本研究透過應用國立成功大學工程科學研究所、資訊工程研究所之研究團之CFD流場模擬與AI人工智慧之預測資料應用於作業場所濃度預測之有效性，並進一步評估是否可以結合人員定位進行暴露評估，對現今暴露監測方法分為：傳統採樣與定量推估，然而會因受限於監測頻率、人力與設備成本，難以即時掌握污染物濃度變化，可能導致勞工暴露風險的低估，而直讀式儀器（本研究係稱感測器）則可以進行實時監測，可彌補監測頻率不足之問題，但實務應用上卻不知如何佈點，然而現今文獻關於雖有使用CFD流場模擬、AI人工智慧應用於稱感測器佈點規劃，但對於實際驗證則探討有限。因此，本研究透過CFD流場模擬與AI人工智慧，藉由現場感測器架設，進行實場驗證並探討CFD流場模擬、AI人工智慧應用於作業場所監測上之有效性與推估人員暴露之可行性。
本研究選定臺南某學校的模擬作業場域進行測試，場域大小為 18.7 × 9.1×3（m3），並設置 16臺感測器 進行不同通風條件與污染源釋放流量，對應改變污染源位置（場域中心、左側），提供AI人工智慧與CFD流場模擬驗證依據。本研究亦透過python建立虛擬勞工與場域濃度分佈以評估勞工暴露時態，透過考量定位誤差( 0.1、0.3、0.5、1公尺 )且結合CFD流場模擬與AI人工智慧預測結果並考量感測器誤差( 1%、5%、10% )計算勞工暴露之相對誤差（誤差標準設為10%），評估CFD流場模擬、AI人工智慧是否可以應用於暴露評估。
現場實驗結果顯示，污染源位置與通風條件對場域內污染物濃度分布是相互影響。污染源位於中心時，場域濃度分布較左側為均勻，進一步計算每組實驗的與GSD，範圍分別為：426.47 ~ 1003.92 ppm ；GSD範圍：1.05~1.31，而CFD流場模擬驗證結果顯示，GSD範圍：1.05 ~ 1.33顯示資料數據與現場實驗分佈數據集中程度相似，而GM範圍：507.59 ppm ~ 1015.30 ppm與T 檢定結果顯示皆具有顯著差異（p < 0.05），本研究進一步計算每一個感測器點位的相對誤差時，範圍∈[0 % ~ 43 % ]，雖然CFD流場模擬能描述現場濃度變化趨勢，但在某些點位誤差較大，而如果剛好該位置為勞工暴露時則有低估的風險，表明CFD流場模擬對於現場濃度預測不具有效性。
AI人工智慧於濃度預測中風速中濃度組別結果顯示，GM範圍：（572.59 ppm ~ 814.98 ppm；現場實驗：578.40 ​~ 783.37 ppm）、GSD範圍：（1.15 ~ 1.24；現場實驗：1.16 ~ 1.21），在預測場域濃度分佈與變異上差異不大，表明AI人工智慧能透過現場實驗建立之暴露資料庫進行學習與濃度預測，而進一步考量每臺感測器點位誤差，在於實場考量定點採樣中，最佳的感測器設置為將16臺感測器減少至最少僅需3～8臺即可推估實場濃度，此時AI驗證每個時間序列中的符合本研究設置標準：相對誤差10％，而隨著使用的感測器的增加，誤差也隨之越低，表明對於作業環境濃度預測具有有效性。
對於CFD與AI結合人員定位結果，CFD流場模擬對於勞工暴露評估在感測器誤差依序（1%、5%、10％）組別中之相對物差誤差範圍分別為：14.88%～15.14%； 15.31%～15.58%； 16.62%～17.02%，而AI人工智慧之暴露評估結果在感測器誤差（1%、5%、10％）組別中之相對誤差依序為：「最大值為10.75 % (定位精度：1 m，感測器臺數：3臺)，誤差最小值為3.78 %（定位精度：0.1 m，感測器臺數：8臺）」；「最大值為10.91 % (定位精度：1 m，感測器臺數：3臺)，誤差最小值為5.39 %（定位精度：0.1 m，感測器臺數：8臺）」；「最大值為13.31 % (定位精度：1 m，感測器臺數：3臺)，誤差最小值為8.97 %（定位精度：0.1 m，感測器臺數：8臺）」，隨著現場架設的感測器臺數數量越多，誤差隨之降低；而隨著定位精度誤差越大，勞工暴露的相對誤差也隨之變大。若以NIOSH感測器準確度標準（±25 %）而言，CFD流場模擬與AI人工智慧的預測勞工暴露的相對誤差皆可以接受，然而以本研究設定的誤差標準（±10%）時，則CFD流場模擬預測勞工暴露之誤差則不符合標準，而AI人工智慧誤差則要選用更多的感測器以選用定位精度更好的系統才可符合標準，而若以環測標準（±5 %）評估時，則感測器誤差僅能選用1%標準下，並且感測器最少僅能選用8臺且定位精度在0.1 ~ 0.5公尺。由此可知，CFD流場模擬對於場域濃度預測上有低估勞工的風險，需準確預測才可結合人員定位應用於勞工暴露評估、AI人工智慧預測結果應用於評估勞工暴露監測為可行的，然而對於實務應用上需謹慎考慮勞工定位精度與感測器精準度，才可精準應用於勞工個人暴露評估。
目前的方法必須事先針對場域縝密佈點後，才後續進行優化，但這並不是有助於實務應用且不符合成本，故透過本研究使用已開發之CFD流場模擬、AI人工智慧方法結果，建議可考慮透過在透過CFD流場模擬能更加符合現場的前提下，進行大量收集資料，涵蓋不同的實驗變異，並提供AI人工智慧預先訓練後，再透過現場少量感測器數據佐以修正模型，以得符合實務應用上之便利性。

