基于混合时间序列卷积神经网络的轴承故障诊断

在传统滚动轴承故障诊断中,绝大多数方法采用了从振动信号提取特征的诊断模式,但是这种模式必然会使原始信号降维进而导致故障信息的丢失。卷积神经网络(CNN)通过权重共享和稀疏连接直接对原始信号进行操作,实现自适应特征提取,最大化保留故障信息。受CNN原理启发,开发出了一种基于工业振动信号特征的新型诊断框架,称之为混合时间序列CNN(HTS-CNN)。首先,利用估计总体比例的方法自适应确定模型训练样本数目;其次,通过对时间序列片段进行随机组合的方式,使模型能够提取非相邻信号特征;最后,利用Softmax激活函数在模型输出端执行多分类任务。通过对凯斯西储大学及CUT-2平台轴承数据进行分析,实验结果表明:该方法能够准确、有效的对滚动轴承故障进行分类。     

Peanut FAD2 Genotype and Growing Location Interactions Significantly Affect the Level of Oleic Acid in Seeds

The level of oleic acid is an important parameter in determining seed nutritional quality and oil stability. The level of oleic acid in peanut is genetically controlled by a pair of fatty acid desaturase genes (FAD2A and FAD2B), but the environmental conditions of the production sites can also have a significant effect. To investigate the effect of gene and environment interaction, 45 accessions were grown at three locations for 2 years. Environmental data were collected; individual plants were genotyped with functional SNP markers from FAD2A and FAD2B; and seed level of oleic acid was determined by gas chromatography. Three FAD2A/FAD2B genotypes (448G/no insertion 442A, 448A/no insertion 442A, and 448A/insertion 442A) were identified and designated as G/W, A/W, and A/A, respectively. A/A genotype averaged the highest level of oleic acid (80.0%), followed by A/W (56.0%), and then G/W (40.7%). Analysis of gene and environment interaction revealed that oleic acid phenotype plasticity could be explained by the interaction of FAD2 genotype and photothermal time, which quantified environmental conditions. The A/W genotype was the most sensitive to photothermal time changes. The oleic acid plasticity revealed in this study would be useful for breeders, farmers, and product processors.     

便携式飞行时间质谱仪的研制及其在石化污水系统挥发性有机物监测中的应用

近年来，国内的大气污染问题越来越严重，挥发性有机物（VOCs）的排放是其中一个重要原因。VOCs是臭氧和气溶胶的前驱物之一，在大气化学反应过程中扮演着极其重要的角色。本实验自行研制了一台基于真空紫外（VUV）灯的高分辨光电离飞行时间质谱仪（TOF MS），建立了一种新的VOCs稀释采样方法，并将其应用于石化行业污水处理系统VOCs的采样与分析。结果表明：隔油系统的3个池子基本没有测到高浓度的挥发性有机物，其中溶气气浮甲苯物质的质量浓度最高，为50.41 mg/L，其余组分的质量浓度大多数在0.1~1.0 mg/L之间；1#加氢和2#加氢隔油池有着类似的组分浓度，以1#加氢隔油池为例，二甲苯物质的质量浓度为10.60 mg/L，乙苯为52.33 mg/L，苯为59.80 mg/L；由于罐区隔油池没有加盖，处于露天状态，挥发性有机物气体容易向大气中扩散，并没有检测出较高浓度组分。可以看出，新式采样方法可有效地稀释定量样品，操作简便，减少了样品因闪蒸过程造成的VOCs损失，检出物质更丰富；同时，使用自行研制的TOF MS仪器能够快速检测不同装置污水处理系统的VOCs种类，将该仪器用于实际样品检测，发现化工企业大多数密闭的池子都含有苯系物、醚类等，但不同工艺的污水处理系统VOCs存在明显差异。该质谱仪适用于石化企业污水系统VOCs监测。     

Validation of Models Predicting Nucleation Rates from Induction Times and Metastable Zone Widths

Three approaches for estimation of nucleation rates from induction time and metastable zone width (MSZW) were validated based on directly measured nucleation rates for paracetamol in ethanol. To quantitatively predict nucleation kinetics using Kubota's methods it is necessary to know the minimum detectable number concentration of nuclei. This was found by determination of light transmission of a series of diluted suspensions of newly nucleated crystals where the size had already been assessed by optical reflection measurement (ORM). The measured nucleation rates strongly depended on both temperature and supersaturation. The Nyvlt method predicted nucleation rates in this system reasonably well; however, it gave slightly low estimates for all temperatures. The methods of Kubota provided nucleation rates that were low by an order of magnitude.     

