In a recent conversation with a politician I was told we needed a lot more pumped storage hydro to store surplus wind power from when the wind blows for use when it is calm.
We have now had several posts on the topic of storing wind power. Recently Roger Andrews was Estimating Storage Requirements At High Levels of Wind Penetration and presented 5 scenarios for the UK where storage varied from 700 to 5000 GWh. I have written a number of posts looking at different pumped hydro storage schemes from FLES O-PAC (8 GWh), Coire Glas (30 GWh) and the concept of Strath Dearn (6800 GWh).
In his post Roger put some numbers on the storage required to transform variable wind into dispatchable uniform baseload. In this post I take a different approach. I calculate how much storage would be required to deliver the diurnal peaks in demand from dispatchable wind – pumped – storage – hydro. I’ve taken this approach for a number of reasons:
- The daily demand peaks fetch the highest prices and supplying these peaks follows the traditional finance model for pumped storage hydro – buying low and selling high
- Servicing the peaks as opposed to base load minimises the amount of storage required (the demand peaks represent 18% of total demand in March 2015)
- Supplying the demand peaks in the UK from wind + storage will allow about 20 GW of conventional generation to be retired
- Allowing the fossil fuel generators to supply base load allows them to run at optimum efficiency and to minimise their CO2 emissions per unit of electricity produced. By way of contingency it leaves the door open for an all-nuclear base load supply.
UK generation for the month of March 2015 was chosen more or less at random to construct this model (Figure 1). The model results are specific to that month and may not be applied to annual UK generation. Peak demand was of the order 50GW. The night time minima follow a baseline of 30GW (Figure 1) and this was chosen as the value dividing dispatchable from base load supply. Subtracting 30GW from total demand provides a picture of the dispatchable supply to be provided by wind + pumped storage hydro (Figure 2).
The wind model is based on metered wind generation for March 2015 * 1.46 to adjust for unmetered wind . The second multiplication factor in the wind model is a variable used to adjust the amount of wind power on the system that was optimised at 1.76 (Figure 3). The reason for this should become apparent below.
The objective is to make the stochastic wind pattern of Figure 3 match the regular demand pattern of Figure 2 using nothing but wind and pumped hydro storage. Subtracting wind supply from demand produces the pattern of surpluses and deficits shown in Figure 4. In constructing Figure 4 the surpluses and deficits are reduced by a factor of 0.9 to account for round trip storage system losses of 20%.
The model is initiated with the pumped storage reservoirs empty and since it is windy at the beginning of the modelled period, they quickly fill up with water (Figure 5). The storage requirement is calculated by summing the surpluses and deficits across time (Figure 5). The multiplication factor for installed wind was adjusted until we made it through the month using wind and pumped storage alone. The model was optimised with an uplift in installed wind capacity of 1.76 on todays fleet. Effectively wind that matches demand gets used with the surpluses and deficits going into and out of storage.
To get through the month of March the best part of 1092 GWh (1.1 TWh) of storage is required. This is a gigantic amount! The problem of matching wind (Figure 3) to demand (Figure 2) may seem trivial. But the problem lies in the 10 day lull that began on day 16. To get through that on wind and storage alone requires a vast amount of storage.
Figure 1 UK electricity supply and demand for March 2015 as recorded by BM reports and Gridwatch . 30GW has been chosen as the boundary between base load that can be provided by continuous generation from nuclear and coal-fired power stations and variable dispatchable power currently provided mainly by CCGTs (gas). Click charts to get a very large version that will open in a new browser window.
Figure 2 Subtracting 30GW from the total supply / demand curve produces this picture of dispatchable supply that in this model needs to be met by wind and pumped storage hydro alone.
Figure 3 Actual UK wind production for March 2015 grossed up by a factor of 1.46 to account for unmetered wind . It is then grossed up by a further 1.76 in the model in order to provide sufficient energy to service the demand peaks (Figure 2).
Figure 4 Deducting the profile of Figure 2 (demand) from Figure 3 (gross supply) produces this picture of surpluses and deficits that are either used to pump water into storage or used to generate electricity at times of deficit. There are two key observations: 1) there are generating opportunities (revenue opportunities) on 23 of the 31 days of March, but the scale of those opportunities is small compared with the total storage required (see Figure 5 and section on Business Model) and 2) the generating opportunities are centred on day time peak demand and the pumping opportunities are centred on night time off peak demand satisfying the current criteria for storage based on buying low and selling high.
Figure 5 Summing the surpluses and deficits (Figure 4) across time produces this picture of storage required. The maximum total of 1092 GWh is a rather large number. The storage history for this one month can be divided into 4 parts: 1) days 1 to 8 when gross wind supply was in excess of demand and storage was filled quite quickly; 2) days 9 to 15 when gross supply and demand were roughly balanced; 3) days 16 to 25 representing a 10 day lull that emptied storage completely and 4) days 26 to 31 where the wind began to blow again leaving storage 75% full by the end of the month.
