A project to create a digital twin of an enterprise based on documentation written in natural language
- Our customer
- A large oil & gas company
- Context
- Need for centralized control means over remote assets
- Need to increase predictability of all processes to avoid major damage
- Need to optimize and synchronize technological operations to increase profitability and efficiency
- Industry
- Oil & Gas
Tasks
Solution
Result
- Create a digital twin of the equipment
- Perform mapping of equipment elements based on technical documentation
- Enrich and detail the digital twin using the information obtained
- Implement and train models on pre-annotated data
Tasks
Solution
Result
- Based on technical documentation in natural language, a digital copy of the equipment has been created
- Based on tags in MES systems, data has been compared and digital twins of the equipment have been enriched
- Entities and relationships have been extracted using NLP tools
Tasks
Solution
Result
An analysis of technical documentation in natural language has been conducted, a working digital model of key processes has been created, the accuracy of compliance with technologies and process safety have been increased, visual control has been provided, and labor costs have been reduced. A digital twin of the oil refining unit has been created.
97,3%
Documentation parsing accuracy
up to 20 times
Reduction in processing time
10 times
Reduction in labor costs