This paper describes a numerical simulation of thermal discharge in the cooling pool of an electrical power station, aiming to develop general-purpose computational programs for grid generation and flow/pollutant tran...This paper describes a numerical simulation of thermal discharge in the cooling pool of an electrical power station, aiming to develop general-purpose computational programs for grid generation and flow/pollutant transport in the complex domains of natural and artificial waterways. Three depth-averaged two-equation closure turbulence models, k-ε, k- w, and k- w, were used to close the quasi three-dimensional hydrodynamic model. The k- w model was recently established by the authors and is still in the testing process. The general-purpose computational programs and turbulence models will be involved in a software that is under development. The SIMPLE （Semi-Implicit Method for Pressure-Linked Equation） algorithm and multi-grid iterative method are used to solve the hydrodynamic fundamental governing equations, which are discretized on non-orthogonal boundary-fitted grids with a variable collocated arrangement. The results calculated with the three turbulence models were compared with one another. In addition to the steady flow and thermal transport simulation, the unsteady process of waste heat inpouring and development in the cooling pool was also investigated.展开更多
Ocean Drilling Program (ODP) Site 807A was recovered from the Ontong-Java plateau, western equatorial Pacific. Quantitative analysis of planktonic foraminifera, combined with oxygen and carbon isotope data, reveals th...Ocean Drilling Program (ODP) Site 807A was recovered from the Ontong-Java plateau, western equatorial Pacific. Quantitative analysis of planktonic foraminifera, combined with oxygen and carbon isotope data, reveals the glacial-interglacial variations of sea-surface temperature and the upper water vertical structure in this region during the late Quaternary. Our results indicate that since 530 ka sea-surface temperature (SST) and the depth of thermocline (DOT) have changed significantly in the western Pacific warm pool (WPWP). The average glacial-interglacial annual SST difference was up to 4.2 ℃, and the DOT fluctuations could exceed more than 100 m, further suggesting the instability of the WPWP. The spectral analyses of SST and DOT reveal two dominating cyclicities—the typical 100 ka cycle and the semi-precessional cycle, which is significant in the tropical spectrum, indicating that late Quaternary paleoceanographic changes in the study area were influenced not only by a high latitude forcing but also by tropic-driving factors.展开更多
经典AOD-Net(All in One Dehazing Network)去雾后的图像存在细节清晰度不足、明暗反差过大和画面昏暗等问题。为了解决这些图像去雾问题,提出一种在AOD-Net基础上改进的多尺度算法。改进的网络结构采用深度可分离卷积替换传统卷积方式...经典AOD-Net(All in One Dehazing Network)去雾后的图像存在细节清晰度不足、明暗反差过大和画面昏暗等问题。为了解决这些图像去雾问题,提出一种在AOD-Net基础上改进的多尺度算法。改进的网络结构采用深度可分离卷积替换传统卷积方式,减少了冗余参数量,加快了计算速度并有效地减少了模型的内存占用量,从而提高了算法去雾效率;同时采用多尺度结构在不同尺度上对雾图进行分析和处理,更好地捕捉图像的细节信息,提升了网络对图像细节的处理能力,解决了原算法去雾时存在的细节模糊问题;最后在网络结构中加入金字塔池化模块,用于整合图像不同区域的上下文信息,扩展了网络的感知范围,从而提高网络模型获取有雾图像全局信息的能力,进而改善图像色调失真、细节丢失等问题。此外,引入一个低照度增强模块,通过明确预测噪声实现去噪的目标,从而恢复曝光不足的图像。在低光去雾图像中,峰值信噪比(PSNR)和结构相似性(SSIM)指标均有显著提升,处理后的图片具有更高的整体自然度。实验结果表明:与经典AOD-Net去雾的结果相比,改进算法能够更好地恢复图像的细节和结构,使得去雾后的图像更自然,饱和度和对比度也更加平衡;在RESIDE的SOTS数据集中的室外和室内场景,相较于经典AOD-Net,改进算法的PSNR分别提升了4.5593 dB和4.0656 dB,SSIM分别提升了0.0476和0.0874。展开更多
