Fabrication and abrasive wear properties of metal matrix composites reinforced with three-dimensional network structure

Reticulated polyurethane was chosen as the preceramic material for preparing the porous preform using the replication process. The immersing and sintering processes were each performed twice for fabricating a high-porosity and super-strong skeleton. The aluminum magnesium matrix composites reinforced with three-dimensional network structure were prepared using the infiltration technique by pressure assisting and vacuum driving. Light interfacial reactions have played a profitable role in most of the ceramic-metal systems. The metal matrix composites interpenetrated with the ceramic phase have a higher wear resistance than the metal matrix phase. The volume fraction of ceramic reinforcement has a significant effect on the abrasive wear, and the wear rate can be decreased with the increase of the volume fraction of reinforcement.

SYNTHESIS AND CRYSTAL STRUCTURE OF THE FIRST THREE DIMENSIONAL NETWORK CU(Ⅱ) COMPLEX BRIDGED BY BOTH OXAMIDE AND AZIDE

A novel three dimensional network complex polymer [Cu_4(oxen)_2(N_3)_3]_n(ClO_4)_n·2nH_2O, where oxen is N,N' -bis(2-aminoethyl)oxamide dianion, has been synthesized. It crystallizes in triclinic system, space group P, with a=11.486(2), b=11.706(3), c=12.291(3) , α=77.42(2), β=67.59(2), γ=77.96(2)°, and z=2. The least-square refinements converged at R=0.047, with 3416 observed unique reflections. The complex has a pronounced three-dimensional character and can be viewed as the tetranuclear asymmetric repeating units through inversion and translation operations to extend a three-dimensional network. The structure of Cu_4 asymmetric unit consists of two square planar and two square pyramidal Cu central atoms linked by both azide ligands in end-on and end-to-end bonding modes, and oxamidate bridge in trans conformation.

A Novel Hydrogen-bonded Three-dimensional Network Complex Containing Nickel

A novel complex, (H 3O) 2[Ni(2,6-pydc) 2]·2H 2O was synthesized in an aqueous solution and characterized by means of single-crystal X-ray diffraction, elemental analyses and IR spectra. The X-ray structural analysis revealed that the novel compound forms three-dimensional(3D) networks by both π-π stacking and hydrogen-bonding interactions. The crystal data for the complex are a=13.853(3) nm, b=9.6892(19) nm, c=13.732(3) nm, α=90.00°, β=115.52(3)°, γ=90.00°, Z=3, R 1=0.0786, wR 2=0.1522.

Analysis of Mean Monthly Rainfall Runoff Data of Indian Catchments Using Dimensionless Variables by Neural Network

This paper focuses on a concept of using dimensionless variables as input and output to Artificial Neural Network (ANN) and discusses the improvement in the results in terms of various performance criteria as well as simplification of ANN structure for modeling rainfall-runoff process in certain Indian catchments. In the present work, runoff is taken as the response (output) variable while rainfall, slope, area of catchment and forest cover are taken as input parameters. The data used in this study are taken from six drainage basins in the Indian provinces of Madhya Pradesh, Bihar, Rajasthan, West Bengal and Tamil Nadu, located in the different hydro-climatic zones. A standard statistical performance evaluation measures such as root mean square (RMSE), Nash–Sutcliffe efficiency and Correlation coefficient were employed to evaluate the performances of various models developed. The results obtained in this study indicate that ANN model using dimensionless variables were able to provide a better representation of rainfall–runoff process in comparison with the ANN models using process variables investigated in this study.

A New Three-dimensional Network Constructed by Heptamolybdate, Sodium Ions and Hexamethylene Tetramine Cations via Hydrogen Bonds

The crystal structure of the title compound [Na2(OH2)5]2+[C6H12N4H2]2-2+ [Mo7O24]6 ?4H2O, prepared from an aqueous solution of Na2MoO4 ?2H2O in the presence of MoCl3 and hexamethylene tetramine, has been determined by single-crystal X-ray diffraction. The crystal is of orthorhombic, space group Pnma with a = 14.6113(2), b = 18.6833(1), c = 15.3712(2), V = 4196.14(8)3, Z = 4, Mr = 1548.13, F(000) = 3016, = 2.157 mm-1 and Dc = 2.451 g/cm3. The final R factor is 0.0526 for 3818 unique observed reflections (I > 2(I)). The structural analysis reveals that heptamolybdate anions in the title compound consist of seven edge-sharing MoO6 octahedra, and are linked into a three-dimensional framework by sodium ions and hydrogen bonds.

