作業環境監測常因樣本數不足，致增加暴露評估結果的不確定性，進而造成無法有效掌握勞工暴露實態。此外，法定以個人直接採樣為主之作業環境監測策略，並無法用以瞭解現場濃度分佈情形，因此亦無法有效掌握逸散源。即時感測裝置為持續監控化學品現場濃度的新興工具，其最大優勢在於能提供詳細的逐時偵測資訊，並且可藉由多處監測瞭解其現場濃度分佈，因此可有效掌握各逸散源，唯如若能結合勞工之時間活動模式，其亦可用於作業場所勞工暴露的管理及控制，並具有節省傳統採樣所衍生之大量人力、物力、財力及時間之優點。唯利用即時感測裝置是否能有效的評估現場之濃度之時空分布，端賴其配置(數量與位置)是否合理，然而目前相關之研究仍極其有限。因此，本研究以即時感測裝置佈點最佳化策略為研究標的，並將其結果應用於現場之濃度分佈監控、暴露評估與警報之規畫等。本研究以石化廠之酚醛樹酯製程( L×W=30×26 )及面板廠之表面塗佈製程( L×W=10×39 )作為實驗測試場域，利用即時感測裝置蒐集場域濃度分佈數據，將即時感測裝置量測點位作為虛擬即時感測裝置點位，透過逐步減少虛擬即時感測裝置點位，以不同內插方法推估點位濃度，並以不同精準度標準(5%、10%、15%、20%)為規範，作為可減少即時感測裝置數目閾值之基礎。本研究所提出之佈點最佳化策略及蒐集之數據，可用於瞭解現場濃度分佈、場域熱點分析、推估勞工暴露，而所獲得之長期暴露資料可再利用貝氏決策分析(Bayesian decision analysis, BDA)技術進行長期暴露評估。本研究先假設於前述石化及面板廠場域分別各布置63及52點。量測濃度範圍分別為0.006 – 1.843 ppm及0.173 - 20.380 ppm，主要逸散源分別為設備之鋼板裂縫及塗佈機。佈點結果顯示精準度5%、10%、15%、20% 不同指標下，石化廠需設置之點位數量(佔原布置之%)，分別為63點( 100 % )、57點( 90 % )、48點( 76 % )、36點( 57 % )，而面板廠則分別為51點( 98 % )、48點( 94 % )、38點( 73 % )、28點( 54 % )。量測濃度結合勞工之時間活動模式後，以貝氏決策分析評估兩場域勞工長期暴露風險，發現石化廠之甲醛勞工暴露量有100%的機率大於50% PEL-TWA；在面板廠中，顯示高風險作業之丙二醇甲基醚醋酸酯勞工有近63%的機率其暴露濃度值大於50% PEL-TWA。此外，為即時掌握現場洩漏狀況，研究亦發現石化廠及面板廠各點位應設置之微洩漏警報限值範圍分別為0.422 – 1.229 ppm及0.492 – 1.473 ppm。本研究所發展之最佳化現場佈點技術，可有效用於現場熱點分析與勞工暴露評估，亦有助於職業衛生日常管理、即時暴露評估及職業病鑑定等，對保障勞工職場健康具積極促進之意義。

Conducting the current regulatory environmental monitoring usually suffers from the insurficient samples which leads to the increase in result uncertainty, and hence becomes inadequate to exposure profiles of workers. Sinec the above approach is on the basis of personal sampling strategy, and hence the results cannot be used for characterize the concentration distributed in the workpace and the identification of fugitive emission sources. Real-time sonsors are emerging instruments for continuously monitoring chemical concentrations with the advantages for providing real-time data of different sites to construct concentration countours of the workplace, which leads to the identification of fugitive emission sources becomes possible. In adition, if the above can be integrated with the time-activity patterns of workers, then evaluation and control of workers’ exposures could be done with the advantage for reducing man-power, cost, and time. However, it should be noted that if the real-time sonsors could provide sufficient data for characterizing the spatial and temporal distribution of concentrations in the workplace mainly depands on how sensors were distributed (including number and sites) in the given workplace. Therefre, the purpose of this study is to develop a strategy for optimizing sensor distribution, and the application of its results in concentration contour establishment, hot-spot detection, exposure assessment, and alarm system setting. The development of the sensor distribution optimization strategy was conducted in the resin process of a petrochemical plant (L×W=30m×26m) and the coating process of a TFT-LCD industry (L×W=30m×26m), and chemical concentration was monitored using real-time sensors. In the present study, virtual sensor grids was first constructed, and the concentration monitored by real-time sensors were used for the gridding optimization purpose. By step-by-step reducing the number of virtual sensor grids, the unknown grid concentrations can be evaluated by different interpolation methods (Linear, Cubic, Spline, Nearest, Makima, Inverse Distance Weighted, Scattered Interpolant). Different precision indicators (5%, 10%, 15%, 20%) were used to determine the optimized gas sensor distribution, respectively. The sensor gridding optimization and collected concentration data can be used for the determination of the concentration contours, field hotspots, alarm setting levels, and for conducting long-term exposure assessments via the used of the Bayesian decision analysis (BDA). The measured