基于卷积神经网络的5G蜂窝网络无线定位方法

5G蜂窝网络发展迅猛，其覆盖面积将逐渐增大，因此使用5G蜂窝网络进行定位是有研究潜力的研究方向。本文提出一种新的深度学习技术来实现高效、高精度和低占用的定位，以代替传统指纹定位过程中繁重的指纹库生成以及距离计算。该方法建立了一个特殊的卷积神经网络，并根据5G天线信号的接收信号强度指示、相位和到达角等特征量，选择合适的输入数据格式构造样本组建训练集，对该卷积神经网络进行训练。训练得到的卷积神经网络可以替代指纹定位中的庞大指纹库，非常有利于直接在5G移动设备端实现定位。虽然卷积神经网络在训练过程中需要大量时间，但在训练完毕后直接进行分类定位的速度非常快，可以保障定位实现的实时性。本文所实现的卷积神经网络权重与偏置所占内存不到0.5 MB，且能够在实际应用环境中以95%的定位准确率以及0.1 m的平均定位精度实现高精度定位。     

Experimental investigation on diamond wire sawing of Si3N4 ceramics considering the evolution of wire cutting performance

When diamond wire saw is used in machining silicon nitride ceramics (Si₃N₄ ceramics), the ultra-hardness of Si₃N₄ causes the saw wire to wear out, which leads to the saw wire cutting performance constantly changing during its life cycle, and thus the machined quality of Si₃N₄ ceramics is affected. Surface roughness and topography are important indicators of the quality of the machined surface. In this paper, the diamond wire saw cutting experiment of Si₃N₄ ceramics was carried out, the effect of the evolution of saw wire cutting performance on the surface roughness and topography of Si₃N₄ ceramics as-sawn slices was investigated based on the analysis of the changes of saw wire wear topography, breaking force, bow angle and kerf loss during the sawing process. The results show that the surface roughness along the saw wire motion direction and the workpiece feed direction tends to decrease and then increase with the evolution of the cutting performance of the saw wire, which accords well with the trend of the as-sawn slices surface morphology. The results of the study can provide experimental reference for the development of high precision diamond wire saw cutting technology for Si₃N₄ ceramics.     

Improved photoluminescence and multi-mode optical thermometry of Er3+/Yb3+ co-doped (Ba,Sr)3Lu4O9 phosphors

Contactless optical thermometers have attracted extensive attentions for applications in scientific research and technological fields due to their apparent advantages. Herein, a novel sequence of Ba_3-xSr_xLu₄O₉ (B_3-xS_xLO):Er³⁺/Yb³⁺ phosphors were successfully prepared to investigate the temperature sensing property. By establishing energy transfer from Yb³⁺ to Er³⁺ and regulating the local lattice environment, up-conversion luminescence of Er³⁺ is dramatically improved when excited by 980 nm laser. This can effectively promote signal-noise ratio and reduce the errors in temperature detection. Furthermore, a multi-mode optical thermometry, which includes the fluorescence intensity ratio (FIR) from two thermally coupled levels of ²H_11/2/⁴S_3/2, FIR based on non-thermally coupled system of ²H_11/2/⁴F_9/2 and fluorescence lifetime of ⁴S_3/2 state of Er³⁺, was explored systematically. The fabricated samples exhibit the superior temperature measurement performances containing wide temperature-sensing range, superior signal discriminability, high sensitivity and favorable repeatability, indicative of the enormous utilization prospects of B_3-xS_xLO:Er³⁺/Yb³⁺ for thermometry.     

The effect of traffic data source on deterioration rates of heavy-duty flexible pavements

The purpose of this study is to assess the effect of traffic data source (estimated vs. actual) on predicted progression rates of roughness and rutting for heavy-duty flexible pavements of rural freeways. Progression rates are predicted using calibrated HDM-4 models. The assessment is performed in terms of variations in maintenance intervention timing associated with the variations in progression rates. Time series pavement condition data (covering 3–5 years) have been collected for 7 sections of rural freeways for use in calibrating HDM-4 deterioration models. They range in length from 10 to 60.8 km and cover different traffic volumes, climate zones and subgrade soil types. For these sections, estimated annual average daily traffic (AADT), growth factors and assumed loading have been extracted from relevant database. Only six segments of these sections have Weigh-in-Motion (WIM) sites so relevant actual AADT, growth factors and axle load distributions have been extracted from WIM reports. The results of running the calibrated HDM-4 deterioration models using different traffic data show that actual traffic data from WIM sites result in higher rates of deterioration to that of estimated data for four sites, resulting in earlier intervention timing and higher present value agency cost. The other two sites have lower rates with actual data due to lower traffic loading than estimated.     

