Motivation

NPAP combines two main steps: network partitioning and aggregation. In the first step, nodes are grouped into clusters based on criteria such as geographic proximity, electrical distance, or structural similarity. A range of clustering methods is available, including K-Means, K-Medoids, DBSCAN, HDBSCAN, and hierarchical clustering.

In the second step, these clusters are aggregated into representative nodes. During this process, properties such as load, generation capacity, coordinates, and line parameters are systematically combined. NPAP provides predefined aggregation methods as well as the flexibility to define custom aggregation strategies.

The library integrates seamlessly into Python-based workflows and works directly with NetworkX graphs as well as tabular data formats such as CSV files. This makes it easy to incorporate NPAP into existing modeling and analysis pipelines.

Getting started with NPAP is straightforward. The library can be accessed via its GitHub repository and integrated into existing Python projects. The documentation provides step-by-step examples, guiding users from loading a network to selecting a partitioning strategy and performing aggregation.

• GitHub Repository
• Documentation

For a first workflow, users typically load a network, choose an appropriate partitioning strategy, and then apply aggregation. The modular design of NPAP allows for easy comparison of different strategies and adaptation to specific modeling needs.