All Our Electricity Generation Is Weather Dependent

The OpenNEM website gives access to several ways of viewing Australia’s electricity usage.  One I like shows Flexibilty.  Figure 1 shows plots for the week to Tuesday 11^th July of consumption sourced from Variable (wind and solar), Fast Flexible (gas, hydro, diesel, battery), and Slow Flexible (black coal and brown coal). I chose this period because it includes Tuesday 4^th, Friday 7th and Saturday 8^th, and illustrates just how variable the Variables can be.

Figure 1: Screenshot of Flexibilty, NEM

Obviously, when Variables are high Flexibles are low and vice versa.  Notice that Variables have one peak at noon each day, as solar generation far exceeds wind- in daylight hours, but Flexibles have two peaks each day, morning and evening.  Figure 2 is my plot for the first week of July, combining all non-variables and showing the total consumption:

Figure 2:  Total Consumption 1-7 July 2023

The total (red) shows the two peaks, morning and evening, and the overnight dip to baseload at about 20,000 MW.  Note that variable generation (green) has a midday peak due to solar that rapidly descends with the sun to a night time level far below baseload.

Note also that on the evenings of the 3^rd and 4^th total consumption is made up almost entirely of flexible generation- variable was only 3.4% of the total on the 3rd and 3.9% on the 4^th.

The next figures contrast conditions on the best day of the week for variables with those on the worst day.

Figure 3:  4^th July Consumption- Worst for Variables

Monday and Tuesday, the 3rd and 4th, saw extensive cloud cover and rain over much of eastern Australia, especially Queensland, while winds were very light in the south. This was the worst day for wind and solar, peaking at about 6,000 MW at 12.30, dropping to 1,170 MW at 6 pm.  Conversely, it was the best day for coal generators who could maintain steady generation all day.  Contrast this with the 7^th:

Figure 4:  7^th July Consumption- Best Day for Variables

The sun was out and the wind was fair. This was a great day for wind and solar, producing 60% of all electricity needed in the middle of the day.  Note how coal fired generators had to decrease then increase very rapidly (nearly 80% increase from 3 pm to 6.30 pm).  That’s not how they were designed to operate.

Electricity statistics for the first week of July show how thoroughly weather-dependent are wind and solar. However, they also show the resilience of non-variable generation, and show the excellent Capacity Factor that coal can achieve.  Capacity Factor is the actual generation as a percentage of nameplate capacity.

Figure 5:  Coal generation Capacity Factor 1 -7 July 2023

Even with Callide C out of action until next year, coal’s capacity factor dropped below 50% only on the 8^th.  It peaked at over 80% on every day of the weak, averaged 73.6% and reached 90% on the 4^th.  While wind’s capacity factor was at 94% on the evening of the 7th, at 5.30 pm on the 3^rd it was only 9.7%, and at that same time solar capacity factor was practically zero.

Thank goodness for variable generation, which adjusts to the vagaries of wind and solar.  In Australia, even fossil fuelled electricity is dependent on the weather.

In my next post I will show why you shouldn’t expect batteries and hydro dams to come to the rescue anytime soon.

(Source: OpenNEM)

Coal Generation Sets New Record After Liddell Closes!

The National Electricity Market lost 2,000 MW of generating capacity last Friday.  In spite of this, coal fired generation increased its share of total generation, to a record for the year to 30 April, of 67.52%, as Figure 1 shows:

Figure 1: Percentage of Total NEM Generation: Coal, Wind, Solar

The other immediate result was that the Capacity Factor of the remaining coal generators suddenly increased by about 5%. 

Figure 2: Running Average Coal Capacity Factor % 1 April -1 May 2023

The remaining coal fired stations ramped up their generation to make up for the shortfall- mainly Eraring in NSW:

Figure 3:  Eraring Electricity Generation 27-29 April: average 69%

Eraring maintained a Capacity Factor of around 95% for most of Saturday until Sunday morning when it dropped to 37% during daylight, then back up Sunday night and most of Monday.

Figure 4:  Eraring Electricity Generation 30 April – 2 May: average 72.1%

Why couldn’t wind and solar fill the gap left by Liddells’s closure?  Because there was not much wind or sunshine!  Figures 5 and 6 show Saturday to Monday generation at Stockyard Hill wind farm and the New England solar farm- two of the biggest:

Figure 5:  Wind Generation at Stockyard Hill: average 3.4% Capacity Factor

Figure 6:  Solar Generation at New England: average 6.7% Capacity Factor

Of course, in the coming winter there will be increased demand, and coal generators will need to be maintained.  We are not out of the woods, but the above graphs show how resilient, reliable, and efficient our much-maligned coal fired power stations are.

