This proposal is about energy storage of a very specific kind to support the electricity grid. The case for energy storage is extremely strong at the moment as we decarbonise electricity generation. The world has reached the very interesting point where the cheapest electricity actually comes from wind and sunshine but these generation forms unfortunately produce electricity only when the primary resource is available - i.e. wind turbines make electricity only when the wind blows and (most) solar power can only make electricity when the sun is shining. So long as these renewables comprise only a small proportion of all of our generation, the intermittency of wind and solar power is no problem at all - because we can control the generation being obtained from coal-fired and gas-fired power stations. However, if we are to generate high fractions of all of our electricity from renewables, we will need to be able to store large amounts of energy.

Now, there are many different ways to store energy. No one form is a solution for all of our needs. Energy storage has to developed to be suitable over a large range of timescales and a large range of sizes. Each system being developed has its own particular set of advantages and disadvantages. Cost is extremely important in all cases: energy storage is extremely expensive. Most people do not realise that even with the best commercial offerings at present, the ratio between the cost of an energy store and the value of the energy that it contains is typically 1000:1. Lifetime is also extremely important. If a given energy store has a lifetime that is only, say, 5000 cycles, then that energy store must be replaced after 5000 cycles and the cost that it will add to the energy that has passed through it will typically be ~20% on this basis. Turnaround efficiency is also important, if you lose 20% of all of the energy that comes into the store, this adds a further cost that could be anything up to 20% (but would usually be more like 10% because the input energy is usually much less valuable than the output).

This proposal sets out to examine a system that appears to offer energy storage over a range of timescales between milli-seconds and tens of hours. The system comprises two distinct energy stores connected in a "serial" fashion in the sense that there is only one output to the grid. One of these energy stores is a very large flywheel. The second is typically a compressed air store but it may also be a high-head pumped hydro store or a pumped-thermal store or an energy store based on liquefied air. The connection to the grid is via a large synchronous generator. These systems are suitable only at medium-to-large scale - powers above 50MW and energy storage capacities in the order of 250MWh and above. They are not suited for urban locations. For those (many) situations where they are suited, these systems appear to offer the potential for extremely high performance at very competitive costs.

Most importantly and also most distinctively, the combination of the flywheel and the rotor of a synchronous machine endows these stores with substantial amounts of "real inertia". Inertia sounds like a bad thing but in the context of electrical power systems it is an extremely good thing and it is present in all of the spinning rotors in steam-turbine-driven power generation. As we move away from generation using coal, oil and gas, we are switching off these big rotating generators and we are losing inertia that was previously present as a free service. With lower inertia, the system responds more suddenly to changes in load or generation. If we allow too much inertia to disappear from our electricity system, we become very vulnerable to uncontrolled system shutdowns from either unexpected weather fluctuations, glitches in communications networks or from mischievous cyber-attacks which can use the system sensitivity to trigger disproportionately large events from relatively small actions.

Potential Impact:
SHyKESS systems could become a single cost-effective energy storage solution operating over a range of timescales from milli-seconds to tens of hours. They marry a fast and highly robust energy storage form (spinning inertia) with unlimited life and superb efficiency (~99%) together with slower secondary energy storage form having lower costs but also lower efficiency (~60% - ~75%). Expected marginal costs per unit of energy storage capacity are ~£250/kWh for the fast storage and <£25/kWh (sometimes much less) for the secondary storage. SHyKESS systems have natural lower-limits of scale: power ratings >50MW, storage durations of tens of minutes for the KE storage and storage durations of a few tens of hours for the secondary store. There are many locations at nodes of the UK electricity transmission and distribution systems where such systems could be installed.

The main overall impact of this development will be that higher penetrations of renewables will be affordable in both the UK and many other regions than would otherwise be the case. Ultimately, this will have a real effect on averting some substantial CO2 release and making reliable electricity systems affordable in many regions where it would not otherwise be so. Obviously, numerous other energy storage propositions would make similar claims but the SHyKESS concept has some real and verifiable attractions over most competing systems:

* Virtually any form of thermo-mechanical energy storage can be integrated directly as the secondary energy store in a SHyKESS system. Specifically, pumped-hydro energy storage, compressed air energy storage, pumped-thermal energy storage and liquid-air energy storage are all attractive candidates. Thus, the real question in these cases is not "SHyKESS or ~ ?" so much as "SHyKESS with ~ ?".

* The SHyKESS design is such that it introduces real inertia back into the system - replenishing the inertia continually being lost as large turbo-alternators are being retired in favour of renewable generation.

* Real inertia is un-hackable . Cyber-security is emerging as one of the primary concerns of utilities and grids.

* Real inertia has no dependency on the integrity of any sensing or communications pathway to act and thus it has potential to be intrinsically very robust.

* Real inertia is always stabilising. Simulated inertia realised via active control can be destabilising at grid level.

* The SHyKESS systems intrinsically deliver the phase-balancing and reactive power compensation functionalities presently provided by "STATCOMs" at zero marginal cost.

* The primary energy store (the flywheel element) has effectively infinite life. An all-steel flywheel rotor of any design would be expected to last at least ten million (107) complete cycles. Since no complete discharge of a SHyKESS flywheel would occur in <5 minutes, this indicates a rotor service lifetime in excess of 200 years. Some elements of SHyKESS systems will certainly require inspection (perhaps annually) to verify continued fitness for purpose. The secondary energy store will naturally see far fewer cycles than the primary store. Most of the natural candidates for secondary energy stores in these systems also have natural lifetimes exceeding many tens of thousands of cycles. Thus, these systems are extremely well suited to long-term infrastructure investments.

* SHyKESS systems are especially well suited to development and manufacture in the UK. The UK has excellent indigenous expertise in hydraulic machinery, electrical machines and realising vacuum vessels. Moreover, SHyKESS systems require no exotic materials whose prices might be volatile in the future for the UK (Lithium, Vanadium, Neodymium, Dysprosium, Palladium, Gallium etc.)