基于BP神经网络优化的离心式冷水机组故障诊断研究

针对离心式冷水机组7种常见故障(包括系统故障和局部故障),建立基于误差反向传播(BP)神经网络的诊断模型。通过调整网络结构及参数、改变训练函数等手段优化模型,并详细讨论优化前后故障诊断的总体正确率及各类故障的诊断结果。结果表明:网络优化后,总体诊断正确率从最初的57.1%(单隐层、10个节点/层)提高到95.5%(四隐层、20个节点/层),采用贝叶斯归一化法训练最佳;与局部故障相比,系统故障由于其对系统运行的广泛影响而更难以被识别。     

SOM离散化和IGA优化神经网络传感器故障诊断

为实现对节点进行实时有效的故障诊断,提出一种基于SOM自组织网映射算法和免疫遗传神经网络的故障诊断方法。首先,设计节点故障诊断的通用模型;然后采用自组织网映射(self organize mapping,SOM)算法对故障特征数据进行离散化以提高诊断模型的泛化性能,并通过主成分分析法(principal component analysis,PCA)对特征数据属性进行约简以减少数据量;最后,建立三层的BP神经网络节点故障诊断模型,并通过免疫遗传算法(immune gene algorithm,IGA)对BP神经网络优化从而得到最终优化的节点故障诊断模型。仿真实验证明该法能有效进行故障诊断,且与其他方法相比,具有诊断效率高和精度高的优点。     

多联机系统制热工况下的制冷剂充注量故障诊断模型——基于多层卷积神经网络

本文提出一种基于卷积神经网络的故障诊断模型,并通过正交试验优化了3层网络的卷积核和神经元数目,利用图形化的多联机(VRF)系统制冷剂充注量故障实验数据训练了多层卷积神经网络,评估了本模型的故障诊断性能。结果表明:该"数据图形化-多层卷积神经网络"方法建立的模型能够有效进行多联机制冷剂充注量故障诊断,20个输入特征时,对9类故障诊断总正确率最大为91%,比传统BP神经网络达到更高的诊断精度。该方法首次利用卷积神经网络完成了VRF制冷剂充注量故障诊断,为相关研究的拓展奠定了基础。     

多联机系统制热工况下的制冷剂充注量故障诊断模型——基于多层卷积神经网络

本文提出一种基于卷积神经网络的故障诊断模型,并通过正交试验优化了3层网络的卷积核和神经元数目,利用图形化的多联机(VRF)系统制冷剂充注量故障实验数据训练了多层卷积神经网络,评估了本模型的故障诊断性能。结果表明:该"数据图形化-多层卷积神经网络"方法建立的模型能够有效进行多联机制冷剂充注量故障诊断,20个输入特征时,对9类故障诊断总正确率最大为91%,比传统BP神经网络达到更高的诊断精度。该方法首次利用卷积神经网络完成了VRF制冷剂充注量故障诊断,为相关研究的拓展奠定了基础。     

基于BP神经网络的盲均衡器设计

为构建用于某语音信号传输系统盲均衡器的BP神经网络模型,编写了基于BP神经网络的盲均衡算法伪代码,计算了算法的时间复杂度,分析了BP神经网络输入层神经元个数、隐含层神经元个数和隐含层层数对盲均衡算法性能的影响,评估了基于Sigmoid的变步长算法、基于误差补偿的变步长算法和基于误差的变步长算法对基于BP神经网络的盲均衡器性能的改进效率,据此设计了一种含双隐层结构的BP神经网络盲均衡器,并对其性能进行了数值仿真分析,明确了其适用范围,为该语音信号传输系统设计提供了技术支撑。     

一种黄金分割优化的极限学习机算法

针对极限学习机(ELM)存在大量隐层神经元个数和随机给定权值导致算法性能不稳定等问题,将黄金分割法(Golden Section)与ELM相结合提出了基于黄金分割优化的极限学习机算法(GS-ELM).首先通过黄金分割法对ELM隐含层节点数进行优化,接着再用该方法对ELM输入层权值和隐含层偏差进行优化.实验结果表明,相比较传统的BP神经网络,支持向量机和极限学习机,GS-ELM算法能获得较高的分类精度.     

