Increasing penetration of renewable energy and energy storage in modern power systems shifts power generation from synchronous to inverter based, introducing instabilities. Grid forming converters offer solutions through local grid supporting functions and by reducing the dependence on the grid in establishing local references.

This project will use machine learning (ML) based methods to achieve multi-objective optimization of control parameters in grid forming converters. Specifically, through the use of a deep reinforcement learning (DRL) assisted framework optimal parameters of grid forming grid forming implementations that meet operational requirements on dynamic performance while also considering stability of the grid forming converter under both strong and weak grids. Analysis will be based on modelling, simulations and implementation of control hardware in the loop.


Electrical Engineering and Telecommunications

Research Area

Energy systems | Power electronics | Real-time simulations | Renewable energy | Power Systems

The student will be hosted at the Real-Time Digital Simulations Laboratory (RTS@UNSW) in the Tyree Energy Technologies Building. He will work together and be supported by a dynamic team of 10 academics, researchers and PhD students specialising in Power Electronics and Real-Time Digital Simulations. He will also receive training on real-time simulation methods and the development of control hardware in the loop solutions.

The expected outcomes of the project are:

  1. Complete implementation of DRL for optimisation of parameters in a grid forming converter.
  2. Training of an "actor" for direct implementation to grid forming converters.
  3. A control hardware in the loop demonstration of the actor and optimisation of converter parameters.
Empty profile image


View Profile

S. Jiang, Y. Zeng, Y. Zhu, J. Pou and G. Konstantinou, "Stability-Oriented Multi-Objective Control Design for Power Converters Assisted by Deep Reinforcement Learning," in IEEE Transactions on Power Electronics, doi: 10.1109/TPEL.2023.3299979. From <>