concentrations were found in the range of 0.006-1.843 ppm and 0.173-20.380 ppm for the studied petrochemical plant and TFT-LCD factory, respectively. The main fugitive emission sources were found to be the steel plate cracks and the coating machine of the equipment for the petrochemical plant and TFT-LCD factory, respectively. The results of sensor gridding optimization show that under precision indicators of 5%, 10%, 15%, and 20%, the number of real-time sensor needed to be placed in the petrochemical plant ( % of original setting number) was 63 ( 100 % ), 57 ( 90 % ), 48 ( 76 % ), 36 ( 57 % ), respectively, and the TFT-LCD plant was 51 ( 98 % ), 48 ( 94 % ), 38 ( 73 % ), 28 ( 54 % ), respectively. Besides, through the use of the concentration contours and workers’ time activity patterns, workers’ long-term exposure profiles analyzed using the BDA show that the exposure of formaldehyde in petrochemical plant is consistently greater than 50% PEL-TWA; and the propylene glycol methyl ether acetate (PGMEA) exposures in high-risk operations has a ~63% probability greater than 50% PEL in the TFT-LCD factory. For early detecting the microleakage from the manufacturing process, the alarm setting limits were recommended as the ranges of 0.422-1.229 ppm and 0.492-1.473 ppm for petrochemical plant and TFT-LCD factory, respectively. In conclusions, the developed optimizing sensor distribution techniques can be effectively using in contour establishment, hot-spot detection, exposure assessment and alarm system setting.

Ali K, Warraich MA. A framework to implement data cleaning in enterprise data warehouse for robust data quality. In: Proceedings of the Information and Emerging Technologies (ICIET), 2010 International Conference on, 2010, IEEE, 1-6.
Allen CT, Young GS, Haupt SEJAE. 2007. Improving pollutant source characterization by better estimating wind direction with a genetic algorithm. 41:2283-2289.
Andersson K, Bakke J, Bjørseth O, Bornehag CG, Clausen G, Hongslo J, et al. 1997. Tvoc and health in non‐industrial indoor environments: Report from a nordic scientific consensus meeting at långholmen in stockholm, 1996. Indoor Air 7:78-91.
Arshak K, Moore E, Lyons G, Harris J, Clifford SJSr. 2004. A review of gas sensors employed in electronic nose applications. 24:181-198.
Badjagbo K, Loranger S, Moore S, Tardif R, Sauve S. 2010. Btex exposures among automobile mechanics and painters and their associated health risks. Human and Ecological Risk Assessment: An International Journal 16:301-316.
Bhasin R, Cepeda J, Lacasse S. Experiences from cost-effective instrumentation for early warning systems installed in some hot-spot areas in south and south east asian countries. In: Proceedings of the Proceedings of the Second JTC1 Workshop on Triggering and Propagation of Rapid Flow-like Landslides, Hong Kong, China, 2018, 3-5.
Blu T, Thévenaz P, Unser MJIToIP. 2003. Complete parameterization of piecewise-polynomial interpolation kernels. 12:1297-1309.
Bochenkov V, Sergeev G. 2010. Sensitivity, selectivity, and stability of gas-sensitive metal-oxide nanostructures. Metal oxide nanostructures and their applications 3:31-52.
Boelter FW, Xia Y, Persky JDJRA. 2017. A bayesian model and stochastic exposure (dose) estimation for relative exposure risk comparison involving asbestos‐containing dropped ceiling panel installation and maintenance tasks. 37:1729-1741.
Borysiewicz M, Wawrzynczak A, Kopka PJFoC, Sciences D. 2012. Bayesian-based methods for the estimation of the unknown model’s parameters in the case of the localization of the atmospheric contamination source. 37:253-270.
Boyd MJ, Allen JR, Ditzler WR, Busse RJ, Barrett MD. 2009. Distributed ground-based threat detection system.Google Patents.
