Fracture enhancement based on artificial ants and fuzzy c-means clustering (FCMC) in Dezful Embayment of Iran

• Published in: Journal of geophysics and engineering 2015-04, Vol.12 (2), p.227-241
• Main Authors: Nasseri, Aynur, Mohammadzadeh, Mohammad Jafar, Tabatabaei Raeisi, S Hashem
• Format: Article
• Language: English
• Subjects: ant colony algorithm (AC); Ant colony optimization; Bays; Clustering; Dezful Embayment of Iran; Discontinuity; Faults; Fracture mechanics; fracture pattern; fuzzy c-means clustering (FCMC); seismic data; Trends; Variance
• ISSN: 1742-2132; 1742-2140
• DOI: 10.1088/1742-2132/12/2/227

This paper deals with the application of the ant colony algorithm (AC) to a seismic dataset from Dezful Embayment in the southwest region of Iran. The objective of the approach is to generate an accurate representation of faults and discontinuities to assist in pertinent matters such as well planning and field optimization. The AC analyzed all spatial discontinuities in the seismic attributes from which features were extracted. True fault information from the attributes was detected by many artificial ants, whereas noise and the remains of the reflectors were eliminated. Furthermore, the fracture enhancement procedure was conducted by three steps on seismic data of the area. In the first step several attributes such as chaos, variance/coherence and dip deviation were taken into account; the resulting maps indicate high-resolution contrast for the variance attribute. Subsequently, the enhancement of spatial discontinuities was performed and finally elimination of the noise and remains of non-faulting events was carried out by simulating the behavior of ant colonies. After considering stepwise attribute optimization, focusing on chaos and variance in particular, an attribute fusion was generated and used in the ant colony algorithm. The resulting map displayed the highest performance in feature detection along the main structural feature trend, confined to a NW-SE direction. Thus, the optimized attribute fusion might be used with greater confidence to map the structural feature network with more accuracy and resolution. In order to assess the performance of the AC in feature detection, and cross validate the reliability of the method used, fuzzy c-means clustering (FCMC) was employed for the same dataset. Comparing the maps illustrates the effectiveness and preference of the AC approach due to its high resolution contrast for structural feature detection compared to the FCMC method. Accordingly, 3D planes of discontinuity determined spatial distribution of fractures in the field in order to assist well planning. Results revealed that the high impedance location probability related to an area in the vicinity of the faults, whilst low impedance location probably could indicate zones of high permeability which indicate flow conduits. Analysis under the present study suggests that the orientation and magnitude of fractures exhibiting the main trend of NW-SE in Dezful Embayment is more susceptible to stimulation and is more likely to open for fluid flow.