2.4 Modelling results

2.4.1 Electricity consumption

A key feature of the Australian deep decarbonisation pathway is that projected growth in electricity consumption is high at around 2.4 percent per annum between 2014 and 2050. This MMRF model outcome is in spite of the assumption of improved end-use electricity efficiency throughout the projection period. The main reason for this strong growth, which is against current trends for stagnant electricity demand,8 is due to the assumed increasing electrification of buildings, transport and industrial processes and the redistribution of economic activity from fossil fuel production and export to non-fossil fuel-related minerals.

The electrification of road transport and shift in economic activity are outputs of ESM and MMRF respectively, while the electrification of buildings and industrial processes is an assumption developed by ClimateWorks. The amount of electricity ESM projects used by the road transport sector is shown in Figure 2.8. It amounts to around 110 TWh or 28 percent of electricity consumption in 2050 but is fairly negligible up until around 2025. More detail on the source of this demand is discussed in the transport part of this report.

Figure 2.8 – Contribution of road transport to national electricity consumption

An interesting outcome of the redistribution of economic activity to non-fossil fuel-related mining is that the share of electricity demand shifts between states, so that between 2014 and 2050 Western Australia moves from being the fourth largest to the number one electricity-consuming state, consuming a third of all electricity generated in 2050.

This outcome raises a number of questions that remain largely unanswered due to time constraints in this project:

  • Western Australia has a high share of off-grid electricity generation. For how long would this remain the least-cost approach at this scale of demand?
  • Western Australia is disconnected from the Eastern States National Electricity Market due to distance and its low share of on-grid electricity consumption. For how long would it remain least-cost to remain disconnected from the eastern states if it reaches a third of consumption (and also given its high-quality solar resources)?

A full investigation of these issues would require detailed modelling of alternative grid/off-grid and long-distance electricity transmission economics. There would also need to be some interrogation of the MMRF projection of growth in non-fossil fuel mining.

For the purposes of the modelling conducted here, Western Australia is assumed to remain disconnected from the eastern states, but with an increasing share of consumption and generation on-grid, implying the current weakly connected grid and distributed off-grid generation becomes more interconnected over time.

Figure 2.9 – State electricity consumption in 2050

The projected level and state distribution of electricity consumption is assumed to be the same across the 100 percent renewable grid, CCS and Nuclear scenarios. However it happens that electricity generation is slightly lower in the CCS and Nuclear scenarios due to lower transmission losses. This reflects shorter assumed transmission distances for CCS and nuclear-based technologies. In comparison, deployment of a 100 percent renewable grid requires accessing renewable resources that are further from end-users.

2.4.2 Electricity generation

100 percent RENEWABLE GRID

Given projected electricity consumption, the abatement incentive assumptions and preference for new generation capacity to be renewable, ESM's projected least-cost electricity generation mix is shown in Figure 2.10. The fairly rapid increase in the abatement incentive in the period to 2020 leads to the closure of conventional brown coal-fired electricity generation capacity (significant expansion of gas-fired power in the short-term is essential to achieving this) and puts black coal-fired power on a relatively steady decline down to a negligible amount by 2035. During that period, black coal and expanded gas-fired power, together with existing hydro, support the stability of the system as most new growth is met by wind and solar panels. As flexible fossil electricity generating capacity is retired, new renewable electricity generation capacity must change in character, including more flexible and load-following plant, such as solar thermal with and without thermal storage, wave and enhanced geothermal power.Any variable renewable technologies are also assumed to be increasingly deployed with some electrical storage.

Biomass could theoretically be another useful contributor to supporting variable renewables, but is not selected by ESM for that task, reflecting its high levelised cost of electricity when competing with other industries, particularly transport, for use of biomass fuel.

The main risk to the ability of the system to accommodate such a high share of renewables and meet system reliability standards is the slow development of enhanced geothermal technology in Australia. On the other hand, battery storage technology does appear to be making significant cost improvements and this project's modelling has not included demand management, which could play a significant role in balancing the system. The King Island off-grid system currently demonstrates a system that is 65 percent renewable over the course of a year.

The high projected share of solar power (either photovoltaic or solar thermal) in the electricity generation mix is a reflection of both its cost advantages (see previous discussion of levelised cost of electricity) and also that a third of consumption has shifted to Western Australia, where solar resources are high quality. The decline in wind power toward the end of the projection period is partly due to a levelling of its competitiveness relative to solar over time and also the increasing cost of supporting wind's variable electricity output as existing flexible fossil fuel capacity is removed.

