应用光谱技术快速测定发动机润滑油的粘度值

提出了一种应用可见近红外光谱技术快速测定发动机润滑油动力粘度值的新方法.对5种不同粘度的润滑油共150个样本进行光谱扣描,经过光谱预处理后应用偏最小二乘法(PLS)建立了润滑油动力粘度值的预测模型,并提取出前6个有效主成分作为最小二乘-支持向最机(LS-SVM)预测模型的输入变量,建市相应的最小二乘-支持向量机(LS-SVM)预测模型,采用径向基函数(RBF)作为核函数,超参数γ和RBF核函数参数σ2的最佧组合为)γ=27.321 2和σ2=3.229 5.用125个样本建模,25个样本验证.实验结果表明,LS-SVM模型比PLS模型能获得更满意的预测效果.说明应用光谱技术可以实现发动机润滑油动力粘度值的快速无损检测.     

量子遗传-神经网络算法的润滑油动力粘度值可见近红外光谱分析

润滑油动力粘度是划分润滑油品质的重要依据之一,高铁变速箱润滑油需要进行实时、快速、无损的检测,因此提出一种基于可见近红外光谱微型模块结合量子遗传-神经网络算法对润滑油粘度值进行定量分析的新方法。不仅实现了高铁变速箱润滑油动力粘度的无损快速实时检测,还进一步提高了对润滑油动力粘度预测的精度。微型光谱仪具有性能优良、体积小巧等优势,在便携式光谱无损检测方面用途越来越多。在这里,选用可见短波近红外和近红外波导光栅两种微型光谱模块进行光纤耦合,实现了330～1 700 nm可见-近红外波段光谱拼接。首先我们采用该组合微型光谱仪对13种不同粘度的润滑油共78个样本进行光谱扫描得到原始光谱数据。原始光谱经过Savitzky-Golay卷积平滑后,再一阶求导,可以有效地消除基线漂移和背景噪声。然后采用主成分分析和马氏距离相结合的方法来识别浓度界外样本,剔除界外样本3个。最后采用BP(back propagation)神经网络和量子遗传神经网络两种回归算法分别建立定量分析模型,并对比分析了两种算法的性能。量子遗传算法是量子计算和遗传算法相结合地一种概率进化算法,采用量子染色体的形式,利用量子逻辑门进行全局搜索,从而可以利用量子遗传算法优化神经网络地权重和阈值,提高建模效率和精度。分别用BP神经网络算法和量子遗传-神经网络算法进行建模仿真,从75个样本随机抽取10个样本作为预测集,其余65个为建模集。在量子遗传寻优算法中,其种群数目设置为40,终止代数为200,寻优结果表明该算法在训练81代后可快速得到最优解。比较两种建模算法的预测结果,采用量子遗传-神经网络算法相比BP神经网络算法得到的粘度预测结果均方根误差从0.345 5降低至0.029 4,决定系数从0.850 4升至0.979 9,可知量子遗传-神经网络算法的预测能力明显强于BP神经网络。为进一步提高润滑油粘度的定量分析准确度和微型光谱模块应用于高铁齿轮箱润滑油在线、实时、快速检测提供了参考方法。     

基于可见-近红外光谱技术的润滑油含水量无损检测方法研究

研究了基于可见-近红外光谱技术的发动机润滑油含水量快速检测方法。在获取光谱信息的基础上,提出了采用不同的光谱建模方法以提高检测精度和简化分析计算。分别采用主成分分析(PCA)和连续投影算法(SPA)方法进行模型输入变量的提取。SPA最终选择了476,483,544,925,933,938,952,970和974nm共9个波长为最优变量。基于SPA选择的变量,分别应用偏最小二乘回归(PLSR)和多元线性回归(MLR)建模。效果均优于全波段PLSR模型和PCA-PLSR模型。说明SPA选择的有效变量能够包含最重要的全波段光谱信息,同时可以去除无用的信息变量。为了进一步提高检测效果,采用LS-SVM分别基于SPA选择后的有效变量和全波段光谱进行建模。两个模型的预测确定系数(Rp2)均在0.9以上。SPA-LS-SVM的效果要优于全波段LS-SVM模型的效果。SPA-LS-SVM模型的Rp2达到了0.983,剩余预测偏差(RPD)值为6.963。表明可见-近红外光谱可以用于发动机润滑油含水量的检测。     

