The identification of interlayer types is the key to improve the geological model to study the characteristics of oil and water distribution. The Donghe Sandstone in Hade Oilfield is used as the research object to calibrate the interlayer type and location by means of the outcrop, core, conventional logging curves and electrical imaging data. However, the logging response characteristics of different types of interlayers are similar, which is difficult to be accurately characterized by the conventional method. Therefore, a Random Forest (RF) intelligent method based on the Ensemble Empirical Mode Decomposition (EEMD) and Hilbert Huang Transform (HHT) is proposed to identify the type of intercalation layer. The model consists of two parts: firstly, EEMD-HHT is used to automatically identify the location of the interlayer, secondly, RF is used to classify the interlayer type. The results show that the EEMD-HHT-RF model corresponds to the results of core identification of interlayer, which can achieve rapid identification and accurate prediction in the block.
Read full abstract