滚齿工艺对齿根粗糙度的影响

针对国内外风电和工业齿轮箱轻量化的设计趋势,较低的安全系数对齿轮齿根弯曲强度提出了更为严格的要求。研究表明,在材料技术和齿形设计难以再优化的情况下,滚齿加工的粗糙度,尤其是齿根部位的粗糙度对齿轮弯曲强度的影响尤为敏感。针对滚齿工艺过程,从滚齿加工参数、刀具前刀面粗糙度和刀具刃口圆角三方面进行分析,验证其对齿根粗糙度的影响。     

某贫磁赤铁矿石选矿工艺试验研究

为了使某矿山贫磁赤铁矿资源得到合理的开发利用，针对现开采部位矿石的性质和可选性问题，进行了工艺矿物学和可选性试验研究。试验结果表明：采用弱磁、强磁—阴离子反浮选工艺可得到精矿铁品位68.55%、精矿产率36.41%、金属回收率79.92%、尾矿品位9.86%的较好选别指标，说明矿石的可选性较好，为该矿石的合理开发利用提供了可靠的技术依据。     

一种波长调制干涉仪光强实时反馈控制系统

提出了一种实时反馈控制系统,保证波长调制干涉仪的激光光强在波长调谐过程中的稳定性,以减小波长调制干涉仪因激光光强的波动产生的相位误差。该控制系统由光电探测器、数据采集卡、光电调幅器和控制软件等组成。实验结果表明,该系统的响应速度超过了干涉仪获取干涉图的速度,不仅能够稳定输出理想的光强值,而且还提高了干涉仪的测量精度。     

Investigations of geometrical structure and morphology of samples made of hard machinable materials after wire electrical discharge machining and vibro‐abrasive finishing

In this article the results concerned influence of processing conditions of the wire electrical discharge machining and vibro‐abrasive machining on the surface layer and morphology of samples made of hard machinable materials used in aircraft industry like: Titanium 5553 β, Inconel 617, Hastelloy X and Magnesium AZ31 have been presented. For this purpose the cubic and cylindrical samples made of hard machinable alloys have been prepared using optimal electric parameters of wire‐cut electrical discharge machining and finally they have been polished using circular vibratory finishing technology and different ceramic shaped stones.     

钢铁生产的水能关系分析

从钢铁生产过程中水资源使用时的能耗角度研究了钢铁联合企业的水能关系，提出了水能强度的概念来评价企业生产过程中水资源利用的节能水平，建立了钢铁企业的水能关系模型。以中国某大型钢铁联合企业为例，计算并分析了该企业的水能关系。该企业总的水能量为55 709 kW·h/h，重复用水水能量占整个钢铁企业全部水能量的82%，补水水能量占16%，排水水能量占1%；各工序中热轧工序占比最大，其次是炼铁、炼钢工序，冷轧、烧结和炼焦工序较低。该企业的吨钢水能强度为0.208 kW·h/m3，炼铁工序的水能强度最高，热轧、炼钢工序次之，烧结、冷轧和炼焦工序较低。最后，从钢铁生产过程水资源利用的角度得到节能的方向及措施。     

Enhancing the Tribological Behavior of Lubricating Oil by Adding TiO2, Graphene,and TiO2/Graphene Nanoparticles

This article investigates the tribological behavior of nanoparticles (NPs) of titanium dioxide anatase TiO₂ (A), graphene, and TiO₂ (A) + graphene added to the pure base oil group ΙΙ (PBO-GΙΙ). The morphology of these two nanostructures of TiO₂ (A) and graphene was characterized by transmission electron microscopy (TEM). Oleic acid (OA) was blended as a surfactant into the formulation to help stabilize the NPs in the lubricant oil. A four-ball test rig was used to determine the tribological performance of six different samples, and an image acquisition system was used to examine and measure the wear scar diameter of the stationary balls. Field emission–scanning electron microscopy (FE-SEM) was used to examine the wear morphology. Energy-dispersive X-ray spectroscopy (EDX), element mapping, and Raman spectroscopy were employed to confirm the presence of (TiO₂ (A) + graphene) and the formation of a tribolayer/film on the mating surfaces. Moreover, a 3D optical surface texture analyzer was utilized to investigate the scar topography and tribological performance. The experiments proved that adding (0.4?wt% TiO₂ (A) + 0.2?wt% graphene) to the PBO-GΙΙ optimized its tribological behavior. These excellent results can be attributed to the dual additive effect and the formation of a tribofilm of NPs during sliding motion. Furthermore, the average reductions in the coefficient of friction (COF), wear scar diameter (WSD), and specific wear rate (SWR) were 38.83, 36.78, and 15.78%, respectively, for (0.4?wt% TiO₂ (A) + 0.2?wt% graphene) nanolubricant compared to plain PBO-GΙΙ lubricant.