针对网络化作战体系中目标重要度评估问题,结合层次分析和复杂网络方法,从目标节点特征、网络指标积极效应、网络指标消极效应3个方面提出新的评估方法.从能力视角描述目标自身特征;基于复杂网络理论,从积极效应层面描述目标的资源关联性、流转时效性、协同性和连通关键性,从消极效应层面描述目标的离散度;构建效能函数计算目标的网络关键度;根据目标自身特征和网络关键度给出目标的评估分.该方法综合考虑了目标自身特征以及目标在作战体系中的地位和作用,实验表明评估结果准确、可靠.
为计算碳纤维增强复合材料(carbon fiber reinforced plastic,CFRP)对称铺层环形法兰连接的拉伸刚度,基于圆板对称弯曲和经典层合板理论提出一种考虑环向支撑、材料各向异性和撬力作用系数的计算方法.在制备过程中,采用预浸料手工铺设和模压成型的方法成型试件,而后开展了 3 组不同法兰板厚度(2.2,3.0,3.8 mm)连接件轴向拉伸试验.通过对比理论计算值与试验结果,验证了计算方法的正确性,误差均不大于 13.61%.通过理论模型分析影响因素,结果表明:拉伸刚度随法兰板厚度、外径和内径的增大而增大,随螺栓所在的半径变大而减小;改变螺栓位置对于拉伸刚度的提升比改变法兰板内外径更显著,当螺栓位置由 90 mm调整至 85 mm时,拉伸刚度提升 59.54%.
深度学习的快速发展给视频压缩感知重构提供了新思路.受网络模型限制,现有的基于深度学习的压缩感知重构方法不能充分利用视频的空时特征,且对于超过 16 帧的视频段重构效果不够理想.采用Transformer网络构建压缩感知重构网络,利用 Transformer 网络在序列信号处理方面的优势构建空时注意力提取模块,学习视频帧间的空时注意力特征,更好地实现对视频连续帧的建模,从而解决长跨度视频段压缩感知重构问题.实验结果表明:所提方法在处理 32 张视频帧的视频分段时,能达到 30 dB以上的重构精度,在处理 96 张视频帧的视频分段时,仍能达到 27 dB以上的良好性能.
语言模型是自然语言处理领域最重要的任务之一,并以迁移学习的方式影响着机器翻译、机器阅读理解、自动文本摘要等诸多下游任务.依托多头自注意力构建的大规模预训练语言模型在特征提取和文本表征能力上相较之前的神经网络模型取得了较大提升,却也极大地增加了模型训练的时空复杂度.为此,从模糊群决策的角度出发,将毕达哥拉斯模糊非偏好函数用作多头注意力跨度范围的先验,提出一种自适应的注意力跨度调节机制,较大程度地改善了模型捕捉长距离文本依赖的能力,并使模型的整体计算复杂度相较原始Transformer结构维持在较低水平.公开语言模型数据集上的实验表明:所提方法在困惑度指标上取得了较好的性能,超越了多种以往的方法.
当前侦察设备波段宽、分辨率高,具备全天候、全天时侦察能力,对现有武器装备的伪装提出了挑战.伪装涂料采用功能颜、填料,通过一定的制备、施工工艺涂覆在目标表面,在不影响武器装备战场机动性能的前提下实现武器装备的伪装,是目前最有效的伪装技术手段之一.在阐述光学伪装涂料、红外伪装涂料、光学兼容红外伪装涂料伪装机理的基础上,梳理了 3 种伪装涂料技术的研究进展,总结了 3 种伪装涂料的技术特点和存在的问题,对伪装涂料技术今后的发展进行了展望,提出了发展光学兼容红外伪装涂料的技术实现途径,为有效对抗日益发展的先进侦察技术,满足战场伪装需求提供了方法指导.
混沌信号具有隐蔽性、不可预测性和高度复杂等特点,特别适用于保密通信和扩频通信等领域.为有效解决有限时间同步控制过度依赖初始条件的问题,基于固定时间稳定性理论,通过设计状态反馈控制器给出了忆阻神经网络实现固定时间同步的充分条件,并对停息时间进行了精确估计.在此基础上,以驱动混沌忆阻神经网络为发射器,以具有不同初值的响应系统为接收器,采用混沌遮掩的方式实施信号加载,实现在固定时间内恢复加密信号,确保保密通信的安全性和时效性.