Real‐time monitoring by proton relaxometry of radical polymerization reactions of acrylamide in aqueous solution

The potential of time‐domain nuclear magnetic resonance (TD‐NMR) for the real‐time monitoring of solution radical polymerizations is demonstrated. A model system composed of a redox‐pair initiator system, acrylamide as monomer and water as solvent was investigated. A second‐generation continuous wave free precession technique was employed to measure the longitudinal relaxation time constant (T₁) of the samples throughout the polymerization reactions. This parameter was shown to be sensitive to the reactant feed free‐radical enhancement of the water molecule relaxation time, making it a good probe to monitor monomer conversion in real time in an automated, non‐destructive fashion. It was found that the T₁ value was better than the transverse relaxation time constant (T₂) for describing the evolution of the polymerization reactions, due to its greater sensitivity to paramagnetic effects. The TD‐NMR signal variation observed was linked to the formation, propagation and termination steps of the radical polymerization kinetics scheme. These first results may contribute to the application of real‐time monitoring of radical polymerization reactions employing low‐cost and robust TD‐NMR spectrometers. © 2018 Society of Chemical Industry     

Functional Causality between Oil Prices and GDP Based on Big Data

This paper examines the causal relationship between oil prices and the Gross Domestic Product (GDP) in the Kingdom of Saudi Arabia. The study is carried out by a data set collected quarterly, by Saudi Arabian Monetary Authority, over a period from 1974 to 2016. We seek how a change in real crude oil price affects the GDP of KSA. Based on a new technique, we treat this data in its continuous path. Precisely, we analyze the causality between these two variables, i.e., oil prices and GDP, by using their yearly curves observed in the four quarters of each year. We discuss the causality in the sense of Granger, which requires the stationarity of the data. Thus, in the first Step, we test the stationarity by using the Monte Carlo test of a functional time series stationarity. Our main goal is treated in the second step, where we use the functional causality idea to model the co-variability between these variables. We show that the two series are not integrated; there is one causality between these two variables. All the statistical analyzes were performed using R software.     

基于LSTM-RF模型的城市道路交通流预测

交通流预测作为信号协调和出行时间预测等任务的基础,成为了交通领域的研究点。对于交通流预测问题,研究人员提出了多种方法,但这些方法大多只使用交通流数据的时域信息进行交通流预测,忽略了空间相关性对于预测目标路段流的影响,导致预测精度不理想。基于组合模型的思想提出了一种称为LSTM-RF的交通流预测模型。在交通流预测过程中,首先使用LSTM模型提取预测目标路段的时序特征,再将其预测值与采集的相邻上下游路段信息同时作为随机森林模型的输入特征,进行交通流时空相关性分析,获得最终的预测结果。并通过贵阳市车牌识别系统采集的城区132条路段的交通流数据进行实验验证。结果表明:该方法在预测精度上优于单一模型,并且预测误差相比单一模型有明显减少。     

Nb和时效时间对热轧中锰TRIP钢中P偏聚的影响

热轧态中锰TRIP钢首先经650 ℃退火2 h,随后在550 ℃进行等温时效热处理,采用场发射扫描电镜(FE-SEM)研究该钢中P的偏聚和时效析出行为的变化情况。结果表明,中锰TRIP钢中P在晶界的偏聚是一种非平衡偏聚现象,临界时间约为50 h,与理论计算结果48 h较为吻合。在局部偏聚区域内,C与P存在共偏聚的关系,即P偏聚量高的地方,C含量也高。而合金元素Nb具有抑制P偏聚的效果,在20~70 h时效时间内,可以相对降低6.57%~19.5%的最大P偏聚量。根据电子背散射衍射(EBSD)菊池线分析,P偏聚量低于2.28at%时,P为固溶状态,高于2.28at.%时,P为析出状态。     

变厚度板中水平剪切波的传播特性分析

为研究超声导波在储罐壁板等变壁厚结构中的传播特性，建立了SH波在空间自由边界中的传播模型，基于各向同性弹性介质Navier-Stokes方程，推导了变厚度板中SH波的频散方程，分析了变厚度板中水平剪切波走时t与倾斜角θ间的相互作用关系；通过简谐点源的远场响应并引入板厚与传播位移的非线性关系，得出了变厚度板中SH波的远场响应方程。在此基础上，仿真分析和实验验证了水平剪切波在不同倾斜角变厚度板中的传播特性，结果表明：在变厚度板的薄端激励水平剪切波并沿结构表面传播时，无频散现象，且无波包分离现象；但随着变厚度板倾斜角θ增大，水平剪切波的走时和接收信号幅值均逐渐递减，且呈非线性关系。