Prediction and Visualization of Reservoir Fracture Based on ANN and GIS
Lianbo Zeng · Geology and Exploration · 2011
Prediction and visualization of fissure distribution in low-permeability reservoirs is an important issue in petroleum exploration and development. Taking the oil-bearing formation Chang 6-1 in the Ordos basin as an example, this work constructed a BP ANN (Artificial Neural Network) based on the known fractal dimension, sand/formation ratio and crack ratio, and designed an optimum network topology and training function through experiments to predict fissure distribution beneath the study area. Then secondary development based on ArcGIS and C-Tech was carried out for 2D 3D visualization of fisure distribution. The experimental result shows the convenience of GIS-based technology in seamless in tegration of professional algorithms.