木材中的水分及其近红外光谱分析

水分是木材的重要特征之一,用近红外光谱分析木材性质,必须考虑水分对预测结果的影响。文章分析了不同含水率状态下木材在近红外光谱区的吸收特性;用近红外光谱法预测木材中的水分含量,相关系数为0.99,校正标准偏差SEC和预测标准偏差SEP分别是0.041和0.043;通过在不同含水率(7%,12%,20%,30%,60%)状态下采集近红外光谱与固定含水率下的木材密度建立关联,结果表明,利用近红外光谱技术可以在不同含水率下采集近红外光谱来预测木材气干密度。     

基于光谱技术和连续投影算法的润滑油品牌快速鉴别研究

为了实现润滑油品牌的快速无损检测,提出了一种基于可见/近红外光谱透射技术与连续投影算法相结合的润滑油品牌快速鉴别新方法。采用连续投影算法对6种润滑油的可见/近红外光谱数据进行波长变量的筛选,再结合偏最小二乘法建立润滑油品牌的鉴别模型。结果表明,鉴别模型的相关系数r为0.9721,预测均方根误差RMSEP为0.4055,鉴别正确率为91.7%。说明提出的连续投影算法结合偏最小二乘算法具有很好的预测效果。     

温度变化对近红外光谱预测木材含水率的影响

为实现温度不稳定环境下木材含水率的近红外光谱检测,探究了不同温度下木材近红外光谱的变化规律及温度变化对近红外预测木材含水率的影响。对从林场采集的樟子松、水曲柳、大青杨和红松原木木块试样各75块,共计300块试样,进行了不同温度和含水率条件下的近红外光谱采集。采用单一温度下的校正集分别与各个温度下的验证集建立偏最小二乘含水率预测模型,探究温度变化对木材含水率模型预测准确性的影响。比较了不同光谱预处理的木材含水率预测温度全局模型。采集相同含水率下不同温度的近红外光谱数据,对光谱进行光谱平均、一次微分、主成分分析和偏最小二乘判别分析,以探究温度变化时,木材近红外光谱的变化规律。结果显示：（1）温度对木材样品光谱存在显著影响;主成分分析和判别分析表明不同温度下的样品有明显聚类趋势,温度判别准确率为96.1%。温度会影响木材的近红外光谱在特定波长吸收峰的位置及吸光度,在含水率相同的情况下,随着温度的升高,特定位置吸收峰有逐渐向高频波段转移的趋势且在零下低温时波峰移动变化更明显。（2）不同温度下的PLS含水率预测模型对温度变动的适应能力有差异,木材含水率预测模型更适应于检测与建模样本相同温度的样品。与单一温度模型相比,PLS温度全局模型对于温度变化具有很好的适应性和应用潜力,RMSEP低于大部分单一温度模型。基于SG平滑+多元散射校正+一次微分预处理联用的PLS含水率温度全局模型有较好的预测效果和温度适应性,RMSEP降为0.074。可见,温度变动是近红外法检测木材含水率的过程中不可忽视的扰动因素;基于光谱预处理的温度全局模型可以显著提高温度适用性。该研究可进一步促进近红外光谱技术在木材生产、加工过程中的应用。     

基于可见-近红外光谱和多光谱成像技术的梨损伤检测研究

提出了利用可见-近红外光谱技术和多光谱成像技术检测鸭梨损伤随时间及程度变化的新方法.利用可见-近红外光谱技术,分别结合偏最小二乘(panial least squares,PLS)和最小二乘支持向量机(least squares-support vector machine,LS-SVM)方法对鸭梨受损程度和受损天数进行预测.结果表明,两种方法在鸭梨损伤后期对损伤程度的判别均具有较好的效果;LS-SVM方法对鸭梨轻度损伤的损伤天数的预测精度较高,但重度损伤天数的预测效果不如PLS方法.然后利用多光谱图像预测鸭梨受损天数.研究发现,利用LS-SVM建立的模型预测效果较稳定,预测结果相关系数均在0.85左右.说明利用可见-近红外光谱分析技术和多光谱成像技术能够快速无损地检测出鸭梨的损伤程度及时间,为鸭梨检测提供了一种新方法.     

