1,6—二磷酸果糖制备研究：Ⅰ.菌种的筛选及培养条件

介绍了１,６－二磷酸果糖产生菌的筛选和最佳培养条件的确定,在所筛选的菌种中９５１８、９５１９号菌株均具较高的产１,６－二磷酸果糖酶系活力。其最佳培养条件为：培养温度２８℃,ｐＨ６．５。     

从蓝色犁头霉中提取壳聚糖的初步研究

报道了一种从蓝色犁头霉细胞壁中提取壳聚糖的简便方法。经７２ｈ摇瓶培养,在１００℃下、５％ＮａＯＨ间歇处理二次,２％ＨＣＬ抽提１２ｈ,天然壳聚糖产量占菌丝干重的１１．７２％,脱乙酰基制备的壳聚糖产量占菌丝体干重的２．５８％。     

Lattice distortion of nanoscale silver particles

The lattice distortion of nanoscale silver particles prepared is studied by using the chemical method. The particles lattice contract was found, and the lattice contraction become greater while the particle size decreases. The results were explained with surface energy and surface tension.     

An all solid-state electrochromic smart window

An all solid-state electrochromic smart window employing prussian blue and electrodeposited WO₃ film with poly (vinyl chloride) (PVC) gel electrolyte that has high conductivity (2 mS/cm) at room temperature has been fabricated for the first time. The smart window has beell found to be excellent for electrochromism and memory characteristics.     

Preparation of InSb nanocrystals embedded in SiO2 thin films

InSb nanocrystals embedded in SiO₂ thin films were prepared by rf magnetron cosputtering technique. THe observation by transmission electron microscope showed that InSb nanocrystals dispersed uniformly in SiO₂ matrices. InSb nanocrystals with different sizes can be obtained by changing the annealing condition. The average size of InSb nanocrystals depended on annealing temperature and time, but not on the t_1/3 rules. X-ray photoelectron spectroscopy and X-ray diffraction were also applied to the analyses of the composite thin films.     

补偿扰动的PWM逆变器无差拍控制

提出了两种ＰＷＭ逆变器的改进无差拍控制算法．这两种算法根据ＰＷＭ逆变器的状态方程,分别采用直接补偿扰动和最优预见控制技术,利用未来信息决定当前的控制输入．使用ＭＡＴＬＡＢ软件对这两种算法进行了仿真．结果表明,当存在随机扰动及电路参数变化时,所提出的两种算法的控制效果均比不考虑补偿扰动的无差拍算法有较大改善．因此,该两种控制算法能够有效地抑制电路参数变化引起的或来自外部的扰动,使输出电压能很好地跟踪参考电压,控制响应快,鲁棒性好．     

一种应用神经网络的实时语音增强方法

提出了一种用神经网络模型和信号子空间特征分解相结合进行语音增强的方法,该方法利用了神经网络并行处理、高速计算的能力和语音信号的短时平稳的特性．它既克服了传统的谱相减法中残留“音乐噪声”的缺陷,又可以实时、有效地增强语音,能够满足一些语音处理与识别系统的需要．     

TS分子筛的催化氧化性能研究：Ⅲ.苯酚氧化制苯二酚

研究了自合成的钛硅分子筛（ＴＳ－１）对Ｈ２Ｏ２氧化苯酚制苯二酚反应的催化活性。对反应的温度、时间、催化剂用量、酚：Ｈ２Ｏ２等因素进行了考察：在催化剂用量为苯酚的１０％、酚：Ｈ２Ｏ２（ｍｏｌ）为３：１时,经５７℃／６ｈ反应,Ｈ２Ｏ２的转化率和对产物苯二酚的利用率可达９０％和８０％,还观察到,分子筛的晶粒大小对苯酚的氧化活性影响较大,大晶粒催化剂的内表面利用率较低,从而影响催化活性。     

基于神经网络的无监督自学习控制系统

所论系统的控制由两个神经网络完成,即对象放识器和系统控制器。并且提出了一种 新的学习算法,它能利用存贮于辨识网络的信息改进控制策略。该控制系统可以对非线性系统,不确定系统等进行无监督的学习控制。