Reduction of CO2 emissions due to wind energy in the Irish electricity system.
A discussion of the SEAI report 2012
Fred Udo and Patrice d’Oultremont
January 25 2016
“One of the primary reasons we don't seek counsel from the wise people around us is that we already know what we are going to hear - and we just don't want to hear it. “[1]
Abstract.
This paper analyses the data presented in the annual report 2012 of the Sustainable Energy Authority of Ireland.
This report admits, that inserting wind energy into the Irish grid leads to appreciable efficiency losses in the fossil fuel generators. A closer analysis reveals, that the fuel savings due to insertion of 15,2% wind energy are less than 40% of the expected savings. Recently the results are confirmed by an independent analysis.
1.Introduction.
The addition of large amounts of wind-generated electricity to the electrical grid causes a loss of efficiency in conventional power stations, as they have to ramp up and down at less than rated output. The amount of these losses is related to the magnitude of the wind contribution, but the losses are increasing faster than proportional to the percentage of wind in the system. The European regulations tell us, that one kilowatt-hour (kWh) of wind electricity replaces one kWh conventional electricity. This is often called the "Nominal Saving", which is to say that, in any particular grid, X% wind electricity production displaces X% of conventional electricity production, irrespective of X.
An earlier paper by Udo, Le Pair and de Groot [2], entirely based on empirical data, showed that the injection of 25% wind electricity in a grid (with no hydropower) will save only 38% of the nominal fuel saving.
It is worrisome, that the EU authorities and all concerned Member States seem to deny this reality and continue to promote the rapid deployment of wind electricity without facing the technical and financial consequences of this major limitation to wind energy use.
As a striking example of this situation, the Dutch government, having discarded the findings of reference 2 formally told the authors that only “models validated by empirical data” could actually quantify the effect of wind power on electrical grids [3]. The present article meets that requirement by calculating the reduction of CO2 emission intensity due to wind in the Irish grid (year 2012) based on model calculations published by the Sustainable Energy Authority of Ireland (SEAI).
The results presented here confirm and reinforce the conclusions drawn in reference 2.
The results of this work have recently been confirmed by an article written by Owen Martin.
He discusses the reduction of fuel use in the Irish electricity system due to the wontribution of wind energy.
2. Discussion of the SEAI calculation.
The SEAI presents in its Annual Report 2012 the fuel and CO2 emission savings with the Irish wind turbines []. The calculation uses the PLEXOS modeling tool, which takes into account many parameters of the grid and its generators, so this calculation describes the actual situation rather closely.
A serious shortcoming of the SEAI report is, that CO2 emissions are calculated from the static efficiency curves of fossil fuel generators.
SEAI uses the same calculation as the grid operator Eirgrid used to publish CO2 emission data every 15 minutes [4].
The report mentions cycling, but does not attempt to translate it into extra fuel consumption.
The SEAI report describes three scenarios, of which only the first two (with and without wind) are used in our analysis.
The Base Model (labelled BM) is a model of the actual Irish grid in 2012 with 80,4% fossil, 15,3% wind en 4,3% hydro and wood.
Scenario 2, labelled NWM1 here, is a model in which wind is replaced by 238 MW extra fossil fuel capacity. With this extra fossil capacity, the grid generates the same annual power (32 280 GWh) as in scenario BM.
Two important features of NWM1 must be emphasized (see Table 2 and figure 13 of the report):
-
• The extra 238 MW of fossil generation are provided by OCGT plants that have a primary energy conversion
-
• factor even lower than coal fired plants.
•The utilisation factor of these OCGT plants is increased from 18% in BM to 28% in NWM1.
The results reported by SEAI are thus obtained by comparing an actual configuration of the grid including wind with a configuration without wind, that artificially emits more CO2 than needed.
The addition of OCGT generation diminishes also another phenomenon:
The removal of the wind contribution improves the fuel efficiency of the Combined Cycle Gas Turbines (CCGT), as the utilisation factor of these efficient machines goes up.
NWM1 is constructed in such a way, that the load on the CCGT’s increases only a few per cent.
In spite of the use of this far from optimal reference case (NWM1) the report states that:
(see the caption of figure 15, page 29) :
Figure 15 shows that the levels of wind on the system in 2012 had the effect of reducing the total electricity system emissions intensity by 12% relative to a case where all wind generation was removed.