This study evaluates the effectiveness of combining Computational Fluid Dynamics (CFD) simulations and Artificial Intelligence (AI) models to predict pollutant concentrations and assess worker exposure in occupational environments. A simulated workspace (18.7 × 9.1 × 3 m³) at a school in Tainan was equipped with 16 sensors to measure concentration under various ventilation conditions and pollutant release scenarios. The study assessed the predictive validity of CFD and AI models and examined their feasibility when integrated with personnel positioning systems, considering sensor errors (1%, 5%, 10%) and positioning inaccuracies (0.1–1 m). While CFD effectively captured overall spatial trends, it underestimated exposure at specific locations, limiting its validity for precise concentration prediction. In contrast, AI predictions closely aligned with field data in terms of geometric mean (GM) and geometric standard deviation (GSD), and maintained exposure errors below 10% using only 3–8 sensors with higher positioning accuracy. When the acceptable exposure error threshold was tightened to 5%, CFD was found to be infeasible under high sensor error (10%), while AI required at least 8 sensors and positioning accuracy between 0.1 and 0.5 m. Field results suggest that AI is more practical for exposure assessment when sensing resources are limited. The study recommends using CFD-generated data for pre-training AI models, followed by minimal on-site data to calibrate the system—offering a cost-effective and scalable approach for sensor deployment and exposure estimation in occupational health applications.

Aksuet, G., Eren, T., & Alakas, H. M. (2024). Using wearable technological devices to improve workplace health and safety: An assessment on a sector base with multi-criteria decision-making methods. Ain Shams Engineering Journal, 15(2), Article 102423. https://doi.org/10.1016/j.asej.2023.102423
Benfradj, A., Thaljaoui, A., Moulahi, T., Khan, R. U., Alabdulatif, A., & Lorenz, P. (2024). Integration of artificial intelligence (AI) with sensor networks: Trends, challenges, and future directions. Journal of King Saud University-Computer and Information Sciences, 36(1), 101892.
