Balancing Act: How Can We Deal with Variability?

I'm in London, looking at the website of the National Grid — owner and operator of the transmission grid for England and Wales — reading the real-time electricity demand figures at noon on this mid-September day: 40,434 MW, close to average. This time yesterday it was 41,482 MW, and I could see that over 8000 MW were being supplied from the north of England to the south. Today that figure is less than 5000 MW, and I see that England is still exporting about 400 MW to France and the Netherlands.

In this country, which uses relatively little air conditioning, demand is likely to hit a low of around 25 GW in the summer wee hours (assuming no heatwave), and a high of up to 60 GW on a winter evening. On a winter workday morning, demand can step up from 35 GW to 47 GW between 6am and 8am — an increase, or "slew," of 12 GW over 120 minutes, so averaging 100 MW per minute. Somehow it works — the nation's kettles boil and toasters toast, and office computers power up daily.

The work of TSOs (transmission service operators) throughout Europe, and their equivalents elsewhere, is fascinating: each conducting its own orchestra of power generating options, then importing or exporting some extra volume, to match fluctuating demand. Grid operators have practically no control over electricity demand. Their job is to respond, yet without ever being 100 percent certain what demand will be. Of course, consumption patterns have been recorded and analysed, so the TSOs know broadly what to expect, and when. In Britain the National Grid keeps a careful eye on the TV schedules; the end of a popular program can account for a surge in demand of over 1,000 MW as the nation rushes to boil kettles for tea, while a pause in live coverage of this year's Royal Wedding led to a remarkable 2,400 MW of kettle boiling.

However, the record remains the 2,800 MW surge at the end of the penalty shoot-out after England's World Cup semi-final against West Germany in 1990.

The forecasting of demand has become a sophisticated business, especially as it also underlies the prices at which electricity is bought and sold. Already highly skilled at dealing with routine and exceptional fluctuations in electric power demand, the TSOs have coped well with accommodating high percentages of variable renewables — mostly wind — on the grids, especially as long- and short-term forecasting of wind has become increasingly accurate. It might even be argued that variable wind power is no more of a problem than inflexible "baseload" generators such as coal-fired plants and nuclear, which maintain a relatively steady output regardless of demand, and regardless of output from renewable plants. (In August however, Vattenfall announced it was commencing R&D to look at the impacts of reducing outputs at times when load is low — for example, at night; France is already balancing its nuclear power output.)

The truly "flexible friend" of the TSO is large hydro (natural and pumped storage) which can be brought online, at scale, extremely rapidly. Natural gas plants are very responsive, offering the power generation system flexibility. To meet real peaks of demand, a whole range of options that normally sit in semi-retirement are brought into play, and these are generally the least efficient, least clean, and most expensive options. It makes more sense to balance using interconnections between the different national/regional transmission systems. Back in 2002, the EU Council set the target for all EU Member States to have electricity interconnections equivalent to at least 10 percent of their installed production capacity by 2005. However, by 2010 nine Member States had still not met this target (including the U.K., where total interconnection amounts to some 1.5 GW, rather than the 8.5 GW minimum that its installed capacity of some 85 GW would require.)

So, given the kinds of demand variability that TSOs routinely encounter, their existing "toolkit" of management options enables them to deal well with current levels of variable renewables, particularly wind power. There is the occasional exception (but all generation technologies present the grid with exceptional circumstances from time to time). The flexibility already built into the system has enabled countries such as Germany, Denmark and Spain to accommodate high percentages of variable renewable generation.

But what about stepping up to the next level? Does the power system need an expanded management toolkit, and if so, which tools will be most effective? There are two challenges to meet, as Professor David MacKay writes in his book Sustainable energy — without the hot air on the one hand short-term fluctuations; on the other, long-term lulls. It can happen that high demand days coincide with days of little wind, for example. How is a system with very high levels of integration to deal with that?

Alstom's upgraded GT26 combined cycle gas turbine achieves an efficiency of over 61% and can deliver more than 500 MW (Source: Alstom)

Plans to expand renewables are being rolled out across, and beyond, Europe. The UK is developing offshore wind power at a rapid pace, and the National Grid (which also — as NETSO — operates the Scottish grid) is anticipating the need to handle 32–33 GW of wind power capacity by 2020. Across the North Sea, Germany plans to phase out its nuclear generation altogether by 2022, and increase its reliance on renewables. Revisions to Germany's Renewable Energy Sources Act for 2012, which were approved in summer 2011, set a requirement for at least 80 percent of the nation's electricity to come from renewable sources by 2050. This will be achieved incrementally, with a target of between 35 and 40 percent of supply within the coming decade. Not all that will be from variable sources, and much will be distributed power, being fed into the distribution grid but nonetheless having implications for the transmission network.

