基于不同测井曲线参数集的支持向量机岩性识别对比

对测井曲线提取的不同参数值,进行支持向量机(support vector machines,SVM)岩性识别对比,找出适合于岩性识别的测井参数集。利用取心段岩心和测井资料,依据M-N值法、曲线重叠法和测井曲线特征值法等3种常用的岩性识别方法,提取能够用于岩性识别的测井曲线参数值,再利用支持向量机进行岩性识别,将识别结果进行对比,按照误差最小原则,找出适合于岩性识别的测井参数集。结果显示,M-N值法、曲线重叠法和测井曲线特征值法3种不同的测井参数集在岩性识别时,测井曲线特征值法和重叠法误差最小,可优先作为识别的基础数据。

Lithology identification methods contrast based on support vector machines at different well logging parameter

To finding the suitable well logging parameter set can use lithology identification through collecting the different well logging parameter set and contrasting the lithology identification results based on one against all support vector machines(SVM) and collected parameter set from well logging. Based on the coring well and well logging data,according to three methods,including M-N value,curve superposition and curve characteristic value,which are often be used on lithology identification,three different well logging curve parameter sets were collected,joining with SVM,lithology identification was fulfilled,after that,to selecting the best well logging parameter set that suitable to used on lithology identification according to error minimum principle through contrasting the results.Two of three different parameter sets indicated the error minimum characteristic on the process,those are curve superposition value and curve characteristic value.The parameter sets of curve superposition value and curve characteristic value methods can be the preferable fundamental data to used on lithology identification from well logging.