A New Eight-connected Three-dimensional Network Based on a Tetranuclear Zinc Cluster Building Block

One interesting coordination polymer, [Zn2（1,2,4-BTC）（OH）（H2O）2]2·2H2O 1, has been synthesized from 1,2,4-BTC （1,2,4-BTC = 1,2,4-bentricarboxylate） under hydrothermal conditions and characterized by elemental analyses, IR, TG and single-crystal X-ray diffraction. Complex I crystallizes in triclinic, space group P^-1, with a = 6.5200（13）, b = 9,0600（18）, c = 10.968（2） A^°, α = 111.55（3）, β = 92.07（3）,γ= 95.03（3）°, C9H10O10Zn2, Mr = 408.91, V= 598.7（2） A^°^3, Dc = 2.268 g/cm^3, F（000） = 408 and Z = 2. X-ray diffraction analysis reveals that complex 1 is a three-dimensional network built from tetranuclear Zn（Ⅱ） building unit. In this complex, the Zn4 unit is an eight-connected knot, while 1,2,4-BTC a four-connected knot. This results in a CaF2 topology. To the best of our knowledge, such Zn4 unit is the first 8-connected building block built from asymmetry ligand.

Prediction of Salinity Variations in a Tidal Estuary Using Artificial Neural Network and Three-Dimensional Hydrodynamic Models

The simulation of salinity at different locations of a tidal river using physically-based hydrodynamic models is quite cumbersome because it requires many types of data, such as hydrological and hydraulic time series at boundaries, river geometry, and adjusted coefficients. Therefore, an artificial neural network (ANN) technique using a back-propagation neural network (BPNN) and a radial basis function neural network (RBFNN) is adopted as an effective alternative in salinity simulation studies. The present study focuses on comparing the performance of BPNN, RBFNN, and three-dimensional hydrodynamic models as applied to a tidal estuarine system. The observed salinity data sets collected from 18 to 22 May, 16 to 22 October, and 26 to 30 October 2002 (totaling 4320 data points) were used for BPNN and RBFNN model training and for hydrodynamic model calibration. The data sets collected from 30 May to 2 June and 11 to 15 November 2002 (totaling 2592 data points) were adopted for BPNN and RBFNN model verification and for hydrodynamic model verification. The results revealed that the ANN (BPNN and RBFNN) models were capable of predicting the nonlinear time series behavior of salinity to the multiple forcing signals of water stages at different stations and freshwater input at upstream boundaries. The salinity predicted by the ANN models was better than that predicted by the physically based hydrodynamic model. This study suggests that BPNN and RBFNN models are easy-to-use modeling tools for simulating the salinity variation in a tidal estuarine system.

Research of Tile Type Transceiver Module Integrating with Two-Dimensional Sum Difference Network

Transceiver module and two-dimensional sum difference network are important components of phased array antenna. In this paper, multilayer printed board is used to integrate millimeter wave multi-channel transceiver circuit and sum difference network. The interconnection between them is realized through RF coaxial vertical transition. At the same time, the heat dissipation design and inter channel shielding design of the module are carried out. The RF and low frequency required by the module are completed through the wiring between and within the dielectric plate layers. Finally, 128 arrays are fabricated and verified by multi-channel passive test. The results show that the type transceiver module integrating with two-dimensional sum difference network has good performance, and 128 channels have excellent amplitude and phase characteristics. The integration technology has the characteristics of lightweight, miniaturization, high integration and low manufacturing cost. It can be widely used in miniaturized phased array antennas.

STUDY OF RECOGNITION TECHNIQUE OF RADAR TARGET'S ONE-DIMENSIONAL IMAGES BASED ON RADIAL BASIS FUNCTION NETWORK

This paper studies the problem applying Radial Basis Function Network(RBFN) which is trained by the Recursive Least Square Algorithm(RLSA) to the recognition of one dimensional images of radar targets. The equivalence between the RBFN and the estimate of Parzen window probabilistic density is proved. It is pointed out that the I/O functions in RBFN hidden units can be generalized to general Parzen window probabilistic kernel function or potential function, too. This paper discusses the effects of the shape parameter a in the RBFN and the forgotten factor A in RLSA on the results of the recognition of three kinds of kernel function such as Gaussian, triangle, double-exponential, at the same time, also discusses the relationship between A and the training time in the RBFN.

