Forecasting techniques for renewable energy sources: performance analysis and economic impact on grid operation
Project Leader: Maria Vrakopoulou
Collaborators: Prof Howard Bondell (School of Mathematics and Statistics)
Primary Contact: Maria Vrakopoulou (firstname.lastname@example.org)
Keywords: Distribution Networks; renewable energy
Disciplines: Electrical & Electronic Engineering
The recent movement of Victoria on subsidising Photovoltaic (PV) power installations on the household level has been increasing the intermittent and uncertain generation in the distribution systems. The unpredictable nature of high penetration of PV power may rapidly deteriorate both the reliability and economic performance of the system.
This PhD project will develop statistical learning methods appropriate for forecasting PV generation. Existing PV power data with spatial and temporal correlation will be utilised to understand the underlying statistical properties of PV generation. Using these properties, short-term time-series forecasting tools will be built in order to provide useful forecasts and uncertainty measures. The resulted high-performance forecasting methods along with the knowledge of basic household load profiles will allow for a more accurate quantification of active power reserve requirements necessitated by the future PV-rich distribution systems. This will be achieved by analysing the impact of the uncertainty on reserve requirements given the technical constraints and control capabilities of a distribution network. This project is a key step towards the development of operational rules for the future distribution systems.
This is a multidisciplinary project that lives in the interplay of power systems, optimisation, probability and startistical learning theory. Interested candidates with an excellent background in at least 3 out of the 4 fields are encouraged to apply.
The project is fully funded for 3.5 years with around $ 30k net income per year and a performance-based top-up of up to % 5k. The starting date can be flexible.