基于随机森林的硬盘故障率预测研究

为了避免硬盘出现故障而造成大量数据丢失,文章提出一种基于随机森林的方法对硬盘的故障进行预测,降低其丢失数据的风险。首先,在数据预处理方面,对所采用的数据做特征映射预处理;其次,通过对决策树进行构建及选取等,构建随机森林预测模型,根据所选取的特征属性预测硬盘故障率所在的区间,并且特征属性的变化能反映出硬盘故障率的变化趋势;最后,对构建的随机森林模型参数进行调优,选取不同的n_estimators参数值进行测试和优化。实验结果表明,与XGBoost(Extreme Gradient Boosting)、LSTM(Long Short-Term Memory)等方法相比,本文方法的F1值(F-Measure)分别提高了0.93%和1.84%,并且对随机森林预测模型的参数值进行不同取值测试,最终准确率达到98.18%,比默认值提高了1.23%,证明该方法能更精确地预测硬盘故障率,反映出硬盘故障率基于特征属性的变化趋势。

黄河水沙监测数据分析模型

为研究黄河小浪底水库下游某水文站水沙通量的变化规律和“调水调沙”效果,以2016-2022年水位、水流量与含沙量的实际监测数据为依据,估算出年均径流量325亿立方米,年均排沙量2亿吨.分别使用双对数回归分析、M-K检验、小波分析、R/S分析对水沙通量的敏感性、突变性、周期性、趋势性进行了分析,得出结论:水位每增加1%,含沙量就增加37.63%,含沙量关于水位非常敏感;水沙通量均在2018年3月发生了突变;水沙通量均以18月为主周期振荡;水通量、沙通量在未来2年分别是增长趋势和下降趋势.使用傅里叶级数叠加线性函数预测未来2年的水沙通量,并建立了0-1规划模型对今后2年的监测方案进行了优化.通过比较“调水调沙”前后水沙通量、河底高程的变化,证明了“调水调沙”效果非常显著.

虚拟筛选和鉴定EGFRL858R/T790M非共价抑制剂

目的:基于计算机虚拟筛选和生物活性评价鉴定新的治疗非小细胞肺癌的EGFR酪氨酸激酶EGFRL858R/T790M非共价抑制剂。方法:基于受体-配体复合物结构(PDB ID:5HCZ)构建药效团模型,选择ROC值为0.849的药效团模型对Specs化合物数据库进行筛选,再经过Libdock对接程序进一步过滤,最终选出20个化合物进行EGFRL858R/T790M酶活测试。结果:当化合物浓度为1μmol·L-1时,有7个化合物能够使EGFRL858R/T790M酶活降至52%以下,且对酶的抑制活性高于对照物吉非替尼[(52.77±0.39)%];当浓度为0.1μmol·L^(-1)时,化合物AN-465/42889633[(65.99±2.94)%]、AG-690/40753617[(68.29±6.18)%]和AN-465/43336751[(78.61±7.23)%]对EGFRL858R/T790M的抑制作用仍优于对照物吉非替尼[(80.42±0.62)%];分子对接结果表明3个分子通过氢键、疏水作用力及静电作用力与EGFRL858R/T790M活性部位相结合。结论:AN-465/42889633、AG-690/40753617和AN-465/43336751可作为非共价EGFRL858R/T790M抑制剂的先导化合物进行研究,同时为新型EG-FR抑制剂的研发奠定基础。

防晒新视野:蓝光时代的肌肤守护

随着数字时代的来临,防晒产品的关注焦点逐渐扩展到抗蓝光领域。本文深入探讨了蓝光对肌肤影响,以及防晒产品在应对这一挑战时的创新应用。通过使用酶标仪测量防晒乳在380~445 nm的OD值得到它们在这一波长内的透光率,研究了亚甲基双苯并三唑基四甲基丁基苯酚(Tinosorb^((R))M)、非纳米二氧化钛对蓝光的防护效果,结果表明:两种原料均具有蓝光防护效果,其中含有Tinosorb?M的防晒乳液对蓝光的防护表现略优。这一综合保护策略为肌肤提供了更全面的防御,使防晒产品在蓝光时代发挥着新的、更为关键的角色。

