技術簡介
工業製程所產生的數據,具有巨量資料的部分特性,多變量資料分析技術探討如何以快速有效的方式進行數據分析並挖掘出可應用的資訊,分析過程除了可應用於找出生產製程的問題所在,也可以作為關鍵品質參數的軟感測器,進一步結合工程師的背景知識,將能去除生產瓶頸與提高生產品質。
Abstract
The data generated from the process of industrial manufacturing possessed of some big-data features. The multivariate analysis technology devoted to investigate how to mine the useful information from these industrial data set in a efficient way. Through this technology it can find the possible problem encountered in the manufacturing process and also be a soft sensor for predicting the key process parameters. Moreover, the improved manufacturing quality and increasing product yield may be reached with the combination of the domain knowledge of engineers.
技術規格
離線MSPC分析建模軟體
1.具備數據前處理功能
2.具備PCA-based之資料分析功能
3.具備PLS-based之數值建模功能
4.包含多種定性與定量之分析圖表
5.具備模型及報表輸出功能
線上MSPC監控預測軟體
1.過OPC等通訊方式擷取即時製程資料
2.線上製程監控診斷功能
3.線上關鍵品質參數預測功能
4.以網頁作為使用者介面,並依使用者群組限定功能及瀏覽頁面
5.具備人員管理功能與模型上線管理機制
Technical Specification
Off-line software for MSPC analysis and modeling
Data preprocessing
Data analysis and outliers deletion based on PCA
Statistics modeling based on PLS
Numerous qualitative and quantitative analysis charts
Model and report export functions
On-line software for MSPC monitoring and prediction
Access to DCS or PLC to obtain real-time process data
Process monitoring and diagnosis in real-time mode
Key quality parameter prediction in real-time mode
Webpage as user interface and which browse functions depends on authority setting of user group
User and Model deployment management
技術特色
1.使用自主開發的MSPC演算法
2.開發簡潔易用的軟體使用介面
應用範圍
石化業、化工業、鋼鐵業、食品業、製藥業、科技業等製程、系統、或設備
接受技術者具備基礎建議(設備)
none
接受技術者具備基礎建議(專業)
具數據分析、統計建模、資訊系統等基礎知識
聯絡資訊
聯絡人:鄭儀誠 電網與電力電子技術組
電話:+886-3-5914930 或 Email:cheng@itri.org.tw
客服專線:+886-800-45-8899
傳真:+886-3-5916457