Know your NEM: AGL’s gas play assumes a lot more renewables

Print Friendly, PDF & Email

The most interesting thing by far at the moment is working out who is going to be short and who is going to be long in retailing space for emissions targets.

The big gentailers have focused on looking after their retail loads and, as a general statement, are likely to have to contract further to meet even the laughably unambitious 26% emissions reduction target for the electricity sector for 2030.

However, if the Queensland and Victorian State Governments have the guts to stick to electorally popular policies they have announced –  that is for 50% renewable generation in QLD by 2030 and 40% in Victoria by 2025 – then the Federal target is irrelevant and the target will be over achieved.

Of course we can add to that whatever is done in NSW to replace Liddell.

All of the above needs to be spelt out in more detail, so look out for another note on this topic shortly.

AGL’s Tomago gas plant

Speaking of NSW, AGL confirmed during the week that it will be building a $400 million, 252MW reciprocating engine, gas plant near Newcastle, potentially we believe around Tomago.

The plant consists of 14 reciprocating engines of 18MW each with a target startup date of 2022 calendar year. AGL has good gas storage facilities near Newcastle.

Historically, reciprocating engines (4 strokes) made mainly by Wartsilla and in smaller sizes by Caterpillar have been used mostly in the off grid market in Australia and in smaller sizes.

They are less efficient than combined cycle plants when the latter are run steadily, and also have higher running costs.

However, they seem to be getting a new lease of life in the utility market because of their better ability to run at part capacity.

That is, you only have to have as many machines going as you have demand, and because they can start up from warm condition in 10 minutes.

The efficiency of combined cycle plants deteriorates quickly when they have to start up or shut down quickly or to run at less than full capacity for extended periods.

We expect further gas plant to be built in NSW, EnergyAustralia is on record as saying its interested and even Snowy might want more gas capacity.

The obvious point to make is that there is no way gas plants are going to provide cheap electricity at anything close to current gas prices. This plant, like Snowy 2.0, only makes sense in a market where the energy is coming from cheap renewables.

During the week

Spot prices had several spikes in South Australia. Near term electricity futures continued to drop about 1% a week but we think this is more aimless trading than anything special.

Oil prices continued to climb making contract LNG exports notionally worth over A$12 GJ.

USA 10 year bonds briefly went over 3% before falling back. Australian 10 year bonds are lower than the USA.  A low 10 year bond rate is a sign of a well-managed economy, in my opinion.

Demand for the week was higher than last year, but basically bang in line with seasonal trends.

We include a new chart showing the REC curve. Renewable investors expecting to get REC returns to supercharge their investment in the 2020s will likely have to replace these with renewables contracts for low emission generation.

There should be a return for the service as the big gentailers will need to sign more contracts.

Gas prices were flat but below last year.

Share prices did not much except in the case of Redflow, which raised more equity.

There appears to be a pause in the selling of the highly volatile lithium shares.

Figure 1: Summary


Figure 2: Commodity prices. Source: Factset

Share Prices

Figure 3 Selected utility share prices


Figure 4: Weekly and monthly share price performance


Figure 5: electricity volumes

Base Load Futures, $MWH

Figure 10: Baseload futures financial year time weighted average

REC Prices

Figure 11 Source: Mercari

Gas Prices

Figure 12: STTM gas prices


Figure 13 30 day moving average of Adelaide, Brisbane, Sydney STTM price. Source: AEMO

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.  

  • Jonathan Prendergast

    I expect the AGL gas plant would take 1-2 years to get approvals, but only around a year to procure and construct. So I don’t think it will be ‘confirmed’ until the procurement and construction capital is committed, in say 2021. They have plenty of time to watch the market until then and decide if the market is suitable for such a project.

    For now, they have made the announcement I think to get the Federal government off their back about Liddell, and also send a signal to their competitors to deter them investigating similar investments.

  • BushAxe

    OC peakers is where the market’s heading to firm up renewables. There’s been a few DA’s go through recently in SA by Origin, EA and Alinta for OCGT’s to replace capacity lost by the impending closure of TIPS A. Ironically the peaks in SA this week was caused by use of OCGT’s to cover the complete failure of Pelican Point CCGT.

    • Jon

      These are reciprocating engines not turbines so peaking plants in the true sense of the word.
      It’d be really interesting to see AGL’s forecast hours per year for these units, you’d think not a lot.