After taking account of some gas-fired generation remaining in the on-site generation sector, the share of renewable generation in total electricity generation in 2050 is 96 percent.

Figure 2.10 – Projected national electricity generation by technology, 100 percent renewable grid, 2010–2050

On-site electricity generation technologies increase their share of total electricity consumption from 12 percent in 2014 to 16 percent by 2050 which, given the total growth in demand, represents a more than tripling of on-site electricity generated. However in share terms this is lower than might have been expected relative to other modelling work that has examined on-site generation uptake, increasing its share to almost 50 percent (Graham and Bartley, 2013). The high rate of industrial electricity consumption growth (less suited to on-site generation), owing to industry electrification, has partially prevented this outcome. For example, if this level of on-site generation was achieved in 'Scenario 1' of Graham and Bartley (2013), where there is far more modest electrification (mainly modest adoption of electric vehicles) and resulting 2050 electricity generation of 295TWh, then the on-site generation share would have been 33 percent.

Another factor is that once the opportunities for including solar panels in buildings and use of available biogas have been exhausted, the remaining on-site generation options (mainly gas and diesel-based) are relatively high emissions and therefore less attractive in the long term under a high abatement incentive, compared to what is available as large-scale generation in the grid (although gas cogeneration and trigeneration options remain very low emissions compared to the current system).

The major source of growth in on-site generation by technology is solar panels, reflecting their strong reduction in costs relative to other on-site generation technologies and their ease of integration into the buildings of all customer types. The attractiveness of solar panels in reducing the need for grid-supplied electricity has been noted as a strong recent national and global trend. It was kick-started by the introduction of feed-in tariffs and renewable energy targets in many different countries, including Australia. While these incentives have since been reduced and may continue to decline, system costs have reduced so significantly that they may not be needed to allow for continued growth, particularly as wholesale and retail prices must rise in order to accommodate the changes in the large-scale grid-connected sector.

Figure 2.11 – Projected national on-site electricity generation mix, 100 percent renewable grid, 2012–2050

The remainder of on-site generation needs are largely met by gas-generation, which may also be deployed as cogeneration and trigeneration, where there are opportunities to use waste heat. Much of this gas-based capacity relates to the use of gas in off-grid applications, such as in mining. As discussed, the extent to which an expanded mining sector continues to have a substantial share of its power supplied off-grid remains unanswered.


The CCS scenario explores the impact of allowing carbon capture and storage technology to be deployed, if it is economically viable to do so, whereas the 100 percent renewable grid scenario preferentially deployed renewable electricity generation technology when new plant was needed. The projected electricity generation technology mix for the CCS scenario is shown in Figure 2.11.

Under the cost assumptions applied in ESM, CCS technology is competitive with renewable electricity generation technology, even more so than might be indicated by levelised cost of electricity (LCOE) data because the market balance constraints within ESM also take into account cost of supporting variable renewable electricity generation technology, rather than simply comparing their LCOE. The result is that the share of electricity supplied by renewable electricity generation is significantly reduced relative to the 100 percent renewable grid scenario. CCS technologies are adopted into the technology mix from 2025 and gradually increase their share to a maximum of 23 percent in 2043. The share declines to 20 percent by 2050 due to two constraints on CCS growth. The first constraint is that during the latter half of the projection period, electricity demand is strongest in Western Australia, which is assumed to have limited carbon dioxide storage opportunities.

The second constraint is that carbon dioxide storage capacity in the east coast states becomes economically and quantity limited. The state with the best storage opportunities, Victoria, cannot make the best use of it due to the poor competitiveness of brown coal with CCS under a high abatement incentive. Consequently, Victorian storage opportunities are utilised by New South Wales black coal generators at additional distribution costs and Victorian gas generators at higher fuel costs.

With CCS opportunities limited, ESM expands use of renewable electricity generation technology, concentrating initially on the lower-cost solar and wind technologies, later supported by enhanced geothermal and a small amount of wave energy. Solar thermal is taken up earlier and greater than in the 100 percent renewable grid scenario. This reflects that wind and large solar panels do not automatically have battery storage included in their development in this scenario and consequently the model has chosen to use more solar thermal with storage (together with peaking gas and enhanced geothermal) to support the management of variable renewables.

