丙碳脱碳溶剂的改进研究

介绍一种由丙碳、叔胺和少量活化剂构成的复合溶剂的开发，并详细地测定了不同压力和操作工况下，这种复合溶剂对ＣＯ２的吸收和解吸性能，以及溶剂的缓蚀性能和抗氧化性能．实验证明这种复合溶剂与纯丙碳相比，具有吸收容量大、速率快、脱碳净化度高、对设备无腐蚀等特点，对改进现行的丙碳脱碳工艺具有应用价值．     

苎麻脱胶废水处理工艺的试验研究

本文针对苎麻脱胶废水的特点，进行了厌氧－好氧方法处理该废水的试验研究。结果表明：在常温下，当Ｎｖ为３．３３ｋｇＣＯＤ／ｍ＾３．ｄ，进行ＣＯＤ为１００００ｍｌ／Ｌ，ＨＲＴ为７２ｈ，厌氧流化床ＣＯＤ去除率为５０％左右。其出水稀释至１００ｍｇ／Ｌ左右，在０．３４ｋｇＣｏｄ／ｋｇＶＳＳ．ｄ负荷下，经６－８ｈ好氧处理即可达到国家排放标准。     

间二甲苯-磷酸-水三元体系的液-液相平衡

实验测定了不同温度下,间二甲苯-水二元体系和间二甲苯-磷酸-水三元体系的液-液相平衡数据。研究了磷酸中的杂质Fe~(3+)、Al~(3+)和Mg~(2+)对间二甲苯-磷酸-水三元体系相平衡的影响,以及杂质在两相中的分配情况。研究结果对共沸精馏浓缩磷酸提供了必要的理论依据和基础数据。     

Formation of novel polymeric nanoparticles

We have found that not only block copolymers but also ionomers can self-assemble in a selective solvent to form surfactant-free nanoparticles. The self-assembly can be induced by chemical reaction, polymer-polymer complexation, and microphase inversion in addition to the temperature. A recently developed microwave method for the preparation of uniform surfactant-free polymeric nanoparticles is also reviewed. Our results have revealed that for a given dispersion, the particle surface area occupied per stabilizer (surfactant, polymer chains, and ionic groups) is close to a constant.     

文本分析的接受美学视角

目前的普通文体学理论只是关注文本本身的分析,难以确保丈本主题把握的正确性以及文本分析过程的科学性和精确度.接受美学认为作者创作过程和读者接受过程是不可或缺的,二者的结合可以形成新的三位一体结构,是对于普通文体学理论的有益补充.     

基于企业生命周期的人力资源经理角色分析

文章从企业生命周期的角度,探讨了人力资源经理在不同阶段主要扮演的角色,为企业选拔和培训人力资源经理提供新的视角。将企业生命周期分为创业期、成长期、成熟期和衰退期四个阶段,依据企业生命周期不同阶段的特点及人力资源管理的策略,得出人力资源经理在不同阶段分别主要扮演战略合作伙伴、HRM专家、员工激励者、变革推动者的角色。     

浅析在VB中使用ADO开发数据库应用程序

本文解析如何通过ADO对象对数据库进行操作,并介绍了ADO对象群中的几个重要对象Connection、Recordset、Command等。     

Tar sand extractions with microemulsions: I-the dissolution of light hydrocarbons by microemulsions using 2-butoxyethanol and diethylmethylamine as cosurfactants

For oil sand extractions with microemulsions it is important to disperse large quantities of light hydrocarbons in an aqueous medium. Fundamental studies on the properties of 2-butoxyethanol (BE) and diethylmethylamine (Et₂McN) in water suggest that these two liquids could be more effective cosurfactants than the usual alcohols used for this purpose. The phase diagrams of microemulsions using BE and Et₂MeN as cosurfactants, combined with typical ionic and non-ionic surfactants and typical aliphatic and aromatic hydrocarbons, were therefore investigated and compared with microemulsions based on n-butanol. Although the phase diagrams depend significantly on the nature of the surfactant and of the oil, the monophasic region generally increases with the cosurfactant in the order n-butanol < Et₂McN < BE. With the active mixture BE-cetyltrimethylammonium bromide, temperature has little effect on the phase diagram and NaCl generally destabilizes the microemulsion.     

Catalytic Combustion of Model Carbon Particles (CPs) over Pt/β-SiC Catalyst

Platinum catalyst supported on a medium surface area -SiC was successfully used for the catalytic combustion of model carbon particles and compared to a catalyst supported on a low surface area -SiC. The -SiC-based catalyst showed no deactivation as a function of cycling tests while a strong deactivation was observed on the -SiC-based catalyst. This deactivation was attributed to the progressive encapsulation of the platinum particles by a layer of silica which built up during the combustion cycle. These results render possible the use of Pt/-SiC catalyst as a diesel carbon particle catalytic filter with continuous regeneration.     

Re-GAN：残差生成式对抗网络算法

目的 生成式对抗网络（generative adversarial network,GAN）是一种无监督生成模型,通过生成模型和判别模型的博弈学习生成图像。GAN的生成模型是逐级直接生成图像,下级网络无法得知上级网络学习的特征,以至于生成的图像多样性不够丰富。另外,随着网络层数的增加,参数变多,反向传播变得困难,出现训练不稳定和梯度消失等问题。针对上述问题,基于残差网络（residual network,ResNet）和组标准化（group normalization,GN）,提出了一种残差生成式对抗网络（residual generative adversarial networks,Re-GAN）。方法 Re-GAN在生成模型中构建深度残差网络模块,通过跳连接的方式融合上级网络学习的特征,增强生成图像的多样性和质量,改善反向传播过程,增强生成式对抗网络的训练稳定性,缓解梯度消失。随后采用组标准化（GN）来适应不同批次的学习,使训练过程更加稳定。结果 在Cifar10、CelebA和LSUN数据集上对算法的性能进行测试。Re-GAN的IS （inception score）均值在批次为64时,比DCGAN （deep convolutional GAN）和WGAN （Wasserstein-GAN）分别提高了5%和30%,在批次为4时,比DCGAN和WGAN分别提高了0.2%和13%,表明无论批次大小,Re-GAN生成图像具有很好的多样性。Re-GAN的FID （Fréchet inception distance）在批次为64时比DCGAN和WGAN分别降低了18%和11%,在批次为4时比DCGAN和WGAN分别降低了4%和10%,表明Re-GAN生成图像的质量更好。同时,Re-GAN缓解了训练过程中出现的训练不稳定和梯度消失等问题。结论 实验结果表明,在图像生成方面,Re-GAN的生成图像质量高、多样性强;在网络训练方面,Re-GAN在不同批次下的训练具有更好的兼容性,使训练过程更加稳定,梯度消失得到缓解。