By Dave Elliott
Several organizations have formulated proposals for transitioning to 100% renewable energy, nationally or globally. In one of the most recent, developing on their earlier 100% global scenario, US academics Mark Jacobson and Mark Delucchi and their team have spelt out how 139 countries can each generate all the energy they will need from wind, water and solar (WWS) technologies by 2050, in substantial detail.
By Dave Elliott
Germany is pushing rapidly ahead with renewables, aiming to get 80% of its electricity from them by 2050. The nation briefly obtained 78% of its electricity from renewables in the summer. But that was obviously a one-off event. Even so, averaged annually, it’s over 30%: http://energytransition.de/2015/07/renewables-covered-78percent-of-german-electricity/
We are always told that it’s vital to have ‘baseload’ – that is ‘always available’ generation capacity – to meet minimum energy demand. Otherwise the lights would go out! Baseload used to be provided mainly by coal plants, these days it’s also nuclear. Indeed, on summers nights when UK demand drops to 20GW or so, it can mostly be nuclear, plus whatever we are getting from our ~6GW of wind and other renewables. But when wind expands (to maybe 40GW!) and nuclear also expands (to say 20GW), then there will be conflicts over which to turn off (‘curtail’), during those periods, especially if there is also, say, 10GW of tidal on the grid. In which case the concept of baseload starts to look unhelpful – the problem being a potential surplus of electricity, not a shortfall.
To avoid ‘curtailment’ problems, we might store some of the excess power or export it to other countries on a supergrid system. That might also help us to balance the variations in output from wind and other renewables – in effect we export excess and then re-import it later when and if there is shortfall. It get stored in, say, large hydro reservoirs in Norway or Sweden (as Denmark does with its excess wind output), although it’s really just ‘virtual’ storage. We don’t get the same electrons back! But for this to be possible we need the grid links.
The same message emerges from recent US National Renewable Energy Laboratory studies of wind curtailment, though their problem is a bit different. A 2009 NREL study concludes that, so far, congestion on the transmission grid, caused mainly by inadequate transmission capacity, is the primary cause of nearly all US wind curtailment.
NREL says that wind curtailment has been occurring frequently in regions ranging from Texas to the Midwest to California. For example it notes that at one point in 2004 nearly 14% of wind generation MWh had to be curtailed in Minnesota, though this fell to under 5% subsequently. It notes that, curtailment has also become a significant problem in Spain, Germany, and the Canadian province of Alberta – up to 60% in Germany in some cases, while in Spain, NREL notes ‘the amount of wind power curtailed as part of the congestion management program has increased steadily over the past two years’ . This is a terrible waste of potential green power…
NREL says that building additional transmission capacity is the most effective way to address wind curtailment, a point also made in its 2010 ‘EWITS’ study of eastern US options – which concluded that wind could replace coal and natural gas for 20–30% of the electricity used in the eastern two-thirds of the US by 2024. That would involve 225–330GW of wind capacity, and an expensive revamp of the power grid. However, like the earlier NREL study, it says that, with an improved grid, especially with long distance HVDC transmission allowing for balancing across the country, the amount of wasted wind energy, and the need for back-up, would decline.
The 2009 study also discusses other possible measures for reducing wind curtailment, including greater dynamic scheduling of power flows between neighbouring regions. That’s moving in the direction of ‘smart grids’ and possibly on to dynamic load management – adjusting demand in line with supply.
It’s already common to reschedule some large load to meet shortfalls in supply – some supply contracts specify interruption options – and reduce prices accordingly. But more sophisticated smart grid systems may have fully interactive load control – switching off some loads temporarily when supply is weak. The classical example is domestic (and retail) freezers, which can happily coast for several hours without power or damage to food stocks.
What we are seeing is a move beyond simple real-time ‘baseload’ thinking and on to balancing supply and demand dynamically, over time. This can involve more than just rescheduling loads or shifting electricity across regions via supergirds, and more than just storing electricity virtually. It can mean shifting to storing some of the electrical energy as heat – heat is much easier to store (e.g. in molten salt heat stores, than electricity). And local heat stores can be topped up with heat from solar and other renewable sources. Most of this heat would be used to meet heating needs directly, but some could be converted back to electricity, for example in steam turbine units, as is planned for the large Concentrating Solar Power plants being built in North Africa – to allow them to carry on generating power from stored solar energy overnight.
A parallel option is conversion of electricity to hydrogen gas via electrolysis, for later use as a fuel, for vehicles, or for heating, or for electricity (and heat) generation in a fuel cell. The efficiency losses from some of these conversion processes may limit how much of this we can use cost-effectively, but we need to start thinking about new optimisation approaches which go beyond simple real-time power links. Neil Crumpton’s scenario tries to do that: see my earlier report.
It’s definitely a challenge to conventional thinking. As Eric Martinot puts it in Renewable World (Green Books): “The radical concept that ‘load follows supply’ on a power grid (i.e. the loads know about the supply situation and adjust themselves as supply changes) contrasts with the conventional concept of ‘supply follows load’ that has dominated power systems for the past hundred years. Storage load represents a variable-demand component of the power system that can adjust itself, automatically within pre-established parameters, according to prevailing supply conditions, for example from renewable power.”
Energy storage is of course expensive, which is the main reason why we don’t have much of it at present. But as we move to a new more interactive energy supply and demand system, then the value of stored energy will increase. You could think of it as ‘virtual’ baseload. But it’s more flexible than that – and flexibility seem likely to be a key requirement in future.