Briggs GAJAT, Laboratory D. 1973. Diffusion estimation for small emissions.83.
Bullock WH, Ignacio JS. 2006. A strategy for assessing and managing occupational exposures:AIHA.
Buringh E, Lanting R. 1991. Exposure variability in the workplace: Its implications for the assessment of compliance. American Industrial Hygiene Association Journal 52:6-13.
Burnett RT, Dales RE, Brook JR, Raizenne ME, Krewski D. 1997. Association between ambient carbon monoxide levels and hospitalizations for congestive heart failure in the elderly in 10 canadian cities. Epidemiology:162-167.
Cao G, Kesidis G, Porta T, Yao B, Phoha SJSNO. 2006. Purposeful mobility in tactical sensor networks.113-126.
Chang T-Y, Huang G-H. 2009. Exposure to volatile organic compounds and kidney dysfunction in one thin-film transistor-liquid crystal display (tft-lcd) company. Epidemiology 20:S110.
Chen B, Zhang L, Guo G, Qiu XJMPiE. 2015. Kd-acp: A software framework for social computing in emergency management. 2015.
Chen C, Campbell KD, Negi I, Iglesias RA, Owens P, Tao N, et al. 2012. A new sensor for the assessment of personal exposure to volatile organic compounds. Atmospheric environment 54:679-687.
Colbeck I, Nasir ZA, Ahmad S, Ali Z. 2011. Exposure to pm 10, pm 2.5, pm 1 and carbon monoxide on roads in lahore, pakistan.
Cook RD. 2007. Concepts and applications of finite element analysis:John Wiley & Sons.
Diamond D, Coyle S, Scarmagnani S, Hayes J. 2008. Wireless sensor networks and chemo-/biosensing. Chemical reviews 108:652-679.
Eberhart R, Kennedy J. A new optimizer using particle swarm theory. In: Proceedings of the Micro Machine and Human Science, 1995 MHS'95, Proceedings of the Sixth International Symposium on, 1995, IEEE, 39-43.
Eck J, Chainey S, Cameron J, Wilson R. 2005. Mapping crime: Understanding hotspots.
Epperson JFJTAMM. 1987. On the runge example. 94:329-341.
Evans DE, Heitbrink WA, Slavin TJ, Peters TMJAoOH. 2007. Ultrafine and respirable particles in an automotive grey iron foundry. 52:9-21.
Fenech A, Strlič M, Cigić IK, Levart A, Gibson LT, de Bruin G, et al. 2010. Volatile aldehydes in libraries and archives. Atmospheric Environment 44:2067-2073.
Feron VJ, Cassee FR, Groten JP, van Vliet PW, van Zorge JA. 2002. International issues on human health effects of exposure to chemical mixtures. Environmental health perspectives 110:893.
Flesch TK, Wilson JD, Yee EJJoam. 1995. Backward-time lagrangian stochastic dispersion models and their application to estimate gaseous emissions. 34:1320-1332.
Fritsch FN, Carlson REJSJoNA. 1980. Monotone piecewise cubic interpolation. 17:238-246.
Gaertner R, Theriault G. 2002. Risk of bladder cancer in foundry workers: A meta-analysis. Occup Environ Med 59:655-663.
Galer D, Leung H, Sussman R, Trzos R. 1992. Scientific and practical considerations for the development of occupational exposure limits (oels) for chemical substances. Regulatory Toxicology and Pharmacology 15:291-306.
GERHARDSSON GJTAooh. 1988. The future of occupational hygiene—an international overview. 32:1-19.
Grain IJG. 1970. Computer interpolation and contouring of two-dimensional data: A review. 8:71-86.
Hanna SR, Briggs GA, Hosker Jr RP. 1982. Handbook on atmospheric diffusion.National Oceanic and Atmospheric Administration, Oak Ridge, TN (USA ….
Hewett P, Logan P, Mulhausen J, Ramachandran G, Banerjee SJJoo, hygiene e. 2006. Rating exposure control using bayesian decision analysis. 3:568-581.
Hutchinson M, Oh H, Chen W-HJIF. 2017. A review of source term estimation methods for atmospheric dispersion events using static or mobile sensors. 36:130-148.
Ihokura K, Watson J. 2017. The stannic oxide gas sensorprinciples and applications:CRC press.
Jones TJM, control. 1989. Trends in the development of gas sensors. 22:176-182.
Kim J, Yong K. 2011. Mechanism study of zno nanorod-bundle sensors for h2s gas sensing. The Journal of Physical Chemistry C 115:7218-7224.