Bustos, D., Guedes, J. C., Baptista, J. S., Vaz, M. P., Costa, J. T., & Fernandes, R. J. (2021). Applicability of Physiological Monitoring Systems within Occupational Groups: A Systematic Review. Sensors, 21(21), Article 7249. https://doi.org/10.3390/s21217249
Byrne, R., & Diamond, D. (2006). Chemo/bio-sensor networks. Nature Materials, 5(6), 421-424. https://doi.org/10.1038/nmat1661
Chein, H. M., & Chen, T. M. (2003). Emission characteristics of volatile organic compounds from semiconductor manufacturing. Journal of the Air & Waste Management Association, 53(8), 1029-1036. https://doi.org/10.1080/10473289.2003.10466239
Chen, C., Campbell, K. D., Negi, I., Iglesias, R. A., Owens, P., Tao, N. J., Tsow, F., & Forzani, E. S. (2012). A new sensor for the assessment of personal exposure to volatile organic compounds. Atmospheric Environment, 54, 679-687. https://doi.org/10.1016/j.atmosenv.2012.01.048
Current and future trends in sensor networks: a survey.
Delgado, A., Aliaga, D., Carlos, C., Vergaray, L., & Carbaja, C. (2021). Artificial Intelligence Model based on Grey Clustering for Integral Analysis of Industrial Hygiene Risk. International Journal of Advanced Computer Science and Applications, 12(4), 389-395. <Go to ISI>://WOS:000648867700050
Di, M., & Joo, E. M. (2007). A survey of machine learning in wireless sensor netoworks from networking and application perspectives. 2007 6th international conference on information, communications & signal processing,
Evers, M., Krzywdzinski, M., & Pfeiffer, S. (2018). Designing wearables for use in the workplace: the role of solution developers.
Fanti, G., Spinazzè, A., Borghi, F., Rovelli, S., Campagnolo, D., Keller, M., Borghi, A., Cattaneo, A., Cauda, E., & Cavallo, D. M. (2022). Evolution and applications of recent sensing technology for occupational risk assessment: A rapid review of the literature. Sensors, 22(13), 4841.
Gudiño-Ochoa, A., García-Rodríguez, J. A., Ochoa-Ornelas, R., Cuevas-Chávez, J. I., & Sánchez-Arias, D. A. (2024). Noninvasive Diabetes Detection through Human Breath Using TinyML-Powered E-Nose. Sensors, 24(4), Article 1294. https://doi.org/10.3390/s24041294
Hawkins, D. M. (1980). Identification of outliers (Vol. 11). Springer.
HSE.The Control of Substances Hazardous to Health Regulations 2002. Legislation.gov.uk; 2002. https://www.legislation.gov.uk/uksi/2002/2677/regulation/7.Huang, F.-C., Shih, T.-S., Lee, J.-F., Chao, H.-P., & Wang, P.-Y. (2010). Time location analysis for exposure assessment studies of indoor workers based on active RFID technology. Journal of Environmental Monitoring, 12(2), 514-523.
Huang, F.-C., Shih, T.-S., Lee, J.-F., Chao, H.-P., & Wang, P.-Y. (2010). Time location analysis for exposure assessment studies of indoor workers based on active RFID technology. Journal of Environmental Monitoring, 12(2), 514-523.
Huang, H., & Haghighat, F. (2002). Modelling of volatile organic compounds emission from dry building materials. Building and Environment, 37(12), 1349-1360, Article Pii s0360-1323(01)00116-0. https://doi.org/10.1016/s0360-1323(01)00116-0
Jekabsons, G., Kairish, V., & Zuravlyov, V. (2011). An Analysis of Wi-Fi Based Indoor Positioning Accuracy. Computer Science (1407-7493), 47.
Kanan, R., Elhassan, O., & Bensalem, R. (2018). An IoT-based autonomous system for workers' safety in construction sites with real-time alarming, monitoring, and positioning strategies. Automation in Construction, 88, 73-86.
Lambert, A. R., Lin, C.-L., Mardorf, E., & O'shaughnessy, P. (2010). CFD Simulation of contaminant decay for high reynolds flow in a controlled environment. Annals of Occupational Hygiene, 54(1), 88-99.
Lee, J., Sayler, S. K., Zhou, M., Zhu, H., Richardson, R. J., Neitzel, R. L., Kurabayashi, K., & Fan, X. (2018). On-site monitoring of occupational exposure to volatile organic compounds by a portable comprehensive 2-dimensional gas chromatography device. Analytical Methods, 10(2), 237-244.
Lin, P., Li, Q., Fan, Q., Gao, X., & Hu, S. (2014). A Real‐Time Location‐Based Services System Using WiFi Fingerprinting Algorithm for Safety Risk Assessment of Workers in Tunnels. Mathematical Problems in Engineering, 2014(1), 371456.
