Advanced Data Driven Solutions for Equinor
Improving efficiency and powering data driven decision making
Partner: Equinor
Description
Applied AI has collaborated with Equinor across several projects to develop machine learning and cloud-based solutions aimed at enhancing their reservoir and well operations. These projects focused on simplifying data usage for both technical and non-technical stakeholders and improving decision-making processes.
One of the key initiatives was the development of a cloud-based benchmarking tool that allows engineers to compare production potential between newly drilled wells and existing ones. This web-based tool, hosted on Azure, integrates seamlessly into Equinor's toolkit, offering a streamlined way to evaluate oil production prospects based on historical data.
Another significant contribution was the creation of a machine learning model to classify well cement logs. This model, trained on historical human interpretations, provides more objective classifications and automatically retrains to ensure accuracy over time. The model runs as a cloud-based service, integrated into Equinor's existing tools for real-time usage, with the entire workflow managed via Azure ML.
Additionally, we developed a system to detect patterns in sensor data related to well production and mechanical risks. By implementing advanced machine learning techniques and ensuring model transparency, this solution supports Equinor’s experts in making informed decisions during drilling campaigns. This AI-based recognition system assists in identifying risks and optimizing production potential during time-critical operations.
Impact
The solutions provided by Simula have enhanced Equinor's ability to make data-driven decisions, improving operational efficiency and safety. The benchmarking tool has become an integral part of their data analysis workflow, aiding in the evaluation of new wells. The cement log classification model has introduced more objectivity and consistency in well log interpretations, contributing to improved accuracy and reliability.
Lastly, the pattern recognition system has supported Equinor in identifying key risks and production opportunities, offering timely insights during critical drilling campaigns. These projects have enabled Equinor to integrate innovative machine learning solutions into their workflows, reducing time spent on manual analysis and improving overall productivity.
Tools and methods
- Cloud based infrastructure
- Explainable AI and Transparency Techniques
- Full-stack development
- Automated model retraining and monitoring (MLOps)
- Advanced data visualisation
Key Personnel
Developed by Simula's Department of Applied AI.