Wind Power Generation Intermittency – It’s Worse Than You Think It Is – Part Two

Posted on Mon 11/30/2020 by


By Anton Lang ~

Power Loss Over Short Time Frames

As I wrote in the introduction, there were large power losses over short time frames and even larger power losses over sustained long time frames. In this second part of the Series, I will show the power losses over three short time frames, (a) in less than one hour, (b) in one hour, and (c) between one and three hours. While that first and second time frame is similar, I split it into two sections, because there were more than one hundred occasions where there was this large fall in power generation in an hour or less. Also, having selected 500MW as the base for the power loss, there were as many in the range 200MW to 450MW as the 107 I have included in those two time frames, effectively more than 300 separate occasions where there has been power losses in short time frames in the 800 days of data collection.

The start date for all of this was in May 2018, and at that time, the Nameplate for all wind plants was less than 5000MW. In that time frame, the total Nameplate has increased on ten occasions, with the addition of new wind plants, and the total Nameplate for wind power is now 7728MW. Most of those those new plants have been constructed in this same area, South Eastern South Australia, and Central Western Victoria.

Both short time frame power losses and sustained power losses are weather related, and using those graphs for power generation I will explain why increasing the number of wind plants has made the problem worse.

With all the images of the graphs and maps shown here, if you click on the image, it will then open on a new page at a larger size so you can better see the detail.

Power Loss in less than one hour

There are 53 entries for this time frame, and they date from early May of 2018 to the end of June in 2020, two years and two months , and even though the first entry shown on this table only dates from August of 2018, this is because those earlier totals were a little lower than 500MW, and mainly, all in the one hour time frame, and when I started to split the time frames into three sections, then the larger power losses did not kick in till a little later in time. (five of those entries listed for power losses in this first table were between 450MW and 500MW)

However, what that does highlight is the fact that as more of those wind plants were constructed, then the frequency of power losses on this scale became more frequent. That I can show you. In the Months between May of 2018 and December 2019, nineteen months, there were 25 occasions where that power loss in less than one hour was greater than 500MW of which 12 were higher than 600MW.  However, in the most recent six months, between January and July of 2020, there were 23 Occasions where that power loss was greater than 500MW, of which 14 were higher than 600MW, so those falls in power became more prevalent, larger in size, and that power loss occurred more quickly. For that earlier 19 Months, the average fall was 581MW, and for the most recent six months since the Nameplate has increased, the average power loss was 710MW. So that is three indicators all showing an increase with the construction of more of those wind plants. What this proves is that this problem has only become worse with the construction of more of these wind plants.

As you view each of these images, be aware of the scale change for the left side vertical axis. So even though the graphs look similar, they have been sized to best fit the original page they were shown on.

This first image is the graph for the first entry on this table, dated 10August2018. I have placed the marker at the top of the power loss, and that loss comes in at 450MW to the bottom of the dip there, and that is over a time frame of 40Minutes. You can see that earlier on, the power generation started to fall, and earlier in the day, it was over 3000MW and at the time of this sudden fall, it was still 2034MW. That 450MW fall constitutes around 240 individual towers, so this is not a few towers here and there, but most probably four or five wind plants in a fairly close proximity.

This second image shows the largest power loss, dated 20 January 2020, when the number of wind plants had increased, as did the total Nameplate. Again, I have placed the marker at the top of the power loss, which in this case was a large 1340MW over a period of 50 minutes, but you can see it got steeper after 20 minutes. The loss started when power generation was at a very high point, 3558MW. That fall of 1340MW is the equivalent of around 700 or more individual wind towers shutting down, almost immediately, and that’s around a dozen or more wind plants, shutting down within that short space of time. You can also see large variability in power levels both before and after this huge drop in power, and even those smaller drops are around 200MW to 400MW, so around 100 to 200 wind towers, going off line, and then, a little later there was a long and constant loss of power as well. Here, Imagine trying to stabilise a grid with all of these variations in power, and these are not small ups and downs.

This third image is the most recent image, the last day of my data collection, dated 30June 2020. Again I have marked the fall in power at the top of the fall, and that is 3658MW. The power generation falls sharply, rises a little, and then falls again. The fall from the top to the bottom of that small ‘blip’ is 540MW and that fall took place in 30 Minutes, most of it in the first ten minutes. As you can then see, there are also a number of times where it rises and falls, and each of those (four at first) is around 200MW. That main fall of 540MW is around five of those wind plants and around 300 individual wind towers. Now, for this last image I have also added the synoptic chart for the area where most of these wind plants are located, and at around the same time as for these falls in power. Note here that in this area, those isobars are closer together as that front starts to come though. Isobars closer together indicates higher wind, and note it is in this same area where the largest number of wind plants are located. Again, as you can easily see from the graph for wind generation, it varies up and down a number of times across those 12 hours from that first almost vertical drop in power generation, and again, imagine trying to stabilise a grid in this situation. These are not small amounts of power.

Power Loss in a one hour time frame

Now, while there may be a small amount of confusion stemming from these first two groups, those less than one hour, and those of one hours duration, when I originally compiled these tables, there were 107 separate entries for power losses of 500MW or more within that one hour duration, some less than an hour, and some right on the one hour time, and rather than have one huge long list, I split it into two separate lists, one showing all those losses in under an hour, and those that occurred on that one hour mark, so in effect there were 107 occasions where the power loss was 500MW or greater in an hour or less, so 53 in that area above and 54 in this area. You can see this a little easier by looking at the Tables, and the links for those tables are at the bottom of the Post here.

For this grouping of power losses in the one hour time frame, there were 54 separate entries in the same 2 years and two months of data collection. In that one hour period, the highest fall was 980MW on 09May 2020, and there were 16 occasions where that loss was 700MW or greater.

