Inclusion of frequency stability constraints in unit commitment problems, two approaches: Analytical and Machine Learning

The unit commitment problem is an optimisation problem for determining the generating resource dispatch in a power system ahead of real-time operation. Yet, the classic form of this problem does not include constraints related to frequency response of the system. This means that the resource dispatch schedule may not necessarily be secure against severe frequency excursions caused by loss of generation events. This is more pressing in low-inertia systems (systems with high levels of renewable generation). In this talk I will give a high-level overview of our recent interdisciplinary research into developing an analytical model for including frequency stability constraints into the unit commitment problem. This is the research conducted in the Department of Engineering with collaboration with the Department of Mathematical Sciences. Our model is purely analytical, however most recently we have been working with a team of researchers in IIT Comillas to develop an alternative approach by using Machine Learning to characterise the same constraint for Islanded systems with low inertia. I will end this talk by briefly touching on this most recent development area.  

Prof Kazemtabrizi's principal research interests are in advanced energy system modelling, for improving operability and reliability of future flexible energy systems. His past and present research encompasses a wide area of topics including but not limited to research in energy systems integration, reliability and optimisation of power systems, and large-scale wind energy integration.

To book a place at this event please visit our EventBrite page.