Energy Storage: What Is the Cheapest Solution?

“Lithium ion would cost $1 million per megawatt for 15 minutes of storage, he added.”
That would be $4,000 per kWhour of storage.  No way.  Carnegie Mellon was using $1,000 per kWh and GM was saying: “No, the Volt Li Ion battery will be significantly less than this.”  Also, there are at least two companies down closer to $300 per kWh for Li Ion.  ...that’s 1/3 of $1,000 or 1/12 of $4,000.  Boy, talking about misinformation.

There is some other obvious storage options being missed here:
1. One of the big power users is for cooling in the sun belt where solar will be most cost effective.  AC units that save money by making ice when power rates are low and then use the ice for cooling when power rates go up, are increasingly available.  (This turns out to be an old trick used to cool theaters back when AC units weren’t large enough to handle such a large load.)  Right now cheaper power is available at night when obody is using the coal generated power.  With more solar available, and as it continues to decline in price, the cheapest power will be available during daylight hours, inspite of increased use during this time and better load matching.  These AC units will then be making ice during the day for use in cooling during the night.  Essentially you’re storing power for cooling in the form of ice.  Very cheap storage used for one of the biggest power loads in the sun belt.
2. Molten salt storage in central tower type Concentrated Solar Thermal (CST) plants can be stored under ground and used later in the day at reasonable cost.
3. We have huge amounts of natural gas available in the shale deposits in the USA.  So forget storage and use solar and wind in combination with natural gas peak power plants.
If you’re going to get into this discussion then provide a table of options and reasonable estimates of their costs.  This article is very conjectural.  $1 million per megawatt for 15 minutes for Li Ion is a way-out-to-lunch number.

Great input mds, please forward any info you have on the cooling concepts. I live in a hot country with huge AC power loads. (JavaScript must be enabled to view this email address).

You’ve done a great job Uci with most of your articles, especially on PV, but your primary source here provided a lot of disinformation.  I’m writing an in-depth analysis of the energy storage options, which I’ll present at the ASME ES conference next spring.  The paper, ES2010-90377, probably won’t be available until shortly before the conference, but the abstract follows:

There is the general perception that increased grid-scale energy storage will facilitate expansion of renewables.  Most discussions of costs and competitiveness of storage options have addressed cycle efficiency and capital costs of energy storage in terms of both $/kW and $/kWhr.  However, likely number of cycles per year, marginal value of delivered energy, impact on GHG emissions, application-specific expected lifetime, discount rate, likely trends in the markets, and other factors have seldom been addressed, at least for grid-scale applications. 

The levelized costs of delivered energy from the leading technologies for grid-scale energy storage are calculated using a model that considers likely number of cycles per year, application-specific expected lifetime, discount rate, duty cycle, and likely trends in the markets.  The expected capital costs of the various options evaluated – pumped hydrostorage, underground pumped hydrostorage (UPHS), hydrogen fuel cells, carbon-lead-acid batteries, advanced adiabatic compressed air energy storage (AA-CAES), lead-acid batteries, lithium-ion batteries, flywheels, sodium sulfur batteries, ultra capacitors, and superconducting magnetic energy storage (SMES) – are based on recent installation cost data to the extent possible.  The marginal value of the delivered stored energy is analyzed using recent grid energy prices from regions of high wind-energy penetration.  The above list is in order from most competitive to least competitive.

Grid-scale energy storage is expected to lead to significant reductions in GHG emissions only in regions where the off-peak energy is very clean.  These areas will be characterized by a high level of wind energy with cheap off-peak and peak prices.  At the expected daily price differentials, the only conventional options expected to be commercially viable are hydro storage and UPHS.  The market value of energy storage for short periods of time (under a few hours) is expected to be minimal for grid-scale purposes.  Only low-cost daily storage is easily justified both from an economic and environmental perspective.