Could we lose 2,000 MW of solar or wind generation and have the rest immediately increase production?  Not likely!

And are Batteries and Hydro capable, and how efficient are they?

Figure 7: Battery Capacity Factor (Percent)

Batteries nearly reached 0.1 % of their stated capacity.

Figure 8: Hydro Capacity Factor (Percent)

Hydro did better- but even when producing over 28% of total NEM generation could only reach a Capacity Factor of nearly 0.4%. 

These dams and batteries are very inefficient for their cost.

Let’s see what the future holds!

(Source: OpenNEM)

Electricity Generation: The Impact of Rooftop Solar

Capacity Factor of an electricity generator is its actual generation as a percentage of its installed capacity.  A generator with an installed capacity of 1,000 Megawatts that generates 500 Megawatts has a Capacity Factor of 50%.  Obviously it is a good idea to have CF as high as possible as that will give a better return for the time, money, and effort used to build and run it.

In this post I am looking at Capacity Factors of all generators in the National Electricity Market (NEM), firstly excluding rooftop solar, then looking at CF when rooftop solar is included.

I use data available from Open NEM for the week from 8^th to 15^th March.

Firstly, Figure 1 shows the total of all major generators in Queensland, New South Wales, Victoria, Tasmania, and South Australia.

Figure 1:  Total NEM Generation 8-15 March

Solar and wind get preference, such that coal is curtailed when the sun is shining, but has to ramp up to meet demand from late afternoon to breakfast time.  Hydro and gas follow the same pattern at a much lower level, while wind generation adds its two bob’s worth at unpredictable times.

Figure 2 shows the Capacity Factor for the whole network (if there was no rooftop solar):

Figure 2: Capacity Factor NEM (excluding rooftop solar)

During this week CF varied in a regular cycle, from 27.9% to 43.8%.  Figure 3 shows this daily cycle:

Figure 3: Capacity Factor by Time of Day- NEM excluding rooftop solar

The NEM is at its most efficient- makes best use of generation resources- between 6pm and 7pm at night.  There is a lower peak in CF at 7am to 7.30am.  There is a drop in CF in the early morning (at baseload time), but the lowest CF is between about 11.30am and 12.30pm on several days.

Capacity Factors for coal, gas, and hydro have cycles reflecting that of the NEM without rooftop solar.

Figure 4: Capacity Factor by Time of Day: Coal, Gas, Hydro

By contrast, wind’s CF, which on the afternoon of the 8^th was briefly over 50%, could be as low as 2.4% and averaged 20.5% for the week.

Figure 5: Capacity Factor by Time of Day: Wind

Decidedly unreliable and inefficient.

Solar generation is much more reliable (in the sense of predictable) as we see in Figure 6.

Figure 6: Capacity Factor by Time of Day:  Solar

Solar CF is between about 40% and 60% in the middle of the day.  Note that utility solar, with tracking panels, reaches close to maximum CF by mid-morning and maintains higher CF than rooftop at nearly every 30 minute period of daylight.  Between sunset and sunrise, CF is zero.  All those millions of panels are useless.

When we include rooftop solar in the generation mix, see what happens to the CF for the whole NEM grid:

Figure 7: Capacity Factor by Time of Day- NEM excluding rooftop solar

Maximum CF is now in the middle of the day.  Figure 8 shows the difference rooftop solar makes to the CF of the whole network:

Figure 8: Change in Capacity Factor by Time of Day with Rooftop Solar

Before 9am and after 3.30pm the system is worse off. While the CF for the whole network has been increased in the middle of the day by between 2% and 6%, the average has been reduced by 4.5%, at baseload times by about 6.5%, and in the evening by nearly 10%.  Every additional panel will reduce CF even further, and this is not even considering the additional network capacity needed to keep the system balanced with such a wildly fluctuating supply.  Not a bad effort for a generating system with an average CF last week of 14.9%.

The final two figures compare actual generation at 12 noon and 4am.

Figure 9: 12 Noon Generation 8-15 March 2023

That’s all the renewables enthusiasts see: solar outperforming coal.  They are willfully blind to baseload needs:

Figure 10: 4:00 a.m. Generation 8-15 March 2023

When the remaining 1,500 MW of Liddell are lost in April, and 2,880 MW at Eraring in August 2025, the 4,330 MW gap in supply at 4:00 in the morning won’t be filled by rooftop solar or by solar farms: it will be made up by the remaining coal units working even harder (giving coal an even higher CF) until the strain is too much and they break down, and by gas and hydro.  Inevitable result: higher prices and probable blackouts (sorry- load shedding).

People of my generation often say we have lived through the best of times.

What will the coming generation say?

(Source: OpenNEM)