The Australian Electricity Sector
Australians consume a lot of electricity – there are many reasons for this but historically cheap power from coal and to a lesser extent hydro was a big factor promoting high levels of residential consumption as well as encouraging investment in energy intensive industries. This trend is reversing, but usage remains high compared with other developed nations.
Australia’s high per capita electricity consumption combined with its reliance on coal as an energy source – as shown in Figure 1 below – means there is a sharp focus on reducing and eventually eliminating CO2 emissions from the electricity sector.
Figure 1 The energy source for Australia’s electricity sector (1).
Increased use of renewables is the obvious answer to reducing CO2 emissions in the electricity sector. As shown in Figure 1, wind and solar make up a relatively small share of our power generation mix but have been growing at about 10% per year over the last 10 years.
So “what happens when the sun doesn’t shine and the wind doesn’t blow?” This is not just a word grab by conservative politicians, it is a reality that needs to be answered if we are to fully decarbonise the electricity sector. The answer is multifaceted and includes reducing overall demand, using smart meters to better match supply and demand as well as leveling out demand peaks and troughs. Major capital investments to install more transmission lines will also be needed. Even with these initiatives, we won’t be able to get to zero carbon without utility scale storage systems that allow electricity generated when the sun is shining and the wind is blowing to be used when generation conditions are less favourable.
Electricity Storage In Australia – the basics
Australia has some natural advantages and disadvantages when it comes to transitioning to zero carbon. On the positive side we are a large country with, in theory, plenty of space for wind farms and solar arrays. We also get plenty of sunshine and have regions with world class wind resources. On the negative side of the leger we are an island. This means that unlike many countries, Australia isn’t able to either import electricity if local generation falls short of demand or export electricity if it has an excess. This has implications for storage – if we reject using gas for backup generation then whatever strategy we use for electricity storage it will have a lot riding on it.
Backing up a step we should explore why storing electricity is not already a well established technology considering that we have been relying on electrical power for a long time. It is not because the value of storage is not recognised – energy storage that allows production and usage to be decoupled has obvious practical advantages. Petrol, whether extracted from local wells or imported from other countries, is stored around the world in tanks and reservoirs so it can be drawn down by motorists and industry as it is needed. The same, more or less, applies to natural gas and coal. Electricity is different – storage is complex and often involves converting electrical energy into another form of energy and converting it back when needed. Inevitably this is an inefficient process with significant losses involved in the conversion processes.
From an investment perspective, over the past 100 years generation technology has advanced so far ahead of electricity storage that pre-renewable grids managed demand fluctuations by installing excess generation capacity. It has been more cost effective to build gas plants that might only run for a few weeks each year than to develop large scale storage capability using batteries or some other device. The transition to renewables means that this strategy is no longer valid. Existing storage technologies need to be optimised and expanded and potentially new storage technologies developed. This will be a major area of global engineering focus and investment over the next 20 years.
The Australian grid
Figure 2 shows a couple of diagrams of the National Energy Market (NEM), which covers the eastern and southern states of Australia and is responsible for about 80% of Australia’s electricity sector. This grid interconnects a relatively small number of large scale generating assets with a huge number of mostly very small electricity consumers in every east coast city, town as well as larger commercial and factory end users. Supply is continuously being ramped up and down and sometimes turned on and off to match a continuously fluctuating in demand.
Figure 2 Two schematics of the National Energy Market (NEM) the power grid that produces over 80% of Australian electricity. By area it is the largest power grid in the world
Western Australia and Northern Territory are not part of the NEM and are supplied by a number of much smaller local grid systems supplying discrete regions within each state. It will be interesting to see if these remain separate from the NEM (and each other) or if they are incorporated into a fully national grid as existing coal and gas plants are replaced by wind and solar backed up with storage.
Figure 3 shows the ownership basis of generation assets in each state. NSW and Victoria are dominated by traditional standalone power plants built and operated solely to produce electricity for the grid. In NSW the majority of the plants are in the Hunter Valley while in Victoria they are in the LaTrobe Valley. In both instances power plant construction was originally carried out by state governments to take advantage of nearby coal reserves. Tasmanian generation is dominated by hydroelectric installations, again mostly built by the Tasmanian government. The industry category shown in Figure 3 are mostly privately owned power plants supplying remote mining facilities with excess output being added to the local grid. Small scale solar are the domestic (and small scale commercial) roof installations seen on many rooftops in most residential areas. Australia has one of the largest per capita rooftop solar systems in the world.