Here the SEAI admits, that the CO2 saving by inserting 15,2% wind is: 12%/15,2% = 79% of the nominal fuel saving.
The SEAI report is to our knowledge, the first document issued by a government body declaring that insertion of large amounts of wind energy in the distribution grid causes substantial efficiency losses in the existing generator system.
They might embellish the facts, but at least they do not conceal them completely.
Reality is still substantially worse as we shall see in the next sections.
3. Modifications made to the SEAI model NWM1.
3A. SEAI data - Year 2012
All calculations in this analysis make use of the following SEAI data :
Total demand (TD) 34 562 GWh
Total import (TI) 2 282 GWh
Total IR production (T = TD - TI)32 280 GWh
Total wind production (W) 5 240 GWh
CO2 intensity in BM 500 g/kWh
CO2 intensity in NWM1 570 g/kWh
CO2 intensity reduction by wind12%
Due to the correction for imports the contribution of wind to the total Irish production is 16,2% instead of 15,3%.
3B. Introduction of scenario NWM2 and corresponding CO2 calculation.
Section 2 shows, that in NWM1 all wind capacity (2109 MW) has been removed and replaced by 238 MW of fossil (gas) capacity.
This OCGT + CCGT capacity mix in NWM1 is suboptimal in two ways:
•CCGT units remain under used
•238 MW superfluous OCGT generation is added and OCGT use is inflated to 28%.
•
Wherefrom the introduction of an alternative scenario (NWM2) with:
•The total production of the gas units is kept equal to the production in NWM1
•The CCGT and OCGT units are the same as in BM.
•The OCGT utilisation factor is equal to 18% as in BM
•The utilisation factor of CCGT is calculated to be 66% instead of 62% in NWM1
•The electricity efficiency of CCGT and OCGT in BM is taken from J.B. Wheatley []
The NWM2 model is constructed so as to fulfil all the constraints given in the report about production figures and utilisation factors of the gas generation. In this way the corrections to the PLEXOS calculations in NWM1 remain small. Doing so, the following quantities emerge:
•A shift of 4,4% of gas generation from OCGT to CCGT generates a saving of CO2 emissions of 29 g/kWh.
•A higher use of CCGT causes a saving of CO2 emissions of 10 g/kWh due to increased CCGT efficiency.
As gas generates 60% of the fossil production in NWM1 as in NWM2, the total of 39 gCO2/kWh corresponds to an effective reduction of the CO2 intensity by 0,6 x 39 = 23.4 g/kWh.
Accordingly:
•The CO2 intensity of NWM2 equals to (570 – 23.4) g/kWh = 547 g/kWh.
•The difference in CO2 intensity between NWM2 and BM is 47 g/kWh
•The displacement of CO2 intensity by wind is 47 g/kWh / 547 g/kWh = 8,5%
The Excel sheet [7] gives the details of the calculation.
It follows from 3A and 3B, that the effective reduction of CO2 emissions by inserting 16,2% wind in the Irish electricity supply is: 8,5% / 16,2% = 53% of the nominal saving.
Given the limitations in the calculation of the CO2 emissions this number has to be considered as an upper bound.
3C. The CO2 emission calculation of Eirgrid.
The analysis presented above relies on the CO2 emission calculation developed by Eirgrid, the Irish grid operator. This calculation neglects the increased fuel consumption due to ramping of the generators. This ramping is necessary to accommodate both the daily variation of the demand and the fluctuations in the amount of wind energy produced. It is the so called load following of the power stations.
The definition originally taken from the website of Eirgrid is:
“The rate of carbon emissions is calculated in real time by using the generators MW output, the individual heat rate curves for each power station and the calorific values for each type of fuel used.”
This defines the CO2 emissions as the emissions of a generator in a steady state, thus the extra fuel consumption due to ramping is not included. The effect of ramping on the CO2 emissions can be seen by comparing CO2 emissions calculated from fuel inputs with the calculated ones published every 15 minutes on the Eigrid website.
The 15’ data disappeared recently from the Eirgrid website, so new data cannot be added anymore.
