A review of modeling approaches for oriented semi-crystalline polymers

Modeling changes in the physical properties of oriented semi-crystalline polymer films is beneficial for understanding the fundamentals associated with structure property relationships and could be used for developing new polymer films with significantly enhanced physical properties. Relating the molecular changes observed in oriented polymer films to inherent polymer characteristics provides valuable insight for the development of new polymers which exhibit enhanced physical properties upon orientation. Modeling efforts will be reviewed that have attempted to use fiber composite theory to explain the transitions seen during the orientation process.     

Environmental Fatigue and Static Behavior of RC Beams Strengthened with Carbon-Fiber-Reinforced Polymer

Many countries around the world have tremendous needs to repair and strengthen their transportation infrastructure. Almost everywhere, traffic loads have reached levels largely exceeding design expectations. Northern countries also experience severe winter conditions that are combined with an extensive use of deicing salts and accelerate structural deterioration. In Canada, the extent of deterioration has prompted many authorities, including the federal and provincial governments, to investigate the potential use of fiber-reinforced polymer products to extend the life of their existing structures. However, it is widely recognized that the large-scale implementation of these products is often impaired by the lack of data on their durability. This paper presents an experimental project undertaken in order to assess the durability of reinforced concrete beams externally strengthened with two types of carbon-fiber-reinforced polymer (CFRP). The beams were first exposed to either wet-dry cycles or continuous immersion in water and then were loaded in fatigue. Finally, they were tested quasi-statically under four-point bending up to failure. The test results presented here provide some insights on the potential long-term performance of CFRP-strengthened beams exposed to severe environmental conditions.     

Performance Comparison of Four Fiber-Reinforced Polymer Deck Panels

In an effort to assess the constructability and performance of bridges with fiber-reinforced polymer (FRP) composite decks, the short-term and long-term responses of a 207 m, five-span bridge retrofitted with four different FRP panel systems were monitored. The overall aspects of the panel systems, connection details, and construction techniques are presented prior to presentation of the observed and measured responses. Key design parameters (impact factors, girder distribution factors, and level of composite action) for FRP and reinforced concrete decks are evaluated. This paper demonstrates that FRP replacement decks are a viable alternative to reinforced concrete decks and identifies the differences in performances of various FRP deck systems. Two of the FRP panel systems were found to perform considerably better than the other deck systems. Issues that may reduce the service life of FRP deck systems are presented and discussed.     

Concrete Contribution to the Shear Resistance of Fiber Reinforced Polymer Reinforced Concrete Members

Seven beams were tested in bending to determine the concrete contribution to their shear resistance. The beams had similar dimensions and concrete strength and were reinforced with carbon fiber reinforced polymer bars for flexure without transverse reinforcement. They were designed to fail in shear rather than flexure. The test variables were the shear span to depth ratio, varying from 1.82 to 4.5, and the flexural reinforcement ratio, varying from 1.1 to 3.88 times the balanced strain ratio. The test results are analyzed and compared with the corresponding predicted values using the American Concrete Institute, the Canadian Standard, and the Japan Society of Civil Engineers (JSCF) fiber reinforced polymer design recommendations. Based on these results and previous experimental data, it is shown that the ACI recommendations are extremely conservative whereas the Canadian and JSCE recommendations, albeit still conservative, are in closer agreement with the experimental data. Overall the Canadian Standard’s predictions are in better agreement with experimental data than the JSCE predictions.     

Bond Between Near-Surface Mounted Carbon-Fiber-Reinforced Polymer Laminate Strips and Concrete

In recent years, a strengthening technique based on near-surface mounted (NSM) laminate strips of carbon-fiber-reinforced polymer (CFRP) has been used to increase the load-carrying capacity of concrete and masonry structures by introducing laminate strips into precut grooves on the concrete cover of the elements to be strengthened. The high experimentally derived levels of strength efficacy with concrete columns, beams, and masonry panels have presented NSM as a viable and promising technique. This practice requires no surface preparation work and, after cutting the groove, requires minimal installation time compared to the externally bonded reinforcing technique. A further advantage associated with NSM CFRP is its ability to significantly reduce the probability of harm resulting from fire, acts of vandalism, mechanical damage, and aging effects. To assess the bond behavior of CFRP to concrete, pullout-bending tests have been carried out. The influences of bond length and concrete strength on bond behavior are analyzed, the tests are described, and the results are presented and discussed in detail. Finally, a local stress-slip relationship is determined based on both experimental results and a numerical strategy.     