      • Peter F

        Because of the short startup times of reciprocating engines, flywheels are also quite a viable form of transition storage

        • Jon

          Very true, which is what they’re looking at doing with a couple of the Liddel units.

          • BushAxe

            I think you’ll find AGL is proposing to convert some of the Liddell generators into synchronous condensers. Reciprocating engines/OCGT both perform the same role as fast start peakers. AGL seems to prefer recip as the new 210MW Barker Inlet at Torrens Island will probably use the same equipment as the proposed Newcastle plant.

  • Peter F

    The difference in efficiency between combined cycle plants and reciprocating plants in very hot weather is relatively minor, because hot weather degrades the performance of gas turbines much more than it does reciprocating engines. This plant will mostly be used in summer
    Combined cycle gas plants are also much less efficient at part load and these plants will spend most of their operating time at part load
    Therefore the lifetime operating efficiency of the reciprocating plant will be higher.
    A multi-unit reciprocating plant is much more reliable because it is modular even though direct maintenance costs per MWh would be slightly higher.
    As you say both start up and ramping times are much faster than CC gas so there is less need for running “just in case” saving more cost
    Finally, capital costs per MW are lower so it really is a no-brainer these days to select reciprocating engines in the Australian environment.
    It must also be noted that there is a strong chance that this politically driven investment will be redundant within a few years of installation if not before, so the lower the capital cost and the more flexible the output the fewer the losses AGL will incur

    • Andy Saunders

      A good CCGT gets over 50% efficiency (thermal energy equivalent in to electric output) at full load (I think 62% was the max ever achieved), whereas an efficient reciprocating engine (Otto cycle) gets more like 45-47%.

      As you say, ramp rates are worse for CCGT (although there are upgrades/optimisation that improves it), and part-load can be worse (in GT-land it’s usually called “turndown”)

      Economics probably more driven by capital cost and infrastructure costs in this case

      • David leitch

        I also think non fuel operating costs are higher for reciprocating engines and capital cost comparable to within 10%

        Reciprocating engine capital costs seem to have fallen in recent years but they are still relatively scarce in utility scale generation.

        • Peter F

          Reciprocating plants are slightly dearer than large OC GT plants but considerably cheaper than CCGT plants

      • Ian

        The 18MW Wartsila generation sets are a relatively good move. Not only is their peak LHV efficiency quoted as 48%, once up to operating temperature they are able to operate for up to 10 hours with a misfire (Caterpillar/MWM, GE/Waukesha, Rolls-Royce and Siemens/Guarscor all shut down within 10 seconds), their life between major overhauls is at least 60,000 hours (minor overhauls 30,000 hrs) and they too can be placed in co-gen operation to generate over 60% efficiency. Turbines have major overhauls at anywhere between 30,000 and 45,000 hours (Siemens SGT-600 is 60,000 hrs). It is also the same generating set that is targeted for the Torrens Island replacement.

        The down side is that reciprocating gensets actually won’t meet the rotating inertia standards for the NEM without inertia ‘ahem’ assistance (aka batteries).

        • Richard Werkhoven

          Well the batteries are going to be a lot better than any ‘inertia’ based regulation system anyway.

          Considering the location of this plant (which is almost certainly going to be sitting on the Tomago Smelter’s distribution lines.) and it’s role is mostly going to be as backup to Tomago to stop the whining about the need for coal power by Tomago, the ability to provide inertia is really not the worry here.

          Liddell is at the other end of one of the transmission lines.

          Liddell is going to be a storage facility providing some regulation also.

          This is largely a strategic measure to remove the excuses that Tomago is putting up to try and get subsidised electricity yet again.

          It does solve the real issues faced by Tomago also, which is the ability to avoid pot-line freezes.

          The best chance to avoid any issues is to provide generation or storage on-site, and the capacity required to back-up a pot-line is over 100MW (without some redesign of the pot-line switchgear)

          This capacity would allow them to keep pot-lines safe under pretty much any situation.

          That removes their best case for their support of the coal industry.

          Which in turn removes their case for lobbying on behalf of coal power, which they do, with $1.5 billion in trade to make their case.

      • Peter F

        62% is achievable in the best H class sets but they are around 700 MW and we don’t need such large beasts on the grid here. Further it is achieved at 15C. Smaller 300-500 MW plants at 40C are about 48% at full load and recip plants about 46% but in effect they can be 45%+ all the way down to 6% load. The CCGT plant won’t even stabilise below 40% load and will be running at about 35% efficiency at that point.
        Inertia can easily be achieved with flywheels or batteries but there are strong arguments that the main reason for high inertia is the slow response of gas and steam turbines. A system with synthetic inertia on wind turbines, batteries and pumped hydro probably doesn’t need extra inertia

        • Andy Saunders

          Peter, I think we have similar experience, and I don’t disagree with anything you write!

