This is a guest post by Roger Andrews. Roger is a British born, naturalised American mining consultant who is now semi-retired and lives on the West coast of Mexico where he spends some of his time sitting under a wavy palm tree blogging and drinking tequila.
In a December 2013 article in the Guardian Nigel Williams, head of electricity systems operations at the National Grid, was quoted as saying “I don’t see an upper limit to how much wind we can accommodate (on the grid)”. This was a curious statement for a grid operator to make because there clearly are limits as to how much intermittent wind power can be accommodated on a grid that continuously has to match supply to demand. But what are the limits for the UK National Grid? Here I will attempt to quantify them.
Methods and assumptions:
The basic procedures used were:
- Download the 5-minute grid data for 2013 from Gridwatch
- Apply generation mixes that contain progressively less coal, oil and gas generation and progressively more wind generation to these data.
- Check whether 2013 peak demand would have been met had these generation mixes been in place in 2013 and evaluate other impacts of the generation mix change.
Using the actual 2013 numbers means that I did not have to make assumptions regarding future demand, power imports or how strongly the wind will be blowing X years from now. Running the generation mix cases, however, required the following assumptions regarding power generation, wind penetration and the “blackout threshold”:
- Maximum Generation: According to the National Grid’s 2013/14 Winter Outlook the UK grid is presently capable of generating up to 60GW. (Although it may be optimistic. The demand peak of 56.7GW on January 16, 2013 came close to causing a blackout and with the capacity retirements since then probably would cause one if the same conditions repeated themselves now).
- Generation Breakdown: About 50GW of the 60GW is generated by facilities that can be cycled to a greater or lesser extent to balance supply and demand during peak and off-peak periods (dominantly gas and coal with a small amount of oil, biomass, hydro and pumped hydro). The remaining ~10GW includes nuclear and wind and also power exports/imports, which often do not follow load.
- Replacing Existing Generation: The 50GW of “peaking generation” is progressively replaced by wind power in the generation mix scenarios. The ~10GW of generation from other sources remains the same as in 2013. Increases in installed wind capacity are simulated by factoring 2013 wind generation upwards in proportion.
- Merit Order: Wind generation displaces peaking generation until no more wind power is available, or until demand is met, or until the curtailment threshold is reached.
- Curtailment threshold: Wind generation is commonly curtailed when it exceeds X percent of demand. I can find no data on X for the UK so I have assumed the 50% value used by the Irish grid for the generation mix options unless otherwise specified.
- Spot price curtailments: Wind curtailments caused by short-term changes in electricity spot prices in 2013 will be projected into the generation mix options.
- Blackout threshold: Ofgem states that a 2.75 GW undersupply is the threshold above which “a large shortfall requiring the controlled disconnection of customers” could be expected. There were, however, several periods in December 2013 when the shortfall exceeded 3GW – briefly reaching 3.7GW on December 4th – without apparent ill effects (the shortfalls were caused by unusually large power exports to France and the Netherlands). I have therefore picked 5GW as the threshold above which blackouts will occur.
- Supply/demand balance: The generation mix options balance supply and demand when sufficient power is available.
2013 Gridwatch data:
Figure 1 plots the Gridwatch data for Febuary 2013, a fairly typical winter month, for illustrative purposes:
The first graph shows demand ranging from around 30GW during off-peak hours to over 50GW during peaks (summer demand is approximately 10GW lower). The grid operates with a fairly consistent 0.5 to 1GW undersupply.
The middle graph shows the power generated by peaking facilities. Peak load cycling was performed mostly by gas but with a significant contribution from coal. Contributions from hydro and pumped hydro were minor.
The bottom graph shows the power generated by non-peaking sources, which consist of nuclear, “other”, power imported or exported via interconnections, and wind. The most important contributor was nuclear. Overall more power was imported than exported but exports exceeded imports on a number of occasions.
During February 2013 coal supplied 44.2% of total UK generation, gas 26.8%, nuclear 18.7%, “Other” (hydro, pumped hydro, interconnections and “other”) 5.3% and wind 5.0%.
Now we will look at various options for expanding UK wind power utilization:
Replace peaking generation with enough wind capacity to replace the power lost:
Table 1 summarizes the impacts of progressively replacing peaking generation with as much additional wind capacity as is needed to replace the lost power. Wind generation is estimated by factoring up actual 2013 wind generation in proportion to the increase in installed wind capacity, which averaged ~10GW in 2013. “Blackout days” are defined by a demand shortfall of more than 5GW, as discussed above:
With this option less than 20% of the UK’s electricity can be supplied by wind before the lights begin to go out.