Analysis of Test Specimens for Cohesive Near-Bond Failure of Fiber-Reinforced Polymer-Plated Concrete

The failure mechanisms of reinforced concrete (RC) members change due to the application of externally bonded fiber-reinforced polymer reinforcement. Although an extensive literature is available describing the failure mechanisms of poststrengthened flexural systems, brittle failure modes caused by bond failure, such as midspan debonding and end peeling, need to be further investigated in order to identify and quantify the fracture processes that result in bond failure. Simplified experimental tests have been designed to idealize the bond between the laminate and the RC member. However, it is unclear how the simplified test results can be related to the actual flexural debonding failures. This paper investigates and compares two bond failure tests: a simplified test (or simple shear test) and a recently proposed shear/normal test. After discussing the characteristics of both tests and how they relate to the midspan debonding and end peeling failures, the shear/normal test is studied in more detail using a nonlinear finite-element fracture mechanics program. The program accounts for cohesive localized and distributed concrete crack damage and is capable of describing the geometrical discontinuities that induce different brittle failure mechanisms. The numerical results compare well with available experimental data and help explain the crack formation and propagation pattern up to specimen failure. Parametric studies are presented to elucidate the influence of different material parameters on the failure mechanisms.     

基于BERT和BiLSTM的语义槽填充

语义槽填充是对话系统中一项非常重要的任务,旨在为输入句子的每个单词标注正确的标签,其性能的好坏极大地影响着后续的对话管理模块.目前,使用深度学习方法解决该任务时,一般利用随机词向量或者预训练词向量作为模型的初始化词向量.但是,随机词向量存在不具备语义和语法信息的缺点;预训练词向量存在"一词一义"的缺点,无法为模型提供具...     

一种考虑强调的英语口语翻译模型

为了打破跨语言交际中的语言障碍,采用翻译系统实现该目的。传统翻译系统不能很容易地处理单词和强调级别之间关系。此外,由于所有组件都是单独训练的,没有进行联合优化。为此,提出一种基于序列到序列模型的处理连续强调级别的方法,并将机器翻译和强调翻译结合到一个单一的模型中,这大大简化了翻译过程,更容易进行联合优化。实验表明,提出的模型与现有方法相比,具有更好的性能,并且能够较好地满足实际需求。     

融合外部语义知识的中文文本蕴含识别

基于神经网络的文本蕴含识别模型通常仅从训练数据中学习推理知识,导致模型泛化能力较弱。提出一种融合外部语义知识的中文知识增强推理模型(CKEIM)。根据知网知识库的特点提取词级语义知识特征以构建注意力权重矩阵,同时从同义词词林知识库中选取词语相似度特征和上下位特征组成特征向量,并将注意力权重矩阵、特征向量与编码后的文本向量相结合融入神经网络的模型训练过程,实现中文文本蕴含的增强识别。实验结果表明,与增强序列推理模型相比,CKEIM在15%、50%和100%数据规模的CNLI训练集下识别准确率分别提升了3.7%、1.5%和0.9%,具有更好的中文文本蕴含识别性能和泛化能力。     

多源异构数据融合驱动的股票指数预测研究

现代信息技术的广泛应用使得资本市场投资者能够获得更及时、更有价值的信息,也更容易受到金融论坛、专业投资网站的影响。融合资本市场的多源异构数据对股票指数进行预测成为该领域的研究热点。提出了一种基于多源异构数据的长短期神经网络（Long Short-Term Memory,LSTM）模型,通过对融合资本市场交易数据、技术指标数据、投资者情绪三种源数据的量化来预测股票指数的走势。提出了一种可以提取深度情感特征的卷积神经网络（Convolutional Neural Networks,CNN）情感分析模型,构建了投资者情绪特征模型。利用“上证50指数”数据进行实验,结果显示：LSTM模型的预测准确率比传统模型更为优秀,数据源的增加也对模型准确率的提升有较大贡献,验证了该方法的可行性和有效性。