隨著高科技產業的快速發展，無塵室已成為半導體、面板等電子產品製造中不可或缺的生產環境。這些產品對微粒污染極為敏感，因此無塵室必須維持極低汙染的作業條件。然而，除了潔淨度的要求外，溫度與濕度的穩定性亦是影響製程品質的重要因素。濕度過低時，環境中容易累積靜電，可能在產品生產或搬運過程中造成靜電放電損傷；濕度過高則可能導致電子元件表面或內部產生腐蝕，使產品在製程中即面臨報廢風險。因此，如何精準掌握並控制溫濕度，是確保產品品質與提升製程穩定性的關鍵課題。 本研究以無塵室空調系統為研究對象，探討其溫濕度控制機制與影響因子。無塵室空氣來源主要分為外氣與內部循環兩部分：外氣經由外氣空調箱處理後導入，負責維持正壓並調節溫濕度；內部循環則透過溫度感測器與乾式冷盤管進行調節，並以閥門開度控制循環氣流比例。兩股氣流在混風區混合後，由 FFU 送入生產區，形成穩定的製程環境。 研究方法方面，本研究利用開源軟體進行資料整合與預處理，並建立模型以分析影響溫濕度的主要因子。透過對空調系統各單元的解析，整理出可能影響環境條件的變數，並以模型運算找出最具影響力的因子與建議設定值。為驗證模型準確性，本研究將模型建議值套入實際空調系統中，並與現場操作人員依經驗調整後的結果比較，以確認模型推論的可行性。 研究結果顯示，所建立之模型能有效協助無塵室空調系統進行溫濕度控制，降低對操作人員經驗的依賴，縮短調整時間並減少人力投入。此方法未來可應用於無塵室工廠之智慧化環境控制，作為提升製程穩定性與產品品質的重要參考。 ;Cleanrooms have become essential environments for the manufacturing of semiconductor, display, and other electronic products, which are highly sensitive to particulate contamination. Beyond cleanliness requirements, maintaining stable temperature and humidity is critical for ensuring process quality. Low humidity can lead to electrostatic discharge damage during production or handling, while excessive humidity may cause corrosion in electronic components. Precise environmental control is therefore vital for product reliability and process stability. This study investigates the temperature and humidity control mechanisms of a cleanroom air conditioning system and identifies key influencing factors. Cleanroom air supply consists of outside air processed through the make up air unit (MAU) and internally recirculated air conditioned by temperature sensors and dry cooling coils. After separate treatment, both air streams mix and are delivered to the production area through fan filter units (FFUs) to form a stable operating environment. Using open source software, this research performs data integration, preprocessing, and model development to determine the primary variables affecting temperature and humidity. The model’s recommended settings are applied to the actual cleanroom system and compared with results achieved through operator experience to validate the model’s accuracy. The findings show that the proposed model effectively supports cleanroom temperature and humidity control, reduces reliance on manual adjustments, shortens tuning time, and decreases labor requirements. This approach demonstrates strong potential for future implementation in smart environmental control systems to enhance process stability and product quality.
