This study develops a three-dimensional heterogeneous numerical model to simulate the water inrush process and predict the water yield for mineral exploration in Yangzhuang Iron Mine in Anhui Province. To identify the...This study develops a three-dimensional heterogeneous numerical model to simulate the water inrush process and predict the water yield for mineral exploration in Yangzhuang Iron Mine in Anhui Province. To identify the hydrogeological parameters of the aquifer in the study area, the model was calibrated and validated using the observed heads through the integrated trial-and-error and automated techniques. Also, the sensitivity analysis of the model was performed to evaluate the uncertainty associated with the calibrated model. According to the mine construction plan at different mining levels of-500 m,-600 m, and-700 m, the calibrated model was then applied to predict the water yields dependent on the different mining levels. As indicated by the prediction results, the numerical simulation model can systematically describe the groundwater system in the mining area and determine the source of water inrush in this iron mine. In conclusion, numerical analyses carried out in this study can provide guidance to decision-makers in balancing the iron ore mining and mine dewatering in the future.展开更多
针对传统图像超分辨重建技术中存在的特征丢失和缺乏高频细节的问题,在生成对抗网络的基础上结合注意力机制对网络进行改进.生成网络中通过多尺度残差注意力模块,学习不同尺度的图像特征,增强对图像高频细节的学习;再通过整体注意力模块...针对传统图像超分辨重建技术中存在的特征丢失和缺乏高频细节的问题,在生成对抗网络的基础上结合注意力机制对网络进行改进.生成网络中通过多尺度残差注意力模块,学习不同尺度的图像特征,增强对图像高频细节的学习;再通过整体注意力模块,进一步捕获更多的信息特征,提高网络对图像细节的还原能力,用于最终重建.判别网络中使用非对称卷积替代传统卷积,减少参数计算量;并引入自注意力机制更精确地获取图像全局信息,提高网络重建性能.实验结果表明,重建后图像和原始图像相比具有更多的高频纹理细节,与7种常见的图像超分辨方法相比,PSNR(Picture Signal to Noise Ratio)平均提升约2.43 dB,SSIM(Structural Similarity Image Measurement)平均提升约0.1.展开更多
基金financially supported by the National Natural Science Foundation of China(No.41402198 and 41372235)Jiangsu Natural Science Fund-Youth Fund(No.BK20131009)the Fundamental Research Funds for the Central Universities(No.2014B03614)
文摘This study develops a three-dimensional heterogeneous numerical model to simulate the water inrush process and predict the water yield for mineral exploration in Yangzhuang Iron Mine in Anhui Province. To identify the hydrogeological parameters of the aquifer in the study area, the model was calibrated and validated using the observed heads through the integrated trial-and-error and automated techniques. Also, the sensitivity analysis of the model was performed to evaluate the uncertainty associated with the calibrated model. According to the mine construction plan at different mining levels of-500 m,-600 m, and-700 m, the calibrated model was then applied to predict the water yields dependent on the different mining levels. As indicated by the prediction results, the numerical simulation model can systematically describe the groundwater system in the mining area and determine the source of water inrush in this iron mine. In conclusion, numerical analyses carried out in this study can provide guidance to decision-makers in balancing the iron ore mining and mine dewatering in the future.
文摘针对传统图像超分辨重建技术中存在的特征丢失和缺乏高频细节的问题,在生成对抗网络的基础上结合注意力机制对网络进行改进.生成网络中通过多尺度残差注意力模块,学习不同尺度的图像特征,增强对图像高频细节的学习;再通过整体注意力模块,进一步捕获更多的信息特征,提高网络对图像细节的还原能力,用于最终重建.判别网络中使用非对称卷积替代传统卷积,减少参数计算量;并引入自注意力机制更精确地获取图像全局信息,提高网络重建性能.实验结果表明,重建后图像和原始图像相比具有更多的高频纹理细节,与7种常见的图像超分辨方法相比,PSNR(Picture Signal to Noise Ratio)平均提升约2.43 dB,SSIM(Structural Similarity Image Measurement)平均提升约0.1.