Multipole Array Acoustic Logging Data Processing Method Based On Rough Set Theory

Sonic logging is one of logging methods developed in recent years. Sonic logging has a very good foreground when it is combined with geologic reconnoissance, especially in respects of lithology judgment, abnormal pressure layer recognition, crack estimation and judgment of liquid characteristic in store layer.The topic of this thesis is based on the project of "an instrument product of digital array sonic logging" taken by university of electronic science and technology of China, which is supported by sea oil company of China. In the thesis, an expatiation of the basic theory and technology of muti-pole sonic logging is given firstly. Then a kind of selection and reduction of significant information features of multi-pole sonic logging based on rough sets is presented. A new approach of data processing and recognition based on fuzzy clustering of rough sets is also put forword as theoretical foundation to process the continued data of muti-pole sonic logging, lithology judgment abnormal pressure layer recognition, crack estimation and judgment of liquid character in store layer.The main works in the thesis are as follows:1. Based on rocky acoustic characteristic, a theory of sonic speed logging, sonic amplitude logging, long-distance sonic logging is intrduced and some main techniques of muti-pole sonic logging are analysed.2. By means of rough sets, a basic conceptioal introduce of information system of muti-pole sonic logging is interpreted, then a kind of improvement algorithm of system parameter's importance is put forward. The reduction of knowledge and presentation of system is investigated by some examples. Thus a new approach of reduction of attributes of multi-pole sonic logging data based on rough sets is presented in the thesis finally.3. Combined with fuzzy cluster techniques, a kind of improvement algorithm of decision-making rules is investigated and a new method of system reduction by information coverage is proposed through some examples. At last, a new system of decision-making rules in multi-pole sonic logging data mining is set up.