A lesser-known energy storage option, Windfuels, is also briefly reviewed.  Here, excess off-peak electrical energy is used to synthesize standard liquid fuels, such as gasoline and jet fuel, from CO2 and H2O.  Simulations have shown that innovations should make it practical to reduce CO2 to CO at 90% of theoretical efficiency limits.  When combined with other process advances, it should then be possible to synthesize hydrocarbons and alcohols from point-source CO2 and off-peak clean grid energy (wind or nuclear) at system efficiencies in the range of 52-61%. The cost of the tanks for storing energy in jet fuel, ethanol, and diesel is only $0.02/kWhr.  The cost of storing vast amounts of energy in batteries, compressed air, or flywheels would be several thousand times greater.

How does flywheel storage compare? I know of one firm which has commercial applications in evaluation by power companies. look at Beacon corp.

Another mixing of problems occurs with this subject, besides mixing short term storage for grid stabilization with longer term energy storage.  Solar and Wind are always lumped together.  I see the two as being different with respect to grid stabilization.  Some areas are already having trouble stabilizing the electrical power supply when large amounts of wind are in use.  The wind can start or stop blowing over wide areas very suddenly.  I can see how this would be a problem.
Solar is different.
1.  Power use on the grid is not stable.  The amount of power drawn is constantly changing.  Distributed PV on houses and businesses will mostly just offset day-time (peak) power demand.  It will just change the already variable demand picture and may actually help reduce peak loads at times.  As PV power reaches grid parity and becomes more common, peak power demand will become a night time phenomena.
2.  One of the major uses of electricity in the sun belt of the USA is for air conditioning.  There are already systems that can make ice when power is cheaper and use the ice later for cooling.  Now, they make ice at night when coal power demand in lower and use it during the day.  In the future, these systems can be used to make ice when PV power is plentiful during the day and use for cooling at night.  In this way PV power will be used to fill the largest power need …even before electrical power storage becomes a factor ...essentially eliminating the current diurnal peak power hump in the USA sun-belt.
3.  Clouds are the only thing that reduce the electrical output of PV, CPV, or CST during the day.  Most of this should be predictable using real-time satellite monitoring and communications.  After all it is the 21st century now.  If you have regions joined by HVDC lines from the East to West coasts in the USA, then power can be shared across the sun-belt.  (The same should work on the Asian continent for China.)  Areas with clouds can get some power from those without clouds.  Changes in PV output should not come as a massive surprise the way changes in wind can.  Balancing of PV power generation across the USA sun-belt should be feasible, if not outright easy.
4.  There are areas where cloud cover is almost never an issue, like the Mohave Desert.
5.  You do have to have over-night power storage or use an alternative source at night.  Some CST plants are already being built with molten salt storage.  Huge amounts of shale gas have just become economical to produce in the USA, so we could use this at night.  (Remember we’re already reducing peak demand for AC by producing solar ice, we’ll be using more LED lights, and other electrical demands at night are lower.)  Finally, you can also consider all the electrical power storage methods mentioned by FDDoty above: “pumped hydrostorage, underground pumped hydrostorage (UPHS), hydrogen fuel cells, carbon-lead-acid batteries, advanced adiabatic compressed air energy storage (AA-CAES), lead-acid batteries, lithium-ion batteries, flywheels, sodium sulfur batteries, ultra capacitors, and superconducting magnetic energy storage (SMES)”  …and also by Charles R. Toca above: “the molten (600 degrees F) sodium sulfur, there is the VRB-ESS - flow battery technology, for about the same price”

Bottom line:
A. Solar power is far more predictable and stable than wind.  The wind can just stop blowing, but there is always some light from the sun shining down during the day.
B. WE WILL BE ABLE TO INSTALL A HUGE AMOUNT OF SOLAR POWER BEFORE ELECTRICAL POWER STORAGE BECOMES AN ISSUE.
C. We can use shale gas and bio-gas at night until electrical power storage can be accomplished at a reasonable cost.  Electrical power storage is now developing much more rapidly.
The truly great news is there really can be a limit to increasing energy costs.

Question for FDDotys - Who is actually working on UPHS?

I would have thought the cost of excavating an underground chamber would be pretty horrible when you look at the energy density. This is the one area where you are better off looking at some form of CAES - you use the water to provide a constant pressure head on the cavern. The reason it makes sense for CAES is that the work done compressing ambient gas to, say 60 bar, is an order of magnitude greater than the energy from pumping water up 600m. I did some rough numbers on this a while ago but sadly can’t lay my hands on them at present.