Making it possible to use paper archives digitally

For our customer OGE (Open Grid Europe GmbH), one of Europe’s leading transmission network operators, we have made it possible to use paper archives digitally, thereby enabling effective planning for the conversion of the transmission network.

OGE faced the challenge of testing its gas transmission network for hydrogen suitability. In order to be able to assess the suitability, the customer needs to analyse the entire technical documentation of all installed materials. The data is only available as scanned documents, some of which date back to 1940. This includes manufacturer information, component lists, material components as well as technical and chemical properties.

Despite the heterogeneous source material, AI applications make it possible to automate large parts of the document evaluation. The key functions are key value detection and table extraction. AI also adds a great deal of value when it comes to text recognition, especially in the case of poor quality scans or documents, or forms with handwritten entries. Experts use key value detection, in particular, to extract specific information.

The AI processes used ensure that the data from the scanned PDF files are available in a structured format. Those responsible for the AI processes developed a data scheme for the information to be extracted and built a database. Internal systems from various departments have access to the database and process and evaluate the data. In addition, all documents are fully searchable via a search interface, which means that even information that was not extracted can be found easily.