A New Reduction Technique for Thermal Models with Fluid Networks

Thermal finite element models are used to simulate the temperature distribution of a body under a certain set of boundary conditions. When short calculation times are necessary (i.e., for parametric study, active control loops, temperature monitoring algorithms, etc.) detailed FE models cannot be used and reduced-order models must be developed. Reduced-order models may have both generalized and physical degrees of freedom (dofs), according to the scope of the reduction. The use of generalized dofs is more suitable to simulate full-field temperature distribution of the body, while inclusion of physical dofs can be advantageous when temperature must be only computed at a small number of locations. In this paper a new reduction technique is proposed for applications where only a small set of temperatures must be computed; it is based on the Guyan reduction method and overcomes the limitations of the original Guyan reduction when applied to thermal models with a fluid network, characterized by non-symmetric matrices. The new technique is applied to a case study, consisting of a dummy turbine disk and its secondary air system in order to show the improvement with respect to the original Guyan approach.     

高炉炉渣滞留量水模试验及分析

模拟高炉炉渣排放情况,设计开展水模试验,对炉渣滞留量影响因素进行分析。结果表明：油的黏度大,油的滞留量大;空隙度大,油的滞留量大;浮起高度大,油的滞留量小。由此可知,炉渣黏度降低、焦炭热强度提高和铁口数量增加有利于降低铁水带渣量;炉底死焦堆浮起高度越大越有利于炉渣排放。     

固定式精炼炉的改进

江西铜业集团公司贵溪冶炼厂(简称贵冶)固定式精炼炉,因其在设计上存在缺陷及生产工艺的改变,不能满足生产的需要。在生产实践中积累经验后,将其进行技术改造。将改造前后的不同状况列出进行比较,从而阐述改造后取得的效果,最后列举出贵冶固定式精炼炉目前还存在的不合理因素,并设计出解决的方案,做到工作持续推进。     

利用自然适应性解决农村配电网电压质量问题的优化规划研究

针对目前农村配电网易出现负荷重时电压偏低、负荷轻时电压偏高的情况,提出一种利用配电网的自然适应性而不必建设通信通道和控制电容器投切的电压治理方案。综合运用增大导线横截面积、改造成三芯架空电缆和安装并联补偿电容器的手段改善农村配电网的电压质量。以规划期内收益最大为目标函数,综合考虑农村配电网改造后节省的损耗费用、加粗或改用架空电缆的费用和加装电容器的建设费用,建立了固定电容器与线路改造相结合的优化规划模型。给出了一个典型算例详细进行了说明,分析结果表明所建议的方法能够有效改善农村配电网电压质量。     

穿透曲线法获取固定床吸附器的传质扩散系数

研究了煤质活性炭对不同浓度苯酚溶液的吸附，应用固定床吸附动力学模型、非线性回归固定床吸附流出曲线数据，获得了酚类在活性炭上的扩散传质系数。结果发现，酚类的表面扩散系数ＤＳ强烈地依赖于入口酚的浓度，随着酚浓度的升高，表面扩散系数增大。采用上述扩散系数理论预测了其他操作条件下的流出曲线，结果表明实验曲线与预测曲线相符甚好，穿透曲线法能可靠地获取液固吸附过程的吸附扩散系数，为吸附工程设计提供有益的帮助     

Neural network analysis of strength and ductility of welding alloys for high strength low alloy shipbuilding steels

Abstract

There are considerable demands for the development of weld metals for high strength low alloy steels. To assist in meeting such demands, a neural network was trained and tested on a set of data obtained on weld metals for steels of the type used for shipbuilding. The input variables for the network were the chemical elements and the weld cooling rate. The outputs consisted of the yield and ultimate tensile strengths, elongation, and reduction of area. Many models were created using different network configurations and initial conditions. An appropriate committee of models was then assembled by testing the ability of the models to generalise on unseen data. The neural network technique used is due to MacKay, with a Bayesian framework, and hence allows the estimation of error bars, which warn the user when data are sparse or locally noisy. The method revealed significant trends describing the dependence of mechanical properties on weld composition and cooling rate.     

A Proposal on Data Treatment In the Feynman-α Experiment

A new algorithm of Monte Carlo criticality calculations for implementing Wielandt's method, which is one of acceleration techniques for deterministic source iteration methods, is developed, and the algorithm can be successfully implemented into MCNP code. In this algorithm, part of fission neutrons emitted during random walk processes are tracked within the current cycle, and thus a fission source distribution used in the next cycle spread more widely. Applying this method intensifies a neutron interaction effect even in a loosely-coupled array where conventional Monte Carlo criticality calculation methods have difficulties, and a converged fission source distribution can be obtained with fewer cycles. Computing time spent for one cycle, however, increases because of tracking fission neutrons within the current cycle, which eventually results in an increase of total computing time up to convergence. In addition, statistical fluctuations of a fission source distribution in a cycle are worsened by applying Wielandt's method to Monte Carlo criticality calculations. However, since a fission source convergence is attained with fewer source iterations, a reliable determination of convergence can easily be made even in a system with a slow convergence. This acceleration method is expected to contribute to prevention of incorrect Monte Carlo criticality calculations.     

基于解析模式分解的密集工作模态参数识别

长大跨度的桥梁结构或者高层建筑的工作环境振动响应中经常包含密集的模态成分,并会出现模态叠混现象,而传统的信号分析方法往往难以识别结构的密集模态参数。提出一种基于解析模式分解理论与随机减量技术相结合的方法识别环境激励下的结构密集模态参数。对于工作环境激励下的结构振动响应,通过随机减量技术可以提取结构的自由振动响应,利用解析模式分解方法对具有密集模态的自由振动响应进行有效的分解,对每一阶自由振动响应利用最小二乘拟合方法识别出频率与阻尼比。通过两层框架的数值模拟以及对密集频率的密集程度指数和信号时程长度等参数分析,其结果表明通过随机减量技术提取的自由振动响应可以有效的减少模态叠混的影响,虽然提取的自由振动响应的时程长度比实际的信号时程要短,然而解析模式分解仍然能够十分有效的对短时程具有密集模态成分的信号进行有效的分解。最后,通过对一具有密集模态的36层框架的数值模拟,以及对一具有密集模态的3层框架的振动台实验,验证该方法可以有效的识别出环境激励下的结构密集模态参数。     

基于多元统计方法的新疆固定资产投资结构分析

运用因子分析法对新疆15个地区14个行业的固定资产投资进行分析,提取出3个公因子（固定资产投资综合因子、服务性投资因子和高级产业投资因子）分析各地区的投资结构和产业布局。在此基础上,运用聚类分析法将15个地区分成5类,找出投资结构类似的地区。最后,利用3个公因子,对新疆各地区投资结构与经济增长之间的关系进行回归分析,结果表明固定资产投资综合因子对经济增长起正向拉动作用,而服务性投资因子和高级产业投资因子对经济的拉动作用是负向的,说明新疆的经济增长还是粗放型的,从而为今后新疆固定资产投资结构与产业布局的优化调整提供理论依据和方向参照。     

基于相似性函数的稀疏成分分析混合矩阵估计方法

稀疏成分分析是信号处理中解决欠定盲源分离问题的新方法,本文研究了稀疏成分分析中的混合矩阵估计问题,提出了无需预知源个数利用一种相似性函数估计混合矩阵的方法。首先,估计相似性函数中的核参数,使得算法适应不同的稀疏信号。然后,给出了估计混合矩阵的不动点算法。最后,实验结果表明提出的算法通过适当地选取参数,能够准确有效地估计出具有不同源个数的混合矩阵,对不太稀疏的源也有令人满意的结果。