If CCS opportunities are less limited, then all else being equal, the share of CCS technologies could increase significantly beyond that projected here. The demonstrated scale of carbon dioxide storage resources in each state remains a significant uncertainty in understanding the potential of this technology.

The inclusion of CCS delays the uptake of other low-emission technologies, particularly large-scale solar compared to the 100 percent renewable grid scenario, delaying the closure of some existing coal capacity to fill the gap before CCS is assumed to be commercially available. In the longer term, relative to the 100 percent renewable grid scenario, the availability of CCS means there is less need to draw on the more marginal fossil and renewable technologies, such as natural gas without CCS and wave power.

Figure 2.12 – Projected national electricity generation technology mix in the CCS scenario, 2010–2050


The Nuclear scenario explores the impact of allowing nuclear electricity generation technology to be deployed. The projected electricity generation technology mix for the Nuclear scenario is shown in Figure 2.13.

Under the cost assumptions applied in ESM, nuclear electricity generation technology is competitive with renewable electricity generation technology. As in the CCS scenario, it is not just the levelised cost of nuclear power that is competitive but also its flexibility relative to variable renewables, which require some additional support. The use of gas peaking, solar thermal with storage and enhanced geothermal to provide flexible generation alongside nuclear power is similar to the CCS scenario.

As in the CCS scenario, we do include some limits on nuclear power deployment. ESM only allows nuclear power in the larger states of New South Wales, Queensland and Victoria, since nuclear power stations generally have to be a minimum size, which would preclude their deployment in the smaller states.9 While Western Australia becomes a large power-using state in aggregate over the course of the projection period, the lack of clarity around how much power remains off-grid and how separate the South East and Pilbara grids remain meant that nuclear was not included as an option for that state. The main concern was that, without more detailed transmission modelling, it would be difficult to estimate the cost of transmitting nuclear power throughout that region.

The result of these assumptions is that all three large states take up nuclear power but are not ultimately dominated by it, indicating nuclear power is competitive under the assumptions but not by a large margin. By 2047 nuclear power has peaked at 86TWh and a 14 percent market share.

ESM was re-run to check what would have happened if Western Australia had been allowed to adopt nuclear power. Under that assumption, the share of nuclear power increases to 27 percent. This could perhaps be achieved with an expanded Western Australian transmission system or alternatively through the use of small-scale nuclear plant, which are not included in ESM's technology set, but were examined in BREE (2012). With the use of small-scale nuclear plant or a more interconnected grid there is no technical reason why nuclear power could not supply a major share of electricity consumption. The 14 percent share projected here should definitely not be interpreted as an upper limit but is rather a necessarily conservative result, given that detailed transmission network planning was outside the scope of this study.

A potential risk to high adoption of nuclear power is that as other countries take up nuclear power the uranium price may experience periods of high prices, increasing the cost of nuclear power relative to alternatives. However unless there is a genuine resource constraint relative to demand, commodity markets tend to revert to prices that reflect the cost of production over the longer term.

Similar to the CCS scenario, the use of nuclear power means that there is reduced adoption of some of the higher-cost renewable technologies.

Figure 2.13 – Projected national electricity generation technology mix in the Nuclear scenario, 2010–2050


In the 100 percent renewable grid scenario, following the recent mothballing of coal-fired electricity generation and reduced electricity demand, the modelling projects a brief period of stability in emissions up to around 2016 (Figure 2.14). Thereafter, the projected greenhouse gas emissions over time reflect the changes in the emission intensity of electricity generation. As coal-fired electricity generation is retired during the period 2017 to 2035, electricity sector greenhouse gas emissions rapidly decline. By 2050, greenhouse gas emissions are negligible in grid-supplied electricity but remain positive in the on-site generation sector, owing to the use of natural gas in on-site generation (particularly off-grid applications).

Figure 2.14 – Projected greenhouse gas emissions from grid and on-site electricity generation, 100 percent renewable grid scenario, 2010–2050

The projected outcome for greenhouse gas emissions for the 100 percent renewable grid, CCS and Nuclear scenarios is compared in Figure 2.15. The projection indicates that all the technology mixes are capable of reducing greenhouse gas emissions from electricity generation by 2050 from between 85 to 94 percent relative to 2010. The greenhouse gas emissions intensity of electricity generation achieved in the 100 percent renewable grid, CCS and Nuclear scenarios are 0.02, 0.05 and 0.04tCO2e/MWh respectively.

