Performance Optimization
Predictive Model for Drilling Problem Anticipation
- Client
- Major Oil Operator
- Country
- Italy
- Year
- 2015-2016
- Industry
- Oil & Gas
- Service
- Industrial Data and AI, Operations & Asset Performance
When it comes to cutting drilling time and cost, predicting Non-Productive Time (NPT) is the best strategy. In a pilot project for a major oil operator, kwantis set out to teach real-time monitoring to recognize patterns in rig site data and warn of problems before they cause downtime.
The challenge
NPT prediction is one of the most topical issues under study for oil and gas operators. Existing real-time monitoring already helps solve operating problems from remote locations.
The huge amount of data coming from rig site sensors, combined with big data analytics, can bring a big improvement in operation monitoring, for both economic and safety reasons.
Our approach
kwantis set out to enhance real-time monitoring by teaching the system to recognize trends and patterns, so that downtime can be avoided by anticipating problems.
The pilot project starts from pipe sticking, one of the most common and risky well problems. For it, kwantis designed an Operational Intelligence solution that:
- integrates different data sources
- combines them on a time and depth basis
- examines and explores them to find trends and patterns
The outcome
The patterns flag an upcoming issue and raise a warning that suggests adjusting the drilling parameters, to avoid NPT.
Once the pilot project proves feasible, Operational Intelligence could be extended to other well problems and applied to both exploration and development wells, across all wells worldwide.
