中空纖維膜之超微粒於PM2.5空氣過濾之3D原位模擬解析

In Situ 3D Simulation for the Filtration Process of Ultrafine Particles Removal from PM2.5 Using Hollow Fiber Membrane

陳珮鵑(Pei-Chuan Chen)

指導教授 ： 童國倫

國立臺灣大學/工學院/化學工程學研究所/碩士(2019年)

https://doi.org/10.6342/NTU201900127

摘要

近年來，中空纖維膜愈來愈受重視，並且廣泛應用於不同的領域。中空纖維膜因為有著較大的比表面積且可以達到較小的孔洞，被認為具有潛力應用於超微粒空氣過濾領域。然而，由於中空纖維膜的複雜結構及過濾程序的複雜機制，要使用計算流體力學軟體建立中空纖維三維虛擬模型並進行過濾效能分析至今仍非常具有挑戰性。因此，本研究的目的為提出使用商業軟體GeoDict建立中空纖維三維虛擬模型以進行過濾效能分析之方法，並進行不同模擬模型與文獻實驗數據之比較。

本研究成功使用GeoDict軟體建立中空纖維三維虛擬模型並進行過濾效能分析。於孔洞分析方面，我們使用孔徑分析法及粒徑分析法兩種不同的分析模型進行孔洞分析，並與文獻實驗數據進行比較，結果顯示，孔徑分析法與文獻實驗數據之平均誤差為3.8 %，粒徑分析法與文獻實驗數據之平均誤差41.5 %，相較之下，孔徑分析法較能貼近文獻實驗數據。另外，本研究於流場分析中，使用了三種不同的解析演算法進行氣體通透量之分析，並與文獻實驗數據進行比較，結果顯示，SIMPLE-FFT解析演算法較LIR解析演算法及EJ解析演算法比起來更為精確，此方法與實驗數據相比能達到2.11 %的平均誤差。本研究亦使用GeoDict軟體進行超微粒之長效過濾模擬解析，並分析流場壓降及過濾效能隨時間之變化，結果顯示，本研究之模擬成果能達到與文獻實驗數據相同趨勢之結果，並於長效壓降結果方面達到2.17 %的平均誤差，於過濾效能方面達到0.01 %的平均誤差。本研究提供了一個方法來建構中空纖維膜之三維虛擬模型並進行過濾效能之模擬解析，模擬結果可以幫助未來更有效地進行實驗設計，此模擬方法也可用於建立其他模組在不同條件下之孔洞分析、流場分析及過濾效能分析之探討。

關鍵字

中空纖維膜 ； 超微粒空氣過濾 ； PM2.5 ； 三維模擬解析

並列摘要

In recent years, hollow fiber membrane module is more and more emphasized and applied to a lot of fields. Due to large surface area and small pore size, hollow fiber membrane provides good potential to remove ultrafine particles. However, it is hard to build 3D (three-dimensional) virtual model of hollow fiber membrane in details because of the asymmetric structure of hollow fiber membrane and the big scale difference between membrane modules and pore morphology. Also, it is still a challenge to analyze the filtration process via Computational Fluid Dynamics (CFD) software because of the complicated fouling mechanism of ultrafine particle filtration. Therefore, the purpose of this study is to develop a methodology for constructing 3D virtual model of hollow fiber membrane and simulate ultrafine particles filtration via commercial GeoDict software.

In this study, a detailed 3D model of hollow fiber membrane was successfully constructed and the filtration process of ultrafine particles removal was simulated. Two different modules in GeoDict software were used to analyze the pore size distribution. Compared to 41.5 % average error of Granulometry module, the results of Porosimetry module achieved only 3.8 % average error, which were more similar to the real data. Besides, three different solvers were provided to simulate the flow in hollow fiber membrane. The results simulated from Semi implicit methods for pressure linked equations with Fast Fourier Transform (SIMPLE-FFT) solver were more accurate than those of Left Identity Right (LIR) solver and Explicit Jump (EJ) solver. Compared to the experimental data, the nitrogen permeance result, which was calculated by combination of Navier-Stokes model and SIMPLE-FFT solver, achieved only 2.11 % error. Moreover, the long-time test was analyzed to get the relationships between process time, pressure drop, and filter efficiency. The simulation results were comparable to the experimental data and showed the same tendency with average error at only 2.17 % for pressure drop and 0.01 % for filter efficiency. This study provides a new methodology to set up a 3D virtual model of hollow fiber membrane precisely and figure out the performance of ultrafine particle filtration. This approach can help followers build models of different membrane structure and simulate the filtration process as well.

並列關鍵字

Hollow fiber membrane ； ultrafine particles air filtration ； PM2.5 ； 3D simulation

參考文獻

REFERENCES

[1] R.-J. Huang, Y. Zhang, C. Bozzetti, K.-F. Ho, J.-J. Cao, Y. Han, K.R. Daellenbach, ___J.G. Slowik, S.M. Platt, F. Canonaco, High secondary aerosol contribution to ___particulate pollution during haze events in China, Nature, 514 (2014) 218.

[2] R. Chen, Z. Zhao, H. Kan, Heavy smog and hospital visits in Beijing, China, American ___journal of respiratory and critical care medicine, 188 (2013) 1170-1171.

[3] L.B. Lave, E.P. Seskin, An analysis of the association between US mortality and air ___pollution, Journal of the American Statistical Association, 68 (1973) 284-290.

[4] Y. Xie, B. Zhao, L. Zhang, R. Luo, Spatiotemporal variations of PM2.5 and PM10 ___concentrations between 31 Chinese cities and their relationships with SO2, NO2, CO ___and O3, Particuology, 20 (2015) 141-149.