Figure 3 Generation asset ownership by state (1)
One advantage of a geographically large grid is that it will often be either sunny or windy (or both) somewhere within the generation footprint. So if the wind isn’t blowing in Victoria it may be possible to use, for example, excess Queensland solar electricity to help supply homes in Melbourne. While this opportunity won’t always exist, locating new wind and solar assets in areas that are widely distributed and ideally with different weather patterns will optimise the grid’s ability to shift excess generation to areas with lower generation. If this can be done it will reduce but not eliminate the requirement for grid storage.
Demand Fluctuations
As discussed above, the demand for electricity comes from a vast number of consumers every time they turn on (or off) any sort of electric appliance. Society expects electricity to be continuously available with effectively no limits or constrictions. An example of the demand curve that is produced from this open access model is shown Figure 4. On a seasonal basis power usage is highest in the summer when air conditioners are turned on and lowest in spring and autumn when milder weather reduces the need for both heating and cooling. Figure 4 also shows typical daily trends – lowest demand between midnight and dawn and highest in the evening driven by increased residential consumption. A final observation is the significant variation around the average demand, particularly in summer when air conditioners can either be running flat out or potentially turned off depending on how hot the day is.
Figure 4. Seasonal demand in South Australia (2008/9)
Note: every location will have a different set of daily averages but Figure 4 illustrates the continuously changing demand the generation system needs to satisfy.
Balancing supply and demand
Using Figure 4 for illustrative purposes, a traditional grid has baseload power plants designed to run continuously to satisfy the first 1000 MW of demand which is assumed to be effectively permanent. Baseload plants in Australia are mostly coal fired power plants, though in other countries they could be large gas, hydro or nuclear facilities. A range of mid and small sized power plants are fired up or down as demand oscillates between 1000 MW and about 2800 MW. In Australia these non baseload plants will be a mixture of mid sized coal units backed up by fast start gas plants (able to switch on quickly to meet unexpected spikes in demand). As mentioned above some generation assets may only be used for a week or two each year (when demand is at its highest) and others will be semi baseload, running for about 60% of the time.
Peak demand is an important metric for electricity generators operating under a traditional grid configuration as this determines the total installed capacity (or in layman’s terms the aggregate size) of the generating fleet. In Australia peak demand occurs in summer after a run of very hot days. In other countries peak demand may occur in winter after a period of sub zero temperatures. The grid needs to have most, but not all, of its available generating assets fully in service on these days. Some standby or reserve capacity must always be available to cover unexpected generation outages.
In moving to a 100% renewable, semi decoupled, storage based grid design the necessity of managing power availability at peak demand will continue to be critical. The focus is likely to shift to moderating peak demand rather than simply ensuring there is always capacity to meet it. Expect to see efforts to reduce load on very hot summer days. This might be done by paying large industrial users not to take power or incentivising domestic consumers, especially those with home battery storage, to decouple from the grid. Expect to see increased focus on residential smart systems that will monitor local demand (and likely power pricing) to determine the optimum time to operate appliances like dishwashers and clothes driers.
Grid scale electricity storage – background
Grid scale storage will inevitably be required, even if the grid is configured to minimise periods when demand exceeds supply. Grid operators will try to quantify the storage needed to manage periods of supply shortfalls. Storage requirements will be expressed in terms of the total amount of electricity that is needed and the rate at which the stored electricity needs to be released back into the grid. This will effectively treat stored electricity as a “de facto” generator when it is required to meet supply shortfalls. The required rate at which it needs to discharge into the grid is analogous to generating capacity and the total amount of stored energy is analogous to on site fuel supply for coal and gas fired plants or water storage for hydroelectric plants.
