انتقال تجربه برای بهبود فرایند

The oil well drilling process is the selected representative of a challenging industrial process. The drilling process is becoming more complex as oil fields mature and technology evolves. At the same time, the amount of information is increasing in volume and frequency. Although technology is advancing, failures occur at almost the same rate as before, leading to loss of valuable time. Whenever the process is failing, or running smoothly, valuable experience is gained. To take advantage of established and continually growing new experience a formalized methodology, knowledge intensive case-based reasoning, was applied for capturing of drilling process experience and for reusing it. Experience was collected from different information sources. Structured cases were used to describe failure episodes; its circumstances and how the failure was repaired. A general domain knowledge model supports the case-based reasoning process. It was demonstrated how the system was able to recommend how to solve problems when they arise, while at the same time bridging the gap between new and experienced personnel. Method performance was tested on 62 selected field cases. The system also identified the failure causes of problems and could thereby suggest more effective repair actions.

A majority of the remaining oil and gas reserves are located inside the continental shelves. Offshore drilling operations are very expensive, and numbers of wells are kept as low as possible. For this reason offshore wells tend to be long and complex. This is also the explanation of why the number of process failures does not diminish much over time. In spite of technology evolution, complexity also increases. A failure during drilling operations is defined as the state of a process when non-productive time (NPT) is occurring. NPT is typically in the range of 15% of total rig time, but can become much higher during drilling of difficult wells (Halliburton Solving Challenge, 2012). The motivation behind the work presented here is to advance computerized methods for helping the petroleum industry in reducing unwanted downtime. Another worry in the industry which supports our motivation is the knowledge gap created by “The Great Crew Change” that exists in most companies (e.g. (McCormack, 2010)). The problem is not one of just filling the gaps. There are sufficient numbers of people entering the workforce to do that. The problem is one of “experience attrition”. The immediate challenge is therefore how to transmit the soft and hard skills necessary to quickly bridge the gaps between new and existing personnel. Companies want a measurable return on investment. They want to achieve a reduction in accidents, an improvement in oil and gas measurement yield, and fewer lost days of production. The ultimate goal of our research is to improve process quality and efficiency by systematically capturing useful human experience during a drilling process and make relevant past experience available on-line when needed. This calls for skills to be transportable between companies and among different industries. Case-based Reasoning (CBR) is a methodology that enables computer systems to assist in achieving these tasks. The goal of the work described in this paper is to investigate how knowledge transfer in the drilling process can effectively be realized by combining the re-use of situation-specific experiences (cases) with justifications and explanations generated by a general domain model (ontology). The combined approach is referred to as knowledge-intensive CBR. We have developed an experimental system that is able to read data from a drilling process and capture interesting parts of it (Aamodt et al., 2012). After having captured and stored human experience accompanied by technical information, we demonstrate different ways of re-using the experience, and point out especially two applications in larger detail: • Determine the optimal repair of a problem • Determine the failure cause of a problem We have studied a novel combination of two technologies that have proved to work well in other industries: Case-based reasoning (Cheetham and Watson, 2005) and Ontologies (Obrst et al., 2003). This combination has also been applied to other problems in the oil and gas domain (Kravis and Irrgang, 2005). Case-based reasoning is a method and a technology for solving new problems in complex processes by comparing a new problem to previous situations, and reusing the experience from the most similar situations in solving the new problem (Aamodt and Plaza, 1994). CBR can be described as adoption of common human problem-solving behavior for computer use (Popa et al., 2008). In practical systems, the technology incorporates different types of information, which empower the system to learn and adapt from an ever-growing case base of new experiences. In order to set our method in a wider perspective, in Section 2 we discuss existing approaches to experience transfer. In Section 3 follows a description of how the case knowledge and general knowledge are represented, and the various sources of information and knowledge are described. The quality of the CBR systems, in terms of finding relevant past cases, has been tested on field data in a small scale, and the results are presented in Section 4, including two applications. A discussion of the results and future plans concludes the paper.

An on-line software tool has been developed and evaluated on the basis of historical real-time drilling data. The results are promising. The system has the capability of recognizing episodes and relating them to situations that have occurred before. The episodes are stored as cases consisting of three separate parts: circumstantial information and gained experience, explanation of why the situation arose and how it was handled, and the outcome of the action taken. • Experience embedded in a case is retrieved not only from data streams and documents, but also from the user and other experts during case generation, ontology building, and case evaluation. • Cases representing complex domains like oil well drilling operations need to be related to a rich ontology since the case space is small. Concept expressing experience is translated into a symbolic language and then stored in the ontology as interrelated entities. • Initial real-time runs have proven the tool's functionality and pointed out potential user support. Tests have showed that the CBR system matched and retrieved cases of the correct class to an acceptable degree, even with the sparse data available for the experiments. Applying the knowledge model is an alternative method of revealing the correct failure cause. This will improve the CBR-method and enhance the quality of advises given.