Scholarlink

<p>Data-driven intelligent manufacturing systems have been extensively investigated over the past decade to address the challenges posed by massive production in smart factories. However, such systems are typically unable to process unstructured data and integrate multi-source knowledge, thereby limiting further efficiency improvements across multi-process manufacturing. In such cases, a knowledge-based intelligent manufacturing system plays a more important role in knowledge acquisition, processing, and integration than data-driven technology. In this paper, a knowledge-based process management model is proposed to establish an integrated framework for multi-source data processing, knowledge extraction, knowledge graph generation, and knowledge application in manufacturing. First, a semantic extraction model based on the Dual-BiGRU-Att-CRF mechanism is developed to extract knowledge from unstructured texts, achieving an F1 score of 0.9168. Second, static and quasi-static knowledge frameworks are established for generating a knowledge graph. Furthermore, by integrating a rule-based reasoning engine, the proposed model enables high-confidence recommendations for machining parameters. A case study on aircraft drop tank bracket manufacturing is conducted to explore the applicability of the proposed model. A prototype system is developed based on this model, demonstrating excellent knowledge extraction performance and achieving 96.7% reliability in welding parameter recommendations.</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

Knowledge-based process management model: A framework for knowledge flow from acquisition to application in aviation component manufacturing

  • 期刊论文

  • 发表日期：

    2026年4月23日

查看全文
  • Boyang Zhang,
  • Hu Shi,
  • Leping Zhou,
  • Yang Hui,
  • Xuesong Mei

摘要

Data-driven intelligent manufacturing systems have been extensively investigated over the past decade to address the challenges posed by massive production in smart factories. However, such systems are typically unable to process unstructured data and integrate multi-source knowledge, thereby limiting further efficiency improvements across multi-process manufacturing. In such cases, a knowledge-based intelligent manufacturing system plays a more important role in knowledge acquisition, processing, and integration than data-driven technology. In this paper, a knowledge-based process management model is proposed to establish an integrated framework for multi-source data processing, knowledge extraction, knowledge graph generation, and knowledge application in manufacturing. First, a semantic extraction model based on the Dual-BiGRU-Att-CRF mechanism is developed to extract knowledge from unstructured texts, achieving an F1 score of 0.9168. Second, static and quasi-static knowledge frameworks are established for generating a knowledge graph. Furthermore, by integrating a rule-based reasoning engine, the proposed model enables high-confidence recommendations for machining parameters. A case study on aircraft drop tank bracket manufacturing is conducted to explore the applicability of the proposed model. A prototype system is developed based on this model, demonstrating excellent knowledge extraction performance and achieving 96.7% reliability in welding parameter recommendations.