Application Research Of Hybrid Intelligent Optimization Algorithm In Coal Bed Methane Reservoir Crack Detection

The development of fractures and the spatial distribution characteristics of coalbed methane reservoirs are very important for the exploration,development and utilization of coalbed methane.The evaluation of fracture parameters is of great significance to the exploration and exploitation of coal and gas in coal measures.Fissures are places where fluids are stored and transported,affecting the safe production of coal mines and the development and utilization of coalbed methane.The petrophysical properties of coalbed methane reservoirs have strong nonlinear characteristics.Through petrophysical experiments and numerical simulations,the nonlinear relationship between the microscopic morphological characteristics of coal and its macroscopic rock physical properties can be established.The thesis first calculates the seismic attribute parameters such as coherent attributes,azimuth inclination attributes,curvature attributes,configuration tensor attributes,and weighted instantaneous frequency attributes,and analyzes the application and effect of each method in coal bed methane reservoir crack detection.Research shows that different seismic attributes can characterize the fracture characteristics of coalbed methane reservoirs from different angles.The azimuth attribute based on the gradient structure tensor reflects the distribution system of faults and pore structures on a large scale.The dip angle attribute reflects the target level and The difference of the fissure distribution surface;the configuration tensor attribute enhances the display of this feature on the seismic profile and the boundary range of the fissure development area;the weighted instantaneous attribute reflects the low-frequency and low-velocity zones of the formation from the perspective of absorption and attenuation Vertical fissures are sensitive.Further research based on neural network,using whale optimization algorithm and genetic algorithm to improve the BP neural network,improve the robustness of the algorithm,find the global optimal solution,and use the optimized and improved neural network method for coal bed methane reservoir crack detection.Extract coherent attributes,azimuth inclination attributes,curvature attributes,configuration tensor attributes,and weighted instantaneous frequency attributes from actual seismic data,analyze and explain them separately,and use them as input data for improved BP neural network for coalbed methane storage Comprehensive detection and analysis of layer fissures.Using well data as the output evaluation criterion of the WOA-BP network,a comprehensive inspection of the coalbed methane reservoir cracks in the study area shows that the WOA-BP network can inherit and develop the advantages of existing attributes,and at the same time The instructions and descriptions are more detailed,the hierarchy is clear,and the directivity is obvious.The whale optimization algorithm and genetic algorithm can optimize the initial parameters,connection weights and thresholds of the BP neural network,improve the global search ability,calculation accuracy and stability of the BP neural network;in terms of calculation accuracy and stability,the whale optimization algorithm The effect of network optimization is better than that of genetic algorithm.The hybrid intelligent nonlinear algorithm based on whale optimization algorithm has achieved the expected results in the comprehensive detection of coalbed methane reservoir cracks,which provides a new method and a new way for the search of coalbed methane reservoir cracks,which in turn provides a favorable enrichment area for coalbed methane Provide basis for prediction and evaluation.