Mahyuddin, N., & Awbi, H. (2012). A review of CO2 measurement procedures in ventilation research. International Journal of Ventilation, 10(4), 353-370.
Mahyuddin, N., & Essah, E. A. (2024). Spatial distribution of CO2 Impact on the indoor air quality of classrooms within a University. Journal of Building Engineering, 89, 109246.
Ministry of Manpower. Workplace Safety and Health Act 2006. Singapore Statutes Online; 2006. https://sso.agc.gov.sg/Act/WSHA2006.
Ministry of Manpower. Workplace Safety and Health (General Provisions) Regulations. Singapore Statutes Online; 2007. https://sso.agc.gov.sg/SL/WSHA2006-RG1.
Ministry of Manpower. GUIDELINES ON SAMPLING STRATEGY AND SUBMISSION OF TOXIC SUBSTANCES MONITORING / SAMPLE ANALYSIS REPORT. Ministry of Manpower; 2010. https://www.mom.gov.sg/-/media/mom/documents/services-forms/safety-health/air-sampling-guidelines.pdf.
Marques, G., & Pitarma, R. (2020). A Real-Time Noise Monitoring System Based on Internet of Things for Enhanced Acoustic Comfort and Occupational Health. IEEE Access, 8, 139741-139755. https://doi.org/10.1109/access.2020.3012919
Mokhtari, F., Cheng, Z., Wang, C. H., & Foroughi, J. (2023). Advances in wearable piezoelectric sensors for hazardous workplace environments. Global Challenges, 7(6), 2300019.
NIOSH. Components for Evaluation of Direct-Reading Monitors for Gases and Vapors. National Institute for Occupational Safety and Health；2012. https://www.cdc.gov/niosh/docs/2012-162/pdfs/2012-162.pdf?id=10.26616/NIOSHPUB2012162
OSHA.Calibrating and Testing Direct-Reading Portable Gas Monitors. Occupational Safety and Health Administration； 2013.https://www.osha.gov/publications/shib093013.Park, H., Jang, J.-K., & Shin, J.-A. (2011). Quantitative exposure assessment of various chemical substances in a wafer fabrication industry facility. Safety and Health at Work, 2(1), 39-51.
Park, H., Jang, J.-K., & Shin, J.-A. (2011). Quantitative exposure assessment of various chemical substances in a wafer fabrication industry facility. Safety and Health at Work, 2(1), 39-51.
Pau, G., Arena, F., Gebremariam, Y. E., & You, I. (2021). Bluetooth 5.1: An analysis of direction finding capability for high-precision location services. Sensors, 21(11), 3589.
Pavlou, A. K., Magan, N., McNulty, C., Jones, J. M., Sharp, D., Brown, J., & Turner, A. P. (2002). Use of an electronic nose system for diagnoses of urinary tract infections. Biosensors and Bioelectronics, 17(10), 893-899.
Ridolfi, M., Kaya, A., Berkvens, R., Weyn, M., Joseph, W., & Poorter, E. D. (2021). Self-calibration and collaborative localization for UWB positioning systems: A survey and future research directions. ACM computing surveys (CSUR), 54(4), 1-27.
Riedmann, R. A., Gasic, B., & Vernez, D. (2015). Sensitivity Analysis, Dominant Factors, and Robustness of the ECETOC TRA v3, Stoffenmanager 4.5, and ART 1.5 Occupational Exposure Models. Risk Analysis, 35(2), 211-225. https://doi.org/10.1111/risa.12286
Shu, H., & Liang, Q. (2005). Fuzzy optimization for distributed sensor deployment. IEEE Wireless Communications and Networking Conference, 2005,
Tiele, A., Wicaksono, A., Kansara, J., Arasaradnam, R. P., & Covington, J. A. (2019). Breath analysis using enose and ion mobility technology to diagnose inflammatory bowel disease—A pilot study. Biosensors, 9(2), 55.
Vázquez-Román, R., Díaz-Ovalle, C., Quiroz-Pérez, E., & Mannan, M. S. (2016). A CFD-based approach for gas detectors allocation. Journal of loss prevention in the process industries, 44, 633-641.
