Potentials and limitations of using artificial intelligence to predict grouting parameters – Results of a case study in a tunnel project in Scandinavia

Research output: Contribution to journalArticleResearch


  • Christian Thienert
  • Michael Ouschan
  • Frank Könemann
  • Christoph Klaproth
  • Patrick Gabriel
  • Robert Pechhacker

Organisational units

External Organisational units

  • Studiengesellschaft für Tunnel und Verkehrsanlagen e. V.
  • eguana GmbH
  • geoteam Ingenieurgesellschaft mbH
  • Züblin Spezialtiefbau Ges.m.b.H.


Great importance is attached to ‘pressure-volume records’ for the execution, documentation and billing of rock grouting. In this context, special digital data management systems are now available which can provide data in a structured and consistent format that is also suitable for artificial intelligence (AI) approaches. Using datasets from a tunnel project in Scandinavia, this paper shows that artificial neural networks can be used to reliably predict the evolution of pressure-volume records or the volume of grout injected at the end in the interests of construction site efficiency. Taking into account the technical feasibility of using AI to support tunnel grouting, we then show which contractual modifications would be required in order to make effective use of corresponding developments.


Original languageGerman
Pages (from-to)525-534
Number of pages10
JournalGeomechanics and tunnelling = Geomechanik und Tunnelbau
Issue number5
Publication statusPublished - 4 Oct 2022