Talking about our generation: Our energy models are out of date

“Talking bout my generation:” Peter Townshend 1965

The challenge facing renewables and the NEM

To a high degree of probability the LCOE [levelised cost of electricity] of wind and PV is at worst comparable with gas and coal. In all likelihood it is lower. However, that is not the full story. The cost of electricity to consumers big and small is not just the cost of producing the power, but also the cost of dispatching it and the cost of making it available as required. We can write this down as a straightforward equation.

Show me the model

Nearly every discussion on new generation takes a partial approach. That is why our primary call is to start with some proper planning and a model. We’ve said before and we’ll say it again: all the existing models for electricity supply in Australia are incredibly out of date and simply focus on LCOE of generation.

As discussed the LCOE is not generally reliable unless there are recent examples to go by. No acceptable private sector or government model exists that models high renewable penetration across the NEM. The only existing formal study done by AEMO is years out of date and frankly hopeless. Several high quality academic studies exist but these have yet to be adopted by the industry.

The academic studies are also to an extent partial and certainly don’t take into account positive externalities  (benefits to the system such as reliability or distribution savings or greater reliability of other plant) than can be estimated in a whole of system approach. The best example of the sort of modelling we are talking is the  Massachussetts state of charge report

Homework before ideology. Let’s help South Australia not make consumers footballs to be punted from one chamber to the next

In our view, both the Federal Govt and the Federal Opposition have not done the required amount of homework. The Federal Govt. bangs on about energy security and the cost of renewables but presents no evidence about the cost of renewables. They simply aim to create a perception that there are renewables in South Australia and South Australia has expensive and unreliable electricity. End of story. That’s not good enough for anything except political point scoring. It just isn’t good enough for South Australians or for  electricity consumers across the country.

At the same time the Federal opposition articulates a 50% renewable target across the NEM (a target we strongly support) partly implemented via an emissions intensity scheme, but it provides no evidence of why that won’t result in exactly the same problems that South Australia is seeing today. In short it’s a target without the implementation foundations required to make it credible to business big and small. A big renewable target needs a good model.

As important as the model is, its not the subject of this note. Today we do want to focus a bit on South Australia because it has become a test case and we don’t want to see the State get off its horse half way across the river.

High renewable penetration grids

It’s easy enough to find a list of countries with high renewable penetration, but the majority, if not all, have relatively high levels of import/export capacity which makes the country percentage less relevant to Australia. What we are really interested in is grids with high renewable penetration.  Moreover, we want to be picky and ignore hydro and biomass, both of which are dispatchable. South Australia has neither. A summary of some selected data is shown in Figure 1:

Some of these countries have targets as high as 50% but for the time being South Australia is well in the lead. California and Germany both have good import/export links.  Also, virtually all the regions except South Australia have hydro, bio mass and/or ample gas generation capacity that can meet any  intermittency requirements. That said, the NEM as a whole, and despite excellent PV and wind resource has the second lowest share of “modern” renewables in the table. Both Texas and California are well ahead of the Australia. Ireland -a grid  comparable in size to South Australia  – has legislated a 50% renewable requirement.

,South Australia Government role

Let’s be clear about a few facts.

1. From a generation planning perspective the South Australian Government has zero responsibility for the generation mix in the State. The Sth Australian Govt. doesn’t own any generation (at the moment) and has no responsibilities for network planning or electricity supply other than as one of the Govt members of COAG.
2. The fact that South Australia has a high percentage of wind in its generation capacity mix is because:
   1. The Federal Govt’s RET policy which we continue to criticize from pillar to post makes no allowance for the impact of renewable generation in the State mix. You get a REC whether you are located in South Australia or QLD. You get an REC whether you the first unit of wind built in the State or the last unit. South Australia doesn’t really have much control over how much wind is built in the State to satisfy a Federal target. That’s not to say they didn’t willingly go along but the motivation was Federal policy not State policy. This would have been a sensible thing to debate during the various RET reviews rather than the incessant ideological debates over the size of the target that actually took place.
   2. South Australia is a great place for wind generation. Its windy, there is enough transmission and relatively little community opposition. For many landowners in the State wind has been a blessing.
3. Wind is not responsible for high prices in South Australia. When the wind blows the price is low. It’s the lack of wind that hands control back to gas generators, and leads to high prices. Wind costs consumers in South Australia a lot of money because its doesn’t generate all the time. However, it’s the gas generators that get the high prices and perhaps profits. We also note in this context that AGL has lost profit relative to last year on buying more gas at high prices to feed mainly TIPS A and TIPS B in South Australia.
4. Not just to South Australia but to the NEM in general, the generator market is getting more concentrated. By and large owners and operators of gas fired generation also own much larger coal fired units. If gas generation is bid high or not at all the bigger coal fired units will get higher pool prices. This only matters to that portion of generation that is unhedged. It will be interesting to see which tale of woe emerges from this from the buyer side.