This image shows the graph from close to the time I started collecting the data, 18May 2018. I actually started out at 250MW, and in the first session of data collection, I quickly changed that to 500MW, as I was getting so many entries, so that’s why this dates from 18May 2018, and is only the lesser total of 450MW, but it shows the general indication of the way things went. I have placed the marker at the high point, and it steadily falls by that 450MW over the next one hour to that small flattening before it continues to fall. This 450MW fall is around 240 towers and four wind plants. Note here on that left side vertical scale only goes to 1600MW, and that this fall starts at a lower power level.

This second image shows the graph from 01September 2019. Note that left side vertical scale now has a maximum of 2800MW, so the fall, while much larger, looks a similar size to the one above. Again I have set the marker at the high point of 2744MW, and over the next hour, it fell by 880MW to that low point. That 880MW is the loss of around 470 individual wind towers, and around 8 or more wind plants. Again, notice the frequency of the intermittency, as it rises, then falls, then spikes up and down, and then falls away over a longer period. There is no way to predict when it is going to happen, so again, grid control becomes so much harder.

This last image for the one hour time frame is again closer to the end of the time I was doing this data, 09May 2020. Again, here note the left side vertical scale change and now the graph has a maximum of 4400MW, and as this is from May of 2020, then overall Nameplate has increased that 2800MW or more, so more power is being generated, and again, even though the scale makes that fall look similar to the other two, this loss of power is the greatest at 980MW, from that marked peak at 4089MW. That loss of 980MW is around 500 or more individual wind towers, and probably eight or nine wind plants. And again, as you can see, the power goes up and down across the whole day, and as the scale is larger, then those falls are greater. Here, again, I have added the synoptic chart for this same day and as you can see, in that area of the most wind plants, there is a larger number of isobars, hence higher wind in that area, and because of that, the upper level wind speed was reached turning off those wind towers in that area.

So here, in this one hour time frame period, there has been an increase in Nameplate in that area where the most plants are located, and the falls have become greater, and shorter in the duration for that size of the fall in power.

Power loss between one and three hours

In this group of power losses, there were 52 separate entries. The largest loss was 1780MW in three hours, and there were 15 occasions where the loss was greater than 1250MW. The quickest fall was 1280MW in an hour and 50 minutes.

This image also dates from around the start date, 27May 2018. I placed the marker at the high point there of 2435MW and there was a loss of power of 1140MW over the next two hours and 50 minutes. Nearly all of that loss was in South Australia, and I can say that with relative confidence, and here, I draw your attention to the coloured lines towards the bottom of the graph. That upper one, in purple shows the output of Macarthur Wind Plant, the largest in Australia with a Nameplate of 420MW, and while the black line above shows a marked power loss, Macarthur is still generating at a high level, and that Macarthur plant is in Central Western Victoria, so while the wind is within parameters at Macarthur, it is outside of those parameters  elsewhere. That loss of 1140MW is around 600 individual wind towers going off line, and is also around ten or so wind plants.

This second image is dated 02October 2019. The marker is again at the top of the power generation, set at 3260MW. Again. note here the scale change of that left side vertical axis, with the upper mark now at 3200MW. Here the power loss was 1280MW over two hours and five minutes. Now, while the first image indicated wind plants in South Australia, this shows a power loss in Victorian plants. Again, look at the coloured lines below, and that upper one is again for Macarthur, which has now changed colours as it moves across the page of the list of wind plants under the graph, as more wind plants were added to the grid. You can see here that as the black line total power falls, so does Macarthur, so this graph indicates Victorian plants. That 1280MW loss is almost 700 wind towers and up to a dozen wind plants. You can see that it is still quite intermittent after this large fall for the next eight hours and it then falls away even further.

This third image is again a more recent one, dated 30June 2020, a repeat of the third image from the first group above, only this was later in the day. Again, note the scale change, with the maximum now at 4000MW, again a s more wind plants were added, shown again by the colour change for Macarthur below, now purple in colour. The marker is set at the upper power level of 3421MW, and there was a fall in power of 1210MW over the next two hours and 25 minutes. I won’t show the synoptic chart here again, as it is the same image as the one in that first group above, as they only archive one synoptic chart each day, at that 10AM time slot, so again, you can see that the wind is still relatively strong in that area where most of those wind plants are located.

So, yet again in this longer of the short time frame groupings we see falls in power generation that have become larger and of shorter time duration with the increase in the number of wind plants and the overall total Nameplate. This is the opposite of what we are told when it comes to wind plants, that the intermittency will become less of a problem with the construction of more of them. What is happening in fact is that the intermittency problem is getting worse, the frequency of occurrences is increasing and the swings between low and high are getting higher. It has made the task of operating the grid in a smooth manner so much harder.

Supporters of wind power have often said that coal fired power is not reliable, because every so often, a Unit of 500MW might go off line. Here we have the same thing happening with wind power, only on a scale much more frequently.

The three tables for each of those time frames are at the links shown below. Each Table is a pdf document. When you click on the link, the table will open in a new window.

(a) Loss of power in less than one hour.

(b) Loss of power in one hour.

(c) Loss of power between one and three hours.

The first column is the date of the fall in power. The second column is the size of that fall in MegaWatts (MW) and the third column is the duration of that drop in power from when it started to when it settled at the new lower level.

Link to Part Three – Wind Power Generation Intermittency – It’s Worse Than You Think It Is – Part Three


Anton Lang uses the screen name of TonyfromOz, and he writes at this site, PA Pundits International on topics related to electrical power generation, from all sources, concentrating mainly on Renewable Power, and how the two most favoured methods of renewable power generation, Wind Power and all versions of Solar Power, fail comprehensively to deliver levels of power required to replace traditional power generation. His Bio is at this link.