and total energy generated (Figure 7).

Figure 6: Optimised generation mix - installed technology capacity

Figure 7: Optimised generation mix - total energy generated

6.1.2 Technology-specific results

Based on the modelled assumption that PV costs continue to fall rapidly in both scenarios, PV generation is likely to be considerable in all four cases. The rapid technology transformation assumed in Scenario 1 means that geothermal and wave technologies are likely to develop more quickly than in Scenario 2. As a result, there is expected to be more geothermal generation in Scenario 1 for both 2030 and 2050, with wave power uptake increasing towards 2050.

Wind generation is a relatively mature technology with wide deployment and would need large investments to drive further cost reductions. In Scenario 1 there is likely to be a greater deployment of other technologies, so the cost of wind generation would not decrease as quickly as other technologies. Consequently, wind is expected to be less prominent than PV in both Scenario 1 cases.

In Scenario 2, however, PV and CST are assumed not to be deployed as quickly, leading to an increased global deployment of wind. As a result, wind generation costs reduce relative to other technologies. In both Scenario 2 cases, wind therefore generates a greater proportion of supply.

Bioenergy generation is modelled to operate as a flexible resource that can respond to system requirements. Bioenergy is included in the modelling both as baseload steam turbines and as biogas-fired open cycle gas turbines (OCGT). This combination emerged as being the most flexible and efficient for this resource. The model assumes storage of biogas in existing gas infrastructure for use when demand peaks or intermittent generation production is low. The fuel requirements of this bioenergy are likely to be significant and there could be challenges to be overcome in collecting and processing the required fuel.

Although CST could be designed to operate as a baseload generator at a high capacity factor (for example the Gemasolar in Spain includes 15 hours of CST storage), this would require much larger solar collection fields (mirrors, in the case of the proposed central receiver technology), which would see a corresponding increase in capital costs. This study assumes nine-hour storage for CST, used primarily during morning and evening peaks to complement PV generation.

6.1.3 Generation capacity requirements

To provide reliable supply that allows for contingencies (from generation or transmission trips), the total generation capacity available to the system must exceed maximum demand. In a conventional power system, excess capacity is typically 15 to 25% above maximum demand, whereas this study indicates that a 100 per cent renewable system requires 100 to 130% of excess capacity to meet the same reliability standards. This is due to the variable, weather-dependent nature of many renewable resources.

In the systems modelled, wind and PV account for a large proportion of the capacity installed, but a smaller proportion of the energy produced as both technologies are intermittent and have relatively low capacity factors. Onshore wind has a capacity factor of around 30-40% at the best sites and rooftop PV around 15%.

Other renewable technologies such as geothermal and biomass are similar to traditional baseload generation in that they can operate most of the year at very high capacity factors (80-90%). For this reason, they account for a lower proportion of installed capacity compared to the energy produced.

The total required generation capacity differs between the cases modelled, primarily due to the different projected maximum demand forecasts in each. The technology mix also affects the total amount of generation capacity required due to variations in how each technology contributes to system reliability. This appears to be particularly the case in 2050 for both scenarios where the higher percentage of intermittent generation (PV, wind and wave) means more reserve is expected to be required.

In all four cases, the optimised capacity reserve levels exceed 100 per cent above the maximum demand. This means that to maintain system security, generation with a nominal capacity of more than twice maximum demand must be built. Table 4 summarises the installed capacity of each technology and the total installed capacity required for each case.

Table 4: Hypothetical generation capacity for each technology (MW)


Scenario 1 2030

Scenario 1 2050

Scenario 2 2030

Scenario 2 2050









































Maximum demand (10% POE)





Reserve level (% generation above max. demand)





6.2 Storage

Energy storage is likely to be required predominantly to meet demand after sunset and to manage the evening peak. It may also be used to cover periods of low wind or solar radiation and to provide backup in case of contingency events such as the loss of a transmission line or a large generator.

The model selected CST with storage as the primary storage technology, extending the use of daytime solar energy by applying it to meet demand at a different time. This technology is expected to be supplemented by hydro and biogas on most days to manage the evening peak.

The modelling did include existing pumped hydro, but no additional pumped hydro was added to the mix as the modelling found it to be an uneconomic option.

