Tuesday, 2 December 2014

SEAI's Quantifying Savings from Renewables Report and the Impact of Wind on Reserve Requirements

Earlier this year, SEAI issued a report, which concluded that there was a saving of € 177 million in fossil fuel due to wind generation in 2012, titled "Quantifying Ireland's Fuel and CO2 Savings from Renewables". The report has a number of flaws. Here is a sample of some of those flaws :

  1.  For the "No Wind" model, it assumes that the wind capacity would be replaced by 180MW of OCGT. OCGT is the most inefficient form of gas generation and would therefore result in larger savings than a CCGT plant in their "2012 Base Model" with wind. They could have replaced this capacity with biomass for example, like what is happening in Edenderry Power Plant which would have resulted in much less emissions, thereby decreasing their final savings figure. It also could have been replaced by the replacement of HFO plant with CCGT plant, as in Great Island, which in that case, resulted in a net increase in capacity of about 200MW.
  2. The report states that "The total quantity of ramping in coal generation is higher with renewable electricity on the system. In contrast, gas CCGTs vary their output by a lesser amount with renewable electricity on the system." This is utter nonsense. You can have a look at the ramping profiles of CCGT plant here [Aghada] and here [Dublin Bay]. With more wind in the system, the ramping of gas plants is multiples of what it is in a system with little or no wind (as the only variable is demand).
  3. Number 2 also disproves the claim made that "Wind generation variability in 2012 was less than electricity demand variability". If this were the case, then the graphs linked above, would show the opposite, i.e. more ramping in the no wind or little wind periods or at least similar ramping but the difference is huge. Aghada CCGT's load profile for example has a flat surface in 2009 when there was only small amounts of wind compared to the jagged surface of 2013 when there was much more wind.
  4. The assumptions made on the inefficiencies of generators at different loads - Gas is responsible for 78% of the quantity of savings in the report. Most of that is made up of CCGT. To allow 200MW of wind in the system, a typical 400MW CCGT gas plant must drop to 200MW. This results in increased CO2 emissions of about 0.05- 0.07 tonnes per MWh, an increase of 20%. To allow more wind in, the efficiency drops further and CO2 emissions increase at a faster rate as explained here. So to allow 4,094GWh of wind in to create the assumed savings in the report, CO2 emissions would rise even more than 0.07 tonnes at times.  but they assume that "With the actual level of renewables on the system in 2012 (the Base Model case), the CO2 emissions intensity of fossil-fuel generators is 5% higher than in the No Wind Scenario." In reality, this would be higher, thereby eating into their savings figure due to wind.  
  5. Likewise there is an efficiency drop of 8% from 58% to 50% in the above scenario in Number 4. To allow more wind in, the efficiency drops further and at a faster rate. So to allow 4,094GWh of wind in to create the assumed savings in the report, the efficiency would most likely fall below 50% at times and even to 40% (a drop in efficiency of 18%). So the average efficiency of the CCGT plant is somewhere between 48%-52% (48% being a conservative base figure). This is a drop in efficiency of between 8% and 10%.  We are not told the efficiency figures they use. If higher efficiencies are assumed, then their fossil fuel saving figure is inaccurate.
  6. Perhaps the biggest flaw is their assumptions on reserve - and more importantly, replacement reserves, which increases when large amounts of wind are allowed into the system. A practical example is shown in a previous blog post here. You can see that CCGT plant had to be kept ticking over (just like a car left on and parked outside your house) to step in instantly as the wind dropped off. As this plant stepped in, replacement reserve needs to be made available in the event of a forced outage of another plant or indeed, a further drop in wind. Therefore, there tends to be a peak demand for reserves and replacement reserves at high levels of wind penetration. Let's look into this important matter in more detail:

The Impact of Wind on Reserve Requirements

 In a 2007 report, prepared for Eirgrid, titled "Wind Variability Management Studies (P.Meibom et al)" , Danish scientists and University researchers concluded that:
 "Generally, the demand for replacement reserves increases with
increasing wind power capacity installed.
The occurrence of high demands for replacement reserves is
mainly driven by a high number of simultaneous forced outages that happens
simultaneously to relatively high wind power or load forecast errors. The value of these peaks tends to increase with increasing wind power capacity installed."
In another study by R. Doherty and M. O’Malley of UCD Dept of Electronic and Electrical Engineering, titled “A new approach to quantify reserve demand in systems with significant installed wind capacity,” it was stated that :

The methodology is applied to a model of the all Ireland electricity system, and results show that as wind power capacity increases, the system must increase the amount of reserve carried or face a measurable decrease in reliability [i.e. increase the risks of a blackout - bloggers note].
[Note that in these reports, they assume that increased wind capacity (i.e. building more wind farms) will result in higher wind penetration in the system (i.e. higher levels of wind relative to demand)]

SEAI took no account in their study of the increased demand for reserves during periods of high wind penetration i.e. more plant burning fuel behind the wind. Instead, they took a fixed amount of reserves during the year based on the minimum amount of reserve permitted by the regulations :

Primary and secondary operating reserves are calculated dynamically in the model for each period based on 75% of the largest unit running at that time in RoI and the largest unit running at that time in NI. Tertiary reserve requirements are included as fixed quantities based on the largest single electricity in-feed. These were 425 MW in NI and 480 MW in RoI for the first 9 months of 2012 and 500 MW for the last 3 months of 2012 [SEAI].

