Capturing Power System Dynamics by Physics-Informed Neural Networks and Optimization

  title={Capturing Power System Dynamics by Physics-Informed Neural Networks and Optimization},
  author={Georgios S. Misyris and Jochen Stiasny and Spyros Chatzivasileiadis},
  journal={2021 60th IEEE Conference on Decision and Control (CDC)},
This paper proposes a tractable framework to determine key characteristics of non-linear dynamic systems by converting physics-informed neural networks to a mixed integer linear program. Our focus is on power system applications. Traditional methods in power systems require the use of a large number of simulations and other heuristics to determine parameters such as the critical clearing time, i.e., the maximum allowable time within which a disturbance must be cleared before the system moves to… 

Figures from this paper

Neural Networks for Encoding Dynamic Security-Constrained Optimal Power Flow to Mixed-Integer Linear Programs
This paper introduces a framework to capture previously intractable optimization constraints and transform them to a mixed-integer linear program, through the use of neural networks, and demonstrates the approach for power system operation considering N-1 security and small-signal stability.
Closing the Loop: A Framework for Trustworthy Machine Learning in Power Systems
This work outlines five key challenges associated with building trustworthy ML models which learn from physics-based simulation data, and implements methods that connect different elements of the learning pipeline through feedback, thus “closing the loop” between model training, performance assessments, and re-training.


Physics-Informed Neural Networks for Power Systems
This work unlocks a range of opportunities in power systems, being able to determine dynamic states, such as rotor angles and frequency, and uncertain parameters such as inertia and damping at a fraction of the computational time required by conventional methods.
Physics-Informed Neural Networks for Non-linear System Identification applied to Power System Dynamics
This paper investigates the performance of Physics-Informed Neural Networks (PINN) for discovering the frequency dynamics of future power systems and monitoring the system inertia in real-time and compared with state of the art algorithms to assess its performance.
Stability and feasibility of neural network-based controllers via output range analysis
  • B. Karg, S. Lucia
  • Computer Science
    2020 59th IEEE Conference on Decision and Control (CDC)
  • 2020
This paper introduces a parametric description of the neural network controller and uses a mixed-integer linear programming formulation to perform output range analysis of neural networks and proposes a novel method to modify a neural network Controller such that it performs optimally in the LQR sense in a region surrounding the equilibrium.
Predicting AC Optimal Power Flows: Combining Deep Learning and Lagrangian Dual Methods
A deep learning approach to the Optimal Power Flow problem that exploits the information available in the prior states of the system, as well as a dual Lagrangian method to satisfy the physical and engineering constraints present in the OPF.
Learning Optimal Power Flow: Worst-Case Guarantees for Neural Networks
This paper introduces for the first time a framework to obtain provable worst-case guarantees for neural network performance, using learning for optimal power flow (OPF) problems as a guiding example, and shows how to systematically reduce the worst- case guarantees by training on a larger input domain than the domain they are evaluated on.
Transient Stability Assessment of Multi-Machine Multi-Converter Power Systems
Transmission faults caused by recent wildfires in California have induced the disconnection of utility-scale converters in photovoltaic (PV) power plants. Postmortem investigations reported that
An Active Distribution Network Equivalent Derived From Large-Disturbance Simulations With Uncertainty
A reduced-order, “grey-box” model of an active distribution network, intended for dynamic simulations of the transmission system, is derived and is able to reproduce with good accuracy the discontinuous controls of inverter-based generators, such as reactive current injection and tripping.
Generic Model of Active Distribution Network for Large Power System Stability Studies
This paper presents the development of the dynamic equivalent model of an active distributed network (ADN) based on the grey-box approach. The equivalent model of an ADN comprises a
Verification of Neural Network Behaviour: Formal Guarantees for Power System Applications
Developing a rigorous framework based on mixed-integer linear programming, these methods can determine the range of inputs that neural networks classify as safe or unsafe, and are able to systematically identify adversarial examples.
Lyapunov Functions Family Approach to Transient Stability Assessment
The approach generalizes the idea of energy methods, and extends the concept of energy function to a more general Lyapunov functions family (LFF) constructed via semidefinite programming techniques, and shows that they can certify stability of a broader set of initial conditions in comparison to the energy function in the closest-UEP method.