灰色—支持向量机的空调负荷预测

对空调负荷进行准确预测不仅对空调优化控制的意义重大,而且也是实现空调经济运行与节能的关键所在。为了提高建筑空调负荷的预测精度,在分析灰色模型和支持向量机建模特点基础上提出了一种空调负荷组合预测算法。该方法综合了灰色建模计算过程简单以及支持向量机自学习和泛化能力强的优点,能够更加有效地利用样本数据的有效信息,提高模型预测精度。首先,通过灰色建模过程弱化了样本数据的随机因素。然后,对灰色模型输出进行归一化处理及数据重构,以作为支持向量机的输入。最后,通过支持向量机模型的预测得到最终预测结果。将本文所提出的方法应用于福州一栋办公建筑的逐时空调负荷预测中,并与灰色模型及支持向量机模型作比较,证明了组合模型的预测值与实际运行值拟合度最高,平均绝对误差比灰色模型和支持向量机模型分别降低了47.84%和17.39%。该组合预测模型具有较高的预测精度和更好的泛化能力,具有较强的可行性和实用性。     

冲压式舵机动态特性测试系统的研制

为了在实验室条件下模拟冲压式舵机的舵面在空中所受的气动载荷,实现舵机系统动态性能指标的综合检测,设计了基于PXI总线的舵机动态测试系统,工作过程中对舵偏角信号、工作电流信号、扭矩信号进行实时采集、显示,并配合测试流程完成对舵机动态性能的测试。设计中重点考虑了定刚度加载系统的小扭矩加载精度、小转动惯量以及同轴度问题。实验结果表明,该测试系统测试精度高、使用维护方便,满足要求。     

基于悬臂梁称重传感器的圆弧刀研磨力测量仪设计

为解决双圆锥形圆弧刀研磨力在线测量难题,设计了杠杆原理为基础的圆弧刀研磨力测量仪,建立了圆弧刀与磨盘之间研磨力数学模型。该仪器采用悬臂梁称重传感器,并以DSP2812为主控芯片,实现了圆弧刀研磨力的在线采集、处理、存储及实时显示等功能。实验表明:仪器的称重传感精度优于0.15 N,圆弧刀研磨力变化范围降低28.8%,满足圆弧刀研制的需求。该仪器性能稳定,具有较好的实用价值,为高质量的圆弧刀研制和光栅刻划技术提供了技术保障。     

平行梁式电容传感器力学性能的仿真与分析

针对S型平行梁式电容称重传感器参数结构设计问题,运用有限元软件ANSYS Workbench对平行梁式弹性体结构进行参数化模拟仿真,基于二次多项式曲面拟合建立其结构力学性能与主要结构参数的函数关系,并通过算例分析验证了拟合公式的有效性。研究表明,拟合公式对复杂仿真计算后的结果有很好的近似,能更好的反应实际情况,公式结构更加简洁,具有较快的收敛速度、较高的精度和较好的推广能力,有利于准确实现平行梁式电容称重传感器结构参数设计。     

基于多GPU集群的编程框架

现如今,GPU作为一种低功耗高性能图形处理器单元,被广泛应用于高度并行化的应用程序中。其线程和内存的层次结构在诸多成功的多线程应用和科学研究中表现出巨大的优势。为了简化多GPU集群的编程模式以及更好地利用GPU的计算性能,设计并实现了一个新的基于多GPU的MapReduce并行编程框架。使用了并行虚拟文件系统（PVFS）来存储数据,考虑了动态的负载平衡和GPU相关的权重要素以达到优化系统的效率、透明性以及系统的可伸缩性的目的。在文中,将演示使用该编程模式解决地质应用的一个典型的偏移应用-叠前时间偏移（PKTM）,并给出实验结果。     

用三维实体单元分析薄壁箱型结构应注意的问题

从四边固定正方形板受均布载荷的经典问题入手,结合不同类型单元的特点,探讨应用三维实体单元分析薄壁箱型结构时应注意的几个问题,纠正三维实体单元分析和应用的误区,为提高结构分析计算结果精度提供参考.     

柴油发动机活塞静强度分析

针对铁路运输向高速重载方向发展对机车发动机性能要求不断提高的问题,为确保柴油发动机工作的可靠性,对其活塞进行静强度分析.用HyperMesh建立某柴油发动机活塞的三维有限元模型,利用ANSYS对其进行有限元分析.综合考虑热负荷与机械负荷的耦合作用,确定活塞最大应力和最大应变分布的危险部位.分析结果可以为铁路机车柴油发动机活塞设计提供参考.     

Nickel foam and carbon felt applications for sodium polysulfide/bromine redox flow battery electrodes

The first use of nickel foam (NF) as electrocatalytic negative electrode in a polysulfide/bromine battery (PSB) is described. The performance of a PSB employing NF and polyacrylonitrile (PAN)-based carbon felt (CF) as negative and positive electrode materials, respectively, was evaluated by constant current charge-discharge tests in a single cell. Charge/discharge curves of the cell, positive and negative electrodes show that the rapid fall in cell voltage is due to the drop of positive potential caused by depletion of Br₂ dissolved in the catholyte at the end of discharge. Cell voltage efficiency was limited by the relatively high internal ohmic resistance drop (iR drop). Polarization curves indicated that both NF and CF have excellent catalytic activity for the positive and negative redox reactions of PSB. The average energy efficiency of the single cell designed in this work could be as high as 77.2% at 40 mA cm⁻² during 48 charge-discharge cycles.     

Comparison of two IT DIR-SOFC models: Impact of variable thermodynamic, physical, and flow properties. Steady-state and dynamic analysis

This paper compares two dynamic, one-dimensional models of a planar anode-supported intermediate temperature (IT) direct internal reforming (DIR) solid oxide fuel cell (SOFC): one where the flow properties (pressure, gas stream densities, heat capacities, thermal conductivities, and viscosity) and gas velocities are taken as constant throughout the system, based on inlet conditions, and one where this assumption is removed to focus on the effect of considering the variation of local flow properties on the prediction of the fuel cell performance. The refined model consists of mass, energy, and momentum balances, and of an electrochemical model that relates the fuel and air gas compositions and temperatures to voltage, current density, and other relevant fuel cell variables. Simulations for steady-state and dynamic conditions have been carried out and the results obtained from the two models compared. For a co-flow SOFC operating on a 10% pre-reformed methane fuel mixture, with 75% fuel utilisation, inlet fuel and air temperatures of 1023 K, average current density of , and an air ratio of 8.5, the results show that, although the error incurred in the prediction of the flow properties in the first model is significant, there is good agreement between both models in terms of the overall cell performance: the maximum difference in the local temperature values is about 7 K and the cell efficiency differs by less than 1%. However, the discrepancies between the two models increase, especially in the fuel channel, when higher current density values are assigned to the cell.     

EFFECT OF PORE DIFFUSION ON DEACTIVATING RATES OF SECOND ORDER REACTION

The research and development of catalysts, the design and operation of catalytic reactors are complicated by the phenomenon of catalyst deactivation which is affected by various factors. In this paper, the effect of internal mass transfer resistance in the second order catalytic reaction system of uniform independent deactivation on both reaction and deactivation rates of straight cylindrical pore model has been discussed. In addition, the quantitative relations between effectiveness factor and Thiele modulus in second order reaction and first order deactivation system have been correlated