股票错误定价、盈利不确定性与企业投资

随着改革的深化,资本市场与实体经济的联系愈发密切,然而失真的价格信号也为资本无序扩张埋下隐患。股价的变动既与未来投资机会有关,也可能源自错误定价。选用2000—2020年我国A股上市公司的数据,将衡量股票错误定价的新旧指标进行对比分析,重新厘清错误定价对公司投资决策的影响。结果显示,当股价被高估时,企业研发投入显著增加,而日常资本支出、并购投资却与股价高估不相关。而且,研发投入增加的80%是通过迎合渠道产生的,且在非国有企业、高换手率的企业中更加显著。由于我国并购大多签署了利润对赌协议,未来盈利不确定性相对较小,因此,资本市场的高估值对企业投资的影响与盈利不确定性密切相关,主要增加了盈利不确定性大的项目投资。这一结论对科学认识我国资本市场环境下资产价格波动与实体企业投资的关系具有重要的参考价值。

JEFFAMINE M-600(R)改性六方氮化硼及其水性环氧复合涂层的腐蚀防护性能

目的提高六方氮化硼的分散性和环氧涂层的腐蚀防护性能。方法利用O-(2-氨丙基)-O'-(2-甲氧基乙基)聚丙二醇(JEFFAMINE M-600(R))对六方氮化硼(h-BN)进行功能化,实现其在水性环氧树脂中的稳定分散和相容性,并将其添加至水性环氧树脂中制备了具有优异阻隔性能的水性环氧复合涂层。利用SEM、TEM、XPS等手段证实了h-BN@JEFFAMINE M-600(R)的成功制备;此外,采用电化学手段对涂层的腐蚀防护性能进行了评价。结果JEFFAMINE M-600(R)成功修饰在h-BN表面,有效提高了h-BN在水性环氧树脂中的分散稳定性和相容性;同时所制备的h-BN@JEFFAMINE M-600(R)/水性环氧复合涂层在浸泡38 d后仍表现出优异的抗腐蚀性能,其阻抗模值(1.05×10^(9)Ω·cm^(2))较水性环氧涂层(1.31×10^(7)Ω·cm^(2))提高了约2个数量级,这主要归因于h-BN@JEFFAMINE M-600(R)的添加可以有效提高水性环氧涂层的阻隔效应,从而延缓腐蚀介质向基底的扩散,进而表现出优异的腐蚀防护性能。结论JEFFAMINE M-600(R)功能化有效提高了h-BN的剥离效率和稳定分散,使得h-BN@JEFFAMINE M-600(R)纳米杂化材料的添加可以有效提高水性环氧基体的致密性和阻隔效应,进而增强复合涂层对金属基底的腐蚀防护。

^(99)Tc^(m)-DTPA-CGRRAGGSC在非小细胞肺癌裸鼠模型的显像研究

目的:探讨放射性核素^(99)Tc^(m)标记环九肽(CGRRAGGSC)对荷非小细胞肺癌裸鼠的显像研究。方法:用^(99)Tc^(m)标记DTPA-CGRRAGGSC,测定其标记率及稳定性,制备荷非小细胞肺癌小鼠模型,经荷瘤鼠尾静脉注射标记物2 h后进行SPECT静态显像,并计算出靶与非靶(T/NT)比值。结果:成功制备出^(99)Tc^(m)-DTPA-CGRRAGGSC,在PBS和人血清中37℃放置6 h时放射性化学纯度在90%以上;经荷瘤鼠尾静脉注射标记物2 h后T/NT比值最高达3.5。结论:^(99)Tc^(m)标记DTPA-CGRRAGGSC制备简便,其对荷非小细胞肺癌裸鼠具有较高靶向性,^(99)Tc^(m)-DTPA-CGRRAGGSC有可能在IL-11Rα高表达的非小细胞肺癌分子显像中发挥重要作用。

Rational and Continuous Measurement of the Emotional Decision Making in Visual Recognition of Facial Emotional Expressions with M.A.R.I.E.: First Half

