一种K值自适应和局部搜索的KNN矩阵修复方法

数据在采集和传输过程中由于多种原因会造成矩阵残缺,因此在数据分析之前需要对残缺矩阵进行修复。常见的KNN修复方法 k值选取不合理,且需在整个矩阵中搜索近邻,影响算法的修复效果。在其基础上提出了一种k值自适应的局部KNN矩阵修复方法,合理考虑了k值的选取和近邻项的搜索范围。实验证明了该方法能有效提高矩阵修复的正确率,且算法的时效性有所提高。     

K值优化的VMD在轴承故障诊断中的应用

由于机械设备实际运行状态下环境噪声的影响,轴承早期非平稳振动信号的故障特征难以有效提取。为此,将K值优化的变分模态分解引入轴承的早期故障诊断方法中。首先利用小波包降噪法对轴承实际振动信号进行降噪;然后利用K值优化的VMD算法,通过合理设置参数K,将降噪信号分解为若干本征模态分量,利用峭度值选取最佳分量;最后提取最佳分量的样本熵和排列熵组成特征向量,利用模糊C聚类识别轴承的故障类型。实验结果表明,该方法避免了信号的过分解,能有效提取振动信号特征,实现轴承的早期故障诊断。     

Application of Artificial Neural Network to Predict K-Value,Inosine Mono-Phosphate,and Hypoxanthine Concentrations of Grass Carp (Ctenopharyngodon idellus) Fillets During Storage

K-value, inosine mono-phosphate, and hypoxanthine concentrations of grass carp (Ctenopharyngodon idellus) fillets were determined during storage at 273, 277, 281, 288, and 293 K. Simultaneously, a feed-forward artificial neural network was developed to predict these changes in grass carp fillets during storage, and a comparative study on K-value prediction between the artificial neural network and Arrhenius model was also performed. The results showed that the K-value and hypoxanthine concentrations increased with storage time, while inosine mono-phosphate reached a peak and then decreased with time. The artificial neural network was successful in predicting changes in the K-value, inosine mono-phosphate, and hypoxanthine concentrations throughout storage, and it was even more effective in predicting K-value with lower relative errors than the Arrhenius model. The high regression coefficient (R²) and low mean squared error indicated that the artificial neural network could be a potential tool in modeling changes in K-value, inosine mono-phosphate, and hypoxanthine concentrations of grass carp fillets within 273–293 K.     

自蔓延高温合成Ti3AlC2陶瓷材料

以Ti、Al、C、TiC粉末为原料,研究掺杂Si及Al含量对自蔓延高温合成Ti3AlC2的影响,合成材料的X射线衍射仪(XRD)及扫描电子显微镜(SEM)分析结果表明:物质的量比n(Ti)∶n(Al)∶n(C)∶n(TiC)∶n(Si)=2∶1.2∶1∶0.9∶0.1的原始混合粉末,经50 MPa压力压制的压坯在空气中...     

锂离子电池自放电的研究

采用三电极的方法分别通过分析锂离子电池正负极的电位变化来研究电池的自放电.锂离子电池存储过程中的自放电的影响因素主要有：存储温度和存储时电池的充电状态,温度越低,电池的充电状态越低（但电压范围有限制：3.8～4.2V）自放电越小.电池充放电过程中的自放电主要是由两部分组成,一是电池内部的副反应;二是内部微短路.内部微短路是由极片表面的颗粒及阴极极片边缘的金属毛刺所引起的电池内部微短路造成的.减小这些影响因素,可以降低电池的自放电.     

Fabrication of monolithic bulk Ti3AlC2 and impurity measurement by K-value method

Polycrystalline bulk Ti3AlC2 material with high purity and density was fabricated by hot pressing from the powder mixture with the starting stoichiometric mole ratios of 2.0TiC/1.0Ti/1.1Al/0.1Si at 1 300-1 500 ℃. X-ray diffraction patterns and scanning electron microscopy photographs of the fully dense samples indicate that the proper addition of silicon is favorable to the formation of Ti3AlC2, consequently results in high purity of the prepared samples. The Ti3AlC2 hot pressed at 1 300 ℃ and 1 400 ℃ is in plane-shape with sizes of 6-8 μm and 15-20 μm in the elongated dimension, respectively. The purities of samples are measured by the K-value method, and the contents of TiC are given by a linear equation.     

An improved conformal solution method based on hard sphere expansion to predict vapor-liquid equilibria

The conformal solution method based on the hard sphere expansion (HSE) theory predicts thermodynamic properties of mixtures by separating the mixture properties into a contribution from repulsion and other contributions from various types of intermolecular attraction. The original HSE, however, has a tendency to show increasing relative error in the prediction of thermodynamics properties as the difference of molecular sizes in mixtures increases since the radial distribution functions of a mixture are represented by that of a pure reference in the mean density approximation (MDA). When the hard convex body (HCB) equation of state was substituted for the hard sphere equation of state in the repulsion term to overcome the shortcoming of the original HSE, better results on the K-value in binary hydrocarbon mixtures were obtained.     

焙烧炉电收尘物料中Al(OH)3及α-Al2O3含量的分析

为了精确快速分析焙烧炉电收尘物料的物相组成,采用X射线衍射定量分析方法及焙烧-灼减法,对其中的Al(OH)3和a-Al2O3进行了测试分析.结果表明,采用焙烧法可以克服因采用掺入内标物质而进行混样的困难,以及因混样不匀而造成的分析误差,测试结果与配制的含量基本吻合,测试过程简单且实用.