数学距离在气测录井资料解释中的应用

Mathematical distance in gas logging data interpretation

基于数理统计的原理,针对气测录井获得的全烃和烃组分数据集,利用多元逐步线性回归方法、多种空间量度关系计算方法以及组间判别临界值计算方法建立了流体预测模型。应用结果表明,天然气各组分间的相对关系对流体识别更为重要,不同量度方法对引入的组分参数有一定影响,识别精度也有不同;由于空间量度计算方法可选,使预测模型具有了更大的灵活性和适应性。

Based on the principles of mathematical statistics and according to data sets of total hydrocarbon and hydrocarbon components′ that were obtained from gas logging, the fluid prediction model was established by using the method of stepwise multiple linear regression, and calculation methods for a variety of space measurement relationship and determining the critical value between groups. The results showed that the relative relationship among components of natural gas is more important for fluid identification, different measuring methods have certain effects on component parameters, the recognition accuracy is also different; the prediction model has greater flexibility and adaptability because of the optional space measurement method.