基于Bi-LSTM和改进残差学习的风电功率超短期预测方法

现有的方法在以风电功率时间序列拟合功率曲线时,难以表达风电功率数据所包含的趋势性和周期性等时间信息而出现性能退化问题,从而导致预测精度下降。为了解决性能退化问题从而提高风电功率时间序列预测的精度,提出了基于双向长短时记忆(Bi-LSTM)和改进残差学习的风电功率预测方法。方法由两个部分组成,第一部分是以Bi-LSTM为主的多残差块上,结合稠密残差块网络(DenseNet)与多级残差网络(MRN)的残差连接方式,并且在残差连接上使用一维卷积神经网络(1D CNN)来提取风电功率值中时序的非线性特征部分。第二部分是Bi-LSTM与全连接层(Dense)组成的解码器,将多残差块提取到的功率值时序非线性特征映射为预测结果。方法在实际运行的风电功率数据上进行实验,并与常见的残差网络方法和时间序列预测方法进行对比。方法相比于其他模型方法有着更高的预测精度以及更好的泛化能力。

双向数据扩充和LSTNet的户用光伏发电预测

整县光伏政策促使小容量屋顶光伏急剧增长,实现屋顶分布式光伏超短期发电功率的准确预测是分析海量细粒户用光伏电站对电力系统影响的前提。然而,屋顶分布式光伏在原有波动性的基础上存在小容量、分散式、离线式经营的特点,同时缺乏准确的气象数据,使得光伏功率预测异常复杂。为此,文章在有限数据下纵向地从光伏系统历史功率数据中搜索相似样本,横向地收集相邻分布式光伏发电用户功率数据,实现双向数据扩充,在一定程度上克服了光伏发电预测对于一些关键输入特征的依赖;在此基础上借助LSTNet(Long-and Short-term Time-series Network)神经网络的短期局部特征捕捉、长期时序信息强化、周期线性成分提取功能实现光伏功率预测。实验结果表明,在缺乏重要辐照数据的情况下,所提模型仍具有较好的预测精度。

A Sensor Awakening Algorithm for Wireless Multimedia Sensor Networks Three Dimensional Target Tracking

For node awakening in wireless multi-sensor networks, an algorithm is put forward for three dimensional tar- get tracking. To monitor target dynamically in three dimensional area by controlling nodes, we constract vir- tual force between moving target and the current sense node depending on the virtual potential method, then select the next sense node with information gain function, so that when target randomly move in the specific three dimensional area, the maximum sensing ratio of motion trajectory is get with few nodes. The proposed algorithm is verified from the simulations.

基于残差神经网络的大地电磁二维反演

本文开展了基于残差神经网络的大地电磁二维反演研究.采用高斯随机场设计并生成了5万个不同规模、不同边界形状(规则边界与光滑边界)、不同电阻率对比度、单个到多个电性异常体模型,通过基于二维交错网格有限差分批量并行正演程序对模型进行正演计算并输出TE和TM极化模式的视电阻率值,对模型正演TM极化响应的60%数据加入5%的高斯误差,其余为无噪数据,将其作为深度学习反演模型的训练样本;在网络结构设计中引入Res Net-50网络结构深度学习模型对样本进行训练;最后通过测试数据的反演验证了大地电磁深度学习反演模型的可靠性.模型反演的结果表明,神经网络反演模型能够实时输出反演结果,反演结果较为准确,可以有效地刻画各类异常体的边界,对异常体电性参数的恢复也较为准确.

基于DCGAN数据增强的樱桃番茄可溶性固形物含量光谱检测方法

针对樱桃番茄在实际检测中样品数不足的特点,本研究提出一种深度卷积生成对抗网络(deep convolutional generative adversarial network,DCGAN)模型以同时扩充光谱数据及可溶性固形物含量(soluble solids content,SSC)标签数据,并建立一维卷积神经网络回归(one dimensional-convolutional neural networks regression,1D-CNNR)模型以提高模型的预测精度和泛化能力。为了比较,分别建立偏最小二乘回归(partial least squares regression,PLSR)模型和支持向量机回归(support vector regression,SVR)模型。将原始80个样品数据集、1000个样品的DCGAN扩充数据集和1080个样品的合并数据集,分别结合1D-CNNR、SVR及PLSR进行建模与预测。为了进一步验证模型的泛化能力,一批新的总数为40个样品的樱桃番茄数据作为上述3个模型的新测试集。结果显示,使用合并数据集划分所得校正集进行1D-CNNR建模后,模型为最优的SSC回归检测模型。此时1D-CNNR面向合并样品测试集的预测准确率最高,预测相关系数r_(p)=0.9807,均方根误差RMSE_(p)=0.1929;与SVR与PLSR对比,1D-CNNR面向新的40个样品数据集的预测准确率也最高,其r_(p)=0.9638,RMSE_(p)=0.2245。本研究可为有效准确检测樱桃番茄的可溶性固形物含量提供一种新思路。

