Applying an Information Retrieval System to Well Data

Abstract

Abstract From the time a well is drilled until it is ultimately abandoned, data "flow" from the well-regardless of any real fluids which may or may not flow. One of the most powerful tools to exploit this immense quantity of complex and valuable data is an information retrieval system. Building and using a well data system requires a merging of the user's needy and modern computer technology. Costs of using the system can be minimized only by adapting the system to the people who use it. Introduction The history of "unpleasant theorems" in mathematical logic has recently been summarized by Brownowski. Brownowski started with Godel's 1931 finding that "a logical system which has any richness can never be complete, yet cannot be guaranteed to be consistent", and ended with Tarski's even more restrictive assertion of 1956 that "every formal language which is at least as rich as arithmetic contains meaningful sentences that cannot be asserted to be either true or false". In other words, we can erect a logical system, but it can never be complete or consistent. That is the modem verdict of mathematical logic. In spite of this harsh verdict, the advent of highspeed digital computers has seen the development of information retrieval systems which do exactly what these theorems state can never be completely and consistently done. Each information system consists of a body of information stored in some order and a method of questioning the body of information to extract answers. No doubt these theorems of mathematical logic apply even here. A particular answer may be incomplete or inconsistent. This might preclude further interest by the strict mathematical logician. The pragmatic engineer or geologist approaches the information system with a more practical viewpoint. He simply wants mostly "good" answers at a cost he can justify. He is ready to trust and use his experience and creativity to recognize the incomplete and inconsistent answers as he was doing before modern computers were available. When can the costs of an information retrieval system be justified? One can identify four properties of the body of information treated, any one of which, if present, can justify an information retrieval system. These properties area large quantity of items of information,a high economic potential value to getting the right answer from the information,a high cost of getting the original items of information, andcomplex inter-relationships among the items of information. The following is a brief review of these properties as they apply to petroleum and natural gas wells.A well is a continuous source of vast volumes of data. In fact, even with modern computers, a design problem in a well data retrieval system is devising a way to discard the less useful data to make room for the more useful data.The business of exploring for and producing in-situ hydrocarbon reservoirs is the world's biggest risk industry. risk means the right answer has a premium value.Those who spend money at the well will agree that no item of information at that point is cheap (some contractors and service personnel may not agree). The information flowing from the wells is expensive and should, therefore, be used as much as possible. The concept that expense requires use is not new, but frequently forgotten. Some years ago, major airlines made heavy investments in expensive jet aircraft in the face of some gloomy predictions from the financial experts. But because the jets flew more people in less time, i.e., were used more, airlines have had an unparalleled period of consistent profits.The inter-relationships between items of well data are both series and parallel. There is a series relationship, for example, in DST recovery data. Here, a particular item of data is useless unless correctly connected with other itemsthe rest of the DST data, the interval depth, the formation, etc. For example, there must be a parallel relationship in formation tops for different wells. This completes our case that well data, as a body of information, can justify an economic information retrieval system. This does not mean that any well data retrieval system will be economic. There was a time in the latter part of the 19 century when the consumer market could justify a horseless carriage. But most attempts to satisfy that market failed-not because the justification was wrong, but because the implementation was not acceptable. Some few implementations proved successful. JPT P. 1005ˆ