6.3 Transmission

The mix and location of renewable generation in each case determines the transmission build and/or augmentation required. While some renewable energy, notably rooftop PV, requires minimal transmission given its proximity to the load, other renewable resources tend to be more remote.

The modelling approach sought to optimise the total combined generation and transmission costs. It considered both traditional alternating current (AC) connections and newer high voltage direct current (HVDC) lines, with the most appropriate technology being used as required.

The existing transmission system was assumed to remain; the transmission requirements in addition to this are shown in Table 5.

Table 5: Hypothetical additional transmission requirements

Scenario 1 2030

Scenario 1 2050

Scenario 2 2030

Scenario 2 2050

New Capacity (MW)


New Capacity


New Capacity (MW)


New Capacity (MW)


East NSW to South QLD









VIC to East NSW


















VIC to Mid/South SA









East NSW to Cooper Basin





Mid/South SA to Flinders/Eyre









South QLD to North QLD









South QLD to Darling Downs



East NSW to Mid NSW









Broken Hill to VIC-SA interconnector



Overall, the capital costs for transmission represent approximately 10% of total capital costs of installing new generation. This means that locating new generation further from existing transmission networks will not have a significant impact on overall costs unless the distances are substantial.

6.4 Land use

AEMO estimated the additional land use requirements for deployment of the technologies identified in each case. These are based on the land use estimates produced by ROAM Consulting and CSIRO[18] and AEMO's experience in carrying out its own transmission planning obligations.

Depending on the case, the estimated total land required varies between 2,400 and 5,000 square kilometres. This does not include any additional land requirements for biomass, as the biomass is assumed to be sourced by redirecting exiting sources of bioenergy. The process to acquire this land could be challenging and the costs could be significant. These costs have not been included in this report.

Biomass energy requirements in this study might be met with a mix of waste, stubble, plantation and native forest resources identified in the CSIRO input report on biomass, and should not require exclusive use of any land currently used for food production.

There is likely to be competing use for some biomass likely to be required for energy production, for example plantation and native forest resources are used by a number of industries. The value of the alternate uses is likely to affect the energy production costs.

6.5 Hypothetical capital costs

AEMO estimated the hypothetical capital costs for generation, storage, and transmission infrastructure in each of the four cases. As previously noted, these costs are hypothetical and must be interpreted with due consideration given to the assumptions and constraints outlined in
Section 2. While these estimates are consistent with the study scope, in practice the cost of building a 100 per cent renewables electricity system would be substantially higher.

These estimates assume building all the new generation and transmission infrastructure using the applicable costs for the target year, 2030 or 2050. This takes full advantage of the cost reductions and performance improvements of renewable energy technologies expected between now and 2030 or 2050.

Due to the nature of the study, no transition costs are included. Furthermore, no allowance has been made for distribution network costs, financing costs, the cost of stranded assets or the research and development expenditure that may be needed to drive the forecast cost reductions.

The hypothetical capital costs below are expressed in 2012 dollars.

Table 6: Hypothetical capital costs ($2012)

Scenario 1 2030*

Scenario 1 2050*

Scenario 2 2030*

Scenario 2 2050*

Generation and storage connection

$197 billion

$257 billion

$235 billion

$311 billion


$22 billion

$28 billion

$17 billion

$21 billion


$219 billion

$285 billion

$252 billion

$332 billion

*Capital costs are based on DCCEE scope assumptions which include: assumed system build in 2030 or 2050 without consideration of the transition path; no allowance for distribution network costs, financing costs, stranded assets, land acquisition costs or R&D expenditure. Cost inputs are based on data provided by the AETA 2012, CSIRO and ROAM Consulting.

6.6 Impact on wholesale prices

Using the hypothetical capital costs presented above and making allowances for O&M, fuel and financing costs, AEMO estimated the hypothetical annualised costs for generation and storage required for each case, including network connection costs. Again, while these estimates are consistent with the study scope, they do not represent what costs might be in practice.

To cover the hypothetical capital and operating cost of generation and storage plant and connections only, wholesale electricity prices in the range of $111/MWh (in Scenario 1 2030) to $133/MWh (in Scenario 2 2050) would be required. These costs are in 2012 dollars. For comparison, this component is over double the average current wholesale electricity spot price of around $55/MWh, Currently many renewable generators receive financial support from outside the electricity market through schemes such as LRET, SRES and feed-in-tariffs. The costs of these schemes have not been estimated and are not included in this comparison.

