Question 6 of 8: Energy Storage

Question 6: Energy Storage (10 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Electricity is the one commodity a utility must manufacture at the instant it is consumed. The consequence is a fleet that is sized for the annual peak but earns revenue at the daily average, and generating plant that is forced to cycle daily — which is expensive for a coal unit, damaging for a nuclear one, and impossible below a plant's minimum stable load. Storing energy overnight and returning it at the peak decouples the two, letting the cheapest plant run flat out around the clock while the store absorbs the swing. The two methods below are the only ones deployed at genuine grid scale in Canada and worldwide, and they are the two the isolated system of Question 7 would be choosing between.

Method 1 — Pumped hydroelectric storage

The form of the energy. Water is pumped from a lower reservoir to an upper one during off-peak hours and released back through the same machines at the peak, so the energy is held as gravitational potential energy: $E = \rho g H V$ for a working volume $V$ at a gross head $H$. A scheme with a 400 m head and 10 million cubic metres of working volume stores roughly $1000 \times 9.81 \times 400 \times 10^7 = 3.9 \times 10^{13}$ J, or about 11 GWh before losses — several hours at 2000 MW. Modern plant uses reversible Francis pump-turbines with a single motor-generator, so one machine hall serves both directions, and variable-speed units can now regulate on the pumping side as well as the generating side.

Limitations. The binding constraint is geography, not engineering: the scheme needs two large reservoirs with a substantial head difference, close together, geologically sound and environmentally permissible. Suitable sites are rare and, in most developed jurisdictions, largely taken. Reservoir construction floods land and alters downstream flow and fish habitat, and the pumping cycle changes water temperature and dissolved gas. Response is fast but not instantaneous — a minute or two from standstill to full load, less if the machine is spinning in air. And, as Question 7 stresses, an off-river scheme cannot be replenished by inflow: whatever is generated must first have been pumped, so the store is an energy transfer device and never an energy source. Evaporation and seepage steadily erode the inventory of a closed-loop scheme in a dry climate.

Recovery efficiency. Losses occur four times — motor, pump, waterway friction, then turbine and generator on the way back. Modern schemes achieve a round-trip efficiency of 75–80 %; older plant with fixed-speed machines and long tunnels sits nearer 70 %. The 75 % assumed in Question 7 is a fair design figure.

Method 2 — Compressed air energy storage (CAES)

The form of the energy. Off-peak electricity drives a train of intercooled compressors that fills a sealed underground void — usually a solution-mined salt cavern, sometimes a depleted gas field or a lined rock cavern — to between about 4 and 8 MPa. The energy is stored as the pressure exergy of the air, i.e. the work recoverable in expanding it back to atmospheric pressure. At the peak the air is withdrawn, heated and expanded through a turbine. In a conventional (diabatic) plant such as Huntorf in Germany or McIntosh in Alabama the heating is done by burning natural gas in a combustor, so the machine is really a gas turbine relieved of its compressor: it delivers about three times the output of a conventional gas turbine for the same fuel, because the compressor work has already been paid for overnight. In an adiabatic scheme the heat of compression is itself stored in a solid or molten-salt thermal store and returned on expansion, removing the fuel entirely.

Limitations. Like pumped hydro, the technology is site-bound: it needs a gas-tight geological void of the right size and depth, which in practice means bedded salt. Cavern pressure falls as air is withdrawn unless a water compensation leg is provided, so the turbine sees a varying inlet condition. Diabatic plant still burns fuel and therefore still emits carbon dioxide and nitrogen oxides, which undermines the argument for storing renewable energy in it, while adiabatic plant is technically demanding and has only reached demonstration scale. Air must be dried before injection or the cavern will corrode and the turbine will ingest brine.

Recovery efficiency. The correct figure depends on what is being counted. On an electricity-in to electricity-out basis a diabatic plant returns ≈ 50 %, but that ignores the fuel; on a fuel-and-electricity basis its effective round trip is around 70 %. Adiabatic designs target 65–70 % with no fuel at all. Both sit below pumped hydro, and the difference is the unavoidable cost of compressing a gas rather than pumping a liquid.

What determines economic viability

Neither scheme generates energy — both consume more than they return — so neither can be justified on efficiency. Viability rests on five things.

• The price spread between off-peak and peak energy. The store buys at the night marginal cost and sells at the peak marginal cost, and it must clear its own round-trip loss before it earns anything. At 75 % efficiency the peak price must exceed 1 / 0.75 = 1.33 times the off-peak price simply to break even on energy. A system whose merit order is flat — all gas, say — offers no spread and no business case; a system with a large must-run nuclear or coal base and expensive peaking plant offers a large one.
• Avoided capacity. The store's largest revenue is usually not arbitrage at all but the peaking plant it makes unnecessary. If 400 MW of storage lets the utility defer 400 MW of open-cycle gas turbines, the avoided capital and the avoided emissions permits dominate the economics.
• Avoided cycling of the base plant. Keeping a coal unit at full load overnight instead of cycling it to minimum load saves fuel through better heat rate, saves thermal-fatigue damage to headers and rotors, and avoids the start-up costs and forced-outage risk that cycling brings. For a nuclear unit, which cannot follow load economically at all, this benefit is decisive.
• Capital cost and utilisation. Both technologies are capital-intensive civil works with 40–80 year lives and very low running costs, so viability turns almost entirely on the discount rate and on how many cycles per year the store actually performs. A scheme cycled daily earns three hundred times a year; one held for winter peaks alone may earn twenty.
• Ancillary services and the changing generation mix. Both schemes provide spinning reserve, frequency regulation, voltage support and black-start capability, and pumped hydro can be paid for these even when it is not shifting energy. As wind and solar displace thermal plant the value of storage rises sharply, because the surplus to be absorbed is no longer a predictable nightly trough but an intermittent one — which also lengthens the duration the store must cover and can favour the very large, cheap volumes of a cavern or reservoir over batteries.

In the Canadian context the practical comparison is stark. Pumped storage is proven here — Ontario Power Generation's Sir Adam Beck scheme at Niagara has pumped since 1957 — and hydraulic reservoirs on the Churchill, Peace and La Grande systems already provide seasonal storage as a by-product of ordinary hydroelectric operation. Compressed air storage has only one Canadian installation, the small Goderich adiabatic plant in the Ontario salt beds. Where conventional hydro with reservoir storage exists, it will almost always out-compete both, because its energy arrives free with the inflow.