Hands-On Dynamic Simulation in Python with ANDES: Training Data Generation for Grid Foundation Models

Hantao Cui – North Carolina State University

This tutorial introduces ANDES, an open-source Python library for symbolic power system modeling and numerical analysis. ANDES employs a symbolic-to-numeric architecture that automatically generates numerical code and analytical Jacobians from declarative model definitions, enabling rapid prototyping of new device models. The library provides over 100 industry-grade models, including conventional synchronous machines, renewable energy converters, and grid-forming inverter models, and supports power flow, time-domain simulation, and state estimation. ANDES interoperates with MATPOWER and pandapower, allowing existing network models to be extended with dynamics. Participants will learn how to load standard cases, run dynamic simulations, programmatically vary operating conditions and fault scenarios, and extract time-series results into NumPy and pandas for ML workflows. The session uses interactive Jupyter notebooks and assumes no prior experience with dynamic simulation.