查詢結果分析
來源資料
頁籤選單縮合
題 名 | 大數據在石化業應用之初探=A Preliminary Application of Big Data Analysis in Petrochemical Industry |
---|---|
作 者 | 趙仁德; 趙仁德; | 書刊名 | 石油季刊 |
卷 期 | 54:4 2018.12[民107.12] |
頁 次 | 頁65-79 |
分類號 | 457.5 |
關鍵詞 | 物聯網; 膨脹機; 大數據分析; 資料視覺化; 資料探勘; 人工智慧; Internet of things; Expander; Big data analysis; Data visualization; Data mining; Artificial intelligence; |
語 文 | 中文(Chinese) |
中文摘要 | 隨著物聯網科技蓬勃發展,帶動企業研發大數據分析,利用人工智慧演算法將多元數據轉為知識,以提升決策效能。中油公司面對內外環境的變遷,兼顧經濟、環保及社會多面向需求,亟需透過「多元數據收集與資料視覺化」協助跨領域溝通,並藉由「數據整合與人工智慧」研發更精準預測模式,以發展智能決策。本文首先說明在巨量化的浪潮下,大數據分析的起因與可能衍生問題,其次介紹企業建置大數據平台的系統架構與資料視覺化功能,最後以石化廠膨脹機設備異常偵測為例,說明資料視覺化設計與數據分析建模的過程。本研究使用五種分類演算法進行分析,研究發現,預測模型平均準確率達99.9%,其中以隨機森林演算法最優,提供後續數據分析應用之參考。 |
英文摘要 | Internet of Things (IOT) revolution drives the Big Data Analysis, enabling companies to use artificial intelligence in turning data into actionable knowledge for improving decision-making. Taking into account the needs of the economy, the environment and the society, CPC Corporation Taiwan proposes "data visualization" to help cross-disciplinary communication and develops more precise "artificial intelligence" models to provide intelligent forecasting. This paper presents an application of Big Data Analysis in the CPC petrochemical factory. Firstly, the evolution and features in big data application are described. Secondly, a Big Data architecture for Big Data Analysis is depicted. Next, this paper illustrates the process of visualization design for data analysis. Then, the preliminary problem of the expander anomaly in petrochemical plants is discussed. Finally, this study tackles different classification algorithms for expander anomaly detection. It is found that the average accuracy of five prediction models is 99.9%. Excellent predictive algorithm is Random Forest, which provides for future research. |
本系統中英文摘要資訊取自各篇刊載內容。