GridPACK is an open-source high-performance computing (HPC) framework for simulating large-scale electrical transmission systems. Built on MPI, Global Arrays, and PETSc, it enables power engineers to develop parallel simulation applications without managing the low-level details of distributed computing, matrix indexing, or solver integration.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/GridOPTICS/GridPACK/llms.txt
Use this file to discover all available pages before exploring further.
Docker Quickstart
Run GridPACK in minutes using the pre-built Docker image with all dependencies included.
Build from Source
Compile GridPACK on Linux or macOS with step-by-step instructions for all dependencies.
Framework Overview
Understand the core architecture: networks, components, mappers, and solvers.
Application Modules
Use the built-in power flow, dynamic simulation, contingency analysis, and state estimation modules.
What GridPACK does
GridPACK provides building blocks that handle distributed network partitioning, parallel data exchange, and algebraic system construction — so you focus on power system physics rather than HPC infrastructure.AC Power Flow
Newton-Raphson solver with Q-limit handling, switched shunts, LTC transformers, and PSS/E input support.
Dynamic Simulation
Full-Y matrix transient stability simulation with DAE solver and extensive generator/load models.
Contingency Analysis
N-1 contingency screening with automatic slack bus transfer and parallel execution.
State Estimation
Weighted least squares estimation with bad data detection and chi-square convergence criteria.
Build your own applications
GridPACK is also a development framework. Extend it to create custom power grid applications that run on workstations or leadership-class supercomputers.Custom Application Guide
Learn how to create bus/branch components, factories, and mappers for new simulation types.
Python Interface
Use the pybind11-based Python wrappers to drive GridPACK simulations from Python scripts.
Math & Solvers
Linear, nonlinear, and DAE solvers backed by PETSc with a clean C++ API.
Parallel Computing
MPI communicators, Global Arrays, task managers, and distributed data structures.
GridPACK supports PSS/E RAW network formats v23, v33, v34, v35, and v36. Multi-architecture Docker images (AMD64 and ARM64) are available at
pnnl/gridpack:latest.