地下水在防护工程消波层中几乎是不可避免的,采取技术措施降低地下水对消波层效能的不利影响对防护工程十分重要.为量化新型防护材料的实际效能,通过分离式霍普金森杆试验研究了在干燥与饱水两种工况下薄壁空心球粒集料对高频高幅应力波的衰减能力,并与标准砂进行对比.试验结果表明:在饱水工况下,被测材料的消波能力均有降低;但是在高幅值冲击荷载作用下,薄壁空心球粒集料在两种工况下均较标准砂表现更强的变形能力,尤其是在饱水工况下,新型薄壁空心球粒集料填充节理的应力波透射率仅为相同条件下标准砂填充节理的 18.99%～37.6%.使用薄壁空心球粒集料替代传统干散砂作为消波材料可以大大提高地下防护结构消波层在饱水不利工况下的消波效能.
介电弹性体驱动器在能源、信息等高新技术领域应用广泛,但其在使用过程中不可避免地会受到机械应力,引起结构破坏,造成使用寿命缩短,而引入自修复性质将极大延长驱动器使用寿命.同时,考虑到现有单一电场驱动已无法满足现实需求,开发多功能响应的自修复介电弹性体驱动器是一种有效应对方法.基于此,利用聚二甲基硅氧烷(PDMS)与聚苯胺(PANI)、四氧化三铁(Fe3 O4)之间的协同效应,成功制备了一种基于电磁响应的PDMS-PANI-Fe3 O4 介电弹性体,分别研究了其聚合物稳定性、自修复性、力学性能以及驱动性能.结果表明:在 200℃以下,PDMS-PANI-Fe3 O4 膜失重质量分数为 0.8%,说明其具有良好的热稳定性;室温自修复 48 h后,PDMS-PANI-Fe3 O4 聚合物修复完整,表明其使用寿命延长;将 PDMS-PA-NI-Fe3 O4 弹性体加工成燕子翅膀,在磁场作用下,翅膀不断被驱动.在此电场条件下,PDMS-PANI-Fe3 O4膜最大驱动形变为 6.04%,进一步弥补单一驱动的弊端.
测控装置测量的数据是电网状态估计、潮流分析、自动电压控制等应用的数据来源,保证数据正确至关重要.针对某电厂测控装置量测数据异常波动事件展开研究,指出电网中存在 35、37 次高次谐波,在特定情况下混叠到基波附近,干扰测控装置的测量功能,造成数据异常波动.分析了混叠干扰量测功能的机理,并通过仿真及宽频测量数据验证了结论的正确性;针对改进措施进行了探讨,从采样频率选择、抗混叠滤波器设计、测量装置检测规范及谐波监测手段等方面给出了解决措施,加强了对高次谐波的监测能力,为分析和抑制此类振荡提供参考.
通过构筑石墨烯/二硫化铼(ReS2)异质结实现了 5 层ReS2 样品在空气中稳定性的提升.其中石墨烯是通过化学气相沉积法(CVD)制备,ReS2 的热稳定性通过荧光强度进行表征.实验发现:对于原始的ReS2 薄膜,当温度升高到 500℃时,荧光峰严重展宽且荧光强度几乎消失,而石墨烯封装的ReS2 样品,采用同样升温方式时,荧光峰依然很对称且荧光强度很高.通过进一步拟合其变化率发现,本征样品的荧光强度随温度的变化率为-224,而石墨烯封装的ReS2 薄膜样品的荧光强度随温度的变化率为-173,说明通过石墨烯的保护,极大地提高了ReS2 在空气中的热稳定性.这主要是因为石墨烯有效地隔绝了空气中的水、氧气等分子,钝化了ReS2 表面的反应活性,从而提高了其化学稳定性.同时测试对比了石墨烯覆盖前后 ReS2的光电性能,发现石墨烯的覆盖大大加快了ReS2 的光电响应速度.该研究丰富了 ReS2 在半导体器件方面的应用.
动力扰动作用下,处于高静载状态的岩石极易沿已有裂缝产生滑移,造成滑移型岩爆等动力灾害.通过能量分析和变分原理对动力扰动诱发高应力矿柱断裂滑移失稳(滑移型岩爆)的力学机理进行分析,得出了动力扰动条件下矿柱失稳的充要条件.通过开展动静组合荷载作用下的矿柱滑移失稳试验,验证了理论分析的正确性.研究表明:矿柱的动力滑移失稳只会发生在峰后阶段,对于处在峰前稳定平衡状态的矿柱,若要发生动力滑移失稳,需要爆破振动、巷道断裂等人工或者自然活动所产生的动力扰动达到一定量值,同时还需满足加载刚度k1 小于峰后的滑移刚度k2.因此,对矿柱断裂滑移型岩爆量级的预测则须准确地确定过剪应力的量值.
