LANS Informal Seminar: Amy Wilson

Seminar Title: Statistical modelling for assessing capacity adequacy in an electricity system
Speaker: Amy Wilson, Research Associate, University of Edinburgh

Date/Time: 2017-07-14 10:00
Location: Bldg. 240, room 1404

Description:
The risk of having insufficient electricity generating capacity to meet future demand needs to be assessed to plan for mid to long term investment in electricity generating units. As the generation mix changes from traditional coal and gas fired plant to wind and solar generation, new statistical models are needed to assess this risk. These statistical models need to account for the variability and seasonal effects seen in wind and solar generation, as well as any dependence between demand, wind generation and solar generation.

In this presentation, a brief introduction to assessing the risk of capacity shortfalls in Great Britain will be given. Existing statistical models for assessing this risk will be described and the disadvantages of these models will be discussed. Existing models generally follow the so-called time-collapsed approach, which focuses on modelling the marginal distribution of excess of electricity supply over demand at a randomly chosen point in time. This is adequate for the calculation of expected value risk metrics (such as the expected number of hours of shortfall over a year) but cannot be used to estimate more complicated metrics such as the distribution of the number of shortfalls over a year. Possible methods for the joint time-series modelling of demand, wind and solar to overcome this issue will be explored.