火山岩岩性识别方法研究

岩性是测井储层评价的基础，在火山岩油藏的测井评价过程中，由于火山岩岩性定名方法的不统一以及岩性识别方法的不完善，使火山岩油藏的测井评价难以开展。采用国际地科联（IUGS）推荐的TAS图方法对火山岩岩样定名,并以此为基础，使用神经网络方法利用测井资料识别火山岩岩性。研究结果表明，利用传统的统计判别方法建立的测井资料与火山岩岩性的判别关系，识别符合率只有65%;而使用神经网络方法建立的判别关系，识别符合率可达81.8%。

A Study on Lithology Identification Methods for Volcanic Rocks

The lithology identification is a basic work in reservoir evaluation by well logging data. It is difficult to evaluate volcanic reservoirs with well logging data because volcanic lithology classification is not unanimous and the recognition methods are not perfect. Volcanic samples should be classified according to TAS (total alkalies vs silica) chart proposed by IUGS. The volcanic lithology can then be recognized by neural network method with well logging data. Only 65 percent volcanic rocks can be identified by conventional methods. With the method proposed, 81.8 percent rocks can easily be recognized.