A submission to the
Independent Review into the Future Security of the National Electricity Market,
Preliminary Report, December 2016[1]
Derek Bolton, 3/3/17
Executive Summary
· On gas, we find that the likely costs in 2025-2030 will see renewables undercut gas power across the spectrum of generation classes. Substantial development of gas power in the interim will become a barrier to achieving the low levels of emissions then required.
· On cleaner coal, we find that it has no conceivable role.
· On network management, we see potential for modelling to evaluate procedural alternatives.
· On industry confidence, we see rival political factions coupled with public misinformation as the ongoing risk.
Detailed Comments
p3
"Domestic gas prices have risen considerably as Australian gas markets have become
linked to international markets..... The need for greater gas supplies for electricity generation is increasingly urgent."
Would increased supply in Australia significantly dent the international price? Ross Gittins[2] thinks otherwise.
Chapter 1
p12
"Figure 1.1: Historical price of wind turbines"
Does this take into account the increasing capacity factor?
p14
"AEMO has projected an uptake of around 1.5 million electric vehicles in the NEM by 2030"
Energeia[3] forecasts 2.2m.
Chapter 3
Fig 3.1
Does this exclude distributed solar, about 2.6% of demand?
Fig 3.2 shows the price cap at around $97 instead of $93, doubling the apparent remaining margin.
p22
"investment in the electricity sector has stalled ... due to policy instability and uncertainty ...
"shortfall charge of $65 payable by retailers if they fail to surrender sufficient certificates"
Certainty would be better sustained if the shortfall charge were to be recast as a late fee, i.e. the shortfall still needs to be made up the next year.
With that arrangement, the total number of RECs required would be more predictable.
p24, Consultation questions
3.1 What role should the electricity sector play in meeting Australia’s greenhouse gas reduction targets?
The electricity sector is the easiest to rein in and accounts for 54% of emissions. Its potential for reducing emissions is increased yet further by electrification of transport, which accounts for 19%. It follows that the government should encourage EV uptake.
It has been suggested that EVs need to be taxed because they escape the fuel levy that helps pay for roads. However, the side benefits of EVs, such as reduced health harm, more than make up for this. Politically, EVs, being unreliant on oil imports, increase energy security. With encouragement, EV uptake (as percentage of today's fuel consumption) could well reach 20%.
With the present carbon-dominated supply of electricity, a switch to EVs has little net effect on emissions, so shifting 20% of transport to electric would make the electricity sector responsible for 59% of emissions. Halving the carbon intensity of electricity would then cut total emissions by 30%.
3.2 What is the role for natural gas in reducing greenhouse gas emissions in the electricity sector?
Fugitive methane emissions
The extent of fugitive methane emissions from gas extraction and distribution is far from clear. Even 3% wipes out any Greenhouse advantage over coal.
Worldwide anthropogenic methane emissions appear to have jumped 30-60% in 15 years[4]. The American experience is that emissions can be several times that claimed by the industry.
There has been little or no independent monitoring in Australia, and no pre-development baseline studies. It would be unwise to invest massively in gas power as a clean enough alternative until this is resolved.
Baseload (CCGT)
Combined-cycle gas offers significant reduction in emissions (ignoring fugitive methane) in the short term, but risks becoming a stranded asset longer term. By 2040, at 350kg/MWh, emissions levels from CCGT power may well be unacceptable.
Peaking (OCGT)
For peaking power, Lazard's[5] calculations for the US have CST (aka CSP) with storage already beating OCGT:
However, it does not clarify whether this is OCGT frame or aeroderivative – presumably frame. Aero has slightly lower CO2 emissions.
The Australian Power Generation Technology Report (APGT)[6] agrees at the low end - AUD170 for CST, AUD178 for OCGT frame – but lists AUD115 for OCGT aero. The choice will perhaps depend on whether the gas-fired peaking plant is required to operate also as combined cycle.
Lazard's CST number appears rather low. Existing unsubsidised PPAs start at USD135[7].
