Using real-time data for automated drilling performance analysis

Drilling performance analysis from the operating company view is part of the workflow for E&P drilling activity and stimulates organizational learning. The derived measurements help to improve safety, cost effectiveness and quality of ongoing and future operations. Today, operating companies rely on performance analysis that is primarily based on daily activity breakdowns, which are defined through a large variety of activity coding schemes. However, these activity reports are subjective human observations. This fact implies a number of limitations such as the level of detail that results from the time-consuming data entry process and the influence of individual barriers for learning. To overcome these limitations, this paper investigates the use of process related data measured in real time for performance analysis while and after drilling. It shows that it is possible to automatically derive activities and events from real-time data, just as it is possible to accomplish an understanding of various events, which result in non-optimal performance or trouble time through visual inspection of dataplots. Quality problems with existing real-time data (revealed during post analysis) are discussed, as well as their origin in the historically developed paradigm of a geology-driven, depth-based view of the drilling process. The importance of a time-based view of the continuously ongoing drilling process is stressed, as well as the changing and challenging nature of the rig environment. This setup requires new concepts for data handling, which are able to cope and scale with varying requirements. An extensible and flexible infrastructure for measurement data handling is introduced that scales with the complexity of wells, fulfilling existing analysis and visualization requirements. An example shows a comparison between automated rule based performance analysis and results obtained from traditional reporting.