牵引网微机保护整定的探讨

通过对牵引网负荷电流、短路电流中的谐波含量的分析和存在的谐波含量的不同点,利用两种电流中的综合谐波含量作为控制量,可以大大提高牵引网距离保护、高阻接地保护躲过牵引重负荷,再生负荷,启动电流及励磁涌流的能力,对于保护整定,通过探讨提出了整定原则和方法。     

人工神经网络在基坑变形性状研究中的应用

基坑工程不仅要保证维护结构本身的安全,而且要保证周围建(构)筑物的安全和正常使用。开展基坑工程变形性状研究具有重要意义。影响基坑变形的因素很复杂,传统的计算方法已无法准确预测基坑的变形。人工神经网络(ANN)有很强的解决非线性问题的能力。近些年来基于ANN的方法在岩土工程中得到了广泛的应用。本文对ANN在基坑工程中的应用研究进行了综述,并对其发展方向进行了分析。     

光纤光栅传感器在分布式测量系统中的应用

介绍了光纤光栅传感器在分布式测量系统中的应用，给出了光纤光栅测量节点的软硬件实现，并对实时性等影响分布式系统性能的因素做出了分析，最后，对于完善方案提出了构想。     

四川南向通道改造及建设的方案研究

研究目的：通过分析西南路网现状和存在问题，对四川出川通道进行规划展望，提出四川南向通道建设的必要性和迫切性，进一步探讨建设川南通道的方案和通道引入相关枢纽及地区的改扩建方案。 研究方法：采用系统工程方法，从西南片区铁路路网的特点分析，逐一分析出川通道的规划和发展；采用技术经济比较法对南向通道方案进行比较论证。 研究结果：在西南片区形成沟通广东广西地区的一条新路网，有利于西南路网布局及发展；带动相关地区社会经济发展；为相关枢纽总图布置特别是成都枢纽客运站布局提供合理的发展前景。 研究结论：四川南向通道宜采用沿成昆通道至峨眉，修建峨眉经宜宾、叙永至贵阳的双线大能力通道，经黔桂铁路至柳州，分别经湘桂线南下广西地区，经柳肇线可达广州及东南沿海地区；成都枢纽成都南站可作为川南通道客车的始发终到站；贵阳枢纽结合川南通道引入需新建枢纽内第二客运站。     

转K2型转向架弹簧载荷识别及其对弹簧寿命影响的研究

用试验的方法得到在垂直载荷作用下弹簧45°方向应力和最大主应力之间的差异,用有限元法计算了弹簧在横向、垂向加载时的载荷识别系数。根据线路实测动应力识别出转K2型转向架弹簧在运用工况下的2种载荷。最后根据载荷计算出弹簧的应力,利用疲劳损伤模型计算出损伤,得出了2种载荷对疲劳寿命的影响。     

电信计费模型的设计

依据电信计费系统存在的问题,本文详细描述了电信计费模型的设计方法和各个模块的设计内容。     

Organic carbon budget for the Gulf of Bothnia

We calculated input of organic carbon to the unproductive, brackish water basin of the Gulf of Bothnia from rivers, point sources and the atmosphere. We also calculated the net exchange of organic carbon between the Gulf of Bothnia and the adjacent marine system, the Baltic Proper. We compared the input with sinks for organic carbon; permanent incorporation in sediments and mineralization and subsequent evasion of CO₂ to the atmosphere. The major fluxes were riverine input (1500 Gg C year^− 1), exchange with the Baltic Proper (depending on which of several possible DOC concentration differences between the basins that was used in the calculation, the flux varied between an outflow of 466 and an input of 950 Gg C year^− 1), sediment burial (1100 Gg C year^− 1) and evasion to the atmosphere (3610 Gg C year^− 1). The largest single net flux was the emission of CO₂ to the atmosphere, mainly caused by bacterial mineralization of organic carbon. Input and output did not match in our budget which we ascribe uncertainties in the calculation of the exchange of organic carbon between the Gulf of Bothnia and the Baltic Proper, and the fact that CO₂ emission, which in our calculation represented 1 year (2002) may have been overestimated in comparison with long-term means. We conclude that net heterotrophy of the Gulf of Bothnia was due to input of organic carbon from both the catchment and from the Baltic Proper and that the future degree of net heterotrophy will be sensible to both catchment export of organic carbon and to the ongoing eutrophication of the Baltic Proper.     

复突触神经网络模型及其稳定性研究

传统的Hopfield网络在联想记忆、优化计算方面发挥了重要作用,但该网络存在一个弊端,即它只能处理线性或二次的最优化问题。因此,本文首先对Hopfield神经网络进行了扩展,将其称为复突触神经网络(或模型),扩展的网络不仅可以解决目标函数或约束中含有高阶的最优化问题,而且还可以处理目标函数含有对数、三角形式等较复杂函数的最优化问题;然后给出了该网络稳定的充分条件并给出了证明,同时对网络中的权矩阵是否对称也进行了讨论;最后结合典型的模糊c均值聚类问题验证了该模型的有效性。     

Experimental ivnestigation on flow fields of viscous fluid around two-dimensional wings and comparison with computational results

To examine the flow field of a viscous fluid around the trailing edge of a wing with respect to the Kutta condition or the Joukowski hypothesis, an experimental investigation into the velocity fields around two-dimensional (2D) wings NACA0012 and NACA4412 was carried out using an X-type hot-wire anemometer in a wind tunnel and the method of colored milk injection in a circulating water channel. The results of these investigations revealed that the flow of a viscous fluid at the trailing edge of a 2D wing is tangential to the face on the pressure side, and that the flow is very slow or reversed on the suction side due to separation when the angle of attack is greater than 7.5°. By flow visualization, a Kármán vortex street was found in the wake of both wings (NACA0012 and NACA4412) when the angle of attack was 5°. These results show that the ordinary Kutta condition with respect to the direction of outflow at the trailing edge is not necessarily satisfied in a viscous fluid, but the Kutta condition with respect to the pressure at the trailing edge is satisfied as usual. Numerical results, which have been obtained as the solution of Reynolds-averaged Navier-Stokes equations, show good agreement with above-mentioned experimental results. Therefore, CFD (Computational Fluid Dynamics) simulation has proved to be very effective for studying steady viscous flow around the trailing edge of a 2D wing.     

铁路行车事故分析中的语音压缩方法

在铁路运输中，安全行车是第一位的。一旦出现行事事故，对事故原因的分析往往十分棘手。采用黑匣子技术能够准确地记录列车运行中的各种信息，但巨大的语音存储量需求是黑匣子技术的一个难关。若采用传统的语音压缩技术，则理论和实现都很复杂，并且成本太高。采用神经网络不仅简单易行，还有提高信噪比的功能。