某油轮锅炉燃烧系统故障分析与排除

文章针对某油船锅炉出现的燃烧系统故障现象, 从燃油系统、供风系统、控制系统等三个方面进行了详细 分析并最终排除了故障, 提出了锅炉的管理建议, 以期为同类事故的预防提供借鉴。     

水下热动力系统半实物仿真研究

水下热动力半实物仿真试验台可提供仅次于鱼雷热动力系统热车功率试验置信度的试验手段。以实时仿真计算机、背压模拟装置、负载模拟装置、调速与驱动装置构成半实物仿真试验系统的核心,以比例和比例积分控制算法进行背压和负载压力的闭环控制,通过低通数字滤波改善数字仿真模型的稳定性,对数字仿真模型进行动态特性修正以实现半实物系统与实际系统特性的良好吻合,系统背压动态模拟精度达到了0.5%,转速和负载动态模拟精度达到了1.5%,整个半实物仿真系统获得了较高的模拟准确度。     

一种水下航行器轨迹发生器的设计与仿真

为了减少水下航行器（AUV）导航系统算法开发过程中跑车试验及半实物仿真等所耗费的资金和时间,提出一种水下航行器轨迹发生器的设计方法。利用所给出的轨迹微分方程,反推得出生成轨迹参数公式和惯性信息公式,同时给出水下航行器在直航、转弯和爬潜等常见的运动状态下轨迹发生器的具体输入,并将生成的惯性量测值（加速度和陀螺值）输入捷联式惯性导航算法。验证结果表明,利用所设计的轨迹发生器可以为研究导航算法提供模拟的数据,具有工程实际意义。     

基于角增量提取改进算法的旋转矢量法

旋转矢量法应用于角速率陀螺时,将角速率转化为角增量产生了较大误差,为使旋转矢量法有效地应用于角速率陀螺的姿态解算,该文提供了一种改进的三子样旋转矢量法。该方法在经典三子样角增量提取算法上,增加了前后各1／3周期的节点,使得采样值更加有效利用。在不增加区问子样数的前提下,使每个有效节点至少被使用5次,大大降低了角速率向角增量转化的误差。仿真结果表明,该算法可以有效应用于角速率陀螺。     

利用星载激光雷达研究云层垂直结构及其分布

云的垂直结构（CVS）是云的重要特征量,在大气模式的研究中有着有着十分重要的影响。利用CALIPSO（Cloud Aerosol Lidar and Infrared Pathfinder Satellite Observations）星载激光雷达2007年到2010年间的Level 2 Version 3的云数据对中国海及其周边海域的云的垂直结构及其分布随年度和区域的变化进行了统计分析。结果表明：层数不同的云在研究区域内发生的概率（COF）差别较大,云顶海拔高度（LTA）不同的云沿纬度方向的分布差异明显,LTA在10 km到20 km的云出现的概率较高而且大多出现于靠近赤道区域。由于有些云CALIPSO信号不能穿透,可能使云层数以及云顶高度的统计结果偏小。     

基于自然选择策略的蚁群算法求解TSP问题

针对蚁群算法收敛速度慢,容易陷入局部最优解的缺陷,提出了一种基于自然选择策略的改进型蚁群算法,改进后的算法利用自然选择中“优胜劣汰”的进化策略,对每次迭代的随机进化因子大于进化漂变阈值的路径信息素进行二次更新,增强满足进化策略路径上的信息素浓度,以加快算法的收敛速度;而随机进化因子的随机性增强了算法跳出局部最优解的概率。将提出的改进型蚁群算法求解经典的TSP问题,并通过实验证明了改进后的蚁群算法在最优解精度和收敛速度等方面均有所提高。     

Balancing Intermolecular Interactions between Acceptors and Donor/Acceptor for Efficient Organic Photovoltaics

Promoted by uninterrupted materials and device innovation, organic solar cells have achieved impressive development. However, the complicated intermolecular interactions inside active layers are less understood. Herein, the intermolecular interactions are studied from the dual perspectives of acceptor/acceptor (A/A) and donor/acceptor (D/A), and how these interactions synergistically control the final efficiencies. Three small molecular acceptors (SMAs) are designed with different end-caps, which manipulate the crystallinity and electrostatic potential (ESP) distributions of acceptors, and accordingly, the A/A and D/A intermolecular interactions. The results show that SMA LA17 with low A/A interactions exhibits inferior performance around 12%, owing to its strong D/A interaction with donor PM6, which shapes too miscible morphology and increases charge recombination. Instead, LA16 with strong A/A interactions and moderate D/A interactions delivers improved bulk-heterojunction (BHJ) networks, and therefore, enhances charge transport and diminishes geminate or trap-assisted charge recombination. Consequently, PM6:LA16 records the competitive efficiency of up to 13.74% among the three systems. Therefore, this study deepens the synergistic or balancing effect of the D/A and A/A interactions on BHJ blends for efficient organic solar cells.     

吊放声呐信号处理机测试设备系统设计

设计了一套实用的吊放声呐信号处理机测试设备。针对吊放声呐信号处理机实际中调试、检测、维修的需求。利用高性能数字信号处理芯片作为控制中心,结合一些必要的外设,充分考虑实际使用的特点,设计了一套稳定性高,集成度高的测试设备。从系统结构、硬件结构、存储数据结构、软件结构等几个方面入手,阐述了设计思路和实现方法。     

基于Direct3D的雷达视频图像实时显示技术

介绍了基于Direct3D的雷达图像PPI显示技术,讨论了利用多核多线程技术和多媒体定时器计时控制技术实现雷达视频图像实时显示,解决Windows GDI显示时屏闪和无法满足实时性要求等问题的方法,给出了某岸基警戒搜索雷达软件化终端信息实时显示处理系统的设计与实现方案。     

Multi-task learning with deformable convolution

Multi-task learning aims to tackle various tasks with branched feature sharing architectures. Considering its diversity and complexity, discriminative feature representations need to be extracted for each individual task. Fixed geometric structures as a limitation of convolutional neural networks (CNNs) in building models, is also exists and poses a severe challenge in multi-task learning since the geometric variations will augment when we deal with multiple tasks. In this paper, we go beyond these limitations and propose a novel multi-task network by introducing the deformable convolution. Our design, the Deformable Multi-Task Network (DMTN), starts with a single shared network for constructing a shared feature pool. Then, we present task-specific deformable modules to extract discriminative features to be tailored for each task from the shared feature pool. The task-specific deformable modules utilize two new parts, deformable part and alignment part, to extract more discriminative task-specific features while greatly enhancing the transformation modeling capability. Experiments conducted on various multi-task learning types demonstrate the effectiveness of the proposed method. On multiple classification tasks, semantic segmentation and depth estimation tasks, our DMTN exceeds state-of-the-art approaches against strong baselines.