Three questions for AEMO’s  handling of South Australia

In our opinion, AEMO “business fitness” is subject to three questions.

1, Should AEMO have known about the “ride through” settings on the South Australian wind farms and had them adjusted? Some background on this issue can be found her.e In our view this is the most serious of the issues. Surely  AEMO has something of a responsibility if not a duty to be aware of what “ride through” settings are appropriate, to be aware of best practice, and to ensure that best practice is followed.

2 Should AEMO have derated the Heywood interconnector prior to the major blackout?  In our view this is more debatable and comes down to an on the day judgement call.

3 Should AEMO have understood that it wasn’t getting enough bids in South Australia to satisfy forecast demand last wednesday and done something about it? Our view is that AEMO could have and should have done something about this. Clearly by the next day they had worked this out and Pelican Point operated. Engie’s 10^th February press release makes it crystal clear it could operate if directed by AEMO. There isn’t much more to be said about this.

Taken as a group of issues there is a sense that AEMO might have been more proactive, and more  alert to what’s happening in today’s market and that South Australian’s have been the loser out of this. It’s not AEMO’s responsibility that South Australia is short on dispatchable generation but some of the worst outcomes over the past year might not have happened if AEMO had been a bit cleverer. Again its easy to be an armchair after the event critic.

Not that much energy was lost in South Australia last week?

About 90,000 houses and businesses lost power for 45 minutes. We estimate that is roughly about 105MW (compared to peak demand of 3055 MW and an average of 1347 MW over the past 12 months). Using AEMO’s somewhat artificial value of lost load [VOLL] of $15,000 MWh the value of the lost electricity is just $1.5 m.

The basic mismatch between wind and demand is shown in the following chart

South Australia and high renewables

The statement that South Australia is a test case for the NEM and increasing for the wider world is true in our view. However, South Australia’s many generation  limitations mean that as a test it faces some severe handicaps compared to most other grids.

Supply = wind + rooftop PV + gas + imports;

The truth is that this is a pretty small list of sources of electricity for a grid particularly when the interconnection limit is fairly low..

If demand exceeds that supply then there is no energy security and we get load shedding. The problem in the short term is that gas supply and imports are limited. The gas generation is also expensive by world standards and inefficient.

Building more wind in South Australia won’t do much to solve the problem because on the evidence to date all the wind farms in South Australia are highly correlated with each other.

Building more PV won’t be any use when the  Sun isn’t shining.

Building more gas is no use without gas  supply. In any event even gas isn’t clean enough any more give the global pace of decarbonization required. Wind+ open cycle gas is not that effective as a carbon reduction strategy. Wind + combined cycle gas may work but is not without issues from the gas plant’s point of view.

What South Australia do to improve energy security short term

The Government of South Aust has only limited policy options open. We broadly see three options:

1. Demand management. Paying people not to consume. That is an effective solution but unsatisfactory. We want a grid where that isn’t necessary. Still it would likely be relatively cheap and easy to do.
2. Contract with Engie for some or all of Pelican Point’s unused capacity. Its become very obious that that is what is going to happen and we cover Pelican Point in a little more detail below.
3. Take the battery option seriously.

Pelican point

We’ve covered this during last year’s high priced South Australian events. Pelican Point doesn’t run at full capacity

Last week it ran as follows:

Figure 3: Pelican point output, half hourly. Source: NEM Review

The load shedding occurred on the 8^th of Feb when PP’s max output was about 237 MW. The next day output somehow lifted to 320 MW. If that output had occurred on the 8^th the load shedding could have been avoided. PP’s capacity is 485 MW. It presently does not have a gas contract to run the second unit.