为研究节理在剪切过程中的方向性和剪切强度规律,利用直剪仪对复制岩石节理试样进行了爬坡和顺坡两个方向的直剪试验.试验结果表明:剪切方向对岩石节理的剪切强度有明显影响,沿爬坡方向的剪切强度大于沿顺坡方向,但粗糙度系数(joint roughness coefficient,JRC)的减小和法向应力的增大可在一定程度上削弱剪切方向的影响.节理剪切破坏过程具有显著的阶段性特征,中低法向应力条件下以爬坡滑移为主要机制,而高法向应力下以相互啃食和凸起啃断为主要机制.采用平均有效倾角差指标参数描述剪切强度的各向异性,并在此基础上提出了考虑剪切方向的改进剪切强度模型,更好地预测节理正、反方向的剪切强度.通过百分位数参数建立了三维形貌指标与端部轮廓线二维高差指标的联系,大大简化了节理形貌指标的获取难度.该模型参数较少容易获取,物理意义明确,且把岩石节理剪切强度的各向异性考虑在内,具有广阔的潜在工程应用前景.
为了合理分配无人机对多个任务区的侦察时间,提出了一种包含问题建模、求解和方案决策的无人机多任务区侦察时间分配方法.首先,建立了包含侦察收益和侦察风险两目标的无人机多任务区侦察时间分配模型,该模型属于带约束多目标优化问题;其次,提出了一种改进的基于分解的约束多目标进化算法,该算法具有简单、灵活、无参等特点,可有效求解;最后,利用优劣解距离法从非支配解集中选择最优方案.选择了 6 种约束多目标进化算法,在 3 个不同雷达强度指数条件下进行对比实验.Hypervolume 指标说明约束多目标进化算法在求解该问题时优于其他算法.实验结果表明:提出的方法在求解无人机多任务区侦察时间分配优化问题中能够实现快速准确决策.
To address the problems of complicated calculations and weak discriminating ability for some existing hesitant fuzzy entropy measures, the hybrid entropy and cross-entropy measures for hesitant fuzzy elements（HFEs） are designed in the form of convex combinations on the basis of comprehensively evaluating the uncertainty of HFEs themselves and the information they express. The hybrid cross-entropy measures for HFEs are used to measure the differentiated degree between two HFEs and can be unified together with the hybrid entropy measures as the uncertainty measures for HFEs.Compared with the existing uncertainty measures for HFEs, the hybrid entropy measures show the advantages of simple structure, clear physical meaning, and strong discriminating ability, with the hybrid cross-entropy measure being naturally symmetrical. Then, based on the proposed entropy and cross-entropy measures for HFEs, a multi-attribute decision-making（MADM） model, under the condition of unknown attribute weights, is established and further applied to the evaluation of UAVs’ combat effectiveness.
It is of practical significance to study the parameter estimation of the frequency hopping（FH） signals in the case of random missing observations. In view of the failure of the linear time frequency analysis in the case of missing observations, a frequency hopping signal parameter estimation method is proposed in this paper, based on orthogonal matching pursuit（OMP） and Kalman filter（KF）. By this method, a sliding window is added to the signals, and the random missing observations in the window is modeled as a signal sparse representation. As an overcomplete dictionary, the Fourier orthonormal bases makes use of OMP to estimate the frequency of the signal in the window without restoring the signals. KF performs smoothly on the estimated signal frequency. When the frequency changes, the frequency prediction value of KF will seriously deviate from the historical value and the frequency estimated value, and the degree of deviation will offer the support for hopping time estimation. The simulation results and comparative experiments show that the parameter estimation of the proposed method outperforms that of other methods, when the observations are not missing. With the sufficient sliding window length and the eligible signal sparsity requirement, effective FH signal parameter estimation results can also be obtained even in the case of missing observations.
Aiming at the modeling problems of the source-load coupling and pulse characteristics of the battery energy storage system with pulsed load（BESS-PL）, the source-load coupling precise model and average model of BESS-PL were respectively constructed by the state space method and the signal average method. Based on these two models, a quasi-precise modeling method was proposed, which can reasonably simplify the high-frequency components and retain the unique pulse characteristics of BESS-PL. Through stability theory analysis and simulation, the characteristics and applications of three BESS-PL source-load coupling mathematical models were compared. The steady-state transfer function of BESS-PL system was deduced based on the averaged model, and the influence laws of pulse duty ratio and DC/DC converter switch duty ratio stability were discussed on the analysis of the Bode diagram. The correlation laws of key parameters and source-load power, pulsation characteristics and the influence laws of duty ratio stability of BESS-PL system were analyzed combined with simulations of the three models. The three BESS-PL source-load coupling mathematical modeling methods can be flexibly used for transient steady state analysis of source-load coupling characteristics and pulse characteristics of BESS-PL system.
