Page 62 - VOL.53
P. 62
廠務季刊 Facility Journal 2024
智慧辦公大樓的數位優化_以台積電為例
Digital Optimization in Smart Office_Example of TSMC Building
文││梁健政│公共設施服務部
摘要
摘要 台積電智慧辦公大樓的演進,分成 5 個發展階段進行數位優
化。➀建立設施管理系統 (m-PUSD):物業管理經驗與運轉
1. 前言
數據進行數位化並開發成系統;➁驅動運轉模式:將維運經
2. 案例探討 驗與知識結合開啟數據驅動的運轉模式;➂人工智慧開發應
用:將維運經驗與領域知識,以數據為基礎開發空調與設施管
2.1 過去「數據」的取得與使用
理的系統的 AI 模型;➃打造虛擬感測器:以原系統架構下,
2.2 空調出風溫度調整的案例
研發跨系統的虛擬感測器以預測人數比例,用於輔助機電系統
2.3 AIoT的應用 的負載調整;➄快速有效率應用 AI:已成功應用的 AI 模型,
開發快速遷移或複製的模式。以數位科技活化設備與系統所產
3. 台積電智慧商辦大樓的五部曲
生的數據,轉換為有效的訊號進行控制或管理,提升了服務效
3.1 第一部_建立設施管理系統 率,兼具運轉經驗數位化傳承,這是一連串透過提升服務而產
(m-PUSD) 生的數位優化的過程。
3.2 第二部_數據驅動的模式 關鍵詞 : 人工智慧、物聯網、大數據
3.3 第三部_人工智慧的應用:
The evolution of TSMC�s smart office building was
以空調出風溫度調整為例
divided into 5 stages of development for digital
3.4 第四部_虛擬感測器 optimization. ➀Establishment of facility management
3.5 第五部_AI建築的分享與商業化 system(m-PUSD): digitize property management
experience and operation data, and develop it
3.6 應用技術概述
into a system ; ➁Drive operation mode: combine
maintenance experience and knowledge to open the
4. 結論與展望
data-driven operation mode ; ➂Artificial intelligence
參考文獻 development and application:develop AI models of
empty and facility management systems based on
作者介紹
data maintenance experience and domain knowledge
; ➃Create virtual sensors: under the original system
architecture, develop cross-system virtual sensors with
a forecast ratio of people to assist the load adjustment
of electromechanical systems ; ➄Apply AI quickly
and efficiently: AI models that have been successfully
applied, and develop models for rapid migration or
replication. Using digital technology to activate the data
generated by the equipment and system, convert it into
effective signals for control or management, improve
service efficiency, and digitally inherit operation
experience, which is a series of digital optimization
processes generated by improving services.
Keywords : Artificial Intelligence(AI), Internet of Things(IoT),
Big Data
��
��