An artificial neural network(ANN) and a self-adjusting fuzzy logiccontroller(FLC) for modeling and control of gas tungsten arc welding(GTAW) process are presented.The discussion is mainly focused on the modeling and c...An artificial neural network(ANN) and a self-adjusting fuzzy logiccontroller(FLC) for modeling and control of gas tungsten arc welding(GTAW) process are presented.The discussion is mainly focused on the modeling and control of the weld pool depth with ANN and theintelligent control for weld seam tracking with FLC. The proposed neural network can produce highlycomplex nonlinear multi-variable model of the GTAW process that offers the accurate prediction ofwelding penetration depth. A self-adjusting fuzzy controller used for seam tracking adjusts thecontrol parameters on-line automatically according to the tracking errors so that the torch positioncan be controlled accurately.展开更多
在机器视觉感知系统中,从不完整的被遮挡的目标对象中鲁棒重建三维场景及其语义信息至关重要.目前常用方法一般将这两个功能分开处理,本文将二者结合,提出了一种基于深度图及分离池化技术的场景复原及语义分类网络,依据深度图中的RGB-D...在机器视觉感知系统中,从不完整的被遮挡的目标对象中鲁棒重建三维场景及其语义信息至关重要.目前常用方法一般将这两个功能分开处理,本文将二者结合,提出了一种基于深度图及分离池化技术的场景复原及语义分类网络,依据深度图中的RGB-D信息,完成对三维目标场景的重建与分类.首先,构建了一种CPU端到GPU端的深度卷积神经网络模型,将从传感器采样的深度图像作为输入,深度学习摄像机投影区域内的上下文目标场景信息,网络的输出为使用改进的截断式带符号距离函数(Truncated signed distance function, TSDF)编码后的体素级语义标注.然后,使用分离池化技术改进卷积神经网络的池化层粒度结构,设计带细粒度池化的语义分类损失函数,用于回馈网络的语义分类重定位.最后,为增强卷积神经网络的深度学习能力,构建了一种带有语义标注的三维目标场景数据集,以此加强本文所提网络的深度学习鲁棒性.实验结果表明,与目前较先进的网络模型对比,本文网络的重建规模扩大了2.1%,所提深度卷积网络对缺失场景的复原效果较好,同时保证了语义分类的精准度.展开更多
基金supported by FAPESP (Foundation for Supporting Research in So Paulo State), Brazil, of the PIPE Project (Grant No. 2006/56475-3)
文摘This paper describes a numerical simulation of thermal discharge in the cooling pool of an electrical power station, aiming to develop general-purpose computational programs for grid generation and flow/pollutant transport in the complex domains of natural and artificial waterways. Three depth-averaged two-equation closure turbulence models, k-ε, k- w, and k- w, were used to close the quasi three-dimensional hydrodynamic model. The k- w model was recently established by the authors and is still in the testing process. The general-purpose computational programs and turbulence models will be involved in a software that is under development. The SIMPLE （Semi-Implicit Method for Pressure-Linked Equation） algorithm and multi-grid iterative method are used to solve the hydrodynamic fundamental governing equations, which are discretized on non-orthogonal boundary-fitted grids with a variable collocated arrangement. The results calculated with the three turbulence models were compared with one another. In addition to the steady flow and thermal transport simulation, the unsteady process of waste heat inpouring and development in the cooling pool was also investigated.
