Cracks and pores are two common defects in metallic additive manufacturing (AM) parts. In this paper, deep learning-based image analysis is performed for defect (cracks and pores) classification/detection based on SEM images of metallic AM parts. Three different levels of complexities, namely, defect classification, defect detection and defect image segmentation, are successfully achieved using a simple CNN model, the YOLOv4 model and the Detectron2 object detection library, respectively. The tuned CNN model can classify any single defect as either a crack or pore at almost 100% accuracy. The other two models can identify more than 90% of the cracks and pores in the testing images. In addition to the application of static image analysis, defect detection is also successfully applied on a video which mimics the AM process control images. The trained Detectron2 model can identify almost all the pores and cracks that exist in the original video. This study lays a foundation for future in situ process monitoring of the 3D printing process.

The hydraulic fracturing process in the oil and gas industry, which can be represented by a borehole subjected to three-dimensional in situ stresses and a fluid discharge over a finite length of its surface, is essentially a coupled boundary value problem. This paper innovatively reformulates the governing equations for this important three-dimensional problem in transversely isotropic rock formations based on Biot's poroelasticity theory and solves the governing equations with the aid of Fourier series and the standard Laplace and Fourier integral transforms. The problem is subsequently decomposed into five subproblems, each of which can be transformed into a set of dual integral equations and can be readily solved by standard procedure. The verification is followed by the study of the influences of material anisotropy on the stress responses of the porous medium around the borehole. It is found that neglecting the mechanical anisotropy of the rocks would result in substantial errors to the predicted stress distributions around the wellbore and thus considerable impacts on the design of hydraulic fracturing.

In this paper, we propose a novel method for demographic classification and present an image retrieval system that can retrieve facial images by demographic information that includes gender, age and ethnicity. The demographic information is extracted from human faces by demographic classifiers that are learned from boosting Haar feature based look-up-table type weak classifiers. The image retrieval system consists of three modules, face detection, facial feature landmark extraction and demographic classification. Experimental results are reported to show its potential in the management of a large facial image database for online retrieval applications.

Fluid injection into the formation through a completed wellbore, as an important process for the oil and gas industry, is commonly seen in well tests and hydraulic fracturing operations. The problem of wellbore breakdown during fluid injection is revisited in this work by a thorough and comprehensive consideration of some essential issues involved in this subject. First, fluid injection may not necessarily start simultaneously with the completion of wellbore drilling, as is usually assumed in the reported analytical solutions. Second, there exists a common limitation that the borehole must orient in a direction normal to the isotropic (bedding) planes in the previous analytical studies addressing the wellbore problem drilled in a transversely isotropic formation. Last, but not least, the implicit boundary condition of radial stress, which is usually ignored but essential to accurately evaluate the breakdown pressure on such fluid injection, also adds to the difficulty of solving the problem. This paper presents a novel numerical scheme enhanced by user subroutines/functions that can work with mainstream simulators to promote the in-depth study of wellbore breakdown under fluid injection with various model settings. On validation, the effects of the idle time t₀ after drilling and the material anisotropy on the poromechanical responses of the wellbore are extensively studied for the most commonly encountered horizontal wells. The idle time has slight effects immediately after fluid injection starts; however, the influences become generally negligible after a short period of time of fluid injection. It is also revealed that the anisotropy in rock properties has a more significant influence on the stress responses for horizontal wells than the previously studied case of vertical wells. Another important finding is that tensile fracture may not always initiate vertically from the top/bottom of the horizontal borehole wall because the maximum tensile stress occurs elsewhere, although vertical in-situ stress is the maximum principle stress.

The research on increased ductility of organic-rich shale and its impacts on hydraulic fracturing has received relatively little attention. According to recent experimental studies, the popular cohesive zone model that only features decreasing traction along with crack separation may not adequately represent the crack behavior in shale due to ample organic matter. This paper starts by proposing a modified cohesive zone model that can represent various traction-separation laws (TSL) within a unified formulation. Then a fully coupled poroelastic XFEM framework to simulate hydraulic fracturing in organic-rich shales was developed in Matlab and comprehensively verified against the latest analytical solutions. The influences of increased ductility in different forms were studied using the modified cohesive zone model in the context of field-scale hydraulic fracturing simulations. Three important conclusions were drawn. First, the shape of TSL does affect the hydraulic fracturing given the same cohesive crack energy and tensile strength. It suggests that ductility is not only controlled by cohesive crack energy and tensile strength, which further indicates the necessity of the newly proposed TSL. Second, the tensile strength, controlling when the cohesive crack starts propagating, has the greatest impact on the hydraulic fracturing, among all TSL shape parameters. The impacts of TSL parameters become less significant as the fracturing fluid viscosity increases. Lastly, Young’s modulus is the only one among four common poroelastic parameters that significantly changes the ductility/brittleness of rock formation and hydraulic fracture lengths. The increase in cohesive energy accompanied by the decrease of Young’s modulus will greatly reduce the induced fracture length.