Future frequency response requirements in low inertia grids

Electricity grids around the world are undergoing tremendous transitions due to the pressing need to decarbonise. Great strides have been made in Great Britain in recent years, wind and solar penetration doubled between 2014 and 2019, but there are concerns about the effect this transition will have on grid stability. Ancillary services are the tools that electricity system operators use to help maintain grid stability. One of these, frequency response, is a service with the goal of maintaining a stable grid frequency. Grid inertia is another important property of the grid, and the level of inertia is inversely proportional to the rate of change of frequency. Traditionally, frequency response and inertia have mostly been provided by large fossil-fuel power stations. Given the situation outlined above, the aim of this thesis is to understand the frequency response requirements in low inertia grids, with a focus on the situation in Great Britain. From an analysis of historic grid data, I find that frequency volatility has increased in Great Britain in recent years and the underlying causes are likely high rates of change of demand and settlement period boundaries. The increasing penetration of renewables may have also contributed. Using the swing equation, I find that to secure the GB grid in the future, for all potential infeed losses, demand, and inertia scenarios, 1600 MW of frequency response capacity at a delay/ramp time of 0.5 s/0.5 s is needed. I find that in normal day-to-day operation, the frequency volatility does not drastically deteriorate until an inertia level around 20% of current levels (inertia from nuclear and demand only). At this low level, a significant portion of the frequency response capacity needs to be fast acting for successful mitigation. Low inertia has a much greater effect on frequency response requirements in a large infeed loss situation compared to normal day-to-day operation.