Sammendrag
The relatively low level of oil prices has in the past years drawn
attention to cost reduction remedies in oil companies. One way of reducing
costs is to improve coordination of activities and resources. The production
of oil offshore involves a substantial number of activities in a set of wells.
These activities have to be coordinated in order to obey various types of
constraints and, at the same time, optimize on criteria such as the production
level. Well Activity Scheduling (WAS) deals with the proactive coordination
of these activities to form Well Activity Schedules, as well as the dynamic
maintenance of these schedules over time.
In a cooperative project affiliated with the Esprit project TRUTH, Saga
Petroleum a.s. and SINTEF Informatics have studied the WAS problem. In the
project we have developed a constraint model, performed problem analysis,
and developed a prototype system for WAS. The prototype system includes
functionality for defining a particular WAS problem and generating a high-
quality, feasible solution. The user may inspect and modify the generated
schedule through interactive Gantt visualizers. Constraint violations are
detected and presented to the user. The WAS prototype has been developed
using a commercial constraint programming tool. The prototype is running
on Unix/X11 and PC/Windows platforms.
Preliminary empirical investigations using the WAS prototype system on
historical data indicates a high potential for improvement, although the
current system is based on relatively straightforward heuristics and no
automatic optimization facilities.
attention to cost reduction remedies in oil companies. One way of reducing
costs is to improve coordination of activities and resources. The production
of oil offshore involves a substantial number of activities in a set of wells.
These activities have to be coordinated in order to obey various types of
constraints and, at the same time, optimize on criteria such as the production
level. Well Activity Scheduling (WAS) deals with the proactive coordination
of these activities to form Well Activity Schedules, as well as the dynamic
maintenance of these schedules over time.
In a cooperative project affiliated with the Esprit project TRUTH, Saga
Petroleum a.s. and SINTEF Informatics have studied the WAS problem. In the
project we have developed a constraint model, performed problem analysis,
and developed a prototype system for WAS. The prototype system includes
functionality for defining a particular WAS problem and generating a high-
quality, feasible solution. The user may inspect and modify the generated
schedule through interactive Gantt visualizers. Constraint violations are
detected and presented to the user. The WAS prototype has been developed
using a commercial constraint programming tool. The prototype is running
on Unix/X11 and PC/Windows platforms.
Preliminary empirical investigations using the WAS prototype system on
historical data indicates a high potential for improvement, although the
current system is based on relatively straightforward heuristics and no
automatic optimization facilities.