The power industry measures delivery rate or capacity in megaWatts (MWs) and storage in megaWatthours (MWhs). As an example of this nomenclature, the Tesla “super battery” installed in South Australia has a capacity of 100 MW with storage of 129 MWhs. This means that if it runs flat out (at 100 MW) it can operate for about 75 minutes (129 WHrs/100MW) before it needs to be recharged. If a battery of this size was recharged every day and discharged every night it would provide 47,000 MWhs/year (129 MWhs/day x 365 days/year). A quick comparison with the now closed Hazelwood power station in Victoria which had 8 x 200 MW coal fired units producing 12,000,000 MWhs each year shows that while the Tesla installation might be big in battery terms, it is a small player in terms of the overall grid requirement. To be fair, it is also a zero carbon source of electricity – a claim that Hazelwood can not make.
Grid scale electricity storage – options
If Australia needs to shift to a generating model that relies on stored electricity,, what technology options should we be considering
- Pumped Hydro: The concept is shown in Figure 5. When power is plentiful (and hence cheap) this is used to pump water from a lower reservoir to an upper reservoir. The electricity used to pump the water is hence converted and stored as gravitational energy which can be converted back into electricity when the water flows through a turbine as it returns under gravity to the lower reservoir.
Figure 5 Pumped Hydro schematic
The amount of electricity that can be stored by a pumped hydro facility is a function of the size of the storage reservoirs.The rate at which this electricity can be released will be a function of the elevation difference between the two reservoirs, the flow rate of water reaching the turbine and the size of the turbine itself. It is worth noting that the charging rate (rate at which water can be pumped from the lower reservoir to the upper reservoir) will be different to the discharge rate. Globally, pumped Hydro is the most well developed storage technology representing over 90% of the existing storage capacity.
Pumped hydro requires a hilly site that can accommodate the required size for both the upper and lower storage reservoirs with an appropriate elevation difference between the two reservoirs. Recent studies suggest Australia has ample such sites (2) – included in this analysis is the proposed Snowy 2.0 project.
Storing electricity by converting it into gravitational energy (or chemical energy as is the case in the battery examples below) and converting it back involves considerable losses. The amount of electricity used to pump the water to the upper reservoir is greater than that which will be discharged back in the grid. Storing electricity is not technically hard – it is just relatively inefficient. Selecting the least most inefficient system will be part of the storage selection process.
- Lithium Ion Batteries: Lithium Ion batteries were first commercialised by Sony in the early 1990’s for small scale, portable electronics. They are the latest in a long line of different battery technologies all based on generating electricity from the electrochemical differences between different, usually metallic, materials. Lead acid, zinc carbon and nickel cadmium are examples of battery combinations that were commercialised 100 years before the Lithium Ion system.
Figure 6 Lithium Ion battery. The battery functions by Li ions moving from the Lithium metal compound which acts as the cathode to the Lithium Carbon matrix anode. During recharging Li ions move in the opposite direction.
Battery design is very complex, not just between different battery types, but also within battery categories. For example the Li – metal oxide used as the cathode of the Lithium Ion battery can contain Cobalt, Manganese, Iron and Aluminium in different ratios and proportions. All of these different metals influence factors such as cost, charging/discharging rate, safety (potential to catch fire or explode), battery life (number of charging cycles) and discharge efficiency (percentage of theoretical discharge achievable). Because of the importance of the Lithium Ion battery to the electric vehicle industry there is a lot of research into the design and optimum chemistry for these batteries. There will inevitably be further optimisation of Lithium Ion batteries but whether this means they will be a major player in grid scale storage remains to be seen.
The key advantages of Lithium Ion batteries over alternative battery formulations is that they are lightweight and have a high energy density, making them the obvious choice for electric vehicles. This is also attractive for residential power storage but is far less critical for utility scale storage applications.
- Flow batteries: Flow batteries use a liquid based electrolyte system for both the oxidation and reduction half cells in the battery. In other words instead of Li ions being oxidised and reduced within a solid matrix as occurs in a lithium Ion battery, these oxidation and reduction reactions occur within a liquid solution which is being continuously circulated around and past both the anode and cathode. There are a number of different solution options with the Vanadium based system, shown in Figure 6, being one of the most advanced options.
Figure 6 Vanadium based flow battery configuration.
Flow batteries have some advantages over Lithium Ion for grid scale storage. The first of these is that they are easier to scale up (make large enough for grid scale storage). Secondly the power discharge rate is effectively independent of the storage capacity – the power of the battery is a function of anode/cathode sizing while storage is governed by the size of the tanks. Finally flow batteries can have very long cycle lives without degradation – a problem that continues to plague Lithium Ion systems. It is relatively easy to drain the tanks and replenish the electrolyte solution should this be necessary to maintain performance.