There is no doubt the South Australian govt can and almost certainly will underwrite the operation of Pelican Point for the next few years. Still, we don’t think it will be cheap.

The gas cost will likely be around $8 GJ, plus pipleline capacity has to be rented (although Engie may have this already). The second unit won’t be required very often, definitionally it was only needed a few times in the last 12 months.  The heat rate (GJ gas required per MWh) for PP overall is 7.5 but for the second unit, if that is what the Government rents,  its probably around 11.5 making the gas cost alone maybe $80-$90 MWh

Engie will likely require a capacity payment (a fixed monthly fee for providing the capacity) plus coverage of all costs.  Engie which also owns Hazelwood with over $1 bn closure cost and LYB (for sale) may not be feeling in a particularly charitable frame of mind and will strike a hard bargain. The annual rental for the 250 MW unit might be $20-$30 m per year (just a guess). The pipeline rental fees another few $ million on top.

More broadly the gas supplier market is concentrated, the gas transport industry is concentrated and gas is not a great fit with wind.

In short the despite the almost universal sentiment that gas is the answer not just in South Australia but in the NEM we think its not the only answer. Gas is a recipe for high prices and high carbon output.

Batteries should be given a serious go

We think utility scale lithium storage is present around US$600-700/KW installed with inverters.  That’s a KW of usable power. We base this on conversations with battery suppliers. 100MW on say a 4 hour basis would cost around  A$90 m.

Our understanding is that grid scale systems are valued on power.  California has shown that in response to the Aliso Canyon gas leak storage can be requested and deployed inside six months. Some background on AES’s portion of the installation can be found at inside the world’s largest lithium battery site

Residential systems are valued more on energy. The Powerwall 2 for instance is rated at 12.5/kWh deliverable energy and 5kW continuous power. This is probably about $1700/kW of power The average house uses an average power load of only 1/kW on a 24 hour basis. Based on the increasingly accepted principle that storage is most valuable when it is placed at the perimeter of the grid we think the South Australia might consider incentivizing say 50-100 MW of household storage. Lots of local employment for a while and it future proofs the grid.  100 MW would cost gross $200 m but households could pay say 2/3 of the cost.  So the Govt pays say $70 m.  20,000 houses would end up with storage. Add that to 100 MW of utility storage and South Australia is well on the way to the future.

As usual with batteries they serve a variety of roles.

1. Manage demand spikes
2. Manage transient intermittency (brief periods of cloud or rain for instance)
3. Provide frequency control
4. Can be charged at times of high wind output and lower demand enabling the gas generation to run at better efficiency
5. Provide some grid related services.

Studies clearly show that storage is already economic in the  Massachusetts and ERCOT markets where electricity prices are way lower. There is no way storage couldn’t be economic factoring in the “externalities” in South Australia.

As another example a Hawaii Island co-op  in December signed a PPA with US utility AES for a 28 MW PV systems with 20 MW (100 MWh) system for US$110 MWh (A$ 144 MWh). Details can be found at Kauai Solar Storage Peaker

South Australia’s average pool price over the past 12 months is A$93 MWh and more capacity would force down the price. The following chart shows the distribution of pool prices in South Australia over the past year. For instance the 90^th percentile is $139 MWh, ie the price is  < than $139 MWh 90% of the time.

Figure 4: South Australian power prices. Source: NEM Review

Levelised cost is hard to measure and not the only game in town

For many years engineers and financiers have thought about generation in terms of the levelised cost of electricity [LCOE]. This is the electricity price a generation unit needs to average over its life in order to cover its all of its fixed and variable costs. Its equivalent to the Economist’s Long Run Marginal Cost.

The advantage of the LCOE is it enables a comparison of wind, PV, coal, gas, nuclear. Wind has (had) high capital costs relative to gas but lower fuel costs.

However we think that there are lots of problems with the LCOE which means that at the very least it needs to be used more carefully than we sometimes see.