Not surprisingly, the CCS scenario has the highest final greenhouse gas emissions intensity of electricity generation and absolute level of emissions. This reflects the assumption that CCS technology can only capture a maximum of 90 percent of emissions from fossil fuel combustion. The cost of capturing the final 10 percent would be prohibitively high.

Nuclear power is 100 percent emissions free, but the Nuclear scenario does not achieve the same greenhouse gas emissions intensity of electricity generation as the 100 percent renewable grid scenario because the limitations on nuclear power towards the end of the projection period and absence of the requirement for all grid power to be renewable encourages slightly more gas-based grid and on-site electricity generation.

While rapid deployment of large-scale CCS or nuclear power plant in the late 2030s temporarily accelerates emissions reduction in the Nuclear and CCS scenarios below that of the 100 percent renewable grid scenario, on balance, across the whole projection period, the 100 percent renewable grid scenario is a more rapid emissions reduction pathway. Consequently, the amount of cumulative emissions for the 100 percent renewable grid scenario in the period between 2010 and 2050 is around 11 percent lower than the CCS and Nuclear scenarios.

Figure 2.15 – Comparison of projected national greenhouse gas emissions for the 100 percent renewable grid, CCS and Nuclear scenarios, 2010–2050


ESM projects the wholesale electricity price as the marginal cost of delivering a balanced electricity market in a given year. In other words, it is a cost-based price projection methodology. In reality, bidding behaviour and other market features could lead to a different outcome, but all models, regardless of what degree of real world market features they include, will eventually assume the price returns to long-run marginal cost over the long term.10

With a strong rising abatement incentive driving the modelling results, it is not surprising that ESM forecasts rising wholesale electricity prices in the 100 percent renewable grid scenario. The increase occurs in two distinct periods. The first is up to 2020, where prices rise to cover the costs of building new gas-fired, wind and solar plant to meet new demand and replace the loss of brown coal-fired plant.

The second steep increase in prices is between 2025 and 2035, where substantial investment must take place to meet new demand and replace black coal-fired electricity generation. On this occasion, the system requires investment in higher cost renewables, such as enhanced geothermal and solar thermal (with and without storage), so that the system has enough flexible plant to support other variable renewable electricity supply.

After 2035, wholesale electricity prices are stable. This reflects the fact that given new renewable capacity has no carbon emissions its costs are fairly stable, despite rising abatement incentives. It also reflects that the costs of some renewable electricity plant are still improving (due to global and local learning), although this can be offset by higher costs in relation to connecting renewables to the grid. In the early 2040s the decreasing costs of renewable plant appears to be the stronger driver but towards 2050, connecting to higher-cost resources is beginning to assert a stronger impact.

Figure 2.16 – Projected national average wholesale and retail electricity prices, 100 percent renewable grid scenario, 2013–2050

Wholesale electricity prices are only the second largest component of retail prices. The largest component is distribution network costs. The major driver for distribution unit costs is the amount of network that must be built to reliably meet peak demand and the volume of electricity consumption using that infrastructure. If peak demand rises faster than volume then, all else equal, the cost of distribution per volume of electricity must rise. There are other factors such as unevenness in the rate of replacement of assets, the cost of finance and changes to reliability standards, which can and have also impacted the rate of change in distribution costs.

In this scenario, distribution costs increase slightly in the short term due to peak demand growing faster than volume but decrease slightly in the long term, as volume growth begins to outpace peak demand growth. Stronger volume growth in the latter half of the projection period is due to the strong uptake of electric vehicles in road transport. This change increases growth in volume without adding to peak demand (we assume vehicle charging is ubiquitously well managed).

The remainder of residential retail costs are made up of state and federal policies, which are generally phased out, retailer's margin (which is held constant) and transmission network costs. Transmission costs are assumed to rise significantly during the period between 2035 and 2050, as renewable are deployed to more remote locations as outlined in the assumptions and based on Graham et al. (2013).

Adding up these individual components leads to a residential retail electricity unit cost projection of around 38c/kWh in 2012 dollars by 2050. This represents an annual average rate of increase of 0.9 percent per annum, or around 40% to 2050. Our modelling suggests that energy efficiency could lead to around 50% reduction in average electricity use per household by 2050 (excluding energy for electric vehicles, which is legitimately excluded as it replaces costs for petrol, diesel or LPG). This implies that the average household electricity bill would be around 30% lower (adjusted for inflation) in 2050 than today, because energy efficiency gains outweigh power cost rises.

