The variable nature of wind has been the Achilles heel of wind power for centuries, making the storage of energy — for example as compressed air in the buoyancy chamber of a floating wind turbine — an attractive proposition.
Of equal importance, however, are the opportunities that become available for fundamental improvements in wind turbines if the turbine-driven generator is separated from the grid, and all the wind energy is delivered to the storage reservoir.
With such an arrangement, where the wind turbine operates in series (Fig 1) with the storage system, rather than in parallel (Fig 2, below), the grid would be supplied only from an expander-driven generator, fed from the compressed air reservoir.
In such a case the energy supplied to the grid would seem to be much less than the wind turbine can deliver directly, because of the inevitable losses in the compression and expansion cycles.
But there are several ways in which the round-trip efficiency of the system could be improved to an acceptable level.
Simplifications and costs savings also become available that could more than compensate for any lost energy. Fig 1 (above) illustrates these simplifications.
In a wind turbine with a compressed air energy storage (CAES) system, the expander generator connected to the grid, and which is driven by compressed air from the storage reservoir, would be located in a machinery room at near sea level, not in a nacelle 100 metres or so above the water.
This generator would be completely controllable and would run a fixed speed, regulated only by throttle valves or inlet guide vanes at each state of the expander.
Consequently, the complex and costly frequency control system of a conventional wind turbine, which has to cater for large variations in wind velocity and generator speed, could be eliminated.
If the wind turbine generator in the nacelle does not have to provide power at grid frequency, a much simpler generator becomes possible. It could, for instance, be a direct-drive low-frequency machine designed to deliver power at 25Hz or even less.
A low-frequency direct-drive AC generator would be considerably smaller, lighter and cheaper than those in existing turbines.
Alternatively, the wind-driven generator could be a low-speed DC machine. In either case, not only is the generator itself smaller and cheaper, but a large three-stage gearbox and its associated complex lube oil system becomes unnecessary.
The wind turbine generator would only have to supply power to a compressor drive motor and this would be a low-frequency AC or DC machine purpose-designed to accept the available variable power.
Any frequency control system would be simpler and cheaper than one required to supply grid-quality power, and any gearbox required for the compressor would be of the normal high-speed low-toque type, which is standard in packaged multi-stage air compressors.
The icing on the cake with regard to these opportunities is that almost all the machinery would be located at or near sea level, where the wind turbine tower is attached to the floating support.
With all the main equipment installed in an easily accessible machinery room near sea level rather than in the nacelle, the turbine would be cheaper to manufacture, and simpler to install and maintain.
With the main equipment in the turbine nacelle reduced to one moderately sized generator, the need to provide internal access for maintenance would be questionable. Without internal lifts or stairways the tower could be made simpler and cheaper.
With regard to the losses incurred by storing and recovering the wind energy, there would be a unique opportunity to recover heat normally lost to the atmosphere from the air compressor cooling system.
By storing the heated water from the compressor intercooler and aftercooler in an insulated vessel (Fig 3), and reusing this heat in the next expansion cycle, the effective efficiency of the compression cycle would be much greater than in most compressors.
Such an opportunity to improve the efficiency of the system does not exist in normal compressor applications because there is no associated expansion process that can make use of the waste heat.
This innovation would be particularly relevant to a grid-based generating system operating on a 12- or 24-hour cycle of storage and recovery.
The purpose of the hot water system is to improve the efficiency of the compressed air storage and recovery cycle. But if wind energy is used to further increase the temperature of the stored hot water, this system becomes an additional method of energy storage.
If the hot water reservoir is pressurised, which would be a feasible thing to do, there is not theoretical limit to the temperature that could be achieved and the energy stored as hot water could easily exceed that of the compressed air system.
Consequently, the power that could be made available to the grid during peak periods would be limited only by the capacity of the grid, and the size and rating of the expander generator.
Electric heating of either the stored hot water and and expander airflow, or both, means that the final discharge temperature of the expander air could be well above atmospheric temperature.
If so, there is a further opportunity to increase the power supply to the grid and further increase efficiency, with a heat exhanger transferring heat from the expander exhaust air to the incoming air from the storage system.
A wind turbine CAES system requires no fundamental research, and little development of any technology. It relies on well-proven machinery and equipment that could be built and installed quickly.
Ian Crossley is the managing director of Isopwer Ltd