|Energy storage is a key enabling technology in stabilizing our grids, but uncertainty over its viability persists
Credit: National Grid
The global energy market is shifting at an unprecedented rate and this change is being driven by renewable generation in the form of wind and solar photovoltaic.
Forecasts for the total installed wind generation in 2020 vary from 586 GW to 1000 GW, while predictions for solar PV in 2020 vary widely from 330 GW to 1000 GW.
Calculating the cost price of a unit of electrical energy was once straightforward. The metric still generally used is the levelised cost of electricity (LCOE) which sums the total capital cost (CAPEX) and lifetime operating costs (OPEX) including fuel inputs, taking into account the financing costs of both. That number is then divided by the lifetime energy output, to give a cost per unit energy.
Unfortunately, the output of renewable assets cannot be matched to suit demand. As a result, renewable electricity, available at zero marginal cost, must be rejected from the system when there is a surplus.
Conversely, expensive gas peaking plant must be used to provide electricity, while weather conditions do not allow for demand to be covered by intermittent renewables. That makes the amount of electricity which each source produces – or rather, the amount of electricity that can be sold from each source – rather unpredictable.
That is the value of storage: it brings the amount of electricity available to the system back under human control. The Californian regulator is convinced and is mandating that utilities deploy 1.3 GW of electricity storage by 2020. However, for everyone to agree, the value must be greater than the cost by a clear margin. The cost itself is far more challenging to calculate than is usually appreciated.
One method is simply to convert the LCOE to what is known as LCOS – the levelised cost of storage. That provides the cost of storing electricity including the CAPEX, OPEX and also the cost in electricity resulting from efficiency losses in the storage system. To get a number comparable with the competitor of the storage concept – gas peaking – also requires that we add the cost of the input electricity to the LCOS. So the really revealing metric is levelised cost of electricity from storage, or LCOE-S – see boxed formula.
LCOE-S: A simple calculation?
A 6kWh battery bank, at a cost of $1000 per kWh, incurs a capital cost of $6000. That covers the load shifting requirement of one household equipped with rooftop PV panels. A commercial battery system will of course provide storage ‘space’ for hundreds of such households.
With a weighted average cost of capital (WACC) of 12 per cent, a levelised cost of electricity (LCOE) from the solar panels of 10¢ per kWh, a system cycling once a day for 13 years, an efficiency of 85 per cent and negligible maintenance costs, the boxed formula gives us a LCOE-S of 85¢ per kWh.
We have assumed here that the size of the storage technology can be matched exactly to the technical requirements. But we actually need to push our capital cost up to around $10,000. This is because almost all batteries have inherent depth-of-discharge limitations: some of their chemical potential energy must remain potential because if it were used it would cause permanent damage to the electrodes. Even if that depth-of-discharge limit is not breached, getting persistently close to it will reduce the number of charge-discharge cycles over which the system can operate, so it might easily not last 13 years.
It does not take an energy economist to realize that without any kind of subsidy the system described above, though it may save the whole electricity system money by peak shaving, does not make financial sense when peak electricity prices are around 4-5¢ per kWh as in Germany. This example, along with the boxed formula allow us to identify the three key variables of the LCOE-S: CAPEX, cycle life and efficiency. Some indicative technology-specific numbers are included in Table 1.
Customers looking for the optimal electricity storage system are faced with a wide range of different technologies entering the market. Many of the facts and figures are changing rapidly, simply not known, taken out of context or else displayed in a misleading manner. It is common for storage manufacturers to describe their CAPEX costs in per kW or per kWh terms, as with our example of a residential battery system. However, many storage technologies have separate power and energy costs, so a $/kWh or $/kW number is useless on its own. Consequently, we believe that a graph showing $/kWh for different numbers of hours is the best way to show the CAPEX cost of storage, as in Figure 1. Such a graph, combined with the cycle life, allows for a much more accurate assessment of the cost of a system for different applications.
|Figure 1: CAPEX per kWh of various technologies at different discharge time requirements. Note that this does not include any replacement costs for degraded assets. Source: Krajacic et al., 2012 ; Arup , n.d. ; ISETA-RWTH|
An example of an accurate but misleading statement might be that a certain battery costs $250/kWh and can achieve 5000 cycles. The CAPEX might assume 100 per cent depth of discharge, but the cycle life is reduced to 1000 cycles. The same battery might deliver 5000 cycles if only 10 per cent of the depth of discharge is used, but now the real CAPEX is $2500/kWh (since it is only possible to use 10 per cent of the storage).