Figure 2 shows the impact of applying the generation mix in which peaking generation is cut from 50 to 35GW and installed wind capacity increased from 10 to 24GW, the point at which blackouts begin to occur, to the February 13 data. The “Other” plot in the generation mix graph shows actual 2013 generation from all sources other than peaking facilities and wind:
February 12th is the only February day with a shortfall exceeding the 5GW threshold – the other five occur in January and March – but there are another five days in February and 19 days in 2013 when the shortfall exceeds Ofgem’s 2.75GW threshold. Obviously the approach of simply replacing peaking generation with an equal amount of wind generation will not allow the large-scale adoption of wind power in the UK.
Replace peaking generation with wind, no limit on added wind capacity:
Table 2 summarizes the impacts of this option:
This option allows the UK to generate 50% of its electricity from wind and still meet 2013 peak winter demand, but 470GW of additional wind capacity is needed to do it, and at this level over 80% of the electricity the wind turbines are capable of generating gets curtailed.
Figure 3 further shows that admitting large amounts of wind power to the grid requires peaking facilities to be cycled at rates that would probably exceed rate-of-load-change limits. And even with 240GW of installed wind capacity there are still three days in 2013 (one in February) with an undersupply exceeding the Ofgem 2.75GW threshold. This option is clearly not viable either.
Import power during peak demand periods:
The UK wind regime is positively correlated with wind regimes elsewhere in Europe, meaning that when the wind is not blowing in the UK it likely will not be turning turbines across the Channel either. In this case everyone will be short of electricity and there will be no surplus power to import.
Store the wind power for re-use:
Battery storage seems to be a long way from large-scale commercialization, and a recent Stanford study concludes that the EORI of battery storage is too low to make it a viable proposition for wind power anyway. CAES, thermal, superconducting magnets and underground hydrogen storage seem equally distant. This leaves pumped hydro as the only alternative.
Estimating pumped hydro storage requirements is a complex exercise and I hesitate to present any firm numbers, but calculations indicate that storage on the order of hundreds of GWh – many times the present UK installed pumped hydro capacity of ~30GWh – would be needed to smooth out the wind power fluctuations in January 2013, the most “unbalanced” month. It is highly unlikely that this much additional pumped hydro storage could be built in the UK within a time-frame short enough to do any good if indeed it could be built at all.
Add wind but keep backup generation:
Under present circumstances this is the only option that allows the large-scale utilization of wind power while at the same time keeping the lights on. Wind capacity can in fact be added indefinitely if there is sufficient backup generation to meet peak demand when the wind is not blowing. As to how much backup is needed, the Table 1 and 2 results indicate that 45GW of peaking generation plus the ~10GW of nuclear and other generation that was on line in 2013 would be required. And with this backup generation in place the lights indeed never go out, but as shown in Table 3 wind generation becomes subject to progressively more curtailment as levels of penetration increase:
Compounding the problem is that at high levels of wind penetration there is a need for very rapid cycling of peaking facilities to balance abrupt changes in wind strength, while in periods when the wind is blowing most of the backup capacity sits idle. These effects are illustrated on Figure 4, which shows the generation mix that results from applying the 283GW of wind capacity contemplated in the DEFRA Level 4 scenario to the February 2013 data at a curtailment threshold of 75%:
The economics of this approach are also problematic. Wind power is effectively “free” (it has no fuel cost) so it makes economic sense to use as much of it as the turbines can generate and the grid can admit. On the other hand, another 200GW of wind capacity would cost over a trillion US dollars at the $5,600 installed capital cost/kw recently estimated by NERL for offshore wind turbines. As shown in Table 4 overall load factors also decline as more wind is added – to 22% with 100GW of additional wind capacity and to only 9% with 300GW of additional wind capacity. At such load factors the generation system would be inefficient to say the least, and while I have made no attempt to estimate generation costs it is reasonable to assume that they would be correspondingly high:
Finally comes the question of the impact of wind power on CO2 emissions. How large a reduction in CO2 emissions would be achieved if, say, half of the UK’s electricity was presently generated from wind? Approximately 75 million tonnes/year, representing about 17% of UK CO2 emissions and 0.2% of global CO2 emissions. It would be offset within about a month by increased CO2 emissions from developing countries at current rates of growth.