一种粒子群RELM的基因表达数据分类方法

正则极限学习机(regularized extreme learning machine,RELM)具有比极限学习机(extreme learning machine,ELM)更好的泛化能力.然而RELM的输入层权值、隐含层偏差是随机给定的,会影响RELM的稳定性.另外,RELM为了获得较理想的分类精度,仍需设置较多的隐层节点.针对此问题,通过分析粒子群优化算法(particle swarm optimization,PSO)的原理,把RELM初始产生的输入层权值、隐含层偏差作为粒子带入PSO进行寻优.通过在Breast和Brain数据集上进行多次10折交叉验证表明,粒子群改进正则极限学习机(PSO-RELM)可以在隐层节点设置较少时获得比BP神经网络(back propagation,BP)、支持向量机(support vector machine,SVM)、RELM更好的分类精度和更佳的稳定性.     

遗传算法和BP神经网络在电机故障诊断中的应用研究

人工智能方法在电机故障诊断中的应用,使得电机故障能够得到及时准确的预测和诊断,保障了电机的安全运行。介绍了BP神经网络及遗传算法的基本原理及组成结构,针对BP神经网络容易陷入局部极小点及收敛速度慢的问题,利用遗传算法对BP神经网络的权值和阀值优化,改善了BP神经网络的诊断性能;通过GA-BP网络对电机的三种故障模式进行了诊断识别,其实验仿真结果表明：无论是在诊断速度上还是诊断精度上,GA-BP神经网络诊断性能都比单独的运用BP网络有了很大提高。     

基于BP神经网络算法的316L不锈钢极薄带热处理力学性能预测

目的 为了预测不锈钢极薄带热处理后的力学性能、优化热处理工艺以及实现热处理工艺的智能控制,构建基于BP算法的神经网络模型。方法 以316L不锈钢极薄带为研究对象,进行热处理试验和拉伸试验,通过以热处理的退火温度、保温时间和取样方向作为输入层参数,以屈服强度、抗拉强度、断后伸长率作为输出层参数,采用BP算法构建了316L不锈钢极薄带力学性能预测的思维进化算法优化BP神经网络模型,并进行模型的预测和应用验证,考虑不同隐含层节点数及不同BP神经网络模型对性能的影响。结果 思维进化算法优化的BP神经网络模型测试集的屈服强度、抗拉强度和断后伸长率的平均相对误差分别为8.92%、5.21%和9.28%,训练集相关系数为0.980 94。思维进化算法优化BP网络单、双隐含层误差总和最低分别为0.578 6和0.546 9,BP网络与思维进化算法优化的BP网络误差总和最低分别为0.579 9和0.546 9。结论 思维进化算法优化BP神经网络模型具有较好的预测能力和泛化能力,以及较高的预测精度。与企业现用生产工艺相比,采用模型优化后热处理工艺的综合力学性能有显著提高。     

基于改进BP的分子泵故障诊断研究

分子泵为EAST装置提供洁净的真空环境,其运行状态影响EAST实验的顺利开展。由于在EAST实验运行过程中,分子泵设备可能会出现异物坠入或者真空泄漏故障,对装置造成次生危害。针对分子泵故障数据集不平衡导致故障诊断精度低以及模型过拟合问题,提出一种基于时域频域预处理与改进BP相结合的算法,实现分子泵故障诊断。通过在BP神经网络的基础上,引入粒子群算法(PSO)并结合五折交叉验证优化模型。首先在模拟分子泵故障的破坏性测试平台上,采集正常态、真空泄漏以及异物坠入故障振动信号,然后对数据进行时域频域特征提取融合,将得到的特征向量集作为优化算法的输入,对模型进行训练,实现分子泵故障诊断。经实验验证,所提出改进BP算法在诊断精确率上可以达到96.84%,优于支持向量机(SVM)、K近邻（KNN）和BP算法。     

Real gas choked flow conditions at low reduced-temperatures

Real gas choked mass flux is calculated for a frictionless stream expanding isentropically until it reaches the speed of sound and without phase changes. The other parameters associated with the choked state are the pressure, density, temperature ratios, and the speed of sound. Departure of the choked mass flux from the ideal gas model is discussed first in absolute terms and then in relative terms, using the Principle of Corresponding States, for gases with boiling points in the low temperature range. Reduced-stagnation pressures are examined up to values of 30 for hydrogen, neon, nitrogen, argon, methane, krypton, xenon, and R-14 and up to 100 for ⁴He. The corresponding reduced-stagnation temperatures go down to 1.4 and in some cases down to 1.2 for nitrogen and argon. Also discussed are the limiting values of stagnation parameters for which no phase change occurs in the choked state. Compared to the ideal gas, the mass flux may almost double and the critical pressure ratio may decrease by an order of magnitude. The relevance of results is discussed qualitatively and quantitatively for Joule-Thomson cryocooling.     

