Wyniki wyszukiwania - BazTech

1

Quantitative characterization of seepage behavior in rough fracture considering hydromechanical coupling effect: an experimental study

Yu Xiang, Zhang Tong, Yang Ke, Yu Fei, Liu Yang, Tang Ming

Acta Geophysica

|

2023

|

Vol. 71, no. 5

2245--2264

EN

To study the effect of fracture morphology and in situ stress on the seepage behavior of rough fractures, hydraulic–mechanical experiments with different confining stresses, pore pressures and fracture geometry were carried out. The dimensionless parameter non-Darcy coefficient factor K and K-based critical Reynolds number model (KCRN) was proposed to characterize the behavior of rough-wall fracture and fluid seepage. The results show that the seepage flow of rough-wall fracture can be well described by Forchheimer equation. As the confining pressure increases from 1 to 31 MPa, the two walls of the rough fracture are compressed, and the fluid flow capacity is weakened, resulting in an increase of 2–3 orders of magnitude in Forchheimer viscosity coefficient A. Also affected by the increase in the confining pressure, the contact area between the two walls of the rough fracture increases, which makes the fluid channel become curved, increases the dissipation of water pressure in the inertial process and causes the inertial term coefficient B to increase by 2–3 orders of magnitude in general. In the whole range of test confining pressure (1 MPa–31 MPa), the flow state of rough fracture fluid is divided into zones based on the critical Reynolds number. The average hydraulic aperture decreases with the increase in the confining pressure, which can be perfectly fitted by hyperbolic function. The calculated critical Reynolds number of six rough fracture samples varies from 0.0196 to 1.0424. According to the experimental data, the K-based critical Reynolds number model (KCRN) is validated, and the validation results prove the accuracy and reliability of the model.

2

Combination of seismic attributes using clustering and neural networks to identify environments with sandstone-type uranium mineralization

Wu Qubo, Li Ziwei, Wang Yanchun, Cao Chengyin, Qiao Baoping, Huang Yucheng, Yu Xiang

Acta Geophysica

|

2023

|

Vol. 71, no. 6

2715--2731

EN

How to effectively and economically estimate the occurrence of sandstone-type uranium deposits in deep metallogenic environments considering the high costs of drilling methods has troubled the uranium geologists for a long time. To address this, we developed a novel workflow using a combination of clustering and neural networks in seismic attributes (SA) analysis to characterize uranium-bearing environments in directly predicting uranium zones. In the workflow, PCA and a clustering method for SA analysis in the target layer were first used to predict favorable zones for uranium; second, supervised neural networks were utilized for quantitative estimation of gamma volume, with subsequent delineation of the zones with high gamma anomalies; finally, results from the two processes were examined for mutual intersection, and intersection results with ordinal ranking of A and B were then extracted. In practical application, several A-level and B-level regions were predicted, which represent very high U-mineralization potential and significant potential, respectively. The prediction results were then essentially verified by logging data of the study area. We consider our developed approach to be a reasonably cost-effective technique for uranium-bearing environment prediction as it can yield useful suggestions for drilling programs with significantly reduced cost and prospecting risk.

3

Abnormality Diagnosis in Mammograms by Transfer Learning Based on ResNet18

Yu Xiang, Wang Shui-Hua

Fundamenta Informaticae

|

2019

|

Vol. 168, nr 2-4

219--230

EN

Breast cancer is one of the common cancers threatening the health of women while the incident rate of it is quite low in men to contribute to a major killer of men. Early syndromes of breast cancer including micro-calcification, mass, and distortion in mammography images can be very helpful for radiologists to make diagnosis of the cancer at early stage, which means the cancer can be treated or even be cured timely and thus make early diagnosis important. To assist radiologists with diagnosis, we set up a computer-aided diagnosis system to make diagnosis decision of breast cancer in this paper. We acquired regions of interests in mammographic images from public database, and labeled regions containing micro-calcification or mass as abnormality while regions without such abnormalities as normality. By transferring the state-of-the-art networks into our quest, we found that ResNet18 performed best and achieved mean accuracy of 95.91%.