To main content

Decision Support for Daily Production Optimization: In Genuine Real-Time


The presence of a complex bottleneck structure makes it non-trivial to find the best production strategy, and the competence of the production engineer makes a big difference. Over the last decade, the number and quality of computer based tools to support the production engineers in their decision, has increased exponentially. The contribution from such tools could be divided in two.
The current state of the system - It exists valuable information about the current state of the production system that are not easily accessible through measurements. The prime example is the flow of oil, gas and water from each well. A software able to estimate flow rates from each well with acceptable accuracy real-time, are of high value to the production engineer.
Improve the systems performance - Further, the largest value of knowing the current state of the production system, is that it acts as a platform for improvement. However, knowing your current state is not enough. One also needs to know how e.g. the change in a choke valve affects the bottlenecks of the system. These effects could be complex, and only relying on human interpretation when searching for the best production strategy is dangerous. Computer based optimization systems that are able to assist the production engineer in this decisions, are therefore of high value. Unfortunately, the time needed to compute recommendations has until know been prohibit high for efficiently contributing to the production engineers workflow.
This paper discusses challenges and opportunities related to development and use of such advanced decision support tools. It highlights the potential of fast response times that enables totally different workflows, since instant re-optimizations could be conducted on-the-fly during e.g. a well optimization meeting. The Petrobras P-35 FPSO, equipped with IO Center and Petrobras optimization tools, is used as an illustrative case example.


Academic chapter/article/Conference paper




  • Stine Ursin-Holm
  • Anders Thoresen Sandnes
  • Vidar Gunnerud
  • Bjarne Anton Foss
  • Andrew R. Conn
  • Federico Zenith
  • Alex Furtado Teixeira


  • Norwegian University of Science and Technology
  • IBM Research
  • SINTEF Digital / Mathematics and Cybernetics
  • Petrobras




Society of Petroleum Engineers


SPE Intelligent Energy Conference and Exhibition, Nederland, 1-3 April 2014 - Conference Proceedings




457 - 463

View this publication at Cristin