So what options are on offer? There's upgrading of the grid to offer new connections; making the grid "smarter" and using demand response; there's increasing the flexibility and efficiency of non-renewable generation to work alongside variable renewables, and storage is still much discussed, particularly in relation to fleets of electric vehicles (storage won't be covered here other than in the context of pumped storage).

Transmission Options

As established earlier, the U.K. contains the least well interconnected of Europe's grids, yet is able to respond to big swings in demand. To get a sense of the impact on the system of high levels of wind power anticipated in Britain by 2020, MacKay calculates a wind scenario: "If we scale British wind power up to a capacity of 33 GW (so that it delivers 10 GW on average), we can expect to have occasional slew rates [rate of increase or decrease in output] of 3,700 MW/h. So we need to be able to either power up replacements for wind at a rate of 3.7 GW per hour... or we need to be able to suddenly turn down our demand at a rate of 3.7 GW per hour." As established earlier, such rates are indeed in line with the 12 GW routinely delivered in Britain between 6am and 8am on winter weekday mornings. Nonetheless, the National Grid's 2030 accelerated growth scenario suggests a peak demand in the British market of 56 GW, with 134 GW of installed generation (of which 50 GW would be variable wind generation). With much of that new wind installed offshore, a substantially upgraded transmission system designed for the new circumstances is required.

Denmark's interconnections with Norway and Germany have been an essential ingredient in the success of its wind power sector. The key role of the international transmission grid in enabling Europe's ongoing development of renewably generated electricity is clear, and the subject of an offshore "Supergrid" to connect North Sea countries with one another and with Norway's hydro resource (and potentially with a future Mediterranean solar resource) has been discussed in recent issues of REW (see volume 14/2, March-April 2011).

TSOs are pressing for better international connections: for example, in submissions made to a Committee of British Members of Parliament investigating the potential for a "Supergrid" the National Grid wrote that "Greater electricity interconnection represents a vital part of the UK's low carbon economy. The optimal level of interconnection must be debated and will depend on the generation mix and demand side factors in the UK and mainland Europe. However it is likely that, for the U.K., somewhere around 10-15 GW of interconnection would enable the UK to transition to a low carbon energy mix in an affordable and secure manner." In another submission to the Committee, Daniel Dobbeni of the European Network of Transmission System Operators (ENTSO-E) agreed that without new offshore grid transmission assets "Member States will not achieve their [renewable energy] target. You cannot have one, the new energy mix, without the other, more transmission capacity."

High levels of renewable generation also make demands on the existing grid — demands that it was not designed to cope with, such as transporting large amounts of wind power within a country, or coping with — as in Germany — big numbers of solar modules all feeding current into the grid (most likely the distribution grid, which was not designed for two-way traffic).

Demand response

The field of demand-side management is becoming increasingly sophisticated. At one end of the spectrum, consumers and building managers are being encouraged to purchase more efficient devices and to use them efficiently, as basic good practice. At the other end comes the smart grid, with projects under development from groups such as, in Germany, the Fraunhofer Network on Smart Grids.

One aspect of the smart grid is already going commercial: demand response. EnerNOC, a U.S. business that is now becoming established in the U.K., is a prominent example. Alongside its other energy efficiency offerings, the company's demand response formula (DemandSMART) entails working with businesses and institutions (a supermarket chain, for example, or a university or manufacturing plant) to establish how they could cut power consumption if called upon to do so (up to, say, 50 fewer hours per year) — for example by switching off the power supply to a cold store (which can maintain safe temperatures for hours without being boosted), or by cutting a percentage of lighting, or delaying a power-hungry process until later.

The client then agrees with EnerNOC on a rate that it will be paid for making this cut, and joins the demand-response pool.When grid managers anticipate peaks in demand, they can call on EnerNOC to deliver significant cuts in demand. The fully automated system can do this within a few minutes. As of September 2011, EnerNOC had more than 7,000 MW under management across approximately 11,150 sites. Its website tells power companies that demand response can provide "capacity that's cleaner, faster, and more cost-effective than traditional peaking power plants."

During a heatwave in North America on 22 July, when grid operators were experiencing record demand peaks and prices had shot up to 10 times the 2011 average — US$530/MWh — EnerNOC provided over 1,200 MW of demand response across several U.S. states and Ontario, effectively delivering "negawatts" into the system.

However, EnerNOC is not focusing exclusively on delivering demand response during demand peaks. It is also starting to offer its service specifically to balance variable renewables. In February, it announced an arrangement (again in the U.S.) to supply the Bonneville Power Administration (BPA) with demand response capacity to facilitate grid stability as BPA integrates more wind-powered generation into its transmission system. EnerNOC says this is the first project of its kind to draw upon demand response capacity from commercial and industrial sites to balance both increases and decreases in supply from renewable resources, as well as traditional generation, responding if necessary to changing system needs in less than 10 minutes. In the UK, the company has now started working with Low Carbon London to sign up members to a demand-response scheme that can help with load balancing and variability of renewables.

