請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/15256
標題: | 使用深度學習與虛實整合系統之工業數據分析用例推薦 Use-Case Recommendation for Industry Data Analytics by Deep Learning and Cyber-Physical System |
作者: | Wel-Jie Liao 廖偉傑 |
指導教授: | 張瑞益(Ray-I Chang) |
關鍵字: | 虛實整合系統,智慧推薦,深度學習,用例, Cyber-Physical System,Intelligent Recommendation,Deep Learning,Use-Case, |
出版年 : | 2020 |
學位: | 碩士 |
摘要: | 為了達到資源優化、品質提升、靈活生產,工業4.0希望能智慧轉型,以工業數據分析創造更高的價值。然而,不同的工廠有不同的工業數據分析需求,新的需求與技術也不斷發生,使得解決問題的專家十分稀缺。本研究探討如何使用智慧推薦技術,解決不同的工業數據分析需求。我們透過蒐集網路上的相關技術文獻,針對文獻的標題、摘要與關鍵字,使用深度學習中包括長短期記憶 (Long Short-Term Memory, LSTM)模型進行訓練,提出一個工業數據分析用例推薦系統。本研究參考虛實整合系統的5C架構 (Smart Connection Level、Data-to-Information Conversion Level、Cyber Level、Cognition Level、Configuration Level),實際建置一個巨量資料平台以蒐集工業數據,提供資料分析、資料視覺化與動態儀表板,當使用者輸入其問題需求之描述文字,本系統嘗試找出適合的用例,並自動推薦系統之相關資料分析模型;同時參考虛實整合系統的5C架構,提出一個工業設備知識關係流程,用以找出設備與設備間的知識關係,建立工業設備詞向量模型以利用例推薦。 To achieve resource optimization, quality improvement, and flexible production line, Industry 4.0 hopes to intelligently transform and create higher value through industrial data analysis. However, different factories have different requirements of industrial data analysis, new requirements are constantly occurring, and experts who can solve problems are scarce. This research explores how to provide users with an intelligent recommendation solution to meet different problem needs. We propose a use-case recommendation system of industrial data analysis by collecting academic papers (with titles, abstracts, and keywords) on the Internet and using deep learning (with the Long Short-Term Memory (LSTM) model). This study refers to the 5C architecture (Smart Connection Level, Data-to-Information Conversion Level, Cyber Level, Cognition Level, and Configuration Level) of the Cyber-Physical System, and builds a big data platform to collect industrial data, provide data analysis, data visualization, and dynamic dashboards. When users enter descriptive text of their problem, the system attempts to find out suitable use-cases and automatically recommend the relevant data analysis model of the system. According to the 5C architecture of the Cyber-Physical System, an industrial equipment knowledge relationship process is proposed to find out the knowledge relationship between equipment and equipment, and establishes a vector model of industrial equipment for use-case recommendation system. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/15256 |
DOI: | 10.6342/NTU202000143 |
全文授權: | 未授權 |
顯示於系所單位: | 工程科學及海洋工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-109-1.pdf 目前未授權公開取用 | 2.01 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。