Energy Storage Considerations

Energy storage is going through a fast transition period of technological advancement due to market demands in the mobile phones, laptop computers and power tools. A nascent market in the electric vehicle field has taken R&D to another level of battery size that will have a place in the power generation application.

Although electric energy storage has had its place in off-grid power generation, it has only been effective where the cost of getting fuel to the site and the generator maintenance have been very high. Depending on the application these batteries have had a limited life.

Energy storage entry costs in main grids

The Smart Grid and the inclusion of renewable energy in the main grids make for the operation of a “flexible” grid and are bringing new and refreshed storage technologies to the fore.

However, historically, storage has been difficult to sell into the market in the main grids, not only due to high costs but also because the array of services it provides and the challenges it has in quantifying the value of these services particularly the operational benefits such as ancillary services.

The value of energy storage is best captured when selling to the entire grid, instead of any single source. Figure 1 shows the different values of storage and the market size. Vertically integrated utilities could more easily capture the whole of these values; disaggregated utilities require a cooperative approach from generation, transmission and distribution sectors.

So when evaluating a battery solution for the grid it is wise, where possible, to determine the aggregated value of the various applications that the battery will support.

The three highest value applications generally require storage support periods of minutes to 2 hrs whilst the other three larger market sectors require storage support from 2 to 8hrs. These applications classified by the storage size and discharge time are shown in figure 2.

Storage Applications and Market Entry (Source: EPRI) Storage Applications and Market Entry
OPS Advanced Storage Figure 2

Wholesale energy prices in large grids powered by conventional fuels such as coal and gas fluctuate between 3c/kWh at night and 9c/kWh in the evening peak. A few times during the year a critical peak can escalate to 30c/kWh and above. It would be hard to justify the battery costs just to satisfy the few occasions that these critical peaks occur in a year. Here only pumped storage and compressed air are possible solutions as their lower cost could be used for energy arbitrage as well.

Deferring distribution capital cost, improving power quality and reducing system losses are other features that could facilitate applications for batteries in large grids. In rural areas with Single Wire Earth Return (SWER), small battery banks (tens of kW) could be a specific application due to the high voltage support costs. Table 1 shows the battery capital cost target entry points for main grid applications.

Utility Applications US$/kWh
Residential load Management 50
Peak Shaving 450
Frequency Regulation 250
Power Quality 450
Wind and Solar Leveling & Ramping 100
Spinning Reserve 50

An aggregate cost based on the frequency of these and other (T&D) events followed by a life cycle cost of the battery (LCCB) will determine if a battery is a cost effective solution in a main grid.

Figure 2a shows that the value of storage increases (blue line) as it goes further away from the generator or in other words closer to the final use. It accumulates value as it gradually covers more applications. However, this requires good communications, a good regulatory system and variable tariffs. If these elements are not available the utility’s highest yield for ancillary services is achieved at the 4 to 34kV transformer level.

Storage Location on the Grid

Energy storage Costs in Diesel, Solar and Wind Grids

Smaller grids powered by diesel generators can have generation cost of 50c/kWh and higher. In island locations where the fuel delivery is expensive these cost could escalate much further.

As the current (2011) energy costs of wind of 10 to 15c/kWH and photovoltaic of 20c to 30c/kW energy are below those of diesel it is important to see if the added cost of batteries is justifiable. The inverter/charger cost is already part of the PV system so it is not necessarily an extra cost to the battery as the case would be in a main grid application.

The battery in these cases will be providing most of the services of table 1 simultaneously and on a continuous basis. Currently the battery is generally attached to the diesel generator building but in future the batteries could be strategically located where they could also provide transmission and distribution support and customer management increasing the value of the battery.

Furthermore these rural grids are generally vertically integrated so the whole value of the battery can be readily implemented.

Battery Technologies

The applications listed in Table 1 have different requirements of power, energy and discharge times. Some storage devices comply with one or two or the three requirements. Figure 3 shows the applications and capabilities of current technologies.

As a first step in the selection process, the most suitable new technologies for new applications are Li-ion, Vanadium Redox (VR) and Zinc Bromide (ZnBr). These last two batteries are referred to as “flow” batteries as the electrolyte is liquid and is pumped to and from tanks. Sodium Sulphide (NaS) batteries operate at >3000C so they are in principle not suitable for remote areas.

System Ratings Installed Systems ESA

The up front purchase price of a battery is important in that it provides the lowest cost price in a tendering process but not necessarily the lowest long term price of the project as the battery life could vary significantly from one technology to another.

