本研究以工業常用之冰水主機（Chiller）為研究對象，探討不同冷媒溫室氣體排放估算方法、逸散率假設及冷媒洩漏動態模型對排放量與碳費估算結果之影響。研究首先系統性整理政府間氣候變遷專門委員會（Intergovernmental Panel on Climate Change， IPCC）《2006 IPCC Guidelines for National Greenhouse Gas Inventories》、2019年《Refinement to the 2006 IPCC Guidelines》、聯合國環境規劃署（United Nations Environment Programme， UNEP）製冷、空調與熱泵技術評估報告及相關文獻，建立冷媒逸散率資料庫與分析架構。其次，透過冷媒原始充填量、逸散率情境及冷媒存量（Refrigerant Bank）概念，建構固定逸散率法與冷媒洩漏動態模型，比較不同估算方法於設備生命週期中累積排放量及碳費估算結果之差異。
本研究建立固定逸散率法與冷媒洩漏動態模型之比較架構，並結合逸散率情境分析與碳費估算，可作為企業進行冷媒排放量盤查、碳費評估及設備管理之參考，亦可提供主管機關建立本土化冷媒逸散率資料庫、修訂冷媒逸散率預設值及研議冷媒排放管理制度之依據。建議未來結合長期實測資料、即時監測技術及設備生命週期分析，以提升冷媒排放量估算之準確性與制度應用之合理性。;In recent years， countries around the world have progressively implemented carbon pricing mechanisms， greenhouse gas (GHG) inventory systems， and carbon reduction policies， leading enterprises to place greater emphasis on the quantification and management of greenhouse gas emissions. In the manufacturing sector， greenhouse gas emissions primarily originate from electricity consumption and fuel combustion. However， refrigerants used in refrigeration and air-conditioning systems generally possess high Global Warming Potential (GWP). Although refrigerant leakage is relatively small in quantity， it may significantly affect organizational greenhouse gas emissions and carbon fee liabilities. At present， different refrigerant greenhouse gas emission estimation methods and leakage rate assumptions remain inconsistent， resulting in uncertainties in emission estimation. Therefore， evaluating the impacts of different estimation methods on refrigerant greenhouse gas emissions and carbon fees has become an important issue in organizational greenhouse gas inventories and carbon management. This study focuses on industrial chillers commonly used in the manufacturing sector and investigates the effects of different refrigerant greenhouse gas emission estimation methods， leakage rate assumptions， and a Dynamic Refrigerant Leakage Model on greenhouse gas emissions and carbon fee assessment. First， relevant literature， including the 2006 IPCC Guidelines for National Greenhouse Gas Inventories， the 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories， technical assessment reports published by the United Nations Environment Programme (UNEP)， and other related studies， was systematically reviewed to establish a refrigerant leakage rate database and analytical framework. Subsequently， based on the concepts of initial refrigerant charge， leakage rate scenarios， and the refrigerant bank， a Fixed Leakage Rate Method and a Dynamic Refrigerant Leakage Model were developed to compare cumulative greenhouse gas emissions and carbon fee assessment results over the equipment life cycle. The results indicate that refrigerant greenhouse gas emission estimates are highly influenced by the selected leakage rates and modeling approach， with leakage rates being the primary factor affecting cumulative greenhouse gas emissions. Under identical equipment conditions， different leakage rate scenarios result in substantial differences in cumulative greenhouse gas emissions， and these differences become increasingly significant as the service life of the equipment increases. The Fixed Leakage Rate Method assumes a constant annual refrigerant leakage rate， resulting in an approximately linear increase in cumulative greenhouse gas emissions. In contrast， the Dynamic Refrigerant Leakage Model incorporates the gradual decline of the refrigerant bank (stock decay)， leading to progressively smaller annual emission increments in later years and consequently lower long-term cumulative greenhouse gas emissions than those estimated by the Fixed Leakage Rate Method. These findings demonstrate that the choice of estimation model has a significant influence on long-term greenhouse gas emission estimates. The study also found that differences among refrigerant greenhouse gas emission estimation methods are directly reflected in carbon fee assessment results. Due to the high GWP of refrigerants， even relatively small differences in leakage rates may accumulate over time and result in considerable differences in carbon dioxide equivalent (CO₂e) emissions and associated carbon fees. Furthermore， when enterprises maintain comprehensive records of refrigerant charging， recovery， and maintenance， the Mass Balance Approach can effectively improve the accuracy of greenhouse gas emission estimation， reduce the uncertainties associated with default leakage rate assumptions， and enhance the reliability of refrigerant management and carbon management. This study establishes a comparative framework integrating the Fixed Leakage Rate Method and the Dynamic Refrigerant Leakage Model with leakage rate scenario analysis and carbon fee assessment. The proposed framework can serve as a practical reference for enterprises conducting refrigerant greenhouse gas inventories， carbon fee assessments， and equipment management. It may also assist government agencies in developing localized refrigerant leakage rate databases， revising default leakage rate values， and improving refrigerant emission management policies. Future studies are recommended to incorporate long-term field measurements， real-time monitoring technologies， and life cycle assessment to further improve the accuracy and applicability of refrigerant greenhouse gas emission estimation.