可见／近红外漫反射光谱无损检测南丰蜜桔维生素C的研究

应用可见/近红外漫反射光谱对南丰蜜桔维生素C含量进行了无损检测研究.在谱区350～1 800nm,应用主成分分析和偏最小二乘法对经过预处理光谱进行数学建模,主成分数和光谱区间选择分别由完全交互验证和回归系数法确定.定标模型对10个未知样品的预测结果是:预测相关系数为0.813,预测均方差为2.112 mg(100 g)-1,预测偏差为-0.810 mg(100 g)-1.文章表明利用可见/近红外漫反射光谱技术无损检测南丰蜜桔维生素C具有可行性.     

水稻土可见-近红外-中红外光谱特性与有机质预测研究

土壤有机质是农田肥力评估的重要指标,要实现快速获取大面积土壤有机质的含量需要建立高效、稳健的预测模型。光谱技术能够快速诊断土壤有机质,以水稻土为例,从校正样本选择方法的对比,研究了可见-近红外、中红外和可见-近红外-中红外三种不同波段光谱对土壤有机质的预测能力。可见-近红外和中红外区域的光谱反射率转换成吸收率后通过Savitzky-Golay平滑法去噪,通过三种校正样本选择方法建立相应的偏最小二乘回归预测模型。通过Rank-KS法建立的三种波段的有机质预测模型均优于Rank法和KS法,中红外波段光谱的模型预测能力强于可见-近红外和可见-近红外-中红外波段的预测模型,基于Rank-KS法建立的中红外波段有机质预测模型取得了最好的预测效果,RMSE_P仅为3.25 g·kg-1,RPD达到4.24,依据VIP得分筛选出可见-近红外和中红外波段的水稻土有机质重要建模波段。因此,中红外光谱建模技术能够对水稻土有机质进行快速定量分析,Rank-KS法可提高模型的预测能力,为今后农田肥力评价和科学施肥提供技术支持。     

一种用于可见-近红外光谱特征波长选择的新方法

提出了一种基于模拟退火(SA)算法和最小二乘法支持向量机(LS-SVM)选择可见一近红外光谱特征波长的新方法(SA-LS-SVM).该方法用LS-SVM作为识别器,用识别率作为SA的目标函数,提取合适的特征波长数以及对应的特征波长.3种不同品牌的润滑油可见-近红外光谱的特征波长分别用SA_LS-SVM,主成分回归分析(PCA)和偏最小二乘法(PLS)进行处理,提取特征波长或主成分,然后结合反向传播人工神经网络(BP-ANN)对各种处理方法进行识别预测.结果发现,SA-LS-SVM只需从751个数据光谱中提取4个特征波长,就可以使三种品牌润滑油的识别率达到了100%,而其他所有的方法发现预测率都达不到100%,由此验证了SA_LS-SVM的优越性.实验结果表明,SA-LS-SVM不仅能有效地减少建模的变量数,而且可以提高预测精度.     

原子吸收光谱法在发动机异响检测中的应用

现有的人工经验听诊法和仪器辅助诊断法都难以快速诊断发动机异响故障原因,为此创新性地提出了利用原子吸收光谱法对汽车发动机异响故障进行检测。对行驶里程分别为2 000～70 000 km,取样间隔约为2 000 km,35批次的奥迪A6发动机润滑油中的主要金属Fe,Al,Cu,Cr,Pb含量利用原子吸收光谱法进行检测,建立同一车型发动机不同行驶里程数润滑油主要金属含量的数据库。研究结果表明主要金属含量随着汽车行驶里程数在某一范围内波动。在实际工程应用中,通过测定有异响发动机的润滑油中主要金属含量,将其与含量变化趋势图进行比对,在不解体状态下帮助检测维修人员确定发动机异响的类型和部位,降低汽车维修成本,提高了发动机异响故障诊断的准确性。     

A study of the influence of cooling water on the structural modes and vibro-acoustic characteristics of a gasoline engine

