基于WT与LSSVM的储层油水识别方法研究

为了依据测井数据迅速而准确地识别油水层性质,提出基于小波变换(WT)与最小二乘支持向量机(LSSVM)相结合的储层油水识别方法.首先,分析了基于测井曲线的常规交会图油水识别方法的不足;其次,以深侧向探测电阻率作为训练样本和测试样本对基于LSSVM的储层油水识别模型的准确性进行分析;最后,将WT与LSSVM相结合建立储层油水识别模型并结合常规测井资料对储层油水进行识别.实验表明,基于WT与LSSVM的储层油水识别模型具有较快的识别速度和较高的识别精度.因此,将WT与LSSVM相结合应用于储层油水识别建模是行之有效的.

Study on oil/water layer identification method based on WT and LSSVM

In order to rapidly and accurately identify oil and water layers,it is proposed that combining wavelet transform(WT) with least squares support vector machine(LSSVM) is used for the identification of oil and water layers.At first,the shortcomings of conventional diagram oil-water recognition method based on logging curves are analyzed.Secondly,the accuracy of reservoir oil-water recognition model based on LSSVM is also analyzed by using deep lateral resistivity logging data as the training samples and test samples.At last,a reservoir oil-water recognition model based on the combination of WT with LSSVM is proposed,and the recognition result is combined with the conventional logging data recognition result to distinguish the reservoir oil and water layers.Experiment results show that the new model can obtain high recognition speed and high recognition accuracy,and it can be used for effectively recognizing reservoir oil and water layers.