Additional investment required in new shared network transmission infrastructure would add another $6 to $10/MWh to the above estimates.

The wholesale electricity price increase and the additional transmission prices would be passed on to consumers via retail prices. The relative impact of these price rises would depend on the other retail price components such as distribution prices, and would be greater for industrial and commercial customers for whom wholesale prices represents a greater proportion of the total
retail cost.

7. Observations

While appreciating the exploratory nature of this study and noting the assumptions and sensitivities that heavily influence the results, AEMO notes the following observations drawn from the modelling results:

A wide range of technologies and locations are likely to be needed. There is unlikely to be a single technology that dominates; rather, reliance on a broad mix of generation technologies is likely to be required to meet the existing reliability standards. The study shows that generation plant is likely to be spread across all NEM regions. This diversity of generation and location is expected to be critical to maintaining the supply/demand equilibrium necessary for system security and reliability.

Total capital cost estimate (hypothetical) are greater than $219 and $332 billion dollars, depending on scenario. These costs are driven primarily by the study assumptions-in particular that all the plant would be built at the future estimated costs rather than progressively over the period. No allowance has been made for the costs of any modifications required to the distribution networks, the cost of acquiring the required land for generation or the costs of stranded assets. The modelling results are highly sensitive to the assumed technology cost reductions, and any changes to these would see corresponding modelling outputs.

Overall land requirement to support a 100 per cent renewable power system may be between 2,400 and 5,000 square kilometres.

The high level operational review found that operational issues appear manageable, but it is noted that several key considerations would require more detailed investigation. Overall, the transmission network would require significant expansion to transport renewable generation to customers and significant management of the transition to 100 per cent renewables.

Considerable PV generation in all four cases drives demand and load patterns changes. Based on the modelled PV generation levels the NEM is likely to become winter-peaking (in contrast to most regions' current summer peak), which means managing heating loads would be more critical than the current air-conditioning loads. The PV contribution levels also (typically) cause generation availability to peak around midday, so DSP would move demand into this period rather than the traditional late night off-peak periods.

More capacity relative to maximum demand is likely to be required. The results indicate that a 100 per cent renewable system is likely to require much higher energy reserves than a conventional power system. It is anticipated that generation with a nameplate capacity of over twice the maximum customer demand could be required. This results from the prevalence of intermittent technologies such as PV, wind and wave, which operate at lower capacity factors than other technologies less dominant in the forecast generation mix.


[2] As defined by DCCEE.

[3] AEMO. Available at: Viewed 20 March 2013.

[4] WICI. Viewed 20 March 2013.

[5] "Biofuels and biosequestration in perspective" Australian Academy of Technological Sciences and Engineering, Focus, April 2012 (171) pp. 35-37.

[6] AEMO. Available from: Viewed 20 March 2013.

[7] AEMO. Available from: Viewed 20 March 2013.

[8] Sourced from the AETA 2012. CSIRO and ROAM Consulting.

[9] AEMO. Available from: Viewed 20 March 2013.

[10] The diversified maximum demand takes into account that maximum demand in each state generally occurs at different times.

[11]AEMO Rooftop PV Information Paper, 2012 Viewed 20 March 2013.

[12] AEMO. Available from: Viewed 20 March 2013.

[13] AEMO. Available from: Viewed 20 March 2013.

[14] WorleyParsons: Cost of Construction New Generation Technology November 2011: Viewed 20 March 2013.

[15] The capital costs in the AETA 2012 differ by region. This table uses New South Wales region costs for comparison purposes.

[16] AEMO. Available from: Viewed 20 March 2013.

[17] AEMO. Available from: Viewed 20 March 2013.

[18]AEMO. Available from: Viewed 20 March 2013.

Supporting documents

In 2011 the Australian Government commissioned the Australian Energy Market Operator (AEMO) to undertake a study and report on potential 100 per cent renewable electricity generation in the National Electricity Market (NEM) for 2030 and 2050.

In addition to the main report and executive briefing available on this page, a number of additional supporting publications formed a part of the study. Copies of supporting publications are available in the Australian Government Web Archive.

Further information