In recent years the energy sector has looked to renewables as a means to reduce emissions. Wind
power is able to provide large amounts of energy at a reasonable cost from presently available
products. Thus the amount of wind generation has risen steeply in recent years, notably in the
countries of northern Europe. However, this rise in wind power has lead to issues regarding the
variability of the wind power output. Wind power is related to the wind speed, which varies
greatly. This variability can cause issues with wind operators' ability to participate in electricity
markets and can also lead to a rise in balancing costs.
The system proposed in this thesis aims to reduce the variation of wind farm output seen in
the minute to minute time-scale and provide controllability in longer time-scales. To do this the
system uses short-term wind speed predictions and the inertial energy storage of the wind turbines
themselves and does so in a co-ordinated fashion across the whole farm. Using short term wind
speed predictions, the amount of energy in the wind is calculated for the next short period. This
energy can be exported in a controlled manner using the inertial energy to cover short-term wind
energy shortfall or excess. The rotor speed must vary for the storage effect to be achieved and this
requires extra control systems to prevent over-speed or turbine stalling.
The system was tested and found to be effective at smoothing the output power in a range of
different wind scenarios. Tests were performed to assess the effects of using co-ordinated control
on the frequency of an example grid and on the use patterns of portfolio generators. Both tests
show that the use of a co-ordination controller at wind farm level reduces the balancing burden on
the remainder of the system in comparison with the common maximum power form of control.