基于电阻抗谱特性与集成神经网络的天然气水合物饱和度软测量模型

含天然气水合物饱和度的计算是储层优选和资源量评估的关键参数,针对目前数据解释模型计算精度低以及模型输入参数少等问题,提出了一种基于电阻抗特性参数和集成神经网络的软测量模型建立方法。在对电阻抗谱数据进行预处理、特征参数提取以及选择的基础上形成了样本集,针对四对传感器分别设计了BP神经网络,采用平均法作为集成策略将四个BP网络作为子网络进行集成得到集成网络模型。模型测试结果表明：通过集成网络模型计算得到的含水合物饱和度值平均相对误差3.33%、平均绝对误差0.0014、均方根误差为6.56%,三项误差指标均低于各个子网络的计算误差。在宽频范围内对含水合物沉积物进行电阻抗谱测试能够获得沉积物的频率响应特性以及特性描述参数,可为神经网络模型提供大量的输入参数;利用集成神经网络能够综合应用位于不同测量方位的多个传感器的测量数据,通过采用适合的集成策略能够克服水合物空间分布不均匀对饱和度计算准确度的不利影响。     

Possible relaxation and conduction mechanism in W6+doped SrBi4Ti4O15 ceramic

Tungsten (W) doped Strontium Bismuth Titanate ceramics with general formula [SrBi_4-2x/3Ti_4−xW_xO₁₅ (x=0.00,0.02,0.04,0.06,0.08,0.1)] were prepared via solid-state reaction route. The X-ray diffraction analysis confirmed a single phase system with orthorhombic structure for all the prepared samples. The relaxation and conduction mechanism were studied by using impedance spectroscopy analysis. The shifting of Z"_max and M"_max peak to the higher frequency side in accordance with temperature for all the sample are found to obey the Arrhenius law, indicates the presence of relaxation process in the material. The complex impedance plots showed that both grain and grain boundary are responsible for the conduction mechanism and the grain and grain boundary resistance decreases with temperature showing NTCR behaviour in all the compositions. The kinetic analysis of frequency dependence ac conductivity have been discussed in terms of polaron mechanism and the dc conductivity shows a thermally activated process. We have observed that the charge carrier hopping and doubly ionized oxygen vacancy (Vₒ••) are mainly responsible for the conduction mechanism in the systems.     

测定炉管蠕变损伤分布的新方法

根据薄板弯曲自由振动理论，利用差分法从数值上计算了悬臂薄析及其附加集中质量时的基本频率，导出了基本频率间的相关关系，给出了用加速度计测定炉管材料的弹性模量的方法，确定了炉管蠕变损伤沿径向的分布。其结果与超声波探伤结论相吻合。     

耐火材料高温弯曲应力-应变关系测试方法及影响因素

介绍了耐火材料弯曲应力-应变测试仪的基本构成以及对试验炉、加荷装置、变形量测量装置和计算机控制系统等各构件的要求;重点介绍了使用该仪器研究耐火材料应力-应变关系的方法以及影响测试结果的相关的因素。该方法可以用来研究耐火材料在高温下承受弯曲应力时的变形问题,确定耐火材料在各温度下的应力-应变关系,测试和判断耐火材料的弹性、塑性和粘滞流动的温度范围,并根据应力-应变关系计算耐火材料在不同温度下的弹性模量。     

Development of Reproducing Alumina-Magnesia-Carbon Bricks

The reproducing alumina-magnesia-carbon bricks were prepared with the dumped bricks as starting materials. The bulk density, apparent porosity, crushing strength, modolus of rupture and slag resistance of the specimen were analyzed. The results show that the used refractories can be reused and recycled by the right method. The reproducing alumina-magnesia-carbon bricks with better abilities were prepared.     

聚丙烯的高性能化研究

在小型反应器内合成了不含及含有成核剂的聚丙烯(PP)、双峰聚丙烯(BMPP)和聚丙烯共聚物(PPc)，并用纳米CaCO3改性PPc。研究结果表明：釜内成核的PP和BMPP的弯曲模量和热变形温度显著增加，结晶温度和结晶速率明显提高，BMPP的拉伸强度也较PP大幅增加；但成核剂对PPc的性能影响不大，而加入纳米CaCO3后则使PPc的力学性能和结晶温度同步显著增加。     

一种改进的变步长双模算法研究

在对常模算法（CMA）和改进常模算法（MCMA）的研究基础上,提出了一种改进的变步长双模算法（Variable Step-size MCMA—DD）。改进算法通过误差信号自动地调节步长,从而调节算法的收敛速度。通过仿真表明,改进算法具有与MCMA—DD算法相近的误码性能,但是很大程度地加快了算法的收敛速度。     

一种联合盲均衡算法

为克服数字基带信号在通过非屏蔽五类双绞线时产生的严重码间干扰（ISI）,常采用自适应均衡技术来减小码间干扰,大大降低接收端信号的误码率。最小均方误差（LMS）算法能有效降低码间干扰,但需要训练序列,因此影响传输效率。基于判决引导的最小均方误差（DDLMS）算法不需要训练序列,但在眼图未睁开的情况下,可能出现误判,甚至引起误收敛。恒模算法（CMA）具有比DDLMS算法更好的盲均衡特性,但是剩余误差较大。本文提出一种新颖的联合盲均衡算法,即优化现有的CMA算法,与DDLMS算法组成新的联合盲均衡算法,利用均方误差（MSE）来控制2种算法的权值。MATLAB建模和仿真结果表明,新的联合盲均衡算法克服了CMA算法剩余误差较大和DDLMS算法误收敛的缺陷,且能有效对非屏蔽五类双绞线中传输的数字信号进行均衡。     

应力波时间仪的研制及其应用

本文阐述了用加速度传感器将应力波能量转换成电信号,放置两个在被检测材料的两端,一个作为“开门”端,一个作为“关门”端。应力波经过两端所用时间可用微秒时间间隔测量单元测得。从而可以得到应力波的传播速率。本文给出了信号放大、信号变换及门信号产生、时间间隔测量电路的设计原理和过程。     

A displacement-based back-analysis method for rock mass modulus and horizontal in situ stress in tunneling – Illustrated with a case study

This paper presents a displacement based back-analysis method for the determination of rock mass modulus (E) and the horizontal in situ stress (P) perpendicular to the axis line of the tunnel excavation in hard and intact rock masses. The rock mass is assumed to respond in an isotropic and linear elastic manner. Essentially, the method is a best-fit solution of back-analysis by comparing the measured displacements near a tunnel face during excavation with those calculated using a three-dimensional finite element method. The method can be applied efficiently and effectively by iterative algorithms such as direct search technique and damping least square technique. In particular, a three-dimensional finite element pattern technique (3-D FEPT) is used to compute the theoretical displacements for saving computation time. The method is further illustrated with a case study. The case study deals with a test adit at Zhanghewan Pumped Storage Power Station in Hebei Province, North China. It is shown that the back-analyzed results of E and P are reliable and representative for the actual rock masses.