The methanol to olefins （MTO） reaction was performed over ZSM‐5 zeolite at 300℃ under various methanol weight hourly space velocity （WHSV） values. During these trials, the catalytic perfor‐mance was assessed, i...The methanol to olefins （MTO） reaction was performed over ZSM‐5 zeolite at 300℃ under various methanol weight hourly space velocity （WHSV） values. During these trials, the catalytic perfor‐mance was assessed, in addition to the formation and function of organic compounds retained in the zeolite. Analysis of reaction effluents and confined organics demonstrated a dual‐cycle reaction mechanism when employing ZSM‐5. The extent of the hydrogen transfer reaction, a secondary reac‐tion in the MTO process, varied as the catalyst‐methanol contact time was changed. In addition, 12C/13C‐methanol switch experiments indicated a relationship between the dual‐cycle mechanism and the extent of the hydrogen transfer reaction. Reactions employing a low methanol WHSV in conjunction with a long contact time favored the hydrogen transfer reaction to give alkene products and promoted the generation and accumulation of retained organic species, such as aromatics and methylcyclopentadienes, which enhance the aromatic cycle. When using higher WHSV values, the reduced contact times lessened the extent of the hydrogen transfer reaction and limited the genera‐tion of methylcyclopentadienes and aromatic species. This suppressed the aromatic cycle, such that the alkene cycle became the dominant route during the MTO reaction.展开更多
为快速获取烟草科技文献中的知识信息,通过交互式迭代学习的烟草知识实体标注与识别方法,构建了面向烟草领域的文本标注语料库,设计了适用于烟草领域的文本标注规范,并利用BERT+CRF(Bidirectional Encoder Representations from Transfo...为快速获取烟草科技文献中的知识信息,通过交互式迭代学习的烟草知识实体标注与识别方法,构建了面向烟草领域的文本标注语料库,设计了适用于烟草领域的文本标注规范,并利用BERT+CRF(Bidirectional Encoder Representations from Transformers+Conditional Random Field)深度学习网络模型实现了烟草命名实体的识别和预标注,结合人工校对扩充了原始语料的规模,优化了模型性能。结果表明:语料标注一致性F1标注达92.4%;BERT+CRF模型识别能力优于常用的CRF、BiLSTM+CRF命名实体识别模型。该技术可为提升烟草领域文本分析和知识挖掘能力提供支持。展开更多
基金supported by the National Natural Science Foundation of China (91545104,21576256,21473182,21273230,21273005)the Youth Innovation Promotion Association of the Chinese Academy of Sciences~~
文摘The methanol to olefins （MTO） reaction was performed over ZSM‐5 zeolite at 300℃ under various methanol weight hourly space velocity （WHSV） values. During these trials, the catalytic perfor‐mance was assessed, in addition to the formation and function of organic compounds retained in the zeolite. Analysis of reaction effluents and confined organics demonstrated a dual‐cycle reaction mechanism when employing ZSM‐5. The extent of the hydrogen transfer reaction, a secondary reac‐tion in the MTO process, varied as the catalyst‐methanol contact time was changed. In addition, 12C/13C‐methanol switch experiments indicated a relationship between the dual‐cycle mechanism and the extent of the hydrogen transfer reaction. Reactions employing a low methanol WHSV in conjunction with a long contact time favored the hydrogen transfer reaction to give alkene products and promoted the generation and accumulation of retained organic species, such as aromatics and methylcyclopentadienes, which enhance the aromatic cycle. When using higher WHSV values, the reduced contact times lessened the extent of the hydrogen transfer reaction and limited the genera‐tion of methylcyclopentadienes and aromatic species. This suppressed the aromatic cycle, such that the alkene cycle became the dominant route during the MTO reaction.
文摘为快速获取烟草科技文献中的知识信息,通过交互式迭代学习的烟草知识实体标注与识别方法,构建了面向烟草领域的文本标注语料库,设计了适用于烟草领域的文本标注规范,并利用BERT+CRF(Bidirectional Encoder Representations from Transformers+Conditional Random Field)深度学习网络模型实现了烟草命名实体的识别和预标注,结合人工校对扩充了原始语料的规模,优化了模型性能。结果表明:语料标注一致性F1标注达92.4%;BERT+CRF模型识别能力优于常用的CRF、BiLSTM+CRF命名实体识别模型。该技术可为提升烟草领域文本分析和知识挖掘能力提供支持。
