Composition of sandstone and heavy minerals implies the provenance of Kuqa Depression in Jurassic, Tarim basin, China

The petrological characteristics and heavy mineral assembly of the Jurassic System in Kuqa Basin are consistent with the tectonic evolution, which is the major factor controlling the filling process of Kuqa Basin. Controlled by the source areas at the both sides of the Basin, the matrixes are dominated by sedimentary rock, high-grade metamorphic rock, acid magmatic rock, and the highest heavy mineral assemblage of zircon-garnet-magnetite in the content. According to heavy mineral indexes, the Jurassic strata can be classified into three rhythmic units of heavy mineral. The lower sequence contains abundant heavy minerals and relatively high content of detritus and its matrixes primarily came form the recycling orogen. The assemblage of “zircon-garnet-magnetite” indicates that the sedimentary area in the early evolution period of the basin is relatively near the provenance. The middle sequence is different from the lower sequence in the heavy mineral assemblage and has the different characteristics from that of source rock. The heavy mineral assemblage is “garnet-zircon” with the highest content of garnet and its main matrices are the high-grade metamorphic rock, acid magmatic rock and hydrothermal vein at the northeastern edge of the basin. The middle and upper sequences are characterized by the clast provenance, which is primarily from the recycling orogen at the northern edge of the basin and secondarily from the southern uplift. As a result of the two provenances, the strata with apparent changes of heavy mineral assemblage and index better coincides with the interfaces of the secondary strata sequences, which reflects the denudation history and the filling process of the orogen at the northern edge of the basin.     

基于梯度加权类激活热力图的卷积神经网络故障诊断模型鲁棒性分析

深度学习近年来在故障诊断领域受到广泛应用,但基于深度学习的故障诊断模型缺乏工程上的物理解释性,难以保证其故障诊断结果的稳定性。以轴承为例,建立了以小波时频图像为故障诊断依据的卷积神经网络模型(convolutional neural network, CNN),提出了一种基于梯度加权类激活热力图(gradient-weighted class activation map, Grad-CAM)的网络模型鲁棒性分析方法,并利用美国凯斯西储大学(Case Western Reserve University, CWRU)轴承数据集进行验证。首先,将故障直径轴承数据以不同方式混合并训练大、小多个模型。其次,利用Grad-CAM方法,建立时频区域与故障模式之间的联系。最后,利用其他工况下的轴承故障数据,以及含噪数据进行测试,并根据结果结合模型最注重的时频区域进行分析。结果表明,基于深度学习的轴承故障诊断模型在参数较少时更加注重低频区域,并能使其具有更好的鲁棒性。     

横向纤维搭桥下矩形脱层热屈曲数值模拟

基于板的小挠度理论以及热屈曲分析的ANSYS有限元法,考虑横向纤维搭桥的作用,对矩形薄板的脱层热屈曲进行数值模拟.有限元分析结果表明,随着纤维搭桥刚度的提高,热屈曲的临界屈曲温度随之增大,脱层屈曲区域逐渐减小.通过数值模拟的结果,能为工程结构中热屈曲问题提供参考.     

瞬变应答法研究丁烯氧化脱氢的反应机理

在铁系催化剂上研究了文题的反应机理。通过瞬变应答实验和对响应曲线的分析,证实丁烯、丁二烯和氮在催化剂上发生竞争吸附,其反应机理是吸附氧和吸附丁烯间进行的Langmuir-Hinshelwood型机理,反应的控制步骤是氧吸附控制。瞬变应答实验也揭示出催化剂表面状态存在滞后特性,反应速率和选择性与预处理条件有关。     

两个Toeplitz矩阵(或Hankel矩阵)相乘的快速算法

本文给出了两个n阶Toeplitz矩阵(或Hankcl矩阵)相乘以及Toeplitz矩阵与Hankel矩阵相乘的快速算法,这些算法的计算复杂性都为6n~2+O(nlog_2n)。     

高温投下法测定硼泥比热容的研究

用高温投下法,通过控制炉内气氛,在铜卡量热计中测定了硼泥及其分解产物的比热容,其值为硼泥 Cp=-1.608+3.279×10~(-3)T+2.847×10~5T~(-2),/(g·x) 硼泥分解产物 CP=0.461+0.633×10~(-3)T+0.510×10~5T~(-2),J/(g·K)其复相关系数分别为0.9920和0.9996.     

苯并三唑类光稳定剂的高效液相色谱分析

采用重叠分辨率图法和多目标逐步优化设计对反相液相色谱分离苯并三唑类紫外光稳定剂的流动相进行了优化,提出了测定了Tinuvin 328及其中间体氮氧化物的HPLC方法。     

辛醇对TritonX-100溶液胶束型态转化的影响

本文用流变学方法测定含辛醇的非离子型表面活性剂Triton X—100溶液的标准摩尔潘化吉氏自由能△_r~≠G_m~■、标准摩尔活化熵△_r~≠S_m~■、标准摩尔活化焓△_r~≠H_m~■,研究辛醇对Triton X—100溶液球型胶束向柱型胶束转化及非球型化轴比a/b的影响,同时探讨伯辛醇、仲辛醇、叔辛醇的影响规律。     

混合表面活性剂存在下盐酸羟胺-硫酸高铁铵法分光光度测定药品制剂中的青霉素

本文报导了青霉素在碱性条件下于72℃恒温水浴中与羟胺形成羟肟酸盐,而后酸化生成羟肟酸,利用羟肟酸在混合表面活性剂吐温——CTMAB等体积混合液的存在下与Fe(Ⅲ)形成稳定的紫红色络合物,从而借此显色反应作分光光度来测定药品制剂中青霉素的含量。紫红色络合物的最大吸收波长位于480nm处,摩尔吸光系数为8.7×10~3l·mol~(-1)·cm~(-1)。青霉素含量介于0.42～2.11mg/25ml范围内确准地遵守比尔定律。紫红色络合物的色泽能稳定4小时以上,方法灵敏、简便,准确度与精密度均较好。     

非晶态Fe_(13.3)Ni_(69.6)Si_(0.9)B_(16.2)合金的Cross-Over效应

本文研究了非晶态Fe_(13.3)Ni_(69.6)Si_(0.9)B_(16.2)合金的居里温度Tc的Cross-Over现象。在宏观态的总动力学是由具有不同驰豫时间常数的微观态动力学的变化的叠加的假定前题下,可用来说明该现象的变化行为。由实验数据得出v_0=10~(12)s~(-1)是合理的参数值,并且Cross-Over效应与微观态的初始值紧密相关。