Degradation of performance at high temperatures

As noted on social media (Twitter and LinkedIn) earlier, the Australia Institute helpfully coined the headline “can’t stand the heat” when investigating performance degradation of fossil-fired thermal plants at high temperatures in recent years.

Given that we can expect to see more extremes of temperatures into the coming years (and decades) and that the plant mix is changing significantly, we’re investing considerable time to explore a diverse range of aspects of high-temperature performance degradation for all operational DUIDs in the NEM, no matter the fuel type.

Here’s one sample from the analysis this week for a “mystery DUID”:


Massive degradation in performance of mystery DUID at high temperatures

This analysis is being included with all the other analysis performed in our Generator Report Card – which will be released shortly after the federal election (to be held on Sat 18th May).


Happy if you want to leave your guesses in comments below as to which DUID the above chart refers to.

Not that I have time to think it through at all currently, but given our prior history in running competitions, I’d like to think we can organize some form of prize later (after the report has been released and we have caught our breath) if anyone were to hit the nail on the head and guess this one correctly…

About the Author

Paul McArdle
One of three founders of Global-Roam back in 2000, Paul has been CEO of the company since that time. As an author on WattClarity, Paul's focus has been to help make the electricity market more understandable.

14 Comments on "Degradation of performance at high temperatures"

  1. Ok so its not wind, which when it is hot normally goes to close to zero output, because there is no wind.
    Solar as it goes over 25c gets a little less, so when hot a bit more.
    Coal plants do
    Gas plants do
    Battery’s once hot after discharge, struggle to recharge if hot, so their 24 hour cycle will go down when its hot
    Inverters suffer when its hot, both ways, the cooling needed in them goes up a lot (Fans), so to charge a battery must get a double whammy, the battery struggles in the heat, and the inverter to charge it and then to re-send its power back out at the set voltage.
    So which one do I pick, hmmmmmm

  2. Steam turbines – constraints due to ambient temperature can manifest in the condenser hotwell (impacting STG condenser vacuum), the cooling tower (also vacuum), the lube oil and jacking oil coolers (heat exchangers). Ambient temperature of 45degC doesn’t have much effect on a boiler operating at hundreds of degrees.

    Gas turbines can be directly constrained by high intake air temperature – like a turbo vehicle GTs are sensitive to density and exhaust temperatures.

    For a steam turbine plant there are some external factors not directly related to the equipment that can force reduced output. The most obvious being cooling water discharge temperature – if that is tightly monitored, as it is on discharges into rivers and lakes, then reducing load might be the only option to keep those temperatures within their regulated environmental thresholds.

    My guess: Loy Yang A or Kogan Ck

  3. Ben it is more likely Millmerrian than KC. Just my opinion. If its one of them and not a battery/ inverter change discharge cycle in hot weather

  4. The behaviour of the Mystery DUID looks very much like what happens at the wind farms around Goulburn in NSW. My own analysis of weather data and AEMO data shows that Taralga, Woodlawn, Gunning, Gullen Range, and Cullerin Range all have maximum power versus temperature curves that look a lot like the Mystery DUID.

    Solar farms show almost constant maximum a.c. power output as a function of temperature. Probably because they tend to have oversubscribed d.c. power and so the a.c. output can still reach registered capacity even on very hot days.

    The Mystery DUID sure ain’t Kogan Creek. It managed to provide about 94% capacity factor at 40C in 2018. Both units at Millmerran have been at about 82% capacity factor when the temperature was 40C in 2018. Keep in mind that Paul’s graph shows MAXIMUM power, and so while some coal DUIDs have been known to suffer from heat stroke, they seem to be able to keep going some of the time when it gets hot 🙂

    You will have to give is a clue so we can narrow the field Paul!

  5. Correction to my previous post: For solar farms the “registered capacity” in AEMO’s Registration & Exemptions List appears to be the d.c. photovoltaic module total capacity. “maximum capacity” appears to be the solar farm’s maximum a.c. (inverter) output power. When I said registered capacity in my earlier post I meant the thing AEMO calls “maximum capacity”.

