結合混合實境與裂隙擴張之混合域有限元素模型研究

The techniques of AI (Artificial Intelligence), including deep learning are often used to explore the unknown fields and make accurate analysis and prediction. The deep learning technique also makes a huge progress in image classification and detection. The important imaging tool, X-ray CT, plays an important role in different fields. In the field of earth science, X-ray CT is often used to investigate geological structure, analyze the density, porosity, permeability, and other important features of rocks. To analyze the important features of the conventionalCT images (digital information), it need some steps including image processing, noise reduction, smoothing, segmentation. We use the deep learning technique to construct the neural network model to learn to analyze CT images in commercial software and build a database of porosity feature analysis. Also, we develop the model of analyzing the sandstone CT images which can generate the binary images automatically. Through the binary images, we calculate the parameters in Kozeny - Carman Equation to get the spatial distribution ofpermeability and porosity of rock. The development of the microscale unstructured mesh generation model.