Figure 1
The 15 minutes CO2 data are averaged over intervals of a year and plotted as the red dots in the graph. The blue line is calculated from the fuel input data corrected for imports.
The exhibit shows the two sets of data for the years 2011 to 2014.
The difference between the two curves is 7%, so this is the effect of load following/ ramping of the variations in the demand due to daily variations of demand together with variations in the production of wind power.
The dip in the curves in 2011 is due to the contribution of two modern Combined Cycle Gas Turbines, which were installed in 2010. In 2012 they were hardly used because the price of coal dropped.
The red arrow indicates the “Nominal Saving” due to 16,2% of wind.
The effect of 16% wind is simply not visible.
The majority of the ramping is due to the daily variation of the demand. The SEAI report states that the presence of wind energy in the system increases the load following by 30%. The total extra fuel use due to ramping was determined from the preceding graph to be 7%. This implies, that wind increases CO2 emissions due to load following by 1,8% above the CO2 emission calculated by Eirgrid. These 1,8% count in full in the comparison between the system with- or without wind, because the daily fluctuations of the demand are equal in the two cases.
The conclusion is, that ramping of the fossil fuel generators due to wind decreases the CO2 emission saving from 8,5% to 6,7%. The division of the two numbers becomes:
6,7% / 16,2% = 0,41.
The efficiency for replacing fossil fuel by wind energy is 41% in the All Island system.
4.Discussion of the “self-energy” contribution to the CO2 intensity.
The CO2 savings due to insertion of wind energy calculated from operational data are still too high, as they neglect the energy needed to build this additional power generation system, consisting of numerous turbines plus the associated power transport lines.
The energy required to build this extra system equals to about 10% of the total electricity production of the turbines [8]. This means, that the turbines are only effective during 90% of the running time. The self-energy reduces the effective fuel saving from 6,7% to 6,0%.
The CO2 emission saving by 16,2% wind energy in the Irish grid becomes
6,0% / 16,2% = 37% of the nominal CO2 saving.
5. Conclusions
•SEAI admits that the fuel saving due to the insertion of 16,2% wind power in the grid is appreciably less than expected according to EU and Belgian rules.
•SEAI minimizes the efficiency losses by inflating the use of OCGT in the reference case. The resulting efficiency of wind energy is 79% of the nominal saving.
•Correcting the SEAI reference model by limiting the role of OCGT reduces the fuel saving by wind to 53% of the nominal saving.
•The ramping of back up generators adds a correction to the CO2 emission calculation of Eirgrid. This brings the fuel saving due to wind down to 41%.
•With the additional correction for the self-energy we find a CO2 saving of 37% of the nominal value at a share of wind = 16,2%.
Larger losses are expected at higher contributions of wind. Although some degree of uncertainty is attached to the corrections discussed in the present paper, the harsh reality is, that after analysis of the SEAI report one must conclude that the fuel saving due to wind energy in the Irish grid is less than 40% of the expected saving.
References
-
1. Andy Stanley, 1958 -, American Christian Pastor and Author
2. F. Udo, C. Le Pair, K. de Groot, A.M. Verkooijen en C van den Berg
Using Wind Energy to save fuel and reduce CO2 emissions.
3.The letter of the minister (in Dutch):
http://www.tweedekamer.nl/kamerstukken/brieven_regering/detail?id=2014Z20731&di=2014D41934
4. Quantifying Ireland’s Fuel and CO2 Emissions Savings from Renewable Electricity in 2012
Energy Modeling Group May 2014
http://www.seai.ie/Publications/Statistics_Publications/
Recently all links to this publication disappeared from the website. It can be found via the search function
Eirgrid recently modified this service. Now only data from one day at the time can be downloaded.
6.J.B. Wheatley - Quantifying CO2 savings from windpower;
Energy Policy, 2013, vol. 63, issue C, pages 89-96.
7 Details of the calculations are in an Excel worksheet. This sheet is available at:
https://www.dropbox.com/s/juozyojpilta00z/Excel%20sheet.xlsx?dl=0
8 F. Udo - Building windturbines costs more energy than you think
http://www.clepair.net/Udo201303payback.html
http://fredudo.home.xs4all.nl/Zwaaipalen/12._Building_windturbines_2013.htm