Figure 4 shows the capital cost comparison of the technologies (this 2008 figure is provided as a reference only as some of the technologies have progressed dramatically as we shall see the further down the text). A proper and realistic way of comparing technologies is using the life cycle cost (LCC). Most installation, operating and maintenance (and IPP) contracts are over periods of 10 to 15 years, therefore the relevance of the LCC.

Figure 4Figure 4

In the next selection step we consider the cost per cycle or LCC of the different technologies (figure 5) and we see again that Li-ion and the flow batteries are suitable for the OPS applications.

Figure 5Figure 5

The LCCS is the most important economic parameter in the battery selection process.

The Flow Batteries

Both these batteries have liquid electrolytes and consist of a membrane stack and relatively large electrolyte holding tanks. In fact increasing the electrolyte volume and size of the tanks can increase the storage capacity. Two independent pumps circulate the electrolyte.

They don’t have the power capacity of a lead-acid battery let alone a Li-ion one but have good energy capability. There are two commercial technologies, Zinc-Bromide (ZN-Br) and Vanadium Redox (VRB).

The efficiency of these batteries is around 80%. This is an important factor particularly if the energy is coming from expensive generators such as PV, wind or Diesel. Both technologies claim to have >3,000 cycle life (@80% DOD), therefore their low LCC. However there are only a handful number of manufacturers of these batteries, one of them Australian (RedFlow, recently floated in the Stock Exchange raising $17.5 Million dollars)). There are only a few installations in the world.

Figures 6 and 7 show the ZnBr. RedFlow is a systems company selling a packaged inverter-battery-communications system using lead-acid and ZnBr batteries. RedFlow claim a 1,000-cycle life extendable to 4,000 by refurbishing of the modules. Recently a 5kW-20kWh system sold for $A 21,000.

Figure 6Figure 6

Figure 7 5kW 10kWhFigure 7 5kw 10kwh

Figure 8 shows the VRB battery from Prudent Energy installed at a telecom site in Kenya (not many pictures available for these systems).

Figure 8Figure 8

Pike Research has produced a graph (Figure 9) showing the market projection of stationary batteries showing the flow batteries, Li-ion and others. The rapid take up of Li-ion is quite remarkable.

Figure 9 - MW vs. Year, Pike ResearchFigure 9

Li-ion Batteries

There have been huge developments in this technology as it is now the preferred battery for the electric vehicle market.

The characteristics of this battery are high efficiency (>90%), high power and energy density and a cycle life well over 3,000 cycles at 80% DOD (see figure 11). The battery is compact, sealed with no maintenance requirements.

The Li-ion can be built to suit high-power or high-energy applications. Most Hybrid Electric Vehicles need high-power and have power to energy ratios of 20/1, whilst electric vehicles need more energy and have ratios of 4/1. Figure 10 shows the cycle life of a Li ion battery suitable for PV solar applications where the major requirement is energy.

Figure 10Figure 10

There are several major brand manufacturers such as Saft, Johnson Controls, Altairnano, Boston Power, International Battery, Electrovaya, Sony, Panasonic, LG, Hitachi, Toshiba, A123, Ener1 as well as a plethora of Chinese manufacturers. Although the previous graphs show relatively high capital cost and reasonable LCC, recent technology advances and escalation of production are showing remarkable drops in prices as shown in a recent graph by the Deutsche Bank (Figure 11).

Figure 11 - Dark blue is 2009 projections; light blue is 2010Figure 11

The graph shows a 30% drop in price expectations by 2012. Also note that the expected price is US$450/kWh and not over >$1,000/kWh as shown in the previous graphs. The fact is that Chinese manufactured Li-ion batteries can now (2011) be purchased for ~A$650/kWh retail price in Perth, Western Australia.

Figure 12 shows some of the battery developments of Altairnano for batteries on the grid that is relevant to diesel grids as well.

When sizing batteries on diesel grids, designers must be aware that batteries will also wear out through “hundreds” of daily shallow cycles providing frequency regulation and “tens” of not so shallow cycles providing ramping to cope with load and wind or solar variations. These cycles are on top of the wear that occurs when supporting the whole load by itself when none of the other energy sources are available. Some of these Li-ion technologies could cope well as seen in the graph.

Figure 12Figure 12

Note: Figure 12 is included as an indicator of how the battery wears out and not to rely on the absolute numbers (no battery will last 50 years).

