Revealing the Cause behind Cement Failures by Means of the Knowledge Model of Oil Well Drilling

Abstract

Safety improvement is a topic under a continuous focus offshore. To reduce the number of accidents, improvements are approached in several ways. Accident reports are analyzed, human factors are studied in depth and the companies are investigated, after the incident has occurred. An improved method of evaluation is now under consideration, which has the potential of being run in real-time and obtain an immediate pre-assessment of the failure. This method is based on a knowledge model of the drilling process, where all concepts involved in the process are structured in hierarchical categories, based on established knowledge combined with situation specific experiences. The concepts are clearly defined and related to other concepts if there exist any causal relation between them.

Cementing issues is one of the greatest challenges regarding well integrity. By applying the knowledge model and expand it to encompass the details of the cementing process, an ability to recognize progressing downhole problems is created, and symptoms seen at the surface can be related to all relevant types of cementing problems. This allows for preventing actions to be initiated at an early stage.

In this report a knowledge model for cementing issues has been constructed based upon textbook knowledge, and further expanded by analyzing five incidents where the cement job failed. The incidents were later tested by means of the knowledge model to find the most probable cause of the failures. Symptoms interpreted from investigation reports were used as input for the model which revealed the main cause and the most probable underlying causes of the incidents.