Lets have a look at the Regulations :

That's right, this is the Minimum amount of reserve that must be maintained. It is then at the discretion of the TSO (Eirgrid) to increase this depending on circumstances.

Reserve requirements are not influenced by wind generation or other renewable electricity generators at current levels of installed capacity. [SEAI]

However, it was still possible that there could have been high levels of wind penetration in 2012 (which would have resulted in increased demand for reserves).
The All-Island grid study showed that additional reserve requirement in hypothetical 2020 scenarios is related to the amount of wind installed but that the largest contributing factor remains the loss of the largest conventional unit [SEAI].
The study SEAI are referring to is the same report by Danish scientists mentioned above. But what the study actually says in relation to this is :

There should be enough spinning reserves to cover an outage of the largest unit in combination with a fast decrease of the current wind power production. However, the capacity of the largest online unit changes dynamically. (Doherty and O’Malley 2005) further demonstrate the dependency of the demand for TR1 [Reserve Type 1] from the installed wind power capacity. 
SEAI also argue that the system in Ireland is flexible enough to accommodate fluctuations in wind anyway without the need for additional reserve. But as you can see in the example previously mentioned (see here), you can see where two CCGT plants, namely Huntstown and Great Island, stepped in from reserve to cover the loss in wind power on the 3rd November. None of the online plants were flexible enough or had enough spare capacity to provide this cover.

We can look at the impact of wind on reserve with a very simple example. Say, that we end up with a system, where 100% of the wind can be allowed in to meet 100% of demand and this occurs for a whole day. By the following day, 20% of the forecasted wind drops off, and then later in the day 50% of the remaining forecasted wind drops off. By the 3rd day, wind is only providing 5% of demand. Now let's pretend that we do what SEAI say we should do and maintain 480MW-500MW of reserve. Remember, Reserve requirements are not influenced by wind generation. As the 20% wind drops off, we will be able to call on the reserve by ramping up the reserve generators to full output to fill the gap. We must now put replacement reserve in place. Again, this is not affected by the wind so 500MW will do fine. But now we come to later in the day where circa 1,500MW wind has dropped off. We use some of our fast acting (but small capacity) units to try to fill the gap along with the other slower replacement reserve. But there is simply not enough of the slower acting, larger capacity reserve ready to step in. Time has run out and we are now facing Blackouts. So it is clear from this example, that you have to increase the minimum reserves to cover nearly every MW of wind in the system, thereby negating most, if not all, of the fossil fuel or CO2 savings (actually potentially increasing them because the back-up plant have to run on low loads). The wind, itself, in effect, becomes the largest in-feed unit. You can get away with lower reserves at lower levels of wind, but you simply can't take this risk at higher levels.

So, there is no debate on this - increasing wind penetration above a certain level, leads to an increase in reserves. I would put the line at 1,000MW in the Irish system, as now we have the equivalent of two large thermal plant that can drop out at any time and in any sequence and is much more likely to drop out than if they were in fact, two thermal plant. The exact amount can be debated but I will now come to a problem with the report, that presents a fundamental issue for those who are now using it to inform energy policy and public debate. It is there, written down exactly as quoted here, on Page 12, and can't be debated or argued over, or fudged, or ignored, or interpreted in some other fashion.

The main problem with this report is that it gets trotted out by SEAI and policymakers as the basis for installing more wind farms. But, like with previous SEAI reports, there is a huge disclaimer in it, that most people who have read it may well have missed. That is why I put it in bold above. But it deserves repeating and repeating again :

Reserve requirements are not influenced by wind generation or other renewable electricity generators at current levels of installed capacity

So I ask the question, why is this report, long outdated and now redundant, since it is based on 2012 levels of wind in the system, trotted out and used as justification for more wind farms when what it really is saying, just like all the other reports mentioned above, (albeit begrudgingly in this case) is that the integration of more wind in the system will require more fossil fuel reserves ?


  1. No surprises there - is current energy policy not based on the 2010 NREAP for which there has been no mid-term review? The DCENR is firmly rooted in the past.

  2. This comment has been removed by the author.

  3. Average demand on an Island is 1gw, but it varies between .4 and 2 gw and is supplied by a 2 gw coal powered station. They need 10 cent per unit sold to survive financially generating at half capacity. 8760 x 1,000,0000 kwh x 10/100 = 876,000,000 euros for used power. There is a 25% fuel saving on unused capacity. Unused capacity must get 876,000,000 - 219,000,000 = 657,000,000 bringing the total to 1533,000,000

    Now say capacity is increased by 1gw gas to balance wind and 1gw wind.Total 4gw. They must get the same money less 25% fuel saving. As only 1gw on average is used, payments are 1533,000,000 +( 657,000,000 x 2) = 2847,000,000,000 when you only need 1533,000,000. This is nearly 2 times the price. Ireland with adequate generating capacity of say 6.2 gw will have 15,1 gw by 2020. It has 10 now, So take out the fuel not used and you can see the huge cost of too much capacity. Remember, 6,1 gw is almost twice average demand. Because wind cannot be relied on and fossil fuel plant is needed for back it up and this must be paid extra if pushed off the grid by wind. Remember all plant carried the same fixed costs and variable costs except fuel, whether used or not. Prices cannot be reduced, except for the tiny fuel price reduction for used plant. Claims that it can are a myth,