The chloride diffusivity of ultra-high performance concrete (UHPC) is the key parameter that determines the service life of UHPC structures in chloride environments. In this study, a multi-scale model, which considered the structural information at micro and meso-scale, was proposed to predict macro property, i.e. chloride diffusivity, of UHPC. At micro-scale, the chloride diffusivity of UHPC paste was predicted based on an extended HYMOSTRUC3D model and a finite element method. At meso-scale, Anm model and finite element method were applied to calculate the chloride diffusivity of UHPC mortar. At macro-scale, the chloride diffusivity of UHPC was determined with a two phase model. The chloride diffusivity at a lower scale was used as input to predict the chloride diffusivity at the higher scale. Non-steady chloride diffusion tests were performed to validate the proposed model. The results showed that the proposed multi-scale model gave a little higher chloride diffusivity of UHPC than experiment results. However, the proposed model is still helpful for the durability design of UHPC structure in chloride environments.@en

The flexural response of ultrahigh performance fiber-reinforced concrete (UHPFRC) was simulated based on the lattice fracture model. Fiber was modelled as separated beam that was connected to the matrix with interface beams. The simulated results were compared with the experimental results. Deviations occurred at the late stage of the strain-softening period. But both the strain-hardening behavior and multicracking phenomenon were observed in the simulation. The effects of fiber orientation and fiber content were studied with the lattice fracture model. The flexural strength and toughness of UHPFRC improved as the fibers were aligned distributed or the fiber content increased. The proposed model has the potential to help with the materials design of UHPFRC, and the limitations of the model were also discussed in the paper.@en

The effects of pore throat structure on gas permeability in tight sand reservoir were investigated using helium-measured porosity, pulse decay permeability, casting thin sections, scanning electron microscopy (SEM), nuclear magnetic resonance (NMR), and constant-rate mercury injection for five sandstone samples from the Upper Triassic Yanchang Formation in the Western Ordos Basin, China. Results show that permeability values are in the range of 0.474-2.290 mD. The strong variation may be due to the effect of mineral composition and pore-throat structure. Generally, sandstone samples of higher content of quartz, chlorite film, and slit-shaped throat tend to have higher permeability. The movable fluid porosity numerically equals to the product of porosity and the movable fluid saturation based on the film bound water, which can comprehensively reflect the influence of fluid flow channels on porosity while neglecting the influence of bound fluids in the fine pores of clay minerals in the tight reservoirs. Larger movable fluid porosity values mean larger pore spaces for the movable fluids, which contribute to larger permeability. It is more reasonable to analyze the movable fluid pore-throat radius as a range. The homogeneous pore throat structure lead to a narrow lower limit range of the movable fluid pore-throat radius, which is favorable for the fluid flow. Meanwhile, if the lower bound of the lower limit range of the movable fluid pore-throat radius is small, pores and throats are less plugged, and thus make the conditions favorable for fluid flow. The contribution of throats to the pore volume is critical in influencing reservoir permeability as it can directly control the lower limit of the movable fluid pore-throat size. The pore-throat radius where throats account for the smallest amount in the pore space increases with the permeability, which indicates that larger proportions of large throats result in larger throat radius and higher pore connectivity, leading to higher permeability.

@en

Image registration is typically formulated as an optimization process, which aims to find the optimal transformation parameters of a given transformation model by minimizing a cost function. Local minima may exist in the optimization landscape, which could hamper the optimization process. To eliminate local minima, smoothing the cost function would be desirable. In this paper, we investigate the use of a randomized smoothing (RS) technique for stochastic gradient descent (SGD) optimization, to effectively smooth the cost function. In this approach, Gaussian noise is added to the transformation parameters prior to computing the cost function gradient in each iteration of the SGD optimizer. The approach is suitable for both rigid and nonrigid registrations. Experiments on synthetic images, cell images, public CT lung data, and public MR brain data demonstrate the effectiveness of the novel RS technique in terms of registration accuracy and robustness.

@en