Xiao, J., Liu, Z., Yang, Y., Liu, D., & Han, X. (2011). Comparison and analysis of indoor wireless positioning techniques. 2011 International conference on computer science and service system (CSSS),
Xu, J., Li, Z., Zhang, K., Yang, J., Gao, N., Zhang, Z., & Meng, Z. (2023). The principle, methods and recent progress in RFID positioning techniques: A review. IEEE Journal of Radio Frequency Identification, 7, 50-63.
Yang, X., Chen, Q., Zhang, J. S., Magee, R., Zeng, J., & Shaw, C. Y. (2001). Numerical simulation of VOC emissions from dry materials. Building and Environment, 36(10), 1099-1107. https://doi.org/10.1016/s0360-1323(00)00078-0
Zaimen, K., Moalic, L., Abouaissa, A., & Idoumghar, L. (2022). A survey of artificial intelligence based wsns deployment techniques and related objectives modeling. IEEE Access, 10, 113294-113329.
Zhang, D., Ding, E., & Bluyssen, P. M. (2022). Guidance to assess ventilation performance of a classroom based on CO2 monitoring. Indoor and Built Environment, 31(4), 1107-1126.
Zhang, Y., & Kacira, M. (2022). Analysis of climate uniformity in indoor plant factory system with computational fluid dynamics (CFD). Biosystems Engineering, 220, 73-86. https://doi.org/10.1016/j.biosystemseng.2022.05.009
中華民國：勞工作業環境監測實施辦法。全國法規資料庫；2016。https://law.moj.gov.tw/LawClass/LawAll.aspx?pcode=N0060033
中華民國勞動部勞動與職業安全衛生研究所 ：先進國家化學性直讀式環境監測儀器之制度探討。中華民國勞動部；2025。 https://www.ilosh.gov.tw/ 。
日本厚生勞動省：労働安全衛生法。e-Gov 法令検索；1940。https://laws.e-gov.go.jp/law/347AC0000000057。
日本厚生勞動省：作業環境評価基準。厚生労働省；1988。https://www.mhlw.go.jp/web/t_doc_keyword?dataId=74088000&dataType=0&keyword=%E4%BD%9C%E6%A5%AD%E7%92%B0%E5%A2%83%E8%A9%95%E4%BE%A1%E5%9F%BA%E6%BA%96&mode=0&pageNo=1。
王鵬堯：智慧型人員多重定位安全監控雛型製作計畫，勞動部勞動及職業安全衛生研究所；1999。https://www.ilosh.gov.tw/ 。
王寶順：以直讀式儀器之監測輔助推估酸洗作業過程中氫氯酸之最高暴露濃度 ，中山醫學大學職業安全衛生研究所碩士；2019。https://hdl.handle.net/11296/6e82n9 。
勞動部勞動及職業安全衛生研究所： 直讀式儀器應用於作業場所暴露管理技術探討(一)，勞動部；2025。 https://gpi.culture.tw/books/1011400714 。
趙哲偉：無塵室環境氣體偵測器佈點模式合理性探討-以某TFT LCD廠房為例，國立交通大學工學院產業安全與防災學程；2014。 https://hdl.handle.net/11296/qtm67r 。
韓國勞動部：작업환경측정 및 정도관리 등에 관한 고시（工作（營商）環境測量和品質控制等的通知）。국가법령정보센터；2020。https://www.law.go.kr/%ED%96%89%EC%A0%95%EA%B7%9C%EC%B9%99/%EC%9E%91%EC%97%85%ED%99%98%EA%B2%BD%EC%B8%A1%EC%A0%95%EB%B0%8F%EC%A0%95%EB%8F%84%EA%B4%80%EB%A6%AC%EB%93%B1%EC%97%90%EA%B4%80%ED%95%9C%EA%B3%A0%EC%8B%9C 。
韓國勞動部：산업안전보건법（職業安全與健康法）。국가법령정보센터；2023。https://www.law.go.kr/%EB%B2%95%EB%A0%B9/%EC%82%B0%EC%97%85%EC%95%88%EC%A0%84%EB%B3%B4%EA%B1%B4%EB%B2%95 。