Autoformer双分支网络下的多元空气质量长时预测研究

空气质量数据复杂多变,现有方法难以捕捉长期依赖关系,且对季节趋势和多变量建模不足。针对以上问题,研究基于Autoformer模型进行改进,创新性地融入了特征渐进挖掘和多维深度联系两个分支。首先,特征渐进挖掘分支通过序列分解模块将空气质量数据分解为季节分量和趋势分量,对季节分量设计了一种特征增强模块(Feature Enhancement,FE)以捕获关键特征。其次,对趋势分量设计了门控-膨胀因果卷积模块(Gated Linear Unit Dilated Causal Convolution,GLU-DCC)来获取高级时序特征。最后,构建了多维深度联系分支,该分支通过引入维度-分段嵌入模块(Dimension-Segment-Wise Embedding,DSW)和两阶段注意力机制(Two Stage Attention,TSA)提取了多元空气质量数据中的跨维度相关性。研究对两个站点进行空气质量指数(Air Quality Index,AQI)预测,试验结果显示:与基线模型相比,研究模型的两个数据集的均方误差(M_(SE))分别平均下降了47.6%和57.5%,平均绝对误差(M_(AE))分别平均下降了15.5%和38.5%,具有更优的预测性能。

基于2D-CNN和Cox-Stuart早停机制的癫痫预测模型

针对如何有效预测癫痫患者是否将要发病这一问题,提出一种基于非独立患者的2维卷积神经网络(2D-CNN)和Cox-Stuart检验法的癫痫预测模型方法。首先对脑电数据做归一化处理,使用陷波滤波器和高通滤波器滤除脑电信号的噪声;将滤波后的信号输入到2D-CNN模型中进行特征提取和分类,使用Cox-Stuart方法检测是否需要早停,从而降低模型的计算复杂度和时间复杂度。此外,分别在发作前期为10、30、60 min的情况下对模型进行测试,结果显示,发作前期为10 min时,模型的效果最优。在测试集上的准确率为97.70%,灵敏度为97.36%,特异性为98.04%,具有良好的性能。

基于光谱降维和特征融合的高光谱目标跟踪

针对现有高光谱视频跟踪算法在目标尺度变化时表现不佳的问题,提出一种基于光谱降维和特征融合的高光谱视频目标跟踪算法。首先,计算目标局部光谱曲线的差值并结合特征值排序和阈值设定获取目标光谱曲线;随后,利用目标光谱曲线与高光谱图像进行光谱角距离计算来实现降维;之后,利用改进的多尺度胶囊网络提取多尺度特征,为利用不同尺度的信息,将降维生成的掩模进行多尺度特征融合;最后,将融合的多尺度特征输入分类和回归胶囊,利用模版更新机制增强跟踪的稳定性和鲁棒性,使得所提算法能够更好地应对尺度变化带来的挑战。实验结果表明,所提算法在应对尺度变化挑战时具有优越性。

基于立体水网的新疆盐碱地降碱体系构建

新疆盐碱地问题较为突出,为进一步推进盐碱地治理,通过对新疆盐碱地成因、分布进行分析,提出建立从上游到下游、从地表到地下的立体水网解决盐碱地问题,即:在灌区上游建设“横坎儿井式”集水、输水工程,在灌区中游建设排水工程,在灌区下游实施冬灌+“干播湿出”“磁化水+渗灌”的灌溉模式,构建“上游控源、中游输排、下游降碱”的立体控碱、降碱水网体系,对改善土壤盐碱化问题,提高土地资源利用率,水资源利用效率和效益提供理论支撑。

Three-Dimensional TIN Algorithm for Digital Terrain Modeling

The problem of taking an unorganized point cloud in 3D space and fitting a polyhedral surface to those points is both important and difficult. Aiming at increasing applications of full three dimensional digital terrain surface modeling, a new algorithm for the automatic generation of three dimensional triangulated irregular network from a point cloud is pro- posed. Based on the local topological consistency test, a combined algorithm of constrained 3D Delaunay triangulation and region-growing is extended to ensure topologically correct reconstruction. This paper also introduced an efficient neighbor- ing triangle location method by making full use of the surface normal information. Experimental results prove that this algo- rithm can efficiently obtain the most reasonable reconstructed mesh surface with arbitrary topology, wherein the automati- cally reconstructed surface has only small topological difference from the true surface. This algorithm has potential applica- tions to virtual environments, computer vision, and so on.

基于神经网络的船舶阻力预报研究

常规代理模型的阻力预报是以主尺度比、船型系数等作为输入,相比于CFD计算时输入完整船型,其较低的信息密度导致代理模型预报精度较低。本文以4108个完整船型几何形状特征张量作为输入,采用神经网络作为代理模型,以船舶的总阻力系数作为输出,研究船型阻力的高维度、高精度预报方法。首先,将船型进行无量纲化处理,并提取特征张量作为输入;然后,建立神经网络模型,搭建输入层、隐藏层和输出层;最后,将船型的特征张量与总阻力系数输入神经网络,通过误差反向传播进行训练,直至损失函数值收敛。本文研究结果可为基于高维代理模型的阻力性能预报提供理论和技术支持。