Engineering guide / model interoperability

Power-system model conversion: what gets lost in translation?

Getting a power-system file to open in another tool is not the same as preserving the model you trusted. The costly mistakes can be hidden in equipment semantics, constraints, units, and assumptions.

Full Load Forecasting · Engineering guide

The word “conversion” can suggest a harmless file operation. In electrical network modeling it may involve a transformation between different data models, simulation assumptions, equipment abstractions, and numerical conventions.

Why file-format support is not enough

Power-system tools encode related but nonidentical concepts. A route that correctly imports buses and branches may still handle controls, equipment attributes, transformer conventions, time-series data, and element availability differently. A file that parses successfully is not proof of equivalent simulation behavior.

Five places fidelity can break

  1. Network topology: missing or merged nodes, incorrect endpoints, island detection, and mismapped connectivity.
  2. Units and bases: power units, per-unit conventions, voltage bases, and sign conventions can change numerical interpretation.
  3. Equipment representation: tap changers, shunts, generators, storage, switched controls, and protection-related information may have no direct target equivalent.
  4. Solver semantics: DC versus AC assumptions, treatment of losses, constraints, slack/reference choices, and objective-function meanings differ.
  5. Time and provenance: snapshots, scenarios, source version, original metadata, and rights information can be silently discarded.

“Lossless” is a high bar. It should be limited to specific data models and versions for which a testable equivalence criterion is defined. Loss reports and documented limitations are often more honest and useful.

MATPOWER, PyPSA, and CIM/CGMES are not interchangeable labels

MATPOWER is widely used for power-flow and optimal-power-flow research workflows. PyPSA supports power-system analysis and optimization, including investment and dispatch representations. CIM/CGMES concerns structured information exchange in grid contexts with its own profiles and version requirements. Even when two formats share core network concepts, a specific export route needs version-aware tests.

What should a trustworthy converter report?

EvidenceWhy it matters
Input and target format versionsCompatibility depends on exact schemas and tool versions.
Element counts and connectivity checksQuickly catches obvious omissions and topology changes.
Units and assumptionsPrevents numerical values from being mistaken for equivalent quantities.
Unsupported and downgraded propertiesShows what was dropped or approximated.
Validation fixtures and reproducible resultsProvides a basis for engineering review rather than trust alone.

A practical validation workflow

  1. Record the original file, source program version, and rights to use it.
  2. Validate input integrity and basic structural consistency.
  3. Use a documented conversion process appropriate to the source and target.
  4. Produce a conversion receipt with warnings, mapping and assumptions.
  5. Compare representative outputs using suitable numerical tolerance and solver-equivalent cases where possible.
  6. Retain the source and receipt for audit or exact reconstruction.

Where Model Exchange fits

Full Load’s Model Exchange program aims to make this workflow approachable through format adapters, a Model diagnostics, and a Conversion report. Prototype routes are not yet a publicly certified compatibility matrix. We would rather disclose the limitation than create a false sense of interchangeability.

Explore Model Exchange · Read the digital-twin guide