| Abstract: | 不更改現有已從事量產的半導體廠冰水主機系統架構下，利用可控制之系統參數設定，乃是以操作為主要手法之節能設計方式。在本研究中首先蒐集與處理工廠中的冰水系統數據資料、工廠的環境條件、總冷卻負載以及設備的操作特性，其次利用部分最小平方法以及模糊邏輯建置冰水主機運轉預測模型，工廠在真實應用的情況下，可利用設備運轉即時資料結合冰水主機運轉預測模型，預估不同操作條件下的全系統耗能，以得到最少耗能的設備操作指引，給予設備操作員一個有效的設備操作參考依據。
Manipulate the controllable settings of a chiller system, without changing its architecture,thus reducing the total chiller system costs. The way of energy saving design is practiced by manipulating technique. In this study, we collect and process the running data of chiller water
system, environment conditions, total cooling load and equipment operating characteristics. Then,use partial least square method and fuzzy logic to build the chillers’ performance prediction models. In practice, using the real-time running data to input prediction model, the total power cost of the chiller system was predicted. Operators can use the prediction results to determine useful chiller operation strategies for minimizing the total operating cost. |