FFT算法作为 OFDM系统的核心算子占用其系统处理的大多数时间,为提高 OFDM系统数据传输速度,提出了一种改进的多路并行流水线型基 22FFT实现架构.在实现过程中着重对旋转因子的存储进行片上缓存优化,减少了乘法运算次数从而减小整体运算复杂度;设计的数据整合模块用于控制时序,从而保证P 路并行流水型架构正确实现,数据运算吞吐率成P 倍提高.RTL 仿真结果表明,与同类架构相比,提出的架构在硬件开销适中的同时使得性能分别提升了 1.27％、2.04％、50.88％,并且具有 FFT 点数可扩展的特点,可满足随着通信标准的不断提高,FFT点数逐渐增大的实际应用需求.
The various types and huge amount of data that characterize mechanical failures, as well as the uneven quality of the data and the low value density of failure information, pose a big challenge to mechanical failure diagnosis. However, the shallow neural network model can not meet the requirements of accurate fault diagnosis due to its week self-learning ability. Therefore, the research group conducted in-depth research on fault feature extraction and fault diagnosis models. With time series permutation transform and continuous wavelet transform, two 2D-CNN fault feature dada sets were constructed based on images of time series permutation and time-frequency images respectively. Combined with the advantages of convolutional neural network（CNN） and gradient boosting decision tree（GBDT） algorithm, two fault diagnosis algorithms of CNN-Z and CNN-F based on GBDT were proposed. Through experiments, the two algorithms have achieved higher fault diagnosis accuracy, as high as 94.33% and 98.62% respectively. Compared with the traditional CNN algorithm, the experimental results show that the two fault diagnosis algorithms proposed in the paper have higher accuracy. In addition, the error convergence time of the fault diagnosis algorithms of CNN-F based on GBDT is only one third of the former, which verifies the effectiveness and accuracy of the algorithm proposed in this paper.
Affected by the plateau environment, the combustion deterioration of diesel engine leads to power reduction, increased fuel consumption and more harmful emissions. The use of exhaust gas turbocharging can solve the problems to a certain extent, but there are also shortcomings such as the lag of turbo response, overspeed, surge, and the reduction of boost efficiency with the increase of altitude. To improve its adaptability to the plateau environment, the method of adding one more stage of electric booster was studied. Using GT-SUITE software, the numerical simulation models of a certain type of exhaust gas turbocharged diesel engine for power generation and an electric combined supercharging were established. The electric combined supercharging on the high-altitude performance of diesel engine for generation was analyzed. The results show that the compound electric supercharging can not only restore the output power and torque of the diesel engine effectively, but also reduce the fuel consumption rate. At an altitude of 5 000 m, the output power and torque are increased by 5.7% and 9.2% respectively, and the fuel consumption rate has been reduced by 8.1% at the altitude of 4 500 m. At the same time, it also has a certain effect on reducing the emissions of CO and HC compounds in the exhaust gas.
On account of the advantages of fast assembly and being less affected by the seabed terrain environment, the floating platform-floating bridge system plays a significant role in the large scale transportation of personnel, materials and equipment between land and sea. In order to explore the motion response characteristics of floating platform-floating bridge system in waves, the numerical model of a floating platform-floating bridge system was established based on the potential flow theory, and the motion responses of the system were calculated. The values of the numerical model were verified by the model test in a wave tank. The results show that the dynamic response characteristics of the multi-floating body system in irregular waves can be relatively accurately reflected by the numerical model. Based on this, further research was conducted on the influence of the platform on the motion response of the floating bridge under different wave spectrum peak frequencies, wave direction angles and the position of the bridge relative to the platform. It is found that the shielding effect of the platform on bridges motion under high-frequency waves is better than that under low-frequency waves, and the motion responses can be reduced by 40%—50%. The shielding effect of the platform on the bridge motions in head waves is better than that in the oblique waves. When the incident direction of the wave is relatively stable and the wave angle is less than 20°,t heshielding effect of the platform on the bridge on the far side of the waves is more significant than thatnear the waves and in the middle position relative to the platform.