文摘Ocean Drilling Program (ODP) Site 807A was recovered from the Ontong-Java plateau, western equatorial Pacific. Quantitative analysis of planktonic foraminifera, combined with oxygen and carbon isotope data, reveals the glacial-interglacial variations of sea-surface temperature and the upper water vertical structure in this region during the late Quaternary. Our results indicate that since 530 ka sea-surface temperature (SST) and the depth of thermocline (DOT) have changed significantly in the western Pacific warm pool (WPWP). The average glacial-interglacial annual SST difference was up to 4.2 ℃, and the DOT fluctuations could exceed more than 100 m, further suggesting the instability of the WPWP. The spectral analyses of SST and DOT reveal two dominating cyclicities—the typical 100 ka cycle and the semi-precessional cycle, which is significant in the tropical spectrum, indicating that late Quaternary paleoceanographic changes in the study area were influenced not only by a high latitude forcing but also by tropic-driving factors.
文摘经典AOD-Net(All in One Dehazing Network)去雾后的图像存在细节清晰度不足、明暗反差过大和画面昏暗等问题。为了解决这些图像去雾问题,提出一种在AOD-Net基础上改进的多尺度算法。改进的网络结构采用深度可分离卷积替换传统卷积方式,减少了冗余参数量,加快了计算速度并有效地减少了模型的内存占用量,从而提高了算法去雾效率;同时采用多尺度结构在不同尺度上对雾图进行分析和处理,更好地捕捉图像的细节信息,提升了网络对图像细节的处理能力,解决了原算法去雾时存在的细节模糊问题;最后在网络结构中加入金字塔池化模块,用于整合图像不同区域的上下文信息,扩展了网络的感知范围,从而提高网络模型获取有雾图像全局信息的能力,进而改善图像色调失真、细节丢失等问题。此外,引入一个低照度增强模块,通过明确预测噪声实现去噪的目标,从而恢复曝光不足的图像。在低光去雾图像中,峰值信噪比(PSNR)和结构相似性(SSIM)指标均有显著提升,处理后的图片具有更高的整体自然度。实验结果表明:与经典AOD-Net去雾的结果相比,改进算法能够更好地恢复图像的细节和结构,使得去雾后的图像更自然,饱和度和对比度也更加平衡;在RESIDE的SOTS数据集中的室外和室内场景,相较于经典AOD-Net,改进算法的PSNR分别提升了4.5593 dB和4.0656 dB,SSIM分别提升了0.0476和0.0874。
基金National Natural Science Foundation of China and Provincial Natural Science Foundafion of Guangdong, China.
文摘An artificial neural network(ANN) and a self-adjusting fuzzy logiccontroller(FLC) for modeling and control of gas tungsten arc welding(GTAW) process are presented.The discussion is mainly focused on the modeling and control of the weld pool depth with ANN and theintelligent control for weld seam tracking with FLC. The proposed neural network can produce highlycomplex nonlinear multi-variable model of the GTAW process that offers the accurate prediction ofwelding penetration depth. A self-adjusting fuzzy controller used for seam tracking adjusts thecontrol parameters on-line automatically according to the tracking errors so that the torch positioncan be controlled accurately.
文摘在机器视觉感知系统中,从不完整的被遮挡的目标对象中鲁棒重建三维场景及其语义信息至关重要.目前常用方法一般将这两个功能分开处理,本文将二者结合,提出了一种基于深度图及分离池化技术的场景复原及语义分类网络,依据深度图中的RGB-D信息,完成对三维目标场景的重建与分类.首先,构建了一种CPU端到GPU端的深度卷积神经网络模型,将从传感器采样的深度图像作为输入,深度学习摄像机投影区域内的上下文目标场景信息,网络的输出为使用改进的截断式带符号距离函数(Truncated signed distance function, TSDF)编码后的体素级语义标注.然后,使用分离池化技术改进卷积神经网络的池化层粒度结构,设计带细粒度池化的语义分类损失函数,用于回馈网络的语义分类重定位.最后,为增强卷积神经网络的深度学习能力,构建了一种带有语义标注的三维目标场景数据集,以此加强本文所提网络的深度学习鲁棒性.实验结果表明,与目前较先进的网络模型对比,本文网络的重建规模扩大了2.1%,所提深度卷积网络对缺失场景的复原效果较好,同时保证了语义分类的精准度.