• Unfortunately the LCOE is framed in terms of cost. It therefore omits the increasingly important variable of Gas electricity gets much higher prices than wind electricity in Australia, so IF gas was higher cost it MIGHT still be a better deal. This is a subset of the broader problem that generation is considered from its own micro point of view and not from its marginal impact on the entire system. Not all of the marginal revenues and costs are considered.
• Generally an “overnight” cost is assumed. Overnight means that the plant is assumed to be constructed instantaneously. If only the QLD LNG plants really had been overnight. However in the real world a nuclear plant would probably take a decade to build. Even an Ultra Super Criticical [USC] coal plant would take maybe five-eight years to permit, finance and construct. On the other hand a PV plant can be up and running in 18 months, Rooftop PV can be done in a month.
• No allowance is made for flexibility or capacity utilization. The economics of a generation unit change markedly with capacity utilization. The more capital intensive it is the more sensitive to utilization.
• Generally a single weighted average cost of capital [WACC] is assumed. Two errors are often made when setting the WACC;
  • The WACC assumed is generally wrong, or at least not justified. There are countless examples. For instance the 2015 CSIRO report, done in collaboration with a range of industry people assumed a nominal cost of debt of 8%. Any CFO paying 8% in recent years in Australia would be instantly fired. The WACC has a much larger influence on the LCOE of renewable energy than it does on fossil fuels. For this reason the finance industry sometimes ranks projects in terms of their IRR [internal rate of return] and inputs an electricity price.
  • The second thing is that not all forms of generation are equally risky. Renewable energy with a fixed PPA, low and fast construction cost and no fuel price risk might be seen as much less risky than a big complex coal plant that takes years to build, and might be subject to future carbon prices. A lower risk project has a lower WACC. Only rarely if ever is this allowed for.
  • Sensitivities are often run on fuel and capital costs but rarely if ever on the WACC.
• No allowance is made for impacts on the network. Actually that’s unfair in the real world the marginal loss factor [MLF] is allowed for, but my point is broader. Distributed generation has significant networt, even social benefits that are generally not allowed for. It’s a pet complaint of mine that all of this national debate about energy security, and price has been done without any proper modelling. A bunch of thought bubbles. Few are more egregious than the recent call for more coal but generally where is the system wide modelling to back up the policy development? The modelling is more out of date than some of the plants that are modelled.  What’s needed is a complete NEM wide model that incorporates transmission, utility generation, distributed generation, network investment and storage. This would be something the Federal Government could commission and be non partisan.
• Finally, of course, much of the work is theoretical. No one has built a coal plant in the NEM for a decade and there are no USC plants. So we don’t know what the capital cost is. In China, where at times they turn out a plant a week, capital costs are easy to calculate and there is a strong learning curve. Not so in Australia. As far as fuel goes, guessing the gas price and perhaps the coal price is just that, a guess. These risks can be factored in with sensitivity analysis or they can be incorporated into the WACC. In the Press recently there has been a bunch of quoting about the LCOE of a new USC plant in Australia. Figures of $80 -$100 MWh have been reported for coal as if they are fact. They are not necessarily any more factual than reports that wind costs $100 MWh when contracts are being signed at $75 MWh.  The coal cost in QLD a long way from South Australia will certainly be different to the coal cost in NSW. Getting a new thermal coal mine up in NSW is no longer as easy as it once was.

The coal numbers from China, just for comparison

China has by far the world’s largest installed quantity of coal generation. New coal plants there have a capital cost of around US$520/MW compared with Australia where based on the CSIRO 2015 assumptions the capital cost is US$2200 MW, around 4X higher. China’s capital cost is low partly because of labour but probably more to do with the quantity of plants they build  having driven them so far down the cost curve. At 85% capacity use and with coal at US$75 /t  such a plant has an LCOE of US$55 MWh. The reality in China is that most new plants will struggle to get 50% utilization.

---

David Leitch is principal of ITK. He was formerly a Utility Analyst for leading investment banks over the past 30 years. The views expressed are his own. Please note our new section, Energy Markets, which will include analysis from Leitch on the energy markets and broader energy issues. And also note our live generation widget, and the APVI solar contribution.