Scholarlink

<p>Thermal behavior recognition is crucial for preventing thermal runaway and optimizing thermally safe and economic operating parameters in semibatch reactors, and thus, many criteria have been developed to date. However, since the category labels of thermal behavior samples are unsuspected, the identification of thermal behavior should be regarded as an unsupervised clustering issue in the field of pattern recognition. Based on this discovery, the recognition of thermal behavior was investigated from the clustering perspective. Consequently, the novel K-means (KM) criterion was proposed, integrating the KM clustering algorithm with the appropriate feature subset. The rationality of the proposed KM criterion was systematically verified by the visualization of feature distribution, the simulation of the samples near the boundaries of specific boundary diagrams, and the comparison with other criteria. Subsequently, an optimization process for designating safe and effective operating conditions was developed and implemented via the designed genetic-ant colony algorithm, which demonstrated remarkable accuracy and global search capability. Finally, the optimum operating parameters for the aromatic nitration of 4-chlorobenzotrifluoride were obtained, enabling the reaction system to perform effectively under thermally safe scenarios.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Clustering-based recognition of thermal behaviors and optimization of safe–effective operating conditions for isoperibolic semibatch reactors

  • 期刊论文

  • 发表日期：

    2025年4月21日

查看全文
  • Bo Zhang,
  • Guang Yang,
  • Qiuyu Wang

摘要

Thermal behavior recognition is crucial for preventing thermal runaway and optimizing thermally safe and economic operating parameters in semibatch reactors, and thus, many criteria have been developed to date. However, since the category labels of thermal behavior samples are unsuspected, the identification of thermal behavior should be regarded as an unsupervised clustering issue in the field of pattern recognition. Based on this discovery, the recognition of thermal behavior was investigated from the clustering perspective. Consequently, the novel K-means (KM) criterion was proposed, integrating the KM clustering algorithm with the appropriate feature subset. The rationality of the proposed KM criterion was systematically verified by the visualization of feature distribution, the simulation of the samples near the boundaries of specific boundary diagrams, and the comparison with other criteria. Subsequently, an optimization process for designating safe and effective operating conditions was developed and implemented via the designed genetic-ant colony algorithm, which demonstrated remarkable accuracy and global search capability. Finally, the optimum operating parameters for the aromatic nitration of 4-chlorobenzotrifluoride were obtained, enabling the reaction system to perform effectively under thermally safe scenarios.