Introduction

PICurv: A Hybrid Eulerian-Lagrangian Framework for Scalar Transport

PICurv is a parallel CFD and particle framework for incompressible flow and scalar transport. It combines:

Eulerian flow solves on curvilinear grids (PETSc DMDA based)
Lagrangian particle transport (DMSwarm based)
A YAML-driven run pipeline through ./scripts/pic.flow

The user workflow is configuration-first: YAML inputs are validated and translated into generated runtime artifacts consumed by the C solver and postprocessor.

Documentation Entry Points (Structural)

Start here first:

Getting Started: fast first-run path for installation, first case setup, validation, and first outputs.

Then use the structural map and reference pages:

Documentation Map: complete categorized index by workflow, artifact, and runtime layer.
Anatomy of a Simulation: end-to-end run structure and generated artifacts.
Configuration Contract (YAML -> Generated Artifacts -> Runtime): YAML contracts and Python-to-C handoff.
Methods and Models Overview: governing methods and model map.
C Runtime Execution Map: solver execution order and C runtime touchpoints.
**Data Structures (annotated_structured.html)** and **File List (files_structured.html)**: organized code-reference views by module and file type.

Quick Preview

PICurv simulation preview

Example visualization output (ParaView)

Core Methodology

PICurv uses a two-way coupled Eulerian-Lagrangian strategy:

Eulerian phase: velocity/pressure evolution on a structured curvilinear grid.
Lagrangian phase: particle motion and particle-carried quantities.
Coupling:
- Grid -> particle interpolation for advection/source evaluation.
- Particle -> grid projection for reconstructed Eulerian particle-derived fields.

Key Features

YAML-driven pipeline (case.yml, solver.yml, monitor.yml, post.yml)
Cluster scheduler integration (cluster.yml) for Slurm job generation/submission
Study/sweep orchestration (study.yml) with Slurm arrays, dependency chaining, and metric plots
Multiple grid ingestion modes: programmatic_c, file, and grid_gen
Named momentum solver strategy with solver-specific option blocks
Structured post-processing pipelines (Eulerian, Lagrangian, statistics)
Dedicated config contract and ingestion mapping docs for maintenance and extension

For release notes and recent contract changes, see Changelog. For solver-method overviews, see Methods and Models Overview.