Scholarlink

<p>For complex multiprocess industrial systems, this paper introduces an adaptive process Operating Performance Assessment (OPA) method based on Hierarchical and Multiblock Self-Maintenance Slow Feature and Convolutional Analysis (SFCA) Recurrent AutoEncoder (HMSMSCRAE). This approach builds a general subprocess layer within a complete process layer assessment framework, where the subprocess layer develops a performance assessment model by coordinating slow and fast variable feature extraction modules, capturing more discriminative essential features. The full process layer uses a Multi-Head Self-Attention (MHSA) mechanism to adaptively combine results from each subprocess assessment, creating a modular full process OPA model. Additionally, a self-maintenance mechanism for the model is designed to dynamically update the weights of the subprocess performance recognition models, ensuring the accuracy and robustness of online adaptive assessment. When applied to the hydrocarbon-liquid preparation process, the method achieved a macro-average F1 score of 97.66%, confirming its effectiveness and superiority.</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

Adaptive assessment of operating performance for multiprocess industrial process based on hierarchical and multiblock self-maintenance model

  • 期刊论文

  • 发表日期：

    2026年3月21日

查看全文
  • Huichao Cao,
  • Honghe Du,
  • Dongnian Jiang,
  • Wei Li,
  • Jianfeng Yang

摘要

For complex multiprocess industrial systems, this paper introduces an adaptive process Operating Performance Assessment (OPA) method based on Hierarchical and Multiblock Self-Maintenance Slow Feature and Convolutional Analysis (SFCA) Recurrent AutoEncoder (HMSMSCRAE). This approach builds a general subprocess layer within a complete process layer assessment framework, where the subprocess layer develops a performance assessment model by coordinating slow and fast variable feature extraction modules, capturing more discriminative essential features. The full process layer uses a Multi-Head Self-Attention (MHSA) mechanism to adaptively combine results from each subprocess assessment, creating a modular full process OPA model. Additionally, a self-maintenance mechanism for the model is designed to dynamically update the weights of the subprocess performance recognition models, ensuring the accuracy and robustness of online adaptive assessment. When applied to the hydrocarbon-liquid preparation process, the method achieved a macro-average F1 score of 97.66%, confirming its effectiveness and superiority.