线搜索(LS)共轭梯度法非单调算法的全局收敛性

非单调线搜索技巧在非线性优化中得到成功的应用与扩展,非单调线搜索下的共轭梯度法则可以提高大规模非线性优化问题的收敛速度。对LS共轭梯度法做了某些变型,在非单调线搜索下,该方法保证每次迭带都会产生下降的方向,在较弱的条件下得到算法全局收敛性。     

碳锰钢热变形行为及动态再结晶模型

 采用单道次热模拟实验分析了各变形参数(初始晶粒尺寸、变形速率、变形温度和变形量)对碳锰钢动态再结晶的影响。结果表明,当初始晶粒越细小、变形温度越高、变形速率越小时,越易发生动态再结晶,同时在能够发生动态再结晶的条件下,变形量越大,动态再结晶越充分。得到了发生动态再结晶时的形变激活能Qdef、临界变形量模型、动态再结晶百分数模型及动态再结晶晶粒尺寸模型。计算得出ε05模型的平均误差为395%,模型预测值与实验数据计算值符合较好。     

合金化和球化工艺对超高碳钢组织和性能的影响

 研究了合金化和球化工艺对超高碳钢组织和性能的影响。用扫描电子显微镜、透射电子显微镜和能谱仪观察了钢的组织形貌和元素分布。结果表明：在碳和铬含量相同的超高碳钢中加入同量的合金元素铝和硅时,铝可明显抑制锻造组织中网状或粗大的颗粒状碳化物的析出、细化珠光体组织和控制石墨形成。UHCS 213Si和UHCS 261Al钢经850 ℃×3 h球化退火处理后,都能得到较好的球化组织,其力学性能分别为：UHCS 213Si钢,Rm=1 033 MPa,Re=734 MPa,A=149%;UHCS 261Al钢,Rm=973 MPa,Re=677 MPa,A=182%。     

稳恒磁场下铁-镍合金箔的制备及磁性能

 研究了施加稳恒磁场(H=0~07 T)对电沉积铁 镍合金箔的影响。测定了垂直电流方向,不同强度磁场下的电流效率、合金组成及合金箔的磁性能,并用扫描电镜和X射线衍射观察合金箔的微观形貌和结构,分析讨论了稳恒磁场对这些性能的影响规律。结果表明：与不施加磁场相比,施加垂直磁场后合金箔表面更均匀、细致、平整;合金箔铁含量提高约14%;合金箔(111)晶面取向程度增强,磁性能有所改善;但铁 镍合金电镀的电流效率呈先增加后下降的趋势。     

拉伸载荷作用下AlN夹杂物对航空用超高强度钢中裂纹萌生的影响

 采用扫描电镜原位观测法,跟踪观察了航空用超高强度钢中长方型夹杂物导致裂纹萌生与扩展的微观行为。结果表明,拉伸载荷作用下,航空用超高强度钢中裂纹萌生的方式与夹杂物尺寸及夹杂物周围孔洞的大小有关。当夹杂物面积小于一定值时,无论夹杂物周围有无孔洞,裂纹均以夹杂物/基体界面开裂的方式萌生;当夹杂物面积大于一定值后,若夹杂物/基体界面基本完好,则裂纹易以夹杂物自身开裂的方式萌生;若夹杂物周围孔洞面积较大,则裂纹易以夹杂物/基体界面开裂的方式萌生。     

基于RBF神经网络的中厚板轧机的温降模型

 基于Matlab神经网络工具箱,采用改进的径向基函数(RBF)网络优化计算4200中厚板轧机的轧制温度。通过径向基层散布常数的人工调整以及神经元的自适应调整,提高了收敛速度,确定了最佳的网络结构形式。网络预测结果与在4200中厚板轧机上实测的轧件温度进行了对比,预测精度很高,表明了该网络模型的优越性。该模型可为以温度参数为主要控制对象实行自动化生产的4200轧机提供可靠的参数,也可为人工神经网络在其它自动控制方面的应用提供参考。     

OFDM系统中信道估计方法的研究

LS算法是OFDM系统信道估计方法中计算复杂度最低性能最差的方法.根据有限长度h的能量分布,提出了一种改进算法.通过仿真,与LS算法和性能较好的LMMSE比较,验证了改进后的LS算法性能优于最初的LS算法,并可在实际中得到较好的应用.     

Numerical simulation of impact behaviour of structures with response adapted to the intensity of shock

This study presents an alternative approach for the absorption of impact energy that uses the internal pressurization of structures in the framework of a crash‐adaptive response. Numerical simulations were conducted on axial impact of thin‐walled tubular structures with circular cross section that serve as an approximation to a front crash box of a motor vehicle. The main objective of this work consists in studying the effect of internal pressurization of tubular structures in a crashworthiness application, as well as the possibility to obtain a reduction in wall thickness thus improving weight efficiency. A numerical study is presented for an internal pressure of 20 bar and tubular structures of circular section and 1.14 mm thickness. Numerical simulations were performed recurring to LS‐DYNA explicit dynamics software while considering for the material a stainless steel alloy that is a material with interest for crashworthiness applications and manufacturing requisites due to its balance between strength, ductility and energy absorption. The results obtained allow concluding that recurring to internal pressurization it is feasible to reduce the wall thickness and have an impact resistance identical to the original while improving overall efficiency.     

一种改进的基于LS的OFDM盲同步算法

在基于LS(lease squares)的OFDM盲同步算法基础上,提出了一种性能优良的OFDM盲同步算法。采样周期取为原算法采样周期的M倍,计算出抽样点处的判决函数值,然后对相邻的判决函数值进行前后移动平均,定时点取为使移动平均值为最小时对应的那2个抽样点的中间位置。仿真结果表明:该算法不仅减少了计算量,而且在复杂的多径衰落信道中,定时和频偏估计性能也有大的改善。     

COMPUTER SIMULATION TECHNIQUES FOR A MOBILE SATELLITE COMMUNICATION CHANNEL

This paper gives a brief and systematic presentation of the computer simulation techniques of satellite communications, especially for a mobile satellite channel, based on a summarization of the author's exploration and experience in this area. It includes the equivalent complex baseband notation (ECBN) method, a very convenient mathematical form for software simulation; development of a multi-functional and expandable channel simulation system on a computer from L-band to Ka-band; channel impulse or frequency response estimation in real time by the technique burst-mode link analyser with least-squares algorithm (BMLA/LS), modelling and simulation of the channel fast fading due to multipath and shadowing; acquisition of the equivalent normalized binary signal-to-noise ratio (E_b/N₀) data at the channel input; and bit-error rate (BER) evaluation by the directly error-counting approach and the computational approach at the channel output in simulation.