内压缩通道几何参数对高超声速进气道性能的影响

用N-S方程模拟了一系列不同收缩比、不同波系配置的内压缩通道内流动,研究了内压收缩通道几何参数对进气道性能的影响,发现对于相同的外压段,内压面积收缩比对进气道内压缩通道温升比、压比和起动性能具有较好的相似规律,且随着内压面积收缩比增加,进气道温升比、压比增加,出口流场畸变下降,起动马赫数增大。通过对相同压比下不同内外压缩比的进气道性能的研究,得到了内外压缩比对进气道效率和起动性能的影响规律,发现压缩程度相同时,进气道效率和起动马赫数均随内外压缩比有先增大后减小再增大的规律。     

基于VPRS理论的航空发动机送修等级决策

应用变精度粗糙集理论研究了发动机状态信息与单元体性能参数之间的关系，提出了一种基于信息熵属性约简的航空公司发动机维修等级决策方法，从而在维修决策时能够更加客观地反映发动机实际损伤程度。最后以CF6型发动机维修等级决策规则生成为例说明了该决策方法的有效性。     

模糊神经元及其计算能力研究

分析了模糊神经元模型,指出了这种神经元模型在计算能力上的缺陷.提出了对模糊神经元定义的两种改进方法,一是限制传递函数为非单调函数;二是修改定义域.对改进后的模糊神经元网络证明了计算能力与图灵机的等价性,从而对传统的神经网络作了进一步推广.还对混合模糊神经网络的特点做了简单的讨论.      

Models of P/Borrelly: Activity and dust mantle formation

Comet 19P/Borrelly was observed by Deep Space One spacecraft on September 22, 2001 (Soderblom et al., 2002).The DS1 images show a very dark and elongate nucleus with a complex topography; the IR spectra show a strong red-ward slope consistent with a very hot and dry surface (345K to 300K). During DS1 encounter the comet coma was dominated by a prominent jet but most of the comet was inactive, confirming the Earth-based observations that <10% of the surface is actively sublimating. We have developed a thermal evolution model of comet PBorrelly, using a numerical code that is able to solve the heat conduction and gas diffusion equations at the same time across an idealized spherical nucleus ( De Sanctis et al., 1999, 2000; Capria et al., 2000; Coradini et al., 1997a,b). The comet nucleus is composed by water, volatiles ices and dust in different proportions. The refractory component is made by grains that are embedded in the icy matrix. The code is able to account for the dust release, contributing to the dust flux, and the formation of dust mantles on the comet surface. The model was applied to a cometary nucleus with the estimated physical and dynamical characteristics of P/Borrelly in order to infer the status and activity level of a body on such an orbit during the DS1 observation. The comet gas flux, differentiation and thermal behavior were simulated and reproduced. The model results are in good agreement with the DS1 flyby results and the ground based observations, in terms of activity, dust coverage and temperatures of the surface.     

基于试验气动力的纵向机动飞行载荷分析

提出了一种基于非线性风洞试验数据(简称试验气动力)同步进行纵向机动飞行过程和飞行载荷的计算分析方法.通过曲面样条插值将试验气动力的升力系数和俯仰力矩系数引入纵向机动飞行过程的计算,同时基于试验气动力的压力分布进行全机飞行载荷的静气动弹性修正.弹性载荷的计算采用线性气动力影响系数矩阵处理.采用偶极子格网法提供气动力(简称理论气动力).将基于这两种气动力的计算结果进行了对比.结果表明:基于试验气动力的机动飞行过程更为合理;基于理论气动力的载荷随机动飞行过程的变化趋势有可能与基于试验气动力的载荷变化趋势不一致.      

前馈神经网络的一种简单共轭梯度学习算法

针对前馈神经网络学习误差函数维数高、计算复杂度大的特点,对梯度下降BP算法加以改进从而构造出一种简单共轭梯度下降算法(MPARTAN算法).该算法计算复杂度不高于动量BP算法, 与FR共轭梯度法相比,该算法的稳定性好,又具有共轭梯度法的优点,收敛速度快.文中给出了该算法的收敛定理,并用2个实验例子比较了动量BP算法、FR共轭梯度法和MPARTAN算法的计算结果.     

γ辐照对铷频标物理部分参数影响的研究

通过γ辐照试验,分析了辐照对铷频标物理部分各项参数的影响。辐照后,振荡回路的电流变化对频标准确度的影响可忽略不计;光电二极管受辐照后会使短期稳定度变差,辐照剂量率小于1 400 rad（Si）/h,不会对频标的短稳性能产生显著影响;控温电路受辐照后会引起控温点移动,从而改变频率准确度。     

基于GNSS的高轨卫星定位技术研究

利用全球卫星导航系统(GNSS)进行导航定位具有全球、全天候、实时和高精度的优点,应用于高地球轨道(HEO)卫星的定位,能够提供精确的轨道和姿态确定,并且可以克服目前主要利用地面测控系统对HEO卫星进行定位的设备复杂、投资高等缺点,使得自主导航成为可能.本文对利用GNSS的高轨卫星定位相关技术进行了研究,分析了单一GNSS系统和多个GNSS组合系统的卫星可见性、动态性和几何精度因子(GDOP).通过仿真分析表明,利用组合GNSS系统并通过提高GNSS接收机灵敏度的方法,可以解决GNSS进行HEO卫星定位的相关问题,并能保证HEO卫星定位精度的要求.      

Countermeasure Against Regenerative and Forced Chatter of Flexible Workpieces in Milling Processs Using Bi-directional Excitation

A novel extended methodology for chatter suppression in milling process by applying external forced vibrations to the workpiece in two orthogonal directions which are the feed and cross-feed directions.Both the regenerative and forced chatter suppression during the milling process of flexible workpieces are investigated.Here,the workpiece is subject to a sinusoidal periodic force in the feed direction to disrupt the regenerative effect.Additionally,to minimize the forced chatter,the workpiece is subject to the periodic excitation force in cross-feed direction.This force is proportional to the magnitude of the estimated cutting force in cross-feed direction and has a phase opposite to the cutting force to minimize the vibration amplitudes.The effectiveness of the proposed method is evaluated numerically and experimentally,for the spindle speed located in both the local minima and local maxima of the stability lobe diagram.The numerical simulations indicate significant suppression effect in terms of vibration amplitudes,resulting in suppression of both the regenerative chatter and the forced chatter.Experiments were conducted by using a workpiece-mounted active stage composed of flexure hinges and driven by piezoelectric actuators.The experimental results agree qualitatively with the numerical simulations.The proposed method indicates a remarkable vibration reduction effect for both regenerative and forced chatters.     

基于NB-IoT技术和GA-BP神经网络的车位预测系统

围绕有效整合城市停车资源,提高现有车位存量利用率的需求,构建了一种基于NB-IoT技术的车位预测系统。该系统采用NB-IoT技术进行信息采集与传输实现车位信息的共享;考虑到车位状态信息实时变化的特性,用历史车位占用数据来建立车位预测模型,推测出未来短时内车位变化趋势。为了提高车位预测的精度,采用遗传算法(Genetic algorithm,GA)优化反向传播(Back propagtion,BP)神经网络建立GA-BP神经网络车位预测模型。以某地下停车场历史数据为例进行仿真实验,研究结果表明:车位预测模型预测值与实际值相近且趋势保持一致,能够有效准确的预测车位状态变化,具有较高的精度。