Figure 13 shows a stationary Li-ion (Saft) battery assembly with a battery management system for a stationary PV installation.

Figure 13 - Saft stack for PV InstallationFigure 13

Figure 14 shows a single LiFePo4 3.2V 90Ah cell of Chinese manufacture.

Figure 14Figure 14

Table 2 shows a comparison of current (2011) technologies that are potentially suitable for OPS mini grid applications:
- Lead-acid
- Li-ion of Western world manufacture (LI.W)
- Li-ion of Chinese manufacture (Li-Ch)
- ZnBr

Table 2 - Storage Cost per kWh, 2011Table 2 Storage Cost kWh

So currently, although the values are very similar, Li-ion still presents the lowest cost of battery energy throughput and is competitive with diesel fuel generation.

If we take into account the cost improvements for Li-ion forecasted in figure 15, the cost would drop by 18% in 2014 to 18.2c/kWh and by 29% in 2015 to 15.7c/kWh taking the cost of battery energy throughputs to 48.2c/kWh in 2014 and 47.5c/kWh in 2015 making it very competitive with other technologies as well as diesel.

Further investments and efforts are needed to develop suitable Li ion technologies that can support increasing penetration of renewable energy and stabilizing of the electrical grid, a particular requirement is that batteries should be able to operate at their rated output for several (2 to 6) hours. Expectations are that some of these developments for stationary batteries will occur by 2015 (1).

Transition to Li-ion in Micro-grids

Li-ion and NaS technologies are being taken up by main-grid utilities such as AES, AEP, JPM and Xcel Energy to avoid expensive and polluting frequency regulation with gas turbines, defer transmission and distribution expenses and reduce intermittency from wind power. These batteries need to provide high power capacity for seconds and minutes with a high cycle lifetime.

Zero-Diesel Daytime (ZDD)

As PV are now easily cost effective when compared to diesel generator sets, the PV array could be oversized to the daytime load (sized for the wet season) supported by a battery of sufficient capacity for 15 minutes or so to hold the full or partial load.

The diesel generators would be off during the day and only started if the low period exceeds 15 minutes. The inverter and battery would be turned off during the nighttime period and charge the battery for 1 or 2 hours before sunlight.

The PV array would require a partial curtailment control feature if their output is higher than the load and the batteries were fully charged. What would be achieved here is much less cycling of the battery (be it lead acid or Li-ion) thus extending its life considerably, particularly for a lead acid battery.

It could avoid the issues of reactive power needs when PV and diesels operate in parallel. This control logic would avoid the very large battery necessary for night or peak loads support, leaving all that for the diesel generators as they normally operate when in a stand alone basis. The economics with the diesels off for 4 to 5 hrs/day (ZDD) should show a significant improvement in the economics and GHG emissions.

Conclusion

It is still hard to justify energy storage costs in main grids and to justify it requires an aggregation of multiple applications, which are hard to evaluate and need the appropriate regulatory frame work to operate. Black out costs would have to be evaluated properly to justify OPS applications in the main grid.

Load Management together with Energy Efficiency is the best lowest cost to reduce energy use and balance the load. It is vital to assess systems on a life cycle cost basis as we do with PV because the new batteries have a much longer life. Using storage is expensive but is justifiable when the costs of diesel generation surpass 50c/kWh.

Although Lead-Acid is a known and proven battery technology, it struggles to compete with diesel and there are no paradigm shifts in sight that will significantly improve the situation in the future. This is a real limitation for technologies such as wind and solar that continues to reduce costs.

The Flow batteries have been in the R&D stage for decades and are just emerging in the commercial world. There are less than handful manufacturers and most of them with limited resources. The latter could be a liability for OPS remote applications. It would need chemical expertise to service these units.

Li-ion has now surpassed the ubiquitous lead acid battery in costs, service and life expectancy its expanding future is reassured as forecasted in figure 9. Li-ion batteries are not ready yet for stationary applications with durations of several hours (load shifting and battery arbitrage) but are expected to do so in the next five years. Main grid utilities are adopting the technology for economic and environmental reasons.

This is one of the few battery technologies that is used in volume markets and therefore has a secure future. The life expectancy of 15 to 20 years of Li-ion batteries will ensure a second use in stationary applications for the batteries that have completed their lives in electric vehicles, as they will do it with 80% of their capacity still remaining. This will bring costs down further in stationary applications.