抗辐射保偏光纤的制备及其特性研究

通过对保偏光纤进行光漂白与低温相结合处理的方式，来提高光纤的抗辐射性能。保偏光纤再经过1.0 kGy的高剂量预辐照处理后，其损耗会明显增大。随后采用980nm的激光对光纤进行光漂白处理,其损耗并没有发生明显变化。但是，处理后的样品在小剂量辐照条件下其光纤传输呈现先减小后增大的趋势。对于通过液氮进行冷处理的预辐照样品，其光纤传输损耗得到明显的降低，同时光纤经过小剂量辐照后的损耗同样也具有先减小后增大的现象。最后，将两种处理方式相结合，一方面可以有效地降低预辐照光纤的传输损耗，另一方面也使得保偏光纤在小剂量辐照的条件下，使其传输损耗具有减小的趋势，大大提高了保偏光纤在小剂量辐照条件下的抗辐照特性。     

磁选态单束铯束管束光学参数的模拟计算和设计

为提高铯束管跃迁信号强度,在束光学参数设计中,用蒙特卡罗方法模拟计算铯原子通过率和未跃迁比例。通过改变模拟计算中的束光学参数,分析铯原子通过率、未跃迁比例和倍增器首极电流的变化趋势,得到了主要束光学参数的合理取值范围。在铯束管实际装配和试验中使用束光学参数设计结果,有效地增加了倍增器首极电流。     

Fault Estimation and Accommodation for a Class of Nonlinear System Based on Neural Network Observer

The problem of fault estimation and accommodation of nonlinear systems with disturbances is studied using adaptive observer and neural network techniques.A robust adaptive learning algorithm based on switchingβsmodification is developed to realize the accurate and fast estimation of unknown actuator faults or component faults.Then a fault tolerant controller is designed to restore system performance.Dynamic error convergence and system stability can be guaranteed by Lyapunov stability theory.Finally,simulation results of quadrotor helicopter attitude systems are presented to illustrate the efficiency of the proposed techniques.     

黄渤海晴空像元下雾霾识别的机器学习算法

为实时有效地识别和监测雾霾，选取2013年10月4日与10月5日的两景MODIS遥感图像，利用不同类型像元的光谱差异，基于神经网络模型，建立海上雾霾的遥感识别机器学习算法。经检验表明，在晴空条件下，机器学习算法反演精度相对较高，已超过98%，利用算法可以实现对中国近海上空雾霾的监测。     

A surface parametric control and global optimization method for axial flow compressor blades

An aerodynamic optimization method for axial flow compressor blades available for engineering is developed in this paper. Bezier surface is adopted as parameterization method to control the suction surface of the blades, which brings the following advantages:(A) significantly reducing design variables;(B) easy to ensure the mechanical strength of rotating blades;(C) better physical understanding;(D) easy to achieve smooth surface. The Improved Artificial Bee Colony(IABC) algorithm, which significantly increases the convergence speed and global optimization ability, is adopted to find the optimal result. A new engineering optimization tool is constructed by combining the surface parametric control method, the IABC algorithm, with a verified Computational Fluid Dynamics(CFD) simulation method, and it has been successfully applied in the aerodynamic optimization for a single-row transonic rotor(Rotor 37) and a single-stage transonic axialflow compressor(Stage 35). With the constraint that the relative change in the flow rate is less than0.5% and the total pressure ratio does not decrease, within the acceptable time in engineering, the adiabatic efficiency of Rotor 37 at design point increases by 1.02%, while its surge margin 0.84%,and the adiabatic efficiency of Stage 35 0.54%, while its surge margin 1.11% after optimization, to verify the effectiveness and potential in engineering of this new tool for optimization of axial compressor blade.     

任务型教学法的任务设计

任务型教学法是新兴的一种强调“在做中学”(learning by doing)的语言教学方法，是我国外语课程教学改革的一个走向，吸收了以往多种教学法的优点而形成。任务是任务型教学法中一个重要概念。要使任务型教学法更好地发挥作用，任务的设计就显得尤为重要。在任务设计中，要注意任务与人本主义、任务与文化性、任务与教师角色之...     

Particle swarm optimization applied to hypersonic reentry trajectories

This paper presents the novel use of the particle swarm optimization(PSO)to generate the end-to-end trajectory for hypersonic reentry vehicles in a quite simple formulation.The velocitydependent bank angle profile is developed to reduce the search space of unknown parameters based on the constrained PSO algorithm.The path constraints are enforced by setting the fitness function to be infinite on condition that the particles violate the maximum allowable values.The PSO algorithm also provides a much easier means to satisfy the terminal conditions by adding penalty terms to the fitness function.Furthermore,the approximate reentry landing footprint is fast constructed by incorporating an interpolation model into the standardized bank angle profiles.Numerical simulations demonstrate that the PSO method is a feasible and flexible tool to generate the end-to-end trajectory and landing footprint for hypersonic reentry vehicles.     

战斗机嵌入式训练系统中的智能虚拟陪练

智能化"实虚"对抗是现代先进战斗机嵌入式训练系统的重要功能需求。自主空战决策控制技术在未来空战装备发展中扮演关键角色。将当前的功能需求和发展中的技术结合起来,得到了空战智能虚拟陪练的概念。先进控制决策技术的引入使得智能虚拟陪练能够帮助飞行员完成复杂的战术训练,而训练中真实的对抗场景为技术的验证提供了理想的环境,大量的训练数据为技术的持续迭代优化提供了保障。作为可学习和进化的空战战术专家,智能陪练在人机对抗和自我对抗中不断优化,当其具备与人相当甚至超越人的战术能力时,可应用于未来的无人空战系统。智能虚拟陪练需要具备4项基本能力：智能决策能力、知识学习能力、对抗自优化能力和参数化表示能力。对其包含的关键技术进行了分析,提出并实现了一个基于模糊推理、神经网络和强化学习的解决方案,展示了其各项基本能力及目前达到的空战水平。未来更多的模型和算法可在智能虚拟陪练的框架中进行验证和优化。     

空心气冷低压涡轮动叶气热耦合数值模拟

以空心气冷低压涡轮动叶为研究对象,采用高质量的流体域和固体域网格控制技术,带转捩模型的双方程SST湍流模型,开展了基于CFD方法的叶片气热耦合问题研究。获得了不同冷气流量比(分别为1.0%,1.38%,1.8%和2.2%)、温比(分别为2.1,2.25,2.3,2.4和2.5)和压比(分别为1.4,1.6和1.8)对叶片换热特性的影响规律,设计状态中截面按弧长平均的叶片壁面金属温度计算值较试验值偏小0.3%,气热耦合计算的叶片壁面温度分布与试验结果吻合良好,验证了气热强耦合计算方法的精度,为涡轮叶片温度场分析提供了一种有效的方法。