Kim SWJQB-S. 2017. Bayesian decision analysis of the exceedance probability in the application of occupational exposure limits. 36:39-44.
Koehler KA, Volckens JJAooh. 2011. Prospects and pitfalls of occupational hazard mapping:‘Between these lines there be dragons’. 55:829-840.
Kumar A, Kingson T, Verma R, Mandal R, Dutta S, Chaulya S, et al. 2013. Application of gas monitoring sensors in underground coal mines and hazardous areas. International Journal of Computer Technology and Electronics Engineering 3:9-23.
Kurgan LA, Cios KJJItoK, Engineering D. 2004. Caim discretization algorithm. 16:145-153.
Kyung C-M. 2015. Smart sensors for health and environment monitoring:Springer.
Lin Y, Timchalk CA, Matson DW, Wu H, Thrall KD. 2001. Integrated microfluidics/electrochemical sensor system for monitoring of environmental exposures to lead and chlorophenols. Biomedical microdevices 3:331-338.
Lippmann MJAo, hygiene e. 1996. Advancing the mission: Occupational and environmental hygiene in the information age. 11:1287-1293.
Lovejoy ML, Peeters JP, Johnson AW. 2000. Worker-specific exposure monitor and method for surveillance of workers.Google Patents.
Ma D, Deng J, Zhang ZJAe. 2013. Comparison and improvements of optimization methods for gas emission source identification. 81:188-198.
Malaguarnera G, Cataudella E, Giordano M, Nunnari G, Chisari G, Malaguarnera M. 2012. Toxic hepatitis in occupational exposure to solvents. World journal of gastroenterology: WJG 18:2756.
Miranda AI, Martins V, Cascão P, Amorim JH, Valente J, Tavares R, et al. 2010. Monitoring of firefighters exposure to smoke during fire experiments in portugal. Environment international 36:736-745.
Moyne J, Iskandar JJP. 2017. Big data analytics for smart manufacturing: Case studies in semiconductor manufacturing. 5:39.
Munn RE. 1981. Design of air quality monitoring networks:Springer.
O'Brien DMJAo, hygiene e. 2003. Aerosol mapping of a facility with multiple cases of hypersensitivity pneumonitis: Demonstration of mist reduction and a possible dose/response relationship. 18:947-952.
Pejcic B, Eadington P, Ross A. 2007. Environmental monitoring of hydrocarbons: A chemical sensor perspective. Environmental science & technology 41:6333-6342.
Peters TM, Heitbrink WA, Evans DE, Slavin TJ, Maynard ADJAoOH. 2005. The mapping of fine and ultrafine particle concentrations in an engine machining and assembly facility. 50:249-257.
Pontiggia M, Derudi M, Busini V, Rota RJJohm. 2009. Hazardous gas dispersion: A cfd model accounting for atmospheric stability classes. 171:739-747.
Pouli AE, Hatzinikolaou DG, Piperi C, Stavridou A, Psallidopoulos MC, Stavrides JC. 2003. The cytotoxic effect of volatile organic compounds of the gas phase of cigarette smoke on lung epithelial cells. Free Radical Biology and Medicine 34:345-355.
Qiu S, Chen B, Wang R, Zhu Z, Wang Y, Qiu XJRA. 2017. Estimating contaminant source in chemical industry park using uav-based monitoring platform, artificial neural network and atmospheric dispersion simulation. 7:39726-39738.
Ramachandran G, Banerjee S, Vincent JHJAoOh. 2003. Expert judgment and occupational hygiene: Application to aerosol speciation in the nickel primary production industry. 47:461-475.
Ramachandran GJAoOH. 2001. Retrospective exposure assessment using bayesian methods. 45:651-667.
Ribeiro M, de Paula Santos U, Bussacos MA, Terra‐Filho M. 2009. Prevalence and risk of asthma symptoms among firefighters in são paulo, brazil: A population‐based study. American journal of industrial medicine 52:261-269.
Roberts SJT. 1959. Control chart tests based on geometric moving averages. 1:239-250.
Sarkhosh M, Mahvi AH, Zare MR, Fakhri Y, Shamsolahi HR. 2012. Indoor contaminants from hardcopy devices: Characteristics of vocs in photocopy centers. Atmospheric environment 63:307-312.
Shepherd R, Beirne S, Lau KT, Corcoran B, Diamond D. 2007. Monitoring chemical plumes in an environmental sensing chamber with a wireless chemical sensor network. Sensors and Actuators B: chemical 121:142-149.
