论文部分内容阅读
渤海海域中生界火山岩广泛分布、岩石类型复杂多样,同时研究区缺少成像、元素等特殊测井资料,岩性识别存在困难.以研究区钻井取心、壁心以及薄片鉴定为基础,探索了基于常规测井资料的火山岩岩性识别方法,并开展了组合应用.结果表明曲线特征法以定性描述为主,受人为因素影响较大;交会图利用较少曲线,难以充分利用曲线信息,识别精度相对较低;决策树法与Bayes判别法综合多条曲线信息进行识别,判别精度较高,但决策树法对单条曲线细微变化响应敏感,岩性识别稳定性较差,而Bayes法综合利用多条曲线识别,判别精度较高的同时具有较强的稳定性.实际应用中通过决策树法与Bayes判别法的组合应用,同时利用曲线形态以及交会图判别结果进行约束,岩性识别精度得到了较大提升.
The Mesozoic volcanic rocks in the Bohai Sea are widely distributed and the types of rocks are complex and diverse. At the same time, there is a lack of special logging data such as imaging and elements in the study area and difficulties in lithology identification. Based on drilling coring, wall center and identification of thin sections in the study area Volcanic lithology identification method based on conventional well logging data and its application in combination show that the curve characteristic method is mainly qualitative description and is greatly influenced by human factors and it is difficult to make full use of curve information to identify The accuracy is relatively low; the decision tree method and Bayes discriminant method can be used to identify multiple pieces of curve information, and the discrimination accuracy is high. However, the decision tree method is sensitive to the subtle changes of the single curve and has poor lithology identification stability. Multiple curve recognition, high accuracy and stability at the same time.Using the combined application of decision tree method and Bayes discriminant in practical application, the curve shape and the cross-plot discriminant results are used to constrain the accuracy of lithology identification A big upgrade.