In addition, the annual rate of growth in per capita income is 1.2 percent, for a total increase of 56% until 2050. As a result of falling absolute expenditure on electricity and rising incomes, the share of electricity expenditure in household income falls by around half, on average.11

However, even if income were not rising, the total electricity bill (again excluding electric vehicle charging costs) is falling, because the 38 percent increase in residential retail electricity prices is more than offset by a 50 percent reduction in household electricity consumptions. This is demonstrated in Figure 2.16, where we show the electricity cost for a typical 6000 kWh/year electric appliance-only household at around $1600 in 2014 and declining to around $1100 in 2050. Households that retain some use of gas appliances may face higher costs under abatement incentives. Also note, of course, that any costs involved in reducing household electricity use via energy-efficiency measures, would need to be included in a full analysis of household costs.

Figure 2.17 – Projected annual cost of electric appliance-only household electricity consumption in 2014 and 2050, excluding cost of EV-charging and energy-efficiency investments, 100 percent renewable grid scenario

While not provided as a separate projection, it can be said qualitatively, and with confidence, that large commercial and industrial retail electricity prices would be rising faster than residential retail prices under this scenario. This is because wholesale prices are a much larger component of their retail electricity bill and wholesale prices are rising faster than any other component. This impacts the manufacturing sector the most, due to their relatively high electricity intensity per dollar of output. MMRF provides more detail on changes in industry output resulting from this increase in electricity input costs.

Moderating the higher volume of electricity use (through substitution of electricity for other energy sources) and increase in retail prices experienced by large electricity customers is an expected 50 percent improvement in energy use per square metre in commercial buildings, and about 40 percent improvement in energy use per volume of production in manufacturing.


A comparison of projected wholesale electricity prices for the 100 percent renewable grid, CCS and Nuclear scenarios is shown in Figure 2.18. The comparison indicates that there is very little difference between the wholesale price outcomes when taken as a whole over the projection period, reflecting that the costs are in a similar range. There are some small differences at various times.

In the early part of the projection period, the CCS and Nuclear scenarios are lower cost, as they are able to make plant choices that are not influenced by the need to build towards a long-term objective of a 100 percent renewable grid by 2050. This means delaying purchases of new, higher-cost renewable plant until later and deploying the minimum and lowest cost plant required to meet the existing (nominal) 20 percent Renewable Energy Target by 2020.

In the 2020s, the CCS and Nuclear scenarios can no longer delay shifting to higher-cost low-emission technology and this new investment increases their costs relative to the 100 percent renewable grid scenario. In the 2030s the CCS scenario breaks away from the Nuclear scenario, indicating a higher rate of CCS technology deployment is delivering some cost benefits. The CCS and Nuclear price paths merge and move apart again later during the 2040s when they both begin to hit competitive or resource constraints to their deployment. Note that the increasing price trend in the last five to ten years of the projection period is common across all scenarios and is mostly related to the cost of connecting to more remote renewable resources.

Figure 2.18 – Comparison of projected national average wholesale electricity prices for the 100 percent renewable grid, CCS and Nuclear scenarios, 2013–2050

8 The view that demand growth would be low has only developed during the last decade. See, for example, page 69 of the DPMC (2004) Securing Australia's Energy Future, which has a 'medium' electricity demand projection of 650TWh by 2050.

9 There have been suggestions in general literature that smaller, modular nuclear power stations could be deployed in smaller grids. This option is not included in ESM but could warrant future consideration.

10This is to avoid one of two logical inconsistencies: 1) that investors would continue to build new plant with the knowledge that they will not make a reasonable return on investment in a period of sustained prices below long-run marginal cost; or, 2) that governments would not intervene to make the market more competitive in the event of a sustained period of prices being above long-run marginal costs. This is not say that ESM does not project shorter periods where prices are above or below long-run marginal costs. These are possible and are fully consistent with electricity markets, where long-lived assets and investment delays mean that supply is slow to respond to market conditions.

11 If there were any costs involved in reducing electricity use per capita via energy efficiency measures, the cost of these measures would need to be included in a full analysis of electricity supply and end-use costs.