  6. Hi Mike, How about the Big Batt, the cells that have discharged in 40C, how much actual power comes out, vs what was put in when it was cool, plus, how hard is it then for them to cool and take a charge. Meaning lets say the 30MW they use (If they in fact only use the same 30mw of the battery) to cycle in say 20C, what is the charge losses, then discharge losses, and speed in which they can recharge and then discharge again over a 24 hours period. Vs a period of time where its 40C for 12 hours and over 30 for the rest.

  7. The Hornsdale Power Reserve battery has only been called upon to deliver its full output power of 100 MW on a few occasions. Most of the time its output is limited to 30 MW so that it has 70 MW in reserve for emergencies.

    You’d expect it to struggle most when it’s at full power because the heat dissipation will max out when the power is at its maximum. It’s possible that they use some of the batteries at 100% when they are delivering 30 MW. Or they might only use 30% of the power capability of all the batteries when they are delivering 30 MW. I don’t know how they manage it.

    These are some of the times when it has jumped to 100 MW and the weather has been warm to hot at the time:
    * 13 Dec 2017 at about 11:30 and about 37C
    * 24 Jan 2018 at about 16:30 and about 32C

    It was delivering 100 MW on both occasions.

    It hasn’t gone to 100 MW in 2019 so far.

    On a typical day its output is mostly zero. It seems to sit there looking for opportunities to make money on the spot market (energy and FCAS). You can often see it jump to 30 MW when wholesale prices are high and then recharge at 30 MW when wholesale prices are low. Sometimes you see it at 30 MW in one 5 minute AEMO SCADA interval and then recharge at 30 MW in the next 5 minute AEMO SCADA interval. It has done a few thousand of these +30/-30 MW cycles since it was commissioned.

    Tesla designs electric cars. Cars have to work at quite high temperatures (air temperature plus insolation). Tesla’s batteries have a liquid cooling system. My Tesla battery at home has the same sort of cooling system. I can get the full 5 kW for my air conditioner on days that are 40C.

    So I reckon the Tesla battery systems and other properly design grid scale batteries will have no problem on hot days. They will be designed not to overheat when they are charging or discharging. You just need a well designed cooling system.

    I haven’t looked at how much energy is lost in the Hornsdale Power Reserve battery but it kinda doesn’t matter if you can buy $50/MWh and sell at $300/MWh.

  8. PS: The best match I’ve been able to get for Paul’s mystery DUID is Waterloo Wind Farm. That’s my final offer! 🙂

  9. Hello Mike, Thank you for your messages. Ok firstly. What i am talking about is simply the maximum rated amount of power the battery can deliver in 24 hours. It does not matter if its the whole 100, or just 30MW, as long as its just the same sections of battery’s that are being used. Just like a gas plant or coal plant have a rated output. Kogan Creek is like 750MW. And as you know that is 750 by 24, to equal what it can do in 24 hours. CF (Capacity factors are done over a year) So its 750 by 24 by 365. would be 100%. On a hot day as we know Kogan creek has to wind down, it cannot run at 750 during hot weather it can only roughly get out 680 to 700. (Roughly)

    So the battery must have a set maximum it can run at in 24 hours. Like it can be charged and discharged lets say 10 times so in 24 hours it can deliver 1000MW / 24 equals an average of 41MW per hour. (I am guessing at that, what is its rated day capacity? 100mw is how much it can hold, which is diffrent than what it can deliver) However on a hot day, its not going to deliver 41mw per hour, its going to suffer from the heat, its not the batterys will be hot, the inverters will be hot, the power lines will be hot. The battery must also have a reduction in its rated capacity. What is it? That is all I am asking, or alluding to.

    Now as for the battery buying low and selling high, that is 100% fine, to make money. Its not the best, if you are short on power because the battery takes in less than 100% to start with, and then lets out less than 100% of whats in it. So there is a loss both ways. The battery produces zero power, it uses power to deliver at a later date, or as you put it to buy low and sell high. But not matter which way you cut it, unless its being charged with excess power that would otherwise be lost, its adding to demand. Which that battery has been doing most the time as it mostly chargers from the grid.

    Your tesla will also have a maximum amount of times in one day it can be charged and discharged I would call it the cars 12 hour driving range or 24 hour driving range.

  10. Think it’s Loy Yang A or Yallourn Power Station.

  11. Hi John, I think you might be trying to compare apples with oranges? Capacity factor does not appear to be a relevant figure of merit for a battery.

