以阻塞過濾定律與三維數位孿生模型探討薄膜微過濾之粒子污染現象

Abstract

顆粒污染是膜過濾技術面臨的重大挑戰，嚴重影響其性能與使用壽命。傳統的阻塞過濾理論將污染機制分為四種模式，針對不同的粒徑與孔徑關係，分別對應特定的阻塞指數。然而，在實際過濾實驗中，觀察到的阻塞指數常無法完全以傳統模型解釋。為此，本研究引入數位孿生概念，透過數值模擬與實驗數據整合，有效驗證觀測現象並預測複雜污染行為。 本研究以掃描式電子顯微鏡影像為基礎，運用GeoDict®軟體建立軌跡蝕刻聚碳酸酯膜的三維模型，並透過孔徑、孔隙率與純水通量的模擬與實驗對比，驗證所建模方法的可行性與準確性。進一步於0.4與0.8微米孔徑的膜上，針對五種粒徑進行過濾測試，探討粒徑與孔徑比值對堵塞與通量的影響。 阻塞指數分析揭示污染機制之轉變趨勢：當指數先下降後回升至零，顯示系統進入濾餅過濾階段。在此過渡階段觀察到的負值指數，與模擬中顯示的表面部分堵塞相互印證，進一步確認污染狀態的演變過程。 針對粒徑孔徑比值進行延伸模擬，結合理論通量與實驗結果，發現當粒徑大於孔徑時，比值越高，通量越大，而當比值接近1時，通量最低；相反地，粒徑小於孔徑的粒子可順利穿透薄膜，污染影響較小，通量較穩定。此外，相較於粒徑孔徑比值，孔徑對通量的影響更為顯著。 本研究建立之模擬平台可用於早期辨識膜堵塞情形，並支援過濾效能的預測性評估。研究成果對於工廠中粒子去除之廢水處理應用具有實用價值，亦為過濾阻塞行為的理論建構與後續模擬發展奠定基礎。

Particle fouling represents a significant challenge in membrane filtration, substantially affecting performance and long-term stability. Based on the relationship between particle size and pore size, traditional blocking filtration laws categorize particle fouling into four mechanisms, each defined by a specific blocking index. However, experimental observations frequently reveal deviations from these idealized values. To address this limitation, this study introduces the Digital Twin concept, which integrates numerical simulation and experimental validation to investigate the dynamics of particle-induced membrane fouling and predict complex fouling behaviors. Three-dimensional (3D) membrane structures were reconstructed from scanning electron microscopy (SEM) images using GeoDict® software. The feasibility and accuracy of the simulation model were validated by comparing pore size, porosity, and pure water permeance between the simulation and experiments. Filtration tests involving five different particle sizes were conducted using membranes with pore diameters of 0.4 and 0.8 μm to investigate the effects of particle-to-pore size ratio on flow rate and fouling behavior. The analysis of the blocking index revealed transitions in fouling mechanisms: a drop followed by a return to zero indicated the onset of cake formation. Negative values observed during the transition consistent with partial surface blockage in the simulation, confirming changes in fouling stages. Extended simulations of different particle-to-pore size ratios, combined with theoretical flow rate analysis and experimental data, showed that when particle size exceeded pore size, a higher ratio resulted in a greater initial flow rate, while the flow rate was lowest when the ratio approached unity. By contrast, particles smaller than the pore size passed through the membrane more easily, leading to minimal fouling and more stable flow rates. Moreover, pore size was found to have a more significant impact on flow rate than the particle-to-pore size ratio. The validated Digital Twin framework enables in situ visualization of fouling development, provides early-stage diagnosis of membrane clogging, and supports predictive evaluation of filtration performance. The findings are particularly applicable to the treatment of wastewater involving nanoparticle removal in the semiconductor industry and contribute to advancing theoretical understanding and simulation of fouling dynamics.