ACIL Allen[8] quotes wide low-to-high LCoEs for each technology. Top end numbers such as $320/MWh for CST presumably correspond to early, experimental or highly subsidised plant. Consequently table 2 below uses the low-end number plus 10% of range.
Future costs
ACIL-Allen[9] foresees gas prices doubling in real terms by 2030.
Renewable sources and storage[10] still experience declining real costs. For PV and batteries, there are technological advances; for CST and wind it comes from growing experience in manufacturing and installation, and in decline of the risk-loading of discount rate as investors become more comfortable.
The International Renewable Energy Agency (IRENA)[11] forecasts (international) LCoE declines to 2025:
· 50% for solar PV
· 30% for onshore wind
· 45% for CST with 7.5h molten salt storage
APGT6 has similar trajectories, on the assumption that international policies drive the market enough to stay under a 550ppm cap.
More conservative analysts would no doubt be less sanguine, but the lesson of the last 20 years is that even the renewables optimists were rarely optimistic enough.
ACIL Allen also provides the sensitivities to fuel cost. Table 3 below extrapolates to the doubling of gas price expected by 2030.
The Mix
Tables 3 and 4 below compare mixes of CCGT, wind, utility PV and CST. They assume that 10% dispatchable (in addition to any hydro) is required to support an otherwise baseload grid, and increasing the variable fraction of supply requires more peaking. Since CSIRO[12] considers that 30% wind + PV is not a problem with the existing back-up and redundancy, the tables allow an extra 10% dispatchable for each extra 10% variable beyond that.
To minimise output correlation, wind and solar are kept about equal contributors.
At a carbon price of $45/t, CST matches OCGT, while wind and PV can displace half the CCGT. At $70/t gas goes down to 20%.
Allowing for 1% fugitive emissions drops the carbon price by $10 for those thresholds.
Table 4 factors in IRENA's predicted declines in renewables costs by 2025:
Gas has disappeared from the mix completely, no carbon price needed.
Conclusion
CST should be developed rather than OCGT for any enhanced reliability.
In the short term, replacement for retired coal plant could be 50-50 CCGT and renewables, but over time renewables and storage will win. When the emissions intensity of the energy industry has fallen by 90% other areas – industrial processes, transport, agriculture – may be more rewarding targets.
3.3 What are the barriers to investment in the electricity sector?
The uncertainty created by the political polarisation around climate science and, consequently, of energy policy.
3.4 What are the key elements of an emissions reduction policy to support investor confidence and a transition to a low emissions system?
Within the limited context of the existing "Direct Action" scheme, the penalty for failing to surrender sufficient certificates should be a late fee. The shortfall should carry over to the next year. In this way, the total number of certificates required over time is not eroded.
In the various schemes Federal Governments have pushed so far (starting with Kevin Rudd's ETS), much thought has gone into providing certainty for power consumers. The ETS projected price-caps well into the future that could only rise gradually, regardless of contingency. There has been scant attempt to provide corresponding certainty for the new energy industry. The Gillard cap-and-trade had a floor price only in the early few years.
The RET could have done better in this regard had it not been so readily rejigged by successive administrations. The multiplier applied for household PV when there was a single RET crashed the price and delayed utility scale. Separating out the LRET restored it, but strong factions have worked to cut it or abolish it ever since.
A bipartisan agreement would be great, but still subject to the whims of future goverments.
The weight of public opinion could fix that, but too many are influenced by the powerful fossil fuel lobby via certain media organisations.
In short, it is not just the policy, but the confidence that the policy will last.
3.5 What is the role for low emissions coal technologies, such as ultra-supercritical combustion?
Apparently none.
APGT6 quotes 0.792tCO2/MWh sent out for IGCC, black coal, and 0.773 for USC; only 2% better.
For LCoE, it gives USC at $80/MWh, but as shown above that will be undercut by renewables quite swiftly. Coal also has considerable health costs in its mining, transport and combustion. Various international studies[13],[14],[15] have put this at upwards of AUD50/MWh, exclusive of health harm through climate change.