Context: The advent of Artificial Intelligence (AI) requires modeling prior to its implementation in algorithms for most human skills. This observation requires us to have a detailed and precise understanding of the interfaces of verbal and emotional communications. The progress of AI is significant on the verbal level but modest in terms of the recognition of facial emotions even if this functionality is one of the oldest in humans and is omnipresent in our daily lives. Dysfunction in the ability for facial emotional expressions is present in many brain pathologies encountered by psychiatrists, neurologists, psychotherapists, mental health professionals including social workers. It cannot be objectively verified and measured due to a lack of reliable tools that are valid and consistently sensitive. Indeed, the articles in the scientific literature dealing with Visual-Facial-Emotions-Recognition (ViFaEmRe), suffer from the absence of 1) consensual and rational tools for continuous quantified measurement, 2) operational concepts. We have invented a software that can use computer-morphing attempting to respond to these two obstacles. It is identified as the Method of Analysis and Research of the Integration of Emotions (M.A.R.I.E.). Our primary goal is to use M.A.R.I.E. to understand the physiology of ViFaEmRe in normal healthy subjects by standardizing the measurements. Then, it will allow us to focus on subjects manifesting abnormalities in this ability. Our second goal is to make our contribution to the progress of AI hoping to add the dimension of recognition of facial emotional expressions. Objective: To study: 1) categorical vs dimensional aspects of recognition of ViFaEmRe, 2) universality vs idiosyncrasy, 3) immediate vs ambivalent Emotional-Decision-Making, 4) the Emotional-Fingerprint of a face and 5) creation of population references data. Methods: With M.A.R.I.E. enable a rational quantified measurement of Emotional-Visual-Acuity (EVA) of 1) a) an individual observer, b) in a population aged 20 to 70 years old, 2) measure the range and intensity of expressed emotions by 3 Face-Tests, 3) quantify the performance of a sample of 204 observers with hyper normal measures of cognition, “thymia,” (ibid. defined elsewhere) and low levels of anxiety 4) analysis of the 6 primary emotions. Results: We have individualized the following continuous parameters: 1) “Emotional-Visual-Acuity”, 2) “Visual-Emotional-Feeling”, 3) “Emotional-Quotient”, 4) “Emotional-Deci-sion-Making”, 5) “Emotional-Decision-Making Graph” or “Individual-Gun-Trigger”6) “Emotional-Fingerprint” or “Key-graph”, 7) “Emotional-Finger-print-Graph”, 8) detecting “misunderstanding” and 9) detecting “error”. This allowed us a taxonomy with coding of the face-emotion pair. Each face has specific measurements and graphics. The EVA improves from ages of 20 to 55 years, then decreases. It does not depend on the sex of the observer, nor the face studied. In addition, 1% of people endowed with normal intelligence do not recognize emotions. The categorical dimension is a variable for everyone. The range and intensity of ViFaEmRe is idiosyncratic and not universally uniform. The recognition of emotions is purely categorical for a single individual. It is dimensional for a population sample. Conclusions: Firstly, M.A.R.I.E. has made possible to bring out new concepts and new continuous measurements variables. The comparison between healthy and abnormal individuals makes it possible to take into consideration the significance of this line of study. From now on, these new functional parameters will allow us to identify and name “emotional” disorders or illnesses which can give additional dimension to behavioral disorders in all pathologies that affect the brain. Secondly, the ViFaEmRe is idiosyncratic, categorical, and a function of the identity of the observer and of the observed face. These findings stack up against Artificial Intelligence, which cannot have a globalist or regionalist algorithm that can be programmed into a robot, nor can AI compete with human abilities and judgment in this domain. *Here “Emotional disorders” refers to disorders of emotional expressions and recognition.

Rational and Continuous Measurement of Emotional-Fingerprint, Emotional-Quotient and Categorical vs Proportional Recognition of Facial Emotions with M.A.R.I.E., Second Half