As-cast microstructures and behavior at high temperature of chromium-rich cobalt-based alloys containing hafnium carbides

Hafnium is often used to improve the high temperature oxidation resistance of superalloys but not to form carbides for strengthen them against creep. In this work hafnium was added in cobalt-based alloys for verifying that HfC can be obtained in cobalt-based alloys and for characterizing their behavior at a very temperature. Three Co–25Cr–0.25 and 0.50C alloys containing 3.7 and 7.4 Hf to promote HfC carbides, and four Co–25Cr– 0 to 1C alloys for comparison (all contents in wt.%), were cast and exposed at 1200 °C for 50 h in synthetic air. The HfC carbides formed instead chromium carbides during solidification, in eutectic with matrix and as dispersed compact particles. During the stage at 1200 °C the HfC carbides did not significantly evolve, even near the oxidation front despite oxidation early become very fast and generalized. At the same time the chromium carbides present in the Co–Cr–C alloys totally disappeared in the same conditions. Such HfC-alloys potentially bring efficient and sustainable mechanical strengthening at high temperature, but their hot oxidation resistance must be significantly improved.     

Degradation behaviour of sputtered Co–Al coatings on superalloy

Degradation behaviour of sputtered Co–Al coatings on Superni-718 substrate has been investigated. Cyclic high temperature oxidation tests were conducted on uncoated and coated samples at peak temperatures of 900 °C for up to 100 thermal cycles between the peak and room temperatures. The results showed that a dense scale formed on the coated samples during thermal cycling at the peak temperature of 900 °C. The external scale exhibited good spallation resistance during cyclic oxidation testing at both temperatures. The improvement in oxide scale spallation resistance is believed to be related to the fine-grained structure of the coating. Nanostructured Co–Al coatings on Superni-718 substrate were deposited by DC/RF magnetron sputtering. FE-SEM/EDS, AFM, and XRD were used to characterize the morphology and formation of different phases in the coatings, respectively. The Co–Al coating on superalloy substrate showed better performance of cyclic high temperature oxidation resistance due to its possession of β-CoAl phase as Al reservoir and the formation of Al₂O₃ and spinel phases such as CoCr₂O₄ and CoAl₂O₄ in scale. The oxidation results confirmed an improved oxidation resistance of the Co–Al coating on superalloy as compare to bare substrate in air at 900 °C temperature up to 100 cycles.     

Capacitance based mass metering for cryogenic fluids

This paper describes a method for measuring the mass of cryogenic fluids in on-board rocket propellant tanks or ground storage tanks. Linear approximations to the Clausius-Mossotti relationship serve as the foundation for a capacitance based mass sensor, regardless of fluid density stratification or tank shape. Sensor design considerations are presented along with the experimental results for a capacitance based mass gage tested in liquid nitrogen. This test data is shown to be consistent with theory resulting in a demonstrated mass measurement accuracy of ±0.75% and a deviation from linearity of less than ±0.30% of full scale mass. Theoretical accuracies are also shown to be ±0.73% for hydrogen and ±1.39% for oxygen for a select range of pressures and temperatures.     

Equipment using a “static-analytic” method for solubility measurements in potentially hazardous binary mixtures under cryogenic temperatures

A new apparatus designed to study, at cryogenic temperatures, thermodynamic equilibria of potentially explosive binary systems such as hydrocarbon-oxygen mixtures is described herein. This equipment has an equilibrium cell which was especially designed to minimize hazards while allowing accurate phase equilibrium measurements. Reliability of results, obtained with this equipment has been verified by working on the nitrogen-propane system, for which data are already available in literature, over a large range of compositions and at various temperatures. Four isothermal curves describing liquid phase compositions at 109.98, 113.77, 119.75 and 125.63 K have been determined. Our experimental data are represented within 2% in compositions and in pressures through the Peng-Robinson equation of state implying Mathias-Copeman alpha function and Huron-Vidal mixing rule. Comparisons to literature allow pointing out: good agreement is observed with Kremer and Knapp data (1983) while the three sets of Poon and Lu data (1974) presenting systematic positive deviation are consequently judged as suspicious.     