Chapter 4
p25
"their capacity to deliver electricity is lower than that of a coal-fired power station of an equivalent size ... for utility-scale solar PV, by approximately a quarter "
I believe it is about a quarter, not lower by a quarter.
p33 Consultation questions
- 4.2 Should the level of variable renewable electricity generation be curtailed in each region until new measures to ensure grid security are implemented?
In view of the more recent issues with the NSW grid, it would not appear that curtailing VRE is that relevant.
· 4.3.2 Should all generators be required to provide system security services or should such services continue to be procured separately by the power system operator?
While there should be minimum standards, free market principles suggest that a secondary market in providing such services would be more efficient.
· 4.4 What role can new technologies located on consumers’ premises have in improving energy security and reliability outcomes?
It is clear that a system by which a price signal is automatically sent to customer equipment, with preset options in the equipment determining whether to curtail load, would have great benefit both in security and cost of power. The costs and benefits of rolling out such a system need to be assessed.
◦ 4.4.1 How can the regulatory framework best enable and incentivise the efficient orchestration of distributed energy resources?
A difficulty is knowing how many "kW metres" are used in matching distributed generation to distributed load. If the relationship between customer meters and substations is known, the price signal for both power fed into the grid and power drawn from the grid could depend on the load at the substation. When a region is largely self-sufficient, both prices would be low.
· 4.6 How could high speed communications and sensor technology be deployed to better detect and mitigate grid problems? and
· 4.7 Should the rules for AEMO to elevate a situation from non-credible to credible be revised?
The manual nature of this elevation is a concern:
p32
"AEMO has the ability to re-classify non-credible contingencies as credible when the power system is forecast to be in an ‘abnormal state’.... This is a manual process, and requires AEMO to publish a justification report after the event "
There would seem to be scope here for some AI support, trained on a variety of scenarios.
Chapter 5
p35
"wind and solar supress (sic) wholesale prices when they are producing. They rely on subsidies under the Renewable Energy Target through the sale of large-scale generation certificates to make up their fixed costs. "
The subsidies are not an inherent part of the problem. As the costs of renewable technologies fall and subsidies are correspondingly withdrawn, the problem arising from the great range of marginal costs remains.
An interesting question is whether the existing market mechanism leads to the optimal capacity profile. It should be possible to model different capacity mixes of the existing technologies against actual demand patterns to find the least overall cost (while keeping the generators viable). This could be compared with the consequences of the existing market mechanism.
p38
"The wholesale price for a megawatt hour of electricity is the same regardless of whether its source is dispatchable or variable "
This implies a fundamental flaw in the bidding system described on p34. The system models the free market view of supply and demand; as demand increases all suppliers enjoy the same elevated price. That model works well when all suppliers are essentially interchangeable; all the market needs to do is to influence total capacity. In reality, peaking plant also serves when it only stands and waits.
p40 Consultation questions
· 5.2 Is liquidity in the forward contract market for electricity adequate for the needs of commercial and industrial consumers and, if not, what can be done?
Just as dispatchable supply (peaking plant, batteries) provides security of supply, it should also be sufficiently bankable to provide financial market liquidity. It may be that the failure of the NEM bidding adequately to reward peaking supply impairs this.
· 5.5 Rule changes are in process to make the bid interval and the settlement interval the same, both equal to 5 minutes. Are there reasons to set them to a longer or shorter duration?
The existing rule appears to be tailored to specific incumbent technology. It should be changed.
· 5.6 What additional system security services such as inertia, as is currently being considered by the AEMC, should be procured through a market mechanism?
Inertia, standby capacity and rapid response.
Chapter 6
p43
"The cost of policies, such as the Renewable Energy Target and premium feed-in-tariffs, are included in all electricity bills. In FY2015 the cost of these policies represented around 6.2 per cent of the average household bill "
Unlikely, if correctly compared with what the household bills would have been without the growth of PV.[16]
p44
"Cost reflective pricing involves charging prices that accurately reflect the cost of providing network services to different consumer groups. "
A key difficulty is the risk of much higher costs to low income families, particularly elderly sole occupants.