Performance Evaluation of Automated Well-to-Well Correlation Approach Based on Hausdroff Parameter and Mean Value of Well Logs Signal

Accurate knowledge of the stratigraphic sequences and their possible discontinuities is so important for the characterization of reservoirs. The purpose of stratigraphic correlation is detecting geological boundaries to construct a three-dimensional model of the reservoir. This process is usually done manually, therefore, depends on interpreter’s experience, multiple interpretation is not unexpected. Using automatic algorithm to detect geological boundaries in addition to facilitating correlation and removing multiple interpretation will lead to save money and time. In this work, Fractal and statistical features are implemented to extract well-logs pattern. Wavelet transform analysis is used to determine fractal dimension. Detection of the pattern in observation well which is similar to the pattern obtained in witness well (a well which depths of all boundaries are available) around the selected boundary is the basis of this novel approach. Automatic Stratigraphic correlation algorithm is implemented on well-logs of 4 wells in one of the oil fields in southwest of Iran. For this purpose, one of the geological boundaries in witness well is selected and the aim is to find the depth of the corresponding boundary in the other wells. The approach was able to detect the depth of the boundary in all the wells. The deviation of the results of the automated approach from the manual ones is about 1.07 meters which confirm effectiveness and capability of this approach.