The influence of the cooling water on the vibration mode and damping characteristics of an engine was evaluated through experimental and simulation methods. The wet mode method of the fluid–structure interaction field was applied in numerical analysis, which led to several significant conclusions. Several additional fluid related modes were present when the effect of cooling water was considered. Meanwhile, the frequencies of other modes were almost the same. On the other hand, the damping characteristics of an engine were changed while the cooling water was used. For Rayleigh damping, the cooling water effect has made the mass matrix scale factor α nearly double and the stiffness matrix scale factor β almost unchanged. The engine surface vibration acceleration magnitude was calculated by means of the multi-body dynamic analysis. It was shown that the calculated vibration acceleration magnitude was close to the measured one when the cooling water was taken into account. Furthermore, acoustic simulation results were estimated by Boundary Element Method, which revealed that the sound power level of the radiation noise of most engine components was lower than that with the fluid structure interaction model being applied in the engine block modeling. The overall sound power level of the whole engine radiation noise was increased by 3.6 dB when the effect of the cooling water was considered, which demonstrated that the effect of the cooling water on the engine vibro-acoustic characteristics should not be neglected during the computation. This paper has disclosed the influences of the cooling water on the engine structural modes, damping, sound power level of the engine radiation noise and a systematic approach for evaluation of the influences.     

太赫兹光谱结合特征谱区筛选算法在发动机润滑油含水量定量分析中应用研究

发动机润滑油是保障汽车发动机持久且稳定运转的基石,准确评定发动机润滑油各项性能指标是其在生产到使用全过程必不可少的步骤。发动机润滑油在一段时间的使用后会因为多种原因引起油品变质,发动机润滑油变质的指标可以用其中非磁性颗粒物浓度、金属屑含量、pH值、粘稠度、含水率等表述。关于发动机润滑油含水量的检测,传统的检测方法存在操作复杂,及时性差等缺点。太赫兹对水吸收强烈,适合用于对样品中微水含量的分析。通过透射式太赫兹时域光谱系统获得1.0～3.5 THz下的六种不同水含量的发动机润滑油的吸收系数谱线,对谱线进行Savitzky-Golay（SG）平滑去噪,剔除奇异样本后,采用Kennard-Stone算法划分样品集,尝试常规区间偏最小二乘法（iPLS）、向后区间偏最小二乘法（BiPLS）和联合区间偏最小二乘法（SiPLS）对其太赫兹时域光谱特征谱区间进行筛选,着重研究区间间隔数、PLS组件数、最佳主因子数和区间选择等因素对PLS模型属性的影响,并且对不同含水量的润滑油建模分析,对不同模型比较选优,建立最优定量分析模型。建模结果表示特征谱区筛选可以提高建模性能、降低模型复杂性,特征谱区筛选算法通过剔除发动机润滑油太赫兹吸收系数谱线中非线性或者无关变量的方式,使建模结果更好的表达吸收系数谱线与其含水量的关系。结果表明：采用BiPLS模型用于发电机润滑油中微量水含量的定量分析时建模效果最佳,模型区间数为26,入选区间为[18 10 4 3 8 12 5 11 24 13 16 21 2],主因子数为10,最优模型的交互验证均方根误差RMSECV为0.003 5,预测均方根误差RMSEP为0.004 6,校正集相关系数r为0.919 3,预测集相关系数r为0.865 7。由此可见,可以采用反向区间偏最小二乘法（BiPLS）用于发动机润滑油水含量的测定,且实验过程简单,建模计算速度快,效果理想,可以适用于非接触式油品含水量的定量分析。     

The influence of water content, cooling and warming rate upon survival of embryonic axes of Poncirus trifoliata (L.)

The present study investigated the relative contributions of water content and non-equilibrium cooling and warming rates to the survival of cryopreserved axes of recalcitrant P. trifoliata seeds. Reducing water contents from 1.7 and 0.26 g water per g dry mass is believed to increase cytoplasmic viscosity. Cooling to -196 degree C was done at rates averaging between 0.17 and 1300 degree C per second, and warming at 600 or 1.35 degree C per second. Survival was assessed after 4 weeks in vitro. Rapid warming resulted in higher survival and normal development of axes at all water contents. The effects of cooling rate were dependent on the water content of axes. Cooling rates resulting in >70 percent normal development ranged between 0.17 and about 1300 degree C per second for axes at a water content of 0.26 g water per g dry mass narrowing with increasing hydration to an apparent optimum at about 686 degree C per second in axes at 0.8 g water per g dry mass At 1.7 g water per g dry mass, axes cooled at 0.17 degree C per second yielded nearly 40 percent normal development, whereas faster cooling was deleterious. Results are interpreted in the context of the effect of water content on cytoplasmic viscosity and the rate of intracellular ice formation. At low water contents, the high intracellular viscosity slows ice crystallization making survival independent of cooling rate. At higher water contents, the reduced viscosity requires faster cooling to prevent ice crystal damage. The ability to cool rapidly with increasing hydration is balanced with an increasing limitation to dissipate heat fast enough to prevent severe damage.     