Singer MB, Harrison LR, Donovan PM, Blum JD, Marvin-DiPasquale MJSotTE. 2016. Hydrologic indicators of hot spots and hot moments of mercury methylation potential along river corridors. 568:697-711.
Singh AK, Bhadauria BJIJMA. 2009. Finite difference formulae for unequal sub-intervals using lagrange’s interpolation formula. 3:815.
Spitzbart AJTAMM. 1960. A generalization of hermite's interpolation formula. 67:42-46.
Stoer J, Bulirsch R. 2013. Introduction to numerical analysis:Springer Science & Business Media.
Ten Berge WF. 2000. Mathematical models for estimating occupational exposure to chemicals:AIHA.
Thomson LC, Hirst B, Gibson G, Gillespie S, Jonathan P, Skeldon KD, et al. 2007. An improved algorithm for locating a gas source using inverse methods. 41:1128-1134.
Torres C, Jones R, Boelter F, Poole J, Dell L, Harper PJJoo, et al. 2014. A model to systematically employ professional judgment in the bayesian decision analysis for a semiconductor industry exposure assessment. 11:343-353.
Trilles S, Belmonte Ò, Schade S, Huerta J. 2017. A domain-independent methodology to analyze iot data streams in real-time. A proof of concept implementation for anomaly detection from environmental data. International Journal of Digital Earth 10:103-120.
Tsujita W, Yoshino A, Ishida H, Moriizumi T. 2005. Gas sensor network for air-pollution monitoring. Sensors and Actuators B: Chemical 110:304-311.
Vadali M, Ramachandran G, Mulhausen JJJoo, hygiene e. 2009. Exposure modeling in occupational hygiene decision making. 6:353-362.
Vink S, Mikkers J, Bouwman T, Marquart H, Kroese EJRT, Pharmacology. 2010. Use of read-across and tiered exposure assessment in risk assessment under reach–a case study on a phase-in substance. 58:64-71.
Wang F, Li C, Liu W, Jin Y. 2014. Potential mechanisms of neurobehavioral disturbances in mice caused by sub-chronic exposure to low-dose vocs. Inhalation toxicology 26:250-258.
Wang R, Chen B, Qiu S, Zhu Z, Qiu XJA. 2017. Data assimilation in air contaminant dispersion using a particle filter and expectation-maximization algorithm. 8:170.
Werner WJMoc. 1984. Polynomial interpolation: Lagrange versus newton.205-217.
Wilson JD, Sawford BL. 1996. Review of lagrangian stochastic models for trajectories in the turbulent atmosphere.Springer.
Wolkoff P, Wilkins C, Clausen P, Nielsen G. 2006. Abstract. Indoor air 16:7-19.
Won D, Schleibinger H. 2011. Commercial iaq sensors and their performance requirements for demand-controlled ventilation.
Xing J, Liu Z, Huang P, Feng C, Zhou Y, Zhang D, et al. 2013. Experimental and numerical study of the dispersion of carbon dioxide plume. 256:40-48.
Xu C, Tamaki J, Miura N, Yamazoe N. 1991. Grain size effects on gas sensitivity of porous sno2-based elements. Sensors and Actuators B: Chemical 3:147-155.
Yamazoe N. 2005. Toward innovations of gas sensor technology. Sensors and Actuators B: Chemical 108:2-14.
Yau Y, Chew B, Saifullah A. 2012. Studies on the indoor air quality of pharmaceutical laboratories in malaysia. International Journal of Sustainable Built Environment 1:110-124.
Zalk DM, Nelson DI. 2008. History and evolution of control banding: A review. Journal of occupational and environmental hygiene 5:330-346.
Zhang H, Tripathi NKJJoCP. 2018. Geospatial hot spot analysis of lung cancer patients correlated to fine particulate matter (pm2. 5) and industrial wind in eastern thailand. 170:407-424.
Zhao F, Guibas LJ, Guibas L. 2004. Wireless sensor networks: An information processing approach:Morgan Kaufmann.
Zheng C, Bennett GD. 2002. Applied contaminant transport modeling:Wiley-Interscience New York.
Zhu Z, Chen B, Reniers G, Zhang L, Qiu S, Qiu XJIjoer, et al. 2017. Playing chemical plant environmental protection games with historical monitoring data. 14:1155.
Zhu Z, Chen B, Qiu S, Wang R, Wang Y, Ma L, et al. 2018. A data-driven approach for optimal design of integrated air quality monitoring network in a chemical cluster. 5:180889.
天下雜誌編輯群，2018，2018天下兩千大調查 智慧生活爭霸戰，台灣：天下雜誌。