WOA-BP algorithm for crack detection in coalbed methane reservoirs and its application

Generation and spatial distribution of cracks in coalbed methane reservoirs play an important role in exploration, development and utilization of coalbed methane. The WOA-BP algorithm for crack detection in coalbed methane reservoirs is a combination of the WOA (Whale Optimization Algorithm) and BP (Back-Propagation) detection method in order to provide a robust and effective approach to detect reservoir cracks with dominant samples and secondary error control. This tool extracts attributes including coherence, azimuth angle, curvature, configuration tensor, and weighted instantaneous frequency of coal seam cracks from actual seismic data, and uses them as the input data of an improved WOA-BP neural network, which permits to conduct comprehensive detection and analysis of cracks in coalbed methane reservoirs. Using well data, known well production data, and core slice analysis results, it comprehensively establishes superior samples as the output evaluation standard of the BP neural network improved by the WOA optimization algorithm. Comprehensive detection in the study area and analysis results show that the WOA-BP network can inherit and develop the advantages of existing attributes, determine the fracture development level S-evlt value and classification standard, and finely describe the development degree of cracks in coalbed methane reservoirs. Four crack blocks are delineated in the study area and high-quality coalbed methane reservoirs are obtained, implying a good geological effect of exploration and development effect. It demonstrates that the WOA-BP method promotes the development of microcrack prediction and improves the micro-fracture prediction to a new level.