Skip to content

Creep Literature Knowledge Graph (CreepLitKG)

Version 1.0.0

CreepLitKG is an LLM+Ontology-driven pipeline and knowledge graph, that converts creep test metadata extracted from scientific literature into ontology-grounded RDF. It makes experimental creep data reported in publications — material identity, chemical composition, heat treatment history, microstructural features, test conditions, and creep results — findable, machine-readable, and queryable through a public SPARQL endpoint, and federates it into the MatWerk Knowledge Graph (MSE-KG).

Live resources

Resource Link
RDF dataset (MaterialDigital Dataportal) https://dataportal.material-digital.de/dataset/creep_literature_knowledge_graph
SPARQL endpoint https://dataportal.material-digital.de/dataset/a5b4edc4-43ef-44ff-a386-5d1f6fbbc439/fuseki/$/sparql
Guided query UI (Sparklis) Open Sparklis on the endpoint
Source code github.com/HosseinBeygiNasrabadi/Creep_Literature_Knowledge_Graph

What is in the graph?

The knowledge graph contains creep datasets extracted from literature. Each dataset describes one creep test reported in a publication and covers:

  • the source publication (DOI),
  • the material (name, chemical composition, sample identifier),
  • the processing and heat treatment history (manufacturing method, solutionizing, aging),
  • microstructural features (grain size, precipitate fractions and sizes),
  • the test conditions (testing standard, temperature, initial stress),
  • and the creep results (stress rupture time, percentage elongation after creep fracture, steady-state creep rate, stress exponent, activation energy, and further ISO 204 extension parameters).