    At the risk of sounding like a Tesla fan boy (and with apologies to everyone for taking up a lot of space) here are a few thoughts regarding your doubts about modern batteries (…ok, I admit it; I am an Elon Musk fan. Anyone who can nail the synchronised landing of two booster rockets deserves some respect if you ask me).

    In an electric car the battery discharges whenever you are on the accelerator and charges whenever you are on the brakes thanks to regenerative braking. That means it might switch between discharging and charging hundreds of times a day if you’re driving in city traffic like Tesla hire car drivers do, or like Toyota Camry Hydrid taxi drivers do.

    As I have already mentioned, the Hornsdale Power Reserve battery switches between charge and discharge many times each day (and night).

    My home battery switches between discharging and charging pretty much instantaneously. On hot afternoons when my solar panels are generating not quite enough to meet the peak power demand of my air conditioner, as the air conditioner cycles between almost no power and peak power, the battery has to switch between charging (using excess solar power that the house is not currently using) and discharging (when the air conditioner needs more power than the solar system is currently producing). This can happen lots of times per day and Tesla guarantees the battery for 10 years. By the way, my battery has carried me through a dozen power interruptions in the nearly two years I’ve had it. The most recent event was 3 hours in January this year in suburban Sydney. Uninterrupted cricket viewing in a nice cool house. That battery is worth its weight in icy cold XXXX Gold. 🙂

    In summary, as far as I can tell there is no _practical_ limit to the number of times per day a Tesla battery can switch between charging and discharging.

    On the subject of the Hornsdale battery acting as a load, that is not necessarily a bad thing. It can save inflexible generators (e.g., thermal ones) from having to rapidly follow fluctions in demand. That sort of help should save the owners of thermal plants some money and _should_ drive down the cost of electricity (sigh). A big battery could have saved Queensland from some pain when it recently disconnected from the rest of the NEM and the sudden loss of a large lump of load caused the frequency in Queensland to go troppo for a while. Also note that when power is in short supply prices tend to be high and so it is unlikely that a battery would be charging and adding to demand.

    I know it’s politically incorrect to say so in Australia but all the data seems to be saying that batteries might just save the world (or at least what remains of the Hunter Valley).

  12. Hi again John, apologies for not addressing one of your important observations. In 2018 the AEMO dispatch SCADA data says that for the Hornsdale battery about 49 GWh went in and about 40 GWh came back out again, so it returned about 82% of the energy that went into it. 82% is comparable to pumped hydro plants (Wivenhoe, Tumut 3, Shoalhaven, Snowy 2.0, etc). See If you’re worried about the energy losses of a battery they pale into insignificance compared to the energy inefficiency of any thermal generator be it coal, gas, diesel, or whatever. Coal is typically only 33% to 40% efficient even for HELE (ref: My home battery returns about 94% if I deep cycle it and about 85% for the shallow cycling that happens most days. You might think I’d be unhappy with 85% but it’s a lot better than the 25% return I’d get if I sold my excess solar energy to the grid at the NSW IPART rate of 8 cents per kWh during the day and had to buy it back again at 32 cents per kWh during the night.

    I think that’s enough from me.

  13. Hi Mike, Thank you for your message. You do like Mr Musk.
    As for a Battery not having a maximum capacity, and not having losses, you seem to some how have that part missing. The way you write is as if no matter what time of day, temp, the battery works the same, and so does the inverter thats between it. Not matter what.
    Really there will be a difference as there is with everything else.
    I am not saying a Battery is not excellent, I am just saying its also suffers in the heat. As do inverters. As do thermal FF plants, as do solar pannels, as do aircons and lastly as do humans.
    As stated on a 40c day, a batterys ability to meet its maximum supplied 24 hour total amount of power, will be less than on a 20c day. Due to the increased losses charging it and increased losses discharging it, both from the battery itself but also from the inverter / as well as the speed to recharge it will be reduced by the battery and also by the inverter.

  14. Hello Mike, I am sorry I did not see your second reply.
    Yes those numbers are correct. The battery lost around 18% of the power that went into it. Thats a average. So as I have said above. On a hot day, its going to be higher than 18% and on a cool day its going to be lower.

Leave a comment

Your email address will not be published.