There is also a vast amount of conflicting information online. For flywheels you might see a figure of $3500/kW listed as the cost for an installation. However, you would have to dig deeper to establish that this is for only 15 minutes of storage and the cost per kWh is actually much higher at $14,000/kWh. What is the correct number with which to budget?
There is also variation for established technologies, like pumped hydro, where the costs will vary from site to site depending upon local geography. There is no easy answer to this problem and the best advice is to ask detailed questions from the supplier to ensure you understand the context of any figures.
Application is king
There exists a variety of storage technologies to suit various commercial applications, each with different advantages and limitations. Table 2 shows examples of different applications, and indicates which of the three key variables will have the most significant impact on the LCOE-S calculation.
A power-intensive application requires a significant amount of power to be provided at short notice, while an energy-intensive application requires a given amount of power to be generated over a period of hours. Load shifting requires high efficiency because the cost of the ‘fuel’ – off-peak electricity – is a key driver of the cost of the electricity that exits the store. The more energy is lost during the cycle, the more it costs to cover the final energy requirement from the store. Efficiency becomes progressively more important the higher the fuel cost, just as is the case for conventional generating technologies.
That is important. A key high-value application for storage in the near term is in reducing the amount of diesel generation required in remote areas by adding a combination of renewable generation and electricity storage. The lower the efficiency of the technology and the higher the input price of electricity, the more stringent the cost target becomes (see Figure 2 and Table 3). The normal cost of diesel generation is 25-35¢/kWh, depending on the location.
|Components of LCOE’s for different technologies (US¢)|
|Figure 2: Representation of the cost breakdown of the LCOE-S with an input electricity price of 10¢/kWh. Source: Institute for Power Electronics and Electrical Systems (ISETA-RWTH), Aachen; Krajacic et al., 2012;|
Reserve services ensure that the grid can adapt to any unexpected events or power shortfalls that might occur when, for example, a 600 MW generator trips off. Depending upon the exact service being supplied there may be a requirement for low power ($/kW) costs for something like Fast Reserve; alternatively, low energy costs ($/kWh) might be needed where several hours of generation are required, that is, where the requirement is for Short Term Operating Reserve. For most applications in this area, there is likely to be a requirement for several hours of power.
In the state of New York, remuneration is now available for providing power or absorbing it during short-term power variations to help maintain the system frequency. This is an application that is dominated by the power cost and the system’s cycle life. It might see as many as 4000 cycles per annum, which would mean 80,000 cycles in a 20-year life. For comparison a pumped storage plant might only see 5000 cycles in 20 years. Both flywheels and lithium-ion batteries have been installed for this application. A very high $-per-kWh CAPEX is affordable in this application because the LCOE-S is still kept low by the very high number of cycles per annum.
The LCOE-S needs to be lower than the LCOE of flexible generation like gas peaking and diesel gensets. But it is worth remembering that gas and renewables are two sides of the same coin. Storage itself can store electricity from gas plant just as easily as it can store intermittent renewable electricity. No matter where the input comes from, storage reduces the amount of generating capacity that any system requires.
It is time to start thinking seriously about how individual storage technologies can serve the applications described above and how much it will cost for them to do so. That means assigning technology-specific numbers to the three key variables of storage (CAPEX, efficiency and cycle life) and working out how much the input electricity will have to cost to make each technology competitive with flexible conventional generation for a given application.
Abolishing subsidies will not stop the massive roll-out of renewables. In this rapidly developing situation storage has a great deal to offer. But only a proper understanding of the way the cost of storage works will lift the fog that obscures the economics behind the concept.
James Macnaghten is chief executive of Isentropic Limited. For more imformation, visit www.isentropic.co.uk.
Power Engineering International Archives
View Power Generation Articles on PennEnergy.com