          “there are strong arguments that the main reason for high inertia is the slow response of gas and steam turbines” – yes. For a start the controllers are often gapped wide open so don’t start corrective action until too late. And then of course the output lags by several seconds and takes several more to ramp up or down.

          Interesting to see the *system* frequency response in the other reneweconomy article yesterday after a big generator trip – the system response isn’t particularly well-damped, which is a little dangerous.

      • solarguy

        Andy, just so you know the facts. the best thermal efficiency of a 4 stroke engine is 30%. 45-47% is a dream!

        • Andy Saunders

          Wartsila in their marketing blurb quote something up over 48% (take the usual pinch of salt, but they in practice achieve the 45-47% range).

          Otto efficiencies of over 30% are possible – you’re probably thinking naturally aspirated car engines or similar. Needs sophisticated fuel controls, ignition management, super-chargers, intercoolers etc. to get the higher efficiencies.

        • Mike Westerman

          Nonsense – I have personally tested a Cat gas spark ignition 4 stroke turbo high speed engine (Cat G3612) with gas consumption <9.5GJ/MWh, typically down around 9.3 ie close to 39%

          • solarguy

            Interesting, In a previous life I was a mechanic and the text book said thermal efficiency 28% and a good diesel 30-33%. Losses from exhaust, friction, pumping and cooling amounted to 60% Heat lost in oil and pumping oil rounded out the remaining 10%.

            What was your methodology in testing?

          • Mike Westerman

            Generator on test, with 3phase watt meter, gas flow meter, a fairly sophisticated engine manangement/monitoring PLC (unit was remote controlled) collecting a swag of data. We had to rely on generator factory test data to back calculate engine power input, but that was dyno test data so fairly accurate.

            Large HFO slow speed diesels do better so I think the Wartsila figures are not overly cooked! Petrol engines probably struggle to get out of the 20s – I remember at uni doing dyno tests on a car engine and the number stuck in my head is about 20%!

  • Malcolm M

    Perhaps AGL see a niche for offering firming contracts as part of corporate renewable power purchase agreements, and as firming for power purchase agreements AGL make with renewable investors. Snowy Hydro’s firming is constrained by transmission limitations, while all other gas generators need to bear the costs of renting pipeline capacity that will rarely be used, and purchasing gas on the spot market. AGL have the synergy of generating power next to their own gas storage facility, so there are no additional pipeline rentals, and gas should always be available.

    I am surprised they have not announced a much larger but staged peaking facility, so plans can be approved for higher transmission capacity, and a larger site footprint. Being adjacent to their own gas storage facility (the only one in NSW), this must surely be the best possible site for gas generation anywhere in the State.

    • Richard Werkhoven

      Yes the distribution capacity at Tomago is impressive.

      So many lines to existing generation facilities.

      Since the Tomago plant can in most cases take the entire load it won’t even nee to use the distribution. It will just reduce load elsewhere. Even smarter.

  • Ian

    This is interesting, Tomago aluminium smelter uses about 1GW, they can’t just shed load so this move by AGL to install reciprocating gas engines makes a lot of sense. But why just 250 MW, why not 1GW to fully cover this company’s electricity requirements when needed? Maybe AGL estimates renewables will only achieve 25% of grid electricity supply, or maybe they have confidence that in a fully renewables supplied grid that the resource would only drop to 75%, or could it be that other demand response factors would cover shortfalls in power supply. How big are the gas storage facilities that they can supply these gas engines for an extended period considering that gas consumption is roughly 1000cf /1kWh?

    The idea of using fossil fuel standby generators to allow renewables to dominate the grid is not a bad strategy if these are only operated very infrequently.

    • This story may explain why. It doesn’t need 1GW most of the time.

    • Peter F

      Aluminium Smelters can and do ramp down to 50-60% capacity given a few hours notice and they have to ramp down for maintenance etc. They also often run well below capacity for extended periods when Aluminium prices are low.
      It is not worth building a power plant that will run at capacity for only a few hours per year when for 8,600 hours per year, wind, solar, hydro and 300-500 MW of gas can supply the load.