Northwestern Polytechnical University

About: Northwestern Polytechnical University is a education organization based out in Xi'an, China. It is known for research contribution in the topics: Microstructure & Computer science. The organization has 47497 authors who have published 56067 publications receiving 657071 citations. The organization is also known as: Xīběi Gōngyè Dàxué & State Northwest Institute of Engineering.

Tensile failure strength and separation angle of FDM 3D printing PLA material: Experimental and theoretical analyses

Abstract: It is discovered in this investigation that there exist two different failure modes and a special separation angle which is the demarcation point of the two different failure modes when FDM (Fused Deposition Modelling) 3D printing materials fail under a tensile load. In order to further understand the mechanical properties of FDM 3D printing materials and promote the use of FDM 3D printing materials, their tensile failure strengths at different printing angles and separation angles are measured and analysed theoretically. A new separate-modes of transversely isotropic theoretical failure model is established to predict the tensile failure strength and separation angle of FDM 3D printing PLA (polylactic acid) material based on the hypothesis of transverse isotropy and the classical separate-modes failure criterion. During this research, the tensile specimens designed according to the current test standard ISO (527-2-2012) for plastic-multi-purpose specimens are fabricated in 7 different printing angles ( 0 ∘ , 15 ∘ , 30 ∘ , 45 ∘ , 60 ∘ , 75 ∘ , 90 ∘ ) and three levels of printing layer thickness (0.1 mm, 0.2 mm, 0.3 mm). Experimental results show that the tensile failure strength increases with the increase of the printing angle or the decrease of the layer thickness. Meanwhile, inter-layer failure tends to occur when the printing angle is small and in-layer failure tends to occur when the printing angle is big. In comparison with the results predicted by the established theoretical model, all values of the Generalized-Relative-Root-Mean-Square Error are close to zero and the experimental separation angles are also between 45 ∘ and 60 ∘ . So the predictive capacity of the theoretical model is affirmed by experimental results.

Structure of the global plastic waste trade network and the impact of China’s import Ban

Abstract: Millions of tonnes (teragrams) of plastic waste are traded around the world every year, which plays an important role in partially substituting virgin plastics as a source of raw materials in plastic product manufacturing. In this paper, global plastic waste trade networks (GPWTNs) from 1988 to 2017 are established using the UN-Comtrade database. The spatiotemporal evolution of the GPWTNs is analyzed. Attention is given to the country ranks, inter- and intra-continental trade flows, and geo-visual communities in the GPWTNs. We also evaluate the direct and indirect impacts of China’s plastic waste import ban on the GPWTNs. The results show that the GPWTNs have small-world and scale-free properties and a core-periphery structure. The geography of the plastic waste trade is structured by Asia as the dominant importer and North America and Europe as the largest sources of plastic waste. China is the unrivaled colossus in the global plastic waste trade. After China’s import ban, the plastic waste trade flows have been largely redirected to Southeast Asian countries. Compared with import countries, export countries are more important for the robustness of GPWTNs. Clearly, developed countries will not announce bans on plastic waste exports; these countries have strong motivation to continue to shift plastic waste to poorer countries. However, the import bans from developing countries will compel developed countries to build new disposal facilities and deal with their plastic waste domestically.

Electrically Tunable Goos–Hänchen Effect with Graphene in the Terahertz Regime

Abstract: Goos–Hanchen (G–H) effect is of great interest in the manipulation of optical beams. However, it is still fairly challenging to attain efficient controls of the G–H shift for diverse applications. Here, a mechanism to realize tunable G–H shift in the terahertz regime with electrically controllable graphene is proposed. Taking monolayer graphene covered epsilon-near-zero metamaterial as a planar model system, it is found that the G–H shifts for the orthogonal s-polarized and p-polarized terahertz beams at oblique incidence are positive and negative, respectively. The G–H shift can be modified substantially by electrically controlling the Fermi energy of the monolayer graphene. Reversely, the Fermi energy dependent G–H effect can also be used as a strategy for measuring the doping level of graphene. In addition, the G–H shifts of the system are of strong frequency-dependence at oblique angles of incidence, therefore the proposed graphene hybrid system can potentially be used for the generation of terahertz “rainbow,” a flat analog of the dispersive prism in optics. The proposed scheme of hybrid system involving graphene for dynamic control of G–H shift will have potential applications in the manipulation of terahertz waves.

Toward Compact ConvNets via Structure-Sparsity Regularized Filter Pruning

Abstract: The success of convolutional neural networks (CNNs) in computer vision applications has been accompanied by a significant increase of computation and memory costs, which prohibits their usage on resource-limited environments, such as mobile systems or embedded devices. To this end, the research of CNN compression has recently become emerging. In this paper, we propose a novel filter pruning scheme, termed structured sparsity regularization (SSR), to simultaneously speed up the computation and reduce the memory overhead of CNNs, which can be well supported by various off-the-shelf deep learning libraries. Concretely, the proposed scheme incorporates two different regularizers of structured sparsity into the original objective function of filter pruning, which fully coordinates the global output and local pruning operations to adaptively prune filters. We further propose an alternative updating with Lagrange multipliers (AULM) scheme to efficiently solve its optimization. AULM follows the principle of alternating direction method of multipliers (ADMM) and alternates between promoting the structured sparsity of CNNs and optimizing the recognition loss, which leads to a very efficient solver ( $2.5\times $ to the most recent work that directly solves the group sparsity-based regularization). Moreover, by imposing the structured sparsity, the online inference is extremely memory-light since the number of filters and the output feature maps are simultaneously reduced. The proposed scheme has been deployed to a variety of state-of-the-art CNN structures, including LeNet, AlexNet, VGGNet, ResNet, and GoogLeNet, over different data sets. Quantitative results demonstrate that the proposed scheme achieves superior performance over the state-of-the-art methods. We further demonstrate the proposed compression scheme for the task of transfer learning, including domain adaptation and object detection, which also show exciting performance gains over the state-of-the-art filter pruning methods.