Flow batteries naturally have some disadvantages which will probably preclude them from domestic applications. Firstly they take up much more area than an equivalent Lithium Ion battery and secondly the chemical constituents can be toxic and require special handling. These issues, however, should be manageable for industrial grid scale operations.
- Hydrogen: A previous blog article discussed the politics associated with Australian support for hydrogen (3). To recap part of this discussion, hydrogen can be produced by using renewable electricity to electrolyse water. The hydrogen produced via this process can be used to generate electricity using a fuel cell (see Figure 7 below). As with the two battery options discussed above, this is an example of converting excess renewable electricity into chemical energy and converting back into electricity when required.
Figure 7 Schematic of a Hydrogen based fuel cell
Hydrogen can also be compressed and transported for use as a zero carbon fuel in heavy duty vehicles and ocean going vessels. It is this opportunity that has the most publicity – the potential to export compressed hydrogen offers a route for Australia to maintain its place as a major energy exporter. If Australia invests in a compressed hydrogen export industry it could use a portion of the hydrogen to support the electricity grid. If international prices for hydrogen become attractive this may preclude it from being a big player in local electricity storage. If there turns out to be no international market for Australian hydrogen its role in supporting the grid will obviously depend on its competitiveness relative to other storage technologies.
5 Electric Vehicle storage; Australia has about 20 million vehicles. When these are all electric and are fitted with a 200kW lithium ion battery capable of storing 75kWhs there will be a very significant storage baseload sitting in residential garages overnight and in commercial car parks during the day. As a hypothetical and with the right hardware, software and incentives, if 5 million cars with 15kWhs of excess storage is made available to the grid every evening they would supply 27,000,000 MWhs per year to the grid. At twice the annual output of the Hazelwood power plant this is a not insignificant amount of storage The concept is described in Figure 8 below
Figure 8 Nissan promotion showing the potential for electric vehicles to take power from the grid
Australia currently has a very small electric vehicle fleet so this is more of a slow burn opportunity. It is, however, very compatible with the smart meter type approach that will become more widespread for domestic demand leveling. This means that the infrastructure for vehicle to grid storage will potentially be in place well before the number of vehicles increases to a point where it can make a utility scale contribution to electricity storage.
Conclusion
It seems inevitable that new generation in Australia will be strongly skewed toward wind and solar. This will bring storage sharply into the spotlight. Fortunately Australia seems to have a number of options to both minimise the amount of storage needed and install sufficient storage to cover anticipated supply shortfalls. This will, however, cost money so subsidies and support (both financial and emotional) currently going to renewable generation may need to be refocussed on storage.
Pumped hydro seems the obvious choice as the backbone of storage infrastructure capable of supporting a 100% renewables grid. The positives are that it is a well proven technology and Australia has plenty of suitable locations, many near population centres along the edge of the Great Dividing Range. There will be local opposition as pumped hydro has a large footprint and can’t be easily located in areas already dedicated to industrial activity. Perhaps we will be lucky and the optimum locations are not close to national parks, prime farmland, heritage areas or within sight of wealthy retirees. I am guessing we won’t be so lucky.
Given the high take up of rooftop solar in Australia it seems likely that governments will be encouraged to promote domestic battery systems. From a grid storage perspective the ideal arrangement will be if they are integrated into the grid so that households with a fully charged up battery can be paid to stay off the grid during periods of high demand (or reduced supply). Not all households will be able to do this but a system that allows consumers to monetise their flexibility to not use their dishwasher or clothes dryer seems inevitable.
I also suspect that new wind and solar installations will be required to “consider” storage in their permit applications. Ultimately this will mean a requirement to install on site storage – potentially a suitable sized flow battery. We are approaching the point where the installation of new wind and solar without consideration of storage is no longer feasible.
The next 20 years will be exciting times for the Australian electricity sector as coal plants continue to close and more renewables are brought on line. Pragmatic conservative groups who have given up on coal will modern gas plants which are technically well suited to backup large widespread renewable generation. Progressive group will reject gas but may have to update their pro-renewables stance to reflect that storage and demand modification programs are becoming more critical to their ultimate goal of zero carbon
