Prof. Juan Yu

Chongqing University

Research Area

Big data analytics, power system analysis

Title

Data-driven Methods for Operational Reliability Assessment of Power Systems Integrated with High-penetration Renewables

Abstract

With a sharp increase of renewables, the strong uncertainty has become a prominent feature in power systems. Online operational reliability assessment and risk warning becomes an important tool that prevents serious power outages. However, existing model-driven algorithms for operational reliability assessment cannot match the online high-accuracy computational requirements in power industries.

To deal with the above issue, this presentation introduces data-driven methods for operational reliability assessment of power systems integrated with high-penetration renewables. The content contains the following three aspects. a) A data-driven method for operational reliability assessment based on deep neural networks is introduced. b) A solution method for operational reliability based on model and data dual-driven deep neural networks is introduced. c) Transfer learning technologies suitable for variant scenarios are introduced, which can be used to update deep neural networks online and guarantee the computational accuracy.

Focusing on the computational efficiency of the operational reliability assessment in power systems that contain high penetration renewables, this presentation lays a theoretical foundation to support practical applications of an online high-accuracy operational reliability assessment.