长余辉材料Sr_4Al_(14)O_(25)∶Eu~(2+),Nd~(3+)的溶胶凝胶法制备及性能研究

首次研究了以Nd~(3+)离子为辅助激活剂,对Eu~(2+)掺杂的发光材料Sr_4Al_(14)O_(25):Eu~(2+)余辉性能的影响.用溶胶凝胶法合成了Eu~(2+), Nd~(3+)共掺杂的Sr_4Al_(14)O_(25):Eu~(2+),Nd~(3+)发光粉末,并用扫描电镜、X射线衍射计、荧光分光光度计、余辉亮度测试仪、热释光剂量计等手段对粉末样品进行了表征.结果表明,在1350℃得到了单一的Sr_4Al_(14)O_(25)相,粉末颗粒平均粒度在1μm左右.Eu~(2+), Nd~(3+)共掺杂的Sr_4Al_(14)O_(25):Eu~(2+),Nd~(3+)发光粉末有402和485nm两个发射峰,与Eu~(2+)单掺杂的Sr_4Al_(14)O_(25):Eu~(2+)相比,发射峰位置没有变化,但适量的掺杂可以大大提高余辉时间和余辉亮度,余辉时间可达18h以上.最后通过对热释光谱的分析解释了双掺杂发光粉余辉性能增强的原因,适宜深度的陷阱可以有效存储光能,增强余辉的时间和强度.     

Percolation and the resistive transition of the critical temperature Tc of Nb3Sn

The relationship between the distribution of the critical temperature, the percolation function, and the resistive transition of the critical temperature is explored for polycrystalline Nb₃Sn. In the neighborhood of the critical temperature, Nb₃Sn is assumed to be a random mixture of superconducting and normal grains. Percolation concepts are applied to a study of the resistivity. A general analysis is made showing that the onset and shape of the resistive transition for composite conductors are determined by the percolation function and the distribution of the critical temperature. An approximate form of the percolation function is determined based on a linear FEM analysis. Example resistive transitions are calculated for an assumed normal distribution of the critical temperature. An argument is presented that relates grain orientation and strain dependence in Nb₃Sn. It is noted that a dependence of the distribution of T_c with strain, in addition to the usual shift in T_c with strain, would be the result of a strain dependence that is a function of grain orientation. The analysis shows the extent to which the slope of the resistive transition is a measure of the distribution of the critical temperature, and therefore a measure of the grain orientation strain sensitivity. Finally, a method is described to determine the percolation function experimentally.     

Liquid helium cryostat for SQUID-based MRI receivers

We describe a liquid helium cryostat, developed to cool SQUID-based receivers in low field MRI systems. The cryostat has a 4 L liquid helium capacity, a hold time of over 3 days and accommodates 10 cm diameter receiver coils. New vacuum insulation methods reduce the noise level by at least an order of magnitude compared to existing commercial designs. The minimum detectable field at 425 kHz, with a 5 cm diameter circular coil, was estimated to be 0.018 fT/Hz^1/2 from Q-factor measurements and 0.035 fT/Hz^1/2 by direct measurement with a SQUID amplifier. Further measurements indicated that most of this field noise probably originates with dielectric losses in the cryostat’s fibreglass shells.     

Gas gap thermal switches using neon or hydrogen and sorption pump

The very low pressure obtained thanks to adsorption phenomenon at low temperature can be used to build cryogenic heat switches, which offer the possibility to make or break thermal contact between two parts of a cryogenic system. The ON (conducting) and OFF (insulating) states of the switch are obtained by varying the gas pressure between two copper blocks separated by a gap of 100 μm. This pressure is controlled by acting upon the temperature of a small sorption pump (activated charcoal) connected to the gap space. For a “high” sorption pump temperature, the gas previously adsorbed in the sorption pump is released to the gap between the two blocks, allowing a good thermal conduction through the gas (ON state). On the opposite, cooling the sorption pump allows a very good vacuum between the copper blocks, which efficiently break the thermal contact (OFF state). Experimental thermal characteristics (Conductance in the ON and OFF state, ON-OFF switching temperature) of such a “Gas Gap Heat Switch” are described using hydrogen or neon as exchange gas and are compared with theoretical calculations.     

Development of No-Vent™ liquid hydrogen storage system for space applications

A number of technological advances required to store and maintain normal-boiling-point and densified cryogenic liquids, including liquid hydrogen, under zero boil-off conditions in-space, for long periods of time, have been developed. These technologies include (1) thermally optimized compact cryogen storage systems that reduce environmental heat leak to the lowest-temperature cryogen, which minimizes cryocooler size and input power, and (2) actively-cooled shields that surround the storage systems and intercept heat leak. The processes and tools used to develop these technologies are discussed. A zero boil-off liquid hydrogen storage system technology demonstrator for validating the actively-cooled shield technology is presented.