We compare the space-time correlations calculated from direct numerical simulation （DNS） and large-eddy simulation （LES） of turbulent channel flows. It is found from the comparisons that the LES with an eddy-visco...We compare the space-time correlations calculated from direct numerical simulation （DNS） and large-eddy simulation （LES） of turbulent channel flows. It is found from the comparisons that the LES with an eddy-viscosity subgrid scale （SGS） model over-predicts the space-time corre- lations than the DNS. The overpredictions are further quantified by the integral scales of directional correlations and convection velocities. A physical argument for the overpre- diction is provided that the eddy-viscosity SGS model alone does not includes the backscatter effects although it correctly represents the energy dissipations of SGS motions. This argument is confirmed by the recently developed elliptic model for space-time correlations in turbulent shear flows. It suggests that enstrophy is crucial to the LES prediction of spacetime correlations. The random forcing models and stochastic SGS models are proposed to overcome the overpredictions on space-time correlations.展开更多
In a real climate system there are multiple time-space scale atmosphere-ocean interactions, ranging from the planetary scale and basin scale to local air-sea interactions. The Zebiak-Cane (ZC) model with one-level atm...In a real climate system there are multiple time-space scale atmosphere-ocean interactions, ranging from the planetary scale and basin scale to local air-sea interactions. The Zebiak-Cane (ZC) model with one-level atmosphere described only local air-sea interaction process. Thus the planetary scale Hadley cell and Walker cell anomalies should be introduced in the model. Including the planetary scale Hadley cell anomaly in the model improved the prediction skill. It showed that the improved model provided satisfactory prediction of the equatorial eastern Pacific SST anomaly with lead time of 9-10 months not only for 1970-1991 but also for 1992-1995.展开更多
目的 图像复原是计算机视觉领域的经典研究问题。选择性状态空间模型(selective state space model,selective SSM)因其高效的序列建模能力,广泛应用于各类图像复原任务。另外,非局部图像块之间存在依赖关系,能够辅助提升复原性能。而传...目的 图像复原是计算机视觉领域的经典研究问题。选择性状态空间模型(selective state space model,selective SSM)因其高效的序列建模能力,广泛应用于各类图像复原任务。另外,非局部图像块之间存在依赖关系,能够辅助提升复原性能。而传统SSM采用确定性的令牌(token)扫描方式,仅能提取令牌序列的单向依赖关系。此时,令牌间的关系建模因在序列中的先后顺序受到因果性制约,这与图像块之间的非因果相互关系形成冲突,限制了复原性能的进一步提升。针对此问题,提出一种面向图像复原的非因果选择性状态空间模型,旨在赋予SSM建模令牌之间非因果依赖关系的能力。方法 为解决SSM在因果性建模与图像内容非因果关系之间的矛盾,提出随机扫描策略,突破了传统扫描方式在因果性和空间限制上的局限,实现了令牌序列之间的非因果建模。具体而言,构建了随机重排和逆重排函数,实现了非固定次序下的令牌扫描,有效建模了不同令牌之间的非因果依赖关系。此外,针对图像退化干扰存在空间尺度变化和形态结构复杂的特点,融合多尺度先验构建了具有局部与全局信息互补性的非因果Mamba模型(non-causal Mamba,NCMamba),实现了对于各类图像复原任务的有效适配。结果 实验分别在图像去噪、去模糊和去阴影任务上进行,验证了所提非因果建模和局部—全局互补策略的有效性。与现有方法相比,所提模型在图像去阴影数据集SRD(shadow removal dataset)上的峰值信噪比提升0.86 dB。结论 面向图像复原任务,构建了非因果选择性状态空间模型,建模了令牌之间的非因果依赖关系,实现了局部与全局信息的有效互补,显著提升了复原性能。所提方法在主客观评价指标上均取得优异性能,为图像复原领域提供了新的解决方案。展开更多
基金supported by the National Basic Research Program of China (973 Program) (2007CB814800)the National Natural Science Foundation of China (10325211 and 10628206)
文摘We compare the space-time correlations calculated from direct numerical simulation （DNS） and large-eddy simulation （LES） of turbulent channel flows. It is found from the comparisons that the LES with an eddy-viscosity subgrid scale （SGS） model over-predicts the space-time corre- lations than the DNS. The overpredictions are further quantified by the integral scales of directional correlations and convection velocities. A physical argument for the overpre- diction is provided that the eddy-viscosity SGS model alone does not includes the backscatter effects although it correctly represents the energy dissipations of SGS motions. This argument is confirmed by the recently developed elliptic model for space-time correlations in turbulent shear flows. It suggests that enstrophy is crucial to the LES prediction of spacetime correlations. The random forcing models and stochastic SGS models are proposed to overcome the overpredictions on space-time correlations.
基金Project supported by the National 95 Scienct & Technology and the National Natural Science Foundation of China
文摘In a real climate system there are multiple time-space scale atmosphere-ocean interactions, ranging from the planetary scale and basin scale to local air-sea interactions. The Zebiak-Cane (ZC) model with one-level atmosphere described only local air-sea interaction process. Thus the planetary scale Hadley cell and Walker cell anomalies should be introduced in the model. Including the planetary scale Hadley cell anomaly in the model improved the prediction skill. It showed that the improved model provided satisfactory prediction of the equatorial eastern Pacific SST anomaly with lead time of 9-10 months not only for 1970-1991 but also for 1992-1995.
