Petrobras, a Brazilian oil company, developed an expert system to help the special operations called fishing. Knowledge was acquired from retiring experts to diagnose the cause of the problem and suggest possible treatments and was represented on sets or rules. The system is delivered on Nexpert/Toolbook/MS-Windows PC platforms. The treatments operations are now being detailed using hypermedia features.
Fishing operations are done to liberate a stuck pipe (or part of it) or to take out any object fallen or left in a well. They are non-scheduled, are done to correct an accident, are costly and undesirable, and consume many rig hours (lost time), so the first step is to avoid them. Specialized experts are needed to perform it (the fishers). Occurring the accident, it must be solved with efficiency and as soon as possible. In order to perform this, there is the necessity of new techniques and technologies absorption, periodics courses for new fishers and specialization and recycling for the more skilled people. In this process there is a knowledge based system development for diagnosis and treatment procedures according to the type of accident, aiming to spread the knowledge and to accelerate the formation of a new fishermen generation.
Petrobras is the Brazilian state-owned oil company, operating in the down and upstream segments, producing almost 900.000 bbl/day. The Special Operations Sector (GESEP) subordinate to Exploration and Production Department is the responsible organ for coordination of fishing operations. It is settled in Rio de Janeiro and connected the 6 regional groups (Belem, Mossor, Aracaju, Sao Sebastiao do Passe, Sao Mateus and Macae). In 1995 there were 63 fishing occasions on drilling wells and 469 on production wells.
Artificial Intelligence is the part of computer science that designs and develops intelligent computable systems. That is to say, that can "think", solving problems in intelligent considered form if accomplished by man.
Knowledge Systems is an Artificial Intelligence branch and their techniques allow to represent the human knowledge in data structures. Therefore they support the decisions taken in a specified knowledge area, in analysis of alternatives and diagnosis elaboration, in face of incomplete, qualitative or distorted data. They are useful tools for advisement less skilled experts in activities that require much practice associated to the knowledge dominion capacity of high quality that have well defined boundaries.
Commercial use of knowledge based systems (KBS) is spreading fast in Brazil, especially with financial applications. Engineering applications are still scarce. Petrobras is one of the pioneers and this project is the principal in the Company.
Since for a while GESEP was wishing to start the development of a knowledge based system for fishing area. Among the main objectives we can stand out:
To Preserve Expert Knowledge. The success of a fishing operation depends a lot on specialized, experimented and competent professionals. The formation of these experts takes many years and the knowledge dissemination among them is hard, since the operations are scattered in time, sometimes very different of the usual kind and executed under great stress. Also, their work/off-work schedules keep them apart, especially on sea rigs. Moreover, by their very operational nature, most of the fishers are not used to do adequated and detailed technical reports of their operations.