Context: The advent of Artificial Intelligence (AI) requires modeling prior to its implementation in algorithms for most human skills. This observation requires us to have a detailed and precise understanding of the interfaces of verbal and emotional communications. The progress of AI is significant on the verbal level but modest in terms of the recognition of facial emotions even if this functionality is one of the oldest in humans and is omnipresent in our daily lives. Dysfunction in the ability for facial emotional expressions is present in many brain pathologies encountered by psychiatrists, neurologists, psychotherapists, mental health professionals including social workers. It cannot be objectively verified and measured due to a lack of reliable tools that are valid and consistently sensitive. Indeed, the articles in the scientific literature dealing with Visual-Facial-Emotions-Recognition (ViFaEmRe), suffer from the absence of 1) consensual and rational tools for continuous quantified measurement, 2) operational concepts. We have invented a software that can use computer-morphing attempting to respond to these two obstacles. It is identified as the Method of Analysis and Research of the Integration of Emotions (M.A.R.I.E.). Our primary goal is to use M.A.R.I.E. to understand the physiology of ViFaEmRe in normal healthy subjects by standardizing the measurements. Then, it will allow us to focus on subjects manifesting abnormalities in this ability. Our second goal is to make our contribution to the progress of AI hoping to add the dimension of recognition of facial emotional expressions. Objective: To study: 1) categorical vs dimensional aspects of recognition of ViFaEmRe, 2) universality vs idiosyncrasy, 3) immediate vs ambivalent Emotional-Decision-Making, 4) the Emotional-Fingerprint of a face and 5) creation of population references data. Methods: M.A.R.I.E. enables the rational, quantified measurement of Emotional Visual Acuity (EVA) in an individual observer and a population aged 20 to 70 years. Meanwhile, it can measure the range and intensity of expressed emotions through three Face- Tests, quantify the performance of a sample of 204 observers with hypernormal measures of cognition, “thymia” (defined elsewhere), and low levels of anxiety, and perform analysis of the six primary emotions. Results: We have individualized the following continuous parameters: 1) “Emotional-Visual- Acuity”, 2) “Visual-Emotional-Feeling”, 3) “Emotional-Quotient”, 4) “Emotional-Decision-Making”, 5) “Emotional-Decision-Making Graph” or “Individual-Gun-Trigger”, 6) “Emotional-Fingerprint” or “Key-graph”, 7) “Emotional-Fingerprint-Graph”, 8) detecting “misunderstanding” and 9) detecting “error”. This allowed us a taxonomy with coding of the face-emotion pair. Each face has specific measurements and graphics. The EVA improves from ages of 20 to 55 years, then decreases. It does not depend on the sex of the observer, nor the face studied. In addition, 1% of people endowed with normal intelligence do not recognize emotions. The categorical dimension is a variable for everyone. The range and intensity of ViFaEmRe is idiosyncratic and not universally uniform. The recognition of emotions is purely categorical for a single individual. It is dimensional for a population sample. Conclusions: Firstly, M.A.R.I.E. has made possible to bring out new concepts and new continuous measurements variables. The comparison between healthy and abnormal individuals makes it possible to take into consideration the significance of this line of study. From now on, these new functional parameters will allow us to identify and name “emotional” disorders or illnesses which can give additional dimension to behavioral disorders in all pathologies that affect the brain. Secondly, the ViFaEmRe is idiosyncratic, categorical, and a function of the identity of the observer and of the observed face. These findings stack up against Artificial Intelligence, which cannot have a globalist or regionalist algorithm that can be programmed into a robot, nor can AI compete with human abilities and judgment in this domain. *Here “Emotional disorders” refers to disorders of emotional expressions and recognition.

中部地区物流业碳排放时空差异及驱动效应研究

应用LMDI模型和M-R模型从时间和空间两个维度对2006—2020年间中部地区物流业碳排放差异及驱动效应进行分析,结果表明,中部地区物流业碳排放总体呈逐步上升的趋势,2020年首次出现了下降,经济发展水平、能源效率和人口规模对中部地区物流业碳排放具有较强的拉动效应,碳密度、能源强度和物流产业结构占比则是抑制中部地区物流业碳排放的有效因素;较高的碳密度和能源强度以及领先的经济发展水平使得湖北成为中部六省中物流业碳排放最高的省份,安徽省物流业得益于其超低的碳密度水平从而一直保持较低的碳排放,江西省基本处于中部六省中物流业碳排放最低水平,主要原因在于其经济发展水平和人口规模都远低于中部其他省份.文章最后对中部六省物流业低碳发展提出了相关建议.

1985-2020年呼伦贝尔市冻土气候变化特征

利用1985-2021年呼伦贝尔市15个国家气象站各层地温、第一冻土层下限、最大冻土深度资料,研究呼伦贝尔市冻土气候演变特征,同时采用重标极差(R/S)和非周期循环分析,统计最大冻土深度等气象要素时间序列的Hurst指数、分维数和非周期循环的平均循环长度,分析最大冻土深度等气象要素变化趋势和记忆周期。研究表明:(1)0 cm地温、40 cm地温和80 cm地温都呈增加趋势,且0 cm地温增加趋势最显著,特别是0 cm地温最小值增大更加明显。(2)冻结持续日数呈缓慢减小趋势,其中中部偏北海拔超过600 m山区持续时间最长,西南部和东南部地区持续时间最短。(3)7月中旬冻土从北部地区开始,9月-10月下旬向西南和东南地区扩展,次年5月上旬-6月下旬自西南和东南地区向北部地区逐渐消失。(4)最大冻土深度呈现逐年减小趋势,突变年份出现在1988年,最大冻土深度在7-9月最浅,次年2-4月最深,10月-次年1月是最大冻土深度不断加深的过程,5-6月是最大冻土深度显著减小的时段,最大冻土深度最大值出现在西部偏南地区。(5)冻结持续日数和最大冻土深度未来减小趋势仍将持续,持续时间分别为10、8 a;0 cm地温、40 cm地温和80 cm地温未来增加趋势仍将持续,持续时间都为12 a。