Viscosity of interfacial water

The effective shear viscosity and frequency-dependent dynamic oscillatory shear spectra of water containing monovalent or divalent ions (ionic strength 25 mM), confined between mica crystals at 1-2 water molecules thickness, oscillated with twist angle with the period expected for the pseudohexagonal surface lattice. The effective viscosity varied by orders of magnitude as the twist angle was changed. Confinement appeared to imprint lateral spatial correlation on the ultrathin liquid, the more so the better the confining lattices were aligned, but the oft-proposed "ice structure" was not observed dynamically.     

Physicochemical properties of ionic liquid mixtures containing choline chloride,chromium (III) chloride and water: effects of temperature and water content

The effects of water addition and temperature on some physicochemical properties of room temperature ionic liquids containing chromium chloride, choline chloride and water in the molar ratio of 1:2.5:x (where x = 6, 9, 12, 15 or 18) have been studied. The density, viscosity, surface tension and conductivity of the liquid mixtures were measured for the temperature range of 25 to 80 °C. Increasing both water content and temperature resulted in decreasing density, surface tension and viscosity and increasing electrical conductivity. The average void radii (hole sizes) for the liquid systems under study were calculated; they were in the range of 1.21 to 1.82 Å. The average hole size was stated to grow with increasing both temperature and water content in the mixture. The variation of the average void radii correlates with the change in viscosity and conductivity. The activation energies of viscous flow and conductivity diminishes with increasing water content in the liquid mixture. There is a strong linear correlation between conductivity and fluidity which indicates that the conductivity of the ionic liquid mixtures is generally controlled by the ionic mobility. A moderate viscosity and higher conductivity of the Cr(III)-containing ionic liquids with extra-water addition (at x > 9) make them suitable for the development of chromium electrodeposition processes.     

Properties of sugar-based low-melting mixtures

Physico-chemical properties of ternary sugar-based low-melting mixtures were determined. Choline chloride, urea and glucose or sorbitol, serving as sugars, were blended in various compositions. The refractive index, density, viscosity, decomposition temperatures and glass transition temperatures were measured. Further, the influence of temperature and water content was investigated. The results show that the mixtures are liquid below room temperature and the viscosity and density are dependent on the temperature and composition. Moreover, the viscosity decreases with increasing water content. These mixtures are biodegradable, low toxic, non-volatile, non-reactive with water and can be accomplished with low-cost materials. In consideration of these advantages and a melting point below room temperature, these low-melting mixtures can be a good alternative to ionic liquids as well as environmentally unfriendly and toxic solvents.     

含水量对TiO_2纳米管形貌控制参数的影响

以不同含水量的乙二醇溶液为电解液,采用阳极氧化法制备TiO2纳米管阵列。通过记录反应过程中电导率、粘度及回路电流随时间的变化曲线,研究含水量对电解液粘度、电导率及电流等过程参数的影响,分析了纳米管形貌尺寸与TiO2溶蚀所耗电荷量的关系。粘度初始值和初始电导率均与含水量呈三次关系,相关系数分别为0.992 5和0.977 8。在反应过程中,溶液粘度值有缓慢增加的趋势。由于不同含水量的电解液粘度的不同,H+和OH-数量不同,F-迁移速率不同,电导率-时间曲线及电流-时间曲线具有不同的变化趋势,并对其进行了理论分析。当水体积分数为4%,5%,6%和10%时,纳米管的形貌较为有序并且TiO2纳米管阵列表面的碎片较少,纳米管直径变化范围为50nm至72nm,长度变化范围为0.85~1.90μm。F-腐蚀氧化膜时所消耗电量与TiO2氧化膜被腐蚀掉的体积呈一次函数关系,即腐蚀电量越大,腐蚀掉的体积越大,为制备一定形貌尺寸的纳米管提供了一定的控制方法。