基于SRAM的通用存算一体架构平台在物联网中的应用

最近,存算一体(IMC)架构引起了广泛关注,并被认为有望成为突破冯诺依曼瓶颈的新型计算机架构,特别是在数据密集型(data-intensive)计算中能够带来显著的性能和功耗优势.其中,基于SRAM的IMC架构方案也被大量研究与应用.该文在一款基于SRAM的通用存算一体架构平台——DM-IMCA的基础上,探索IMC架构在物联网领域中的应用价值.具体来说,该文选取了物联网中包括信息安全、二值神经网络和图像处理在内的多个轻量级数据密集型应用,对算法进行分析或拆分,并将关键算法映射到DM-IMCA中的SRAM中,以达到加速应用计算的目的.实验结果显示,与基于传统冯诺依曼架构的基准系统相比,利用DM-IMCA来实现物联网中的轻量级计算密集型应用,可获得高达24倍的计算加速比.     

乌兰矿采矿现状及技术应对方案

本文分析了乌兰矿投产前期采矿现状及存在的主要问题,针对该矿所处蒙古国经济落后、投资风险大的现实状况,为避免生产中断、规避投资风险,早日回收前期投资考虑,采取了双斜坡道开拓、全尾胶结充填、高端壁空场嗣后充填采矿、多中段组合式连续开采等系列技术应对方案。大大降低了一次性投资规模及投资风险,前期投资得以快速回笼的同时,矿山产能也充分释放,确保了矿山的持续稳定,取得了较好的经济和社会效益。为海外地下近地表矿体开采矿山规避投资风险提供了很好的技术方案借鉴。     

Tempering–preservation treatment inactivated lipase in wheat bran and retained phenolic compounds

Wheat bran is rich in functional ingredients, but the high level of lipase limits its applications. Tempering–preservation treatment (at 70–90 °C with moisture of 20%–40% for 1–4 h) was exploited for stabilising wheat bran and its effect on polyphenols was investigated. The results showed that more lipase was inactivated at higher tempering moisture, temperature and longer time. The optimum condition for inactivation of wheat bran lipase was 30% moisture and 90 °C for 4 h. The inactivation rate reached 93.8% with a residual enzyme activity of 0.264 U g⁻¹. Under the optimum condition, the sum of free phenolic acids rose from 25.4 to 55.8 µg g⁻¹. As for bound phenolic acids, there was a slight increase of hydroxybenzoic acid derivatives but a slight decrease of hydroxycinnamic acid derivatives. The total contents of phenolic acids before and after stabilisation were not significantly different. This study showed the possibility of using tempering–preservation as an efficient method for inactivation of wheat bran lipase while maintaining its phenolic compounds, which could be used in the production of whole wheat flour.     

基于电机效率的纯电动汽车传动系统参数匹配研究

为降低能耗、提高续航里程,本文提出了以提高电机效率为目标的纯电动汽车三档变速传动方案.首先,在分析电机工作特性的基础上提出了三档传动系统参数匹配方法;然后以某纯电动汽车为研究对象,对其驱动系统进行了改动:方案一在不改变原车型动力系统参数的情况下将传动系统由一档改为三档,并运用所提出的方法确定各档位的传动比;方案二在满足整车动力性指标要求下对电机重新选型并匹配三档传动系统;最后利用ADVISOR软件进行NEDC工况仿真分析.结果表明:在不改变原车型动力系统参数的情况下,采用三档传动可使整车的动力性和经济性都有一定的提升,续航里程提高了 9.2％,百公里电耗降低了 1.28 kW·h;重新选型电机后,采用三档传动续航里程提高了 14.1％,百公里电耗降低了 1.87kW·h.     

中药桂枝总黄酮的提取工艺研究

优化桂枝总黄酮的提取工艺，建立桂枝的提取和含量测定方法。通过单因素试验，考察提取方法、提取溶剂浓度、提取体积、提取时间对药材总黄酮含量的影响，采用正交试验，优化提取工艺条件，筛选出桂枝总黄酮的最佳提取工艺。总黄酮最佳提取工艺为75%乙醇、料液比1:40、回流提取时间60 min。该方法能有效测定桂枝药材总黄酮的含量，为该药材的质量标准研究提供一定的科学依据。本方法重复性较好、方法稳定、可行。     

境外炼化工程项目投标竞争者分析报价策略

全球炼化工程总承包市场呈现出垄断竞争型特征,欧美和日韩国际工程公司由于技术和报价竞争力较强,一直占据本行业高端市场,中国炼化工程企业尚有差距.文章综合考虑投标报价阶段竞争者分析路径,提出了随机型和不完全信息静态博弈综合集成分析的投标报价最优决策模型,并以中东海湾地区某炼化工程总承包项目为例,进行多维度的实证分析,提出模型的有效性和可行性,为中国炼化工程企业"走出去"高质量发展提供有益的决策支持和信息参考.     

Synthesis of nanopowders with low agglomeration by elaborating Φ values for producing Gd2O3-MgO nanocomposites with extremely fine grain sizes and high mid-infrared transparency

Refining ceramic microstructures to the nanometric range to minimize light scattering provides an interesting methodology for developing novel optical ceramic materials. In this work, we reported the fabrication and properties of a new nanocomposite optical ceramic of Gd₂O₃-MgO. The citric acid sol-gel combustion method was adopted to fabricate Gd₂O₃-MgO nanocomposites with fine-grain sizes, dense microstructures and homogeneous phase domains. Nanopowders with low agglomeration and improved sinterability can be obtained by elaborating Φ values. Further refining of the microstructure of the nanocomposites was achieved by elaborating the hot-pressing conditions. The sample sintered at 65 MPa and 1300 °C showed a quite high hardness value of 14.3 ± 0.2 GPa, a high transmittance of 80.3 %–84.7 % over the 3?6 μm wavelength range, due mainly to its extremely fine-grain size of Gd₂O₃ and MgO (93 and 78 nm, respectively) and high density.     

基于双目标优化遗传算法和支持向量机的旋转机械退化状态识别

退化特征提取是机械健康状态监测的重要组成部分,伴随旋转机械长时间连续运转,退化特征出现性能波动甚至下降,给退化特征提取和选择造成了困难.首先利用一个特征映射算法库对振动信号提取特征,并基于Kolnogorov-Smirnov (KS)检验和Benjamini-Yekutieli过程对原始特征集进行过滤,然后利用双目标优化遗传算法(Bi-objective Optimization Genetic Algorithm,BOGA)结合支持向量机分类器(Support Vector Classifier,SVC),在有监督的环境下搜索出最佳特征子集,其中BOGA设置了SVC分类精确度和特征子集维数两个目标函数,前者进行最大化,后者进行最小化.通过在液压泵退化状态数据集上进行实验和在凯斯西楚大学轴承数据集与FRESH_PCAa、ReliefF、JMIM三种方法进行对比,验证了该方法在退化状态识别上的较好性能.