The Cost of 1.1 TWh of Storage
In recent posts I’ve examined a number of different pumped storage concepts and this allows us to put some numbers on the cost:
FLES : O-PAC Flat land large scale electricity storage is an underground concept with a unit storage volume of approximately 8.4 GWh and a cost of €1.8 billon . 130 such units would be required at a cost of €234 billion (£167 billion)
Coire Glas is a conventional pumped storage concept located on the Great Glen of Scotland using a natural lake (Loch Lochy) as the lower reservoir. Storage is 30 GWh and cost £800 million . 37 facilities like Coire Glas would be required at a total cost of £29.6 billion.
Strath Dearn is a concept for a mega storage project on the Great Glen. With a capacity of 6800 TWh this single unit would do with bucket loads of capacity to spare . Cost is unknown but the engineering is on the scale of the Three Gorges Dam that cost $26 billion (£16.8 billion).
The Tesla Battery has a capacity of 10 kWh and a cost of £2,267 each . 110 million batteries would be required at a cost of £249 billion.
Throughout this discussion it is important to bear in mind that we are discussing the wind power and storage needs to meet peak demand and not total demand. In March 2005, total UK demand for electricity was 26.9 TWh. Peak demand (that part over 30 GW) was 4.9 TWh or 18.2% of the total. To service that requires 31.18GW (12.133*1.46*1.76 ) installed wind capacity and 1.1 TWh of storage. The 31.18 GW of installed wind is almost double today’s figure. Figure 3 shows that it is frequently producing 15 GW output (operating at about 50% load) and that having sufficient storage allows this to service up to 20 GW of peak demand.
It is creating a lucrative business model for the 1.1 TWh of storage that is the problem. The model allows for 1.8 TWh of generation from storage over the 31 days of March from 1.1 TWh of storage capacity. The average is 0.058 TWh per day yielding notional utilisation of 5.3% per day while the current pumped storage model aims for utilisation closer to 70% per day. Current pumped storage operates on the diurnal supply demand cycle offering the opportunity to make money to repay high CAPEX every day. Adapting this to a new supply model of storing electricity for weeks clearly requires a new business model that the current “free market” is not able to deliver.
Looking at the storage options detailed in the previous section, the mega concept like Strath Dearn may appear most attractive on the basis of cost, but it is unlikely to be able to respond to the rapid swings of the diurnal demand cycle. It is also extremely unlikely that any organisation in the UK would be willing or able to sink so much capital into a single facility such as this. I’m unsure if a facility like Coire Glas would be able to respond rapidly to the diurnal demand cycle while FLES O – PAC is designed specifically with rapid response in mind. It is clear that a combination of Coire Glas and FLES O – PAC could do the job where large storage like Coire Glas does the heavy carrying while FLES O-PAC provides the rapid close to market response.
The Scottish, UK and EU parliaments have set in motion a set of energy policies that require a large amount of energy storage if the policies are going to make any form of rational sense. Having intervened in the market to create significant opportunity for wind and solar power in particular, the business opportunity for storage quite simply is not there. That is why the storage situation in the UK has not changed for decades. It is naive and quite simply wrong to assume that the existing finance model for storage based on the diurnal demand cycle can be adapted to fit the multi week cycle presented by the passage of cyclonic weather systems across Europe.
The various parliaments mentioned above have created the situation in the UK where we already have about 17.7 GW of installed wind and no means of managing this without using flexible fossil fuel generating plant as the balancing mechanism. The policy to date has created a dependency on fossil fuels that it was designed explicitly to end.
Governments need to confront the high cost of energy storage that they have created through misguided policy decisions and decide how this cost is going to be met. The capital sums involved are substantial and the number of companies able to raise this capital is limited. And there are substantial risks to making an investment for 100 years when policy may change on a whim
An argument can be made that energy storage is of strategic importance and should therefore be built by the State. The cost of 1.1 TWh that is required today and which will take over a decade to build will be of the order £30 to £170 billion. This investment would be in strategic infrastructure that may serve the nation for a century or more. It would reduce the nation’s dependency upon imported fossil fuel. This would enhance energy security and the nation’s trade balance. It is unrealistic to expect the private sector to pay for these services to the State, especially since the current business model cannot enable the investments required. It is time for politicians to either step up to the plate and finance the misguided policies properly in order to make them effective. Or to recognise the mistakes of the past and to follow a different path.
 Clive Best Untangling UK Wind power production
 Leo Smith Gridwatch
 Energy Matters, Euan Mearns Flat-land Large-scale Electricity Storage (FLES)
 Energy Matters, Euan Mearns The Coire Glas pumped storage scheme – a massive but puny beast
 Energy Matters, Euan Mearns The Loch Ness Monster of Energy Storage
 Energy Matters, Roger Andrews How Much Battery Storage Does a Solar PV System Need?