非极性GaN薄膜及其衬底材料

本文分析了在不同衬底上生长无极性GaN薄膜的情况,这些衬底主要包括-γL iA lO2、r面蓝宝石等。通常在蓝宝石上制备的GaN外延膜是沿c轴生长的,而c轴是GaN的极性轴,导致GaN基器件有源层量子阱中出现很强的内建电场,发光效率会因此降低,发展非极性面外延,有望克服这一物理现象,使发光效率提高。

简支槽型钢梁-混凝土曲线组合梁施工阶段扭转效应解析法研究

采用M/r法将竖向均布荷载作用下梁轴曲率的影响等效为分布扭矩,并基于符拉索夫理论,采用初参数法计算曲线开口钢梁的扭转效应,分别通过离散法和连续法分析横隔板对其产生的影响。研究结果表明:采用离散法和连续法得到的截面扭转角和扭率非常接近,扭矩完全相同,但双力矩和翘曲应力在横隔板位置差别较大;增加横隔板对曲线开口钢梁的扭矩无影响,但使截面扭转角、扭率、双力矩和翘曲正应力均明显减小,翘曲正应力随横隔板数量N的增加非线性下降,随曲线跨度与半径比值的增大线性增长。离散法可明确反映横隔板的影响,但比连续法计算复杂;横隔板数量越多,2种方法的结果越接近。实际工程中的曲线组合梁桥,可采用连续法进行分析。

柠檬精油的活性成分分析

目的通过GC/MS分析LEO的活性成分。方法将LEO溶于正己烷,配成2%的溶液,每次进样1μL,流经DB-5毛细管柱,采用GC/MS技术并与NIST标准谱库进行比对确定LEO中各化学成分,按峰面积归一化法得出各组分百分含量。结果 GC/MS分析发现LEO的12种化学成分,占总LEO含量的98.85%,其中含量最高的是m-Mentha-6,8-diene,(R)-(+)-达到77.04%,其次为-Terpinen占14.90%。结论该试验建立的方法简便、快捷、准确,可为EO检测的标准操作流程提供参考,并为建立精油指纹图谱库提供可能。

M系列与R系列染料配伍性能对比与探究

通过滤纸渗圈试验、测试染料的直接性值(S)、吸尽值(E)、固色值(F)、固色速率(R)值、拼色试验、残液中各染料浓度比例、色差分析等方法,分析两组常用活性染料配伍性能,用于指导生产实践。结果表明:M系列染料的亲和力相差较小,上染率、固色率相近,反应时残液中染料浓度比例变化不大,配伍性好。相比而言,R系列染料扩散性好,但染料上染固色曲线规律差别大,配伍性差,因此,残液中染料比例及色光有较大改变。

四种参考作物蒸散量综合法的比较

参考作物蒸散量(ET0)的估算是作物需水量计算的关键,诸多估算方法在不同地区具有不同的适应性。本文利用中国农业主产区6个代表站点的气象数据,以FAO 56 Penman-Monteith(PM)为标准,对常用的1963 Penman(Pen63)、FAO 1979 Penman(FAO 79)、FAO 24 Penman(FAO 24)及1996Kimberly Penman(Kpen)共4种参考作物蒸散量综合方法进行比较评价。结果表明:(1)Pen63、FAO 79及Kpen的日估算值均比PM估算值偏高,FAO 24偏低,其平均偏差分别为0.28、0.52、0和-0.17mm?d-1,相对偏差为16.0%、25.2%、2.4%、-5.3%,相对均方根误差为12.1%、22.4%、14.2%和13.5%。(2)Pen63、FAO79的月估算值显著高于PM值,在高估最大的5月份平均偏高12.5mm(10.8%)和28.2mm(22.6%)。FAO24表现为低估,低估最大的月份平均偏低11.4mm(8.1%),但在南方站点多数月份的估算值与PM估算值无显著差异。Kpen月估算值与PM估算值相比,既有高估(5-10月),也有低估,高估最大的月份平均偏高19.7mm(14.5%),且在南方站点的秋冬季有近6个月与PM无显著差异。(3)Pen63和FAO79的年值均显著大于PM年值,平均偏高103.8mm(11.8%)和191.5mm(21.3%)。FAO24年平均低估PM值60.9mm(6.3%),Kpen则平均高估50.5mm(5.8%)。(4)时间尺度对评价结果具有一定影响,4种综合法依据日、年值的评价效果排序分别为Pen63>FAO24>Kpen>FAO79和Kpen>FAO24>Pen63>FAO79。在日尺度下4种方法更适于湿润气候,但年尺度下仅FAO79和FAO24较适于湿润气候。可见,4种综合法以Pen63普适性最好,FAO79最低,因此使用FAO79前对其进行适应性评价尤为重要。

