Abstract
Abstract
It is widely known that data associated with oil well drilling is often noisy and otherwise of bad quality. It is also known that
solutions need to draw on both people, processes and technology to make traction. However, both the severity of the matter
and the complex causes of bad data are still not completely understood. In this paper we offer a new vantage point, by
presenting the data quality issues as they appear in drilling operations which utilize a real-time wellbore model for supervision
and decision support. We summarize experience from several pilot studies carried out by SINTEF and the Center for
Integrated Operations in the Petroleum Industry, together with industry partners. We find that bad data quality is not only a
cost driver, but a serious drilling hazard in its own right.
It is widely known that data associated with oil well drilling is often noisy and otherwise of bad quality. It is also known that
solutions need to draw on both people, processes and technology to make traction. However, both the severity of the matter
and the complex causes of bad data are still not completely understood. In this paper we offer a new vantage point, by
presenting the data quality issues as they appear in drilling operations which utilize a real-time wellbore model for supervision
and decision support. We summarize experience from several pilot studies carried out by SINTEF and the Center for
Integrated Operations in the Petroleum Industry, together with industry partners. We find that bad data quality is not only a
cost driver, but a serious drilling hazard in its own right.
