Challenge

The system was developed for specialists in a large oil company in charge of oil well exploitation and maintenance planning. Before the system implementation, specialists had to do all the operation planning manually, from oil field prospecting to starting extraction operations. Very often the initial plan contained lots of conflicts: one and the same resource was allocated to several operations, time necessary for working teams to get to their working site was overlooked, incompatible operations were planned at the wells at the same time, etc.

Detection of these conflicts and calendar and resource replanning with further verification were not automated. This required a lot of time and effort. Besides, the human factor was not reckoned with: it was hard to detect all conflicts when there were tens and hundreds of planned operations.

But even if a conflict-free plan was produced with all due consideration for all specified restrictions, there was no guarantee it was optimal. This plan allowed meeting deadlines in operations while using the available resources. However, it did not provide for maximal oil recovery from the oil field. That is why the company needed calendar and resource planning automation.

Solution

To optimize planning the following task was completed: a system was developed that facilitated conflict-free resource and calendar planning and ensured maximal oil recovery from a given oil field within a specified period.

In the first place, all the variables concerned were defined, namely, the distances between the objects at the oil field, the travelling speed of vehicles and working teams, oil debit for every well, costs of resources and teams, technological periods between operations, etc. At the next stage all possible restrictions were defined, e. g. operation compatibility, simultaneous operations and correct operation order, roads and their condition, resource and team availibility, etc. The сumulative oil recovery from the oil field was defined as the performance function.

To deal with the task, the users were provided with an interface allowing them to upload all raw data into the system, define calculation parameters and see results in reports. The adjustable calculation parameters included planning starting date, number of available working teams, and valid distances for simultaneous operation. A genetic algorithm was used to solve the task. The user can set its parameters at the task launch.

Results

The key business effect is to bring down company costs thanks to the planning efficiency. The solution provided the users with conflict-free calendar and resource planning with maximum oil recovery. It is also more time- and energy-efficient.

To visualize optimization results report forms were developed. They contained not only the optimal calendar and resource plan, but also such KPIs as cumulative oil recovery, cumulative shortage, cumulative resource off-time, etc.