小柴胡汤对心肌炎小鼠IL-2R及FasL表达的影响

以CVB3m腹腔注射诱导BALB/c小鼠心肌炎模型 ,予小柴胡汤 (XCT)经灌胃后 ,于第 7、14天采用免疫组化法检测各组心肌浸润细胞IL - 2R及FasL表达。结果 ,XCT在病毒性心肌炎急性期能明显提高心肌浸润细胞IL - 2R的表达 ,与病毒对照组相比有显著差异 (P <0 0 5 )。

河北野生狗牙根遗传多样性的SSR分析

利用SSR标记技术分析了34份河北省野生狗牙根种质资源的遗传多样性及亲缘关系。从108对SSR引物组合中筛选出条带清晰、多态性较高的21对引物组合进行遗传多样性分析,共产生226条清晰条带,其中185条为多态性条带,平均每对引物产生8.81条,多态性位点比率为81.9%。UPGMA聚类分析显示,所有供试的34份狗牙根材料相似系数GS=0.484~0.968,平均为0.679。当GS=0.67时,可将34份供试材料分为4大类群。本研究结果表明河北野生狗牙根种质存在一定的遗传变异,可为河北狗牙根种质的评价、优异基因挖掘和抗性材料选择等提供科学依据。

甘蓝型油菜萝卜质不育系的研究及恢保关系测定

用我国油菜种质资源与甘蓝型油菜萝卜质不育系测交,发现不同类型油菜以及同一类型内不同品种测交后代的缺绿反应不同,找到了基本不缺绿的保持系。从广泛测交后代的变异株中,定向选择出N_(86)不育材料。N_(86)不育材料花瓣平展、育性受温度影响,属不育隐性核基因和不育细胞质共同控制的核质互作类型。用^(60)co—γ射线处理萝卜质不育系,其M_1出现了苗期不缺绿的不育株和可育株,得到了两份中间材料。陕2A不育系、波里马不育系、萝卜质不育系和N_(86)不育材料的恢保关系测定表明:陕2A不育系、波里马不育系的恢保关系相同。陕2A不育系和波里马不育系的恢复系不能恢复萝卜质不育系。陕2A不育系和波里马不育系的恢复系,也不能很好恢复N_(86)不育材料,大部分植株为半不育。

应用TG-FTIR技术研究黄土庙煤催化热解特性

用浸渍法制备过渡金属氧化物担载型催化剂MOx/USY(M=Co、Mo、Co-Mo),用热重红外联用技术考察了MOx/USY催化剂对黄土庙(HTM)煤热解失重特性和热解产物生成规律的影响。热重实验结果表明,MOx/USY催化剂可使HTM煤热解的二次脱气条件更为温和,热解峰温分别提前14、23和9℃。动力学分析结果表明,MOx/USY催化剂可降低HTM煤样热解的活化能。FT-IR研究表明,MOx/USY催化剂可有效改善HTM煤热解产物的组成和分布,CoOx/USY催化剂能显著提高HTM煤热解产物中高热值气体(CO、CH4)和轻质芳烃以及脂肪烃类化合物的含量;MoOx/USY催化剂没有明显改善HTM煤热解产物组成和分布;MoOx-CoOx/USY催化剂可促进CO、CH4、轻质芳烃和脂肪烃类化合物的生成,却使热解产物的生成向高温区移动,说明USY负载的不同过渡金属氧化物对煤样热解行为和热解产物有较大影响。