Detection of Failures and Interpretation of Causes During Drilling Operations

Abstract

Abstract A majority of the remaining oil and gas reserves are located on continental shelves. Offshore drilling operations are expensive, and the numbers of wells are therefrore held low, and they tend to be long and complex. The number of process failures seen in recent years does not exhibit the intended declining tendency. A failure during drilling operations is defined as the state when non-productive time is occurring. The motivation behind the work presented here is to advance a specific computerized method for helping the petroleum industry in reducing unwanted downtime. More up-time is needed. The ultimate goal of our research is to improve the drilling process quality and efficiency. This is achieved by first detecting process deviations (symptoms) during the drilling process, and produce explanations generated by a general knowledge model (ontology). The symptoms have been translated into formal concepts and related into cause-effect relationships with errors and failures. Our experimental system is able to read data from a drilling process and apply the on-line detected and predefined static symptoms to capture a probalistic understanding of the downhole process. We will demonstrate the tool's ability during selected drilling processes to determine which failure type is the most threatening during problem situations and which errors are causing the failure. The failures revealed by the tool on basis of detected symptoms are compared with the reported failure in the field.