Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Scalable Physics-Informed Neural Networks for Accelerating Electromagnetic Transient Stability Assessment

Ignasi Ventura Nadal, Mohammad Kazem Bakhshizadeh, Petros Aristidou, Nicolae Darii, Rahul Nellikkath, Spyros Chatzivasileiadis

TL;DR

The paper tackles the computational burden of EMT-based stability studies in grids with heavy inverter-based resources. It introduces a modular PINN framework that targets bottleneck components, notably replacing a nonlinear PLL in a type-4 wind turbine EMT model with a fast surrogate, achieving 4-6x speedups validated against PSCAD. Training blends data-driven and physics-based losses to ensure accuracy across a wide operating domain, enabling case-independent reuse. The results demonstrate substantial speedups while preserving fidelity, supporting scalable, plug-and-play acceleration for EMT (and potentially RMS) simulations in modern power systems.

Abstract

This paper puts forward a framework to accelerate Electromagnetic Transient (EMT) simulations by replacing individual components with trained Physics-Informed Neural Networks (PINNs). EMT simulations are considered the cornerstone of transient stability assessment of power systems with high shares of Inverter-Based Resources (IBRs), and, although accurate, they are notorious for their slow simulation speed. Taking a deeper dive into the EMT simulation algorithms, this paper identifies the most computationally expensive components of the simulation and replaces them with fast and accurate PINNs. The proposed novel PINN formulation enables a modular and scalable integration into the simulation algorithm. Using a type-4 wind turbine EMT model, we demonstrate a 4--6x simulation speedup by capturing the Phase-Locked Loop (PLL) with a PINN. We validate all our results with PSCAD software.

Scalable Physics-Informed Neural Networks for Accelerating Electromagnetic Transient Stability Assessment

TL;DR

The paper tackles the computational burden of EMT-based stability studies in grids with heavy inverter-based resources. It introduces a modular PINN framework that targets bottleneck components, notably replacing a nonlinear PLL in a type-4 wind turbine EMT model with a fast surrogate, achieving 4-6x speedups validated against PSCAD. Training blends data-driven and physics-based losses to ensure accuracy across a wide operating domain, enabling case-independent reuse. The results demonstrate substantial speedups while preserving fidelity, supporting scalable, plug-and-play acceleration for EMT (and potentially RMS) simulations in modern power systems.

Abstract

This paper puts forward a framework to accelerate Electromagnetic Transient (EMT) simulations by replacing individual components with trained Physics-Informed Neural Networks (PINNs). EMT simulations are considered the cornerstone of transient stability assessment of power systems with high shares of Inverter-Based Resources (IBRs), and, although accurate, they are notorious for their slow simulation speed. Taking a deeper dive into the EMT simulation algorithms, this paper identifies the most computationally expensive components of the simulation and replaces them with fast and accurate PINNs. The proposed novel PINN formulation enables a modular and scalable integration into the simulation algorithm. Using a type-4 wind turbine EMT model, we demonstrate a 4--6x simulation speedup by capturing the Phase-Locked Loop (PLL) with a PINN. We validate all our results with PSCAD software.

Paper Structure

This paper contains 25 sections, 9 equations, 9 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Accelerating Time-Domain Simulations
  3. The vision
  4. RMS vs. EMT
  5. EMT Simulations
  6. Generic formulation
  7. EMT simulation problem
  8. EMT simulation program
  9. Electrical systems
  10. Control systems
  11. Interface network-control solutions
  12. Accelerating the control systems solution with PINNs
  13. Neural Networks Integration
  14. General formulation
  15. PINN formulation
  16. ...and 10 more sections

Figures (9)

  • Figure 1: Represented is the PINN integration vision for faster simulations. We first identify which components have the biggest computing burden in the simulation, and replace them with a PINN.
  • Figure 2: Companion circuits of an inductance and a capacitance.
  • Figure 3: The two main approaches to solve control systems.
  • Figure 4: Interface overview between the network and control systems solution emtpbookog.
  • Figure 5: PINNs can accurately capture the dynamics of nonlinear closed-loop control systems.
  • ...and 4 more figures