Natural gas for rapid response

Natural gas has proven credentials as a flexible power generating option, though high efficiency levels are only achieved in CHP mode, or combined cycle. The German energy agency, DENA, has acknowledged the importance of including efficient and responsive fossil fuel generation in future plans. This year has seen some leading gas turbine manufacturers promoting new products as the ideal partner to wind or large solar in both the 50 Hz and 60 Hz markets.

GE, for instance, launched its new FlexEfficiency 50 Combined Cycle Power Plant in May. Explaining that power plants generally provide either flexibility or high efficiency, GE believes this power plant will deliver an unprecedented combination of both. The FlexEfficiency 50 Combined Cycle Power Plant is rated at 510 MW and offers fuel efficiency greater than 61 percent. It is designed to ramp up at a rate of more than 50 MW per minute, "twice the rate of today's industry benchmarks," it says.

GE's FlexEfficiency 50 Combined Cycle Power Plant is rated at 510 MW and offers fuel efficiency greater than 61 percent (Source: GE)

Imagine if the 3.6 GW/hour ramp up to compensate for falling wind output (in the scenario envisaged by Mackay) were to be delivered solely by this technology: then eight of these plants would together ramp up at over 400 MW per minute, or 4 GW in just 10 minutes.

GE says that the ability to ramp up and down in response to fluctuations in wind and solar power "is essential if renewable power is going to cost-effectively integrate into power grids around the world on a large scale." Ricardo Cordoba, president of GE Energy for Western Europe and North Africa, says, "This innovation can have a dramatic effect on CO2 emissions and offers a nimble, efficient and cost-effective way for us to help EU countries in their pursuit of 20-20-20 energy goals."

Meanwhile, competitor Alstom is promoting its upgraded GT26 gas turbine to enable integration of intermittent renewable sources of energy. In a GT26 combined cycle achieves an efficiency of over 61 percent and can deliver more than 350 MW to the grid in less than 15 minutes. The GT26, in a one-on-one configuration, is capable of producing more than 500 MW of power.

While such a plant could be tied in to a specific wind or solar farm, essentially creating a hybrid plant, it seems more likely that it would work more effectively at transmission system level — but market conditions will play a role in determining that.

European power generation is heavily reliant on natural gas. Used in highly responsive and efficient turbines, it would appear to be the perfect partner for renewables and a bridge to a low-carbon future. Yet there has to be a caveat — the carbon footprint of the natural gas needs to be taken into account.

Storage

It is often claimed that electricity storage is a prerequisite for high levels of integration of renewable energy as it "absorbs" excess generation and delivers it when needed. There's no means of storing electricity, of course — it needs to be converted into mechanical energy (pumped hydropower, compressed air, flywheel), chemical energy (battery, hydrogen), or heat. That stored energy is then used to generate electricity when needed, incurring some losses each time the conversion is made. De facto storage can also be implemented via the smart grid, effectively shifting the load from mainly heating or cooling procedures to boost demand.

Also being investigated is the opportunity to use an increasingly large fleet of electric or hybrid vehicles to absorb excess generation, recharging their batteries at these times (and even discharging them temporarily). Currently, the best established form of storage is pumped hydro, and it appears unlikely that any other measure will be applied at scale in the near-to-medium future. A study by the German energy agency DENA backs this up.

Following a 2005 study by DENA to investigate the grid upgrading needed to reach the target of generating 20 percent of all energy from renewable sources by 2015, measures were put in place to enhance specific power lines, plus the requirement to add 850 km of new routes within the German transmission grid by 2015. A second study, the DENA Grid Study II, has investigated the additional measures that will be needed (up to 2020 with an outlook to 2025), to fully integrate a level of 39 percent electricity from renewables into Germany's power grid while guaranteeing the security of supply and taking into account the effects of the liberalised European energy market.

As well as looking at additional grid-based solutions and demand response, the study examined storage. Implying the use of hydro storage, it states that "in particular, increased use of energy storage capacity in Southern Germany, the Alpine countries and possibly in Scandinavia means that the grid infrastructure must be expanded." However, it concluded that "construction of additional compressed air and hydrogen storage facilities will not occur as driven by the market by 2020 due to economic aspects and the existing market regulations, in spite of the increasing volatility of generation and the associated electricity price fluctuations. Even with an additional consideration of free storage of generation capacity which could otherwise not be integrated, these storage facilities proved uneconomical."

So how do we best deal with variability? Integrated energy policy and planning is probably the answer, rather than any specific technological solution, and the answers will vary depending on how far we want to go with renewables. Improving the grid is at the heart of any solution, along with big advances in demand-side management. More efficient and flexible non-renewable solutions have an important part to play, but we need to keep a careful eye on the fuel that they use, or apparent efficiencies and CO2 reductions might become meaningless.