建置空氣品質無線監測系統暨室內二氧化碳濃度預測模型

Developing An Air Quality Wireless Monitoring Network System and Indoor Carbon Dioxide Concentration Prediction Model

王愛傑(Ai-Jie Wang)

指導教授 ： 曾昭衡

國立臺北科技大學/工程學院/環境工程與管理研究所/碩士(2012年)

摘要

本研究開發一套空氣品質無線監測系統 (Air Quality Wireless Monitoring Network, iAM)，開發之系統應用於以國立台北科技大學三實場作為案例應用，採用二氧化碳、溫度及相對濕度感測器、ZigBee無線傳輸技術監測軟體等。系統驗證方式包含實驗室準確度誤差測試、現場與直讀式儀器比對，並提出空氣品質無線監測系統設備自動查核機制，以利後續無線監測系統維護。建立二氧化碳濃度質量平衡，預測模型以監測值推估場所室內外換氣量並驗證室內二氧化碳濃度與模型可行性，最後以室內空氣品質監測系統與二氧化碳濃度預測模型進行預測室內最大容許人數及評估場所在不同換氣量下改善室內空氣品質情況。
研究顯示，二氧化碳感測器測試，相對準確度為NDIR原理的二氧化碳感測器較固體電解質原理感測器佳，故選擇NDIR感測器進行佈建。空氣品質無線監測系統與直讀式儀器比對之相關係數為室內R2=0.9722~0.9962、室外R2=0.6257~0.8595，顯示室內干擾較少而準確性高。以二氧化碳濃度預測模型推估三場所室內外換氣量為ZG01模組17.21、17.22、14.02 m3/min；AZ7722感測器16.36、18.88、11.66 m3/min。驗證二氧化碳濃度預測三場所結果MAPE平均值分別為ZG01模組13.1 %、21.15 %、6.93%，AZ7722感測模組11.64 %、18.58 %、7.03%，代表模型預測準確性佳至可接受。應用二氧化碳濃度預測模型推估三場所室內最多容許人數約為17、18和12人，超過人數時，CO2可能高於室內空氣品質第二類場所建議值1000 ppm；本研究另以室內增加抽風扇提升場所室內外換氣量評估室內二氧化碳改善率，預測結果為增加1個抽風扇室內CO2濃度降低約10.4%；增加2個抽風扇室內CO2濃度降低約17.3%。

關鍵字

空氣品質 ； 無線監測系統 ； 感測器 ； 室內二氧化碳預測

並列摘要

A wireless sensor network (WSN) for carbon dioxide (CO2) and a simple prediction model for indoor air quality based on mass balance and indoor ventilation volume are developed in this study. Moreover, the prediction model is evaluated and validated by the WSN and used to construct a warning system to poor indoor air quality. The WSNs, which are incorporated with the ZigBee wireless technology, gas sensors (CO2, temperature, and relative humidity), and air quality monitoring software, were tested at three sites of the National Taipei University of Technology (NTUT). Methods of validation for the prediction model includes laboratory accuracy test and on-site comparison of measurements of gas sensors and direct-reading instruments. In addition, a mechanism of auto maintenance for the WSN is developed for on going research.
From the calibration of CO2 measurements with the direct-reading instrument, the non-dispersive infrared (NDIR) CO2 sensors are applied in the WSNs due to that their measurements were more accurate than the solid electrolyte CO2 ones. The correlation coefficients (R2) between the measurements of WSNs and direct-reading instrument for three sites of NTUT are 0.9722–0.9962 for indoors and 0.6257–0.8595 for outdoors, which means there was certain degree of positive correlation between them. In the prediction model for ZG01 gas sensors, average indoor ventilation volumes for the three sites were estimated to be 17.21, 17.22, and 14.02 m3/min, respectively. On the other hand, in the prediction model for AZ7722 gas sensors, average indoor ventilation volumes for the three sites were estimated to be 16.36, 18.88, and 11.66 m3/min, respectively. The maximum percentage errors (MAPEs) of the prediction model for indoor CO2 concentrations at the three sites were 13.1%, 21.15%, and 6.93 for ZG01 gas sensors and 11.64%, 18.58%, and 7.03% for AZ7722 gas sensors, which show the accuracy of prediction model is acceptable. The maximum capacities for the three sites are determined to be 17, 18, and 12 persons by the prediction model for indoor CO2 concentrations, which will exceed the Level 2 Indoor Air Quality Standard 1000 ppm if there were more than these numbers of people at the three sites. Furthermore, by the prediction model, the indoor CO2 concentrations are estimated to be lowered by 10.4% and 17.3% by the installation of one and two 14-inch ventilation fans.

並列關鍵字

Air Quality ； Wireless Sensor Network ； Sensor ； Prediction Model for Indoor Carbon Dioxide Concentration

參考文獻

[35] 黃志清，室內空氣品質CO、CO2及O3監測儀器及直讀儀器相對準確度測試查核，碩士論文，國立臺北科技大學環境工程與管理研究所，台北，2010。

[75] 劉佳鑫，應用遺傳規劃法於空氣污染物涵容能力之分析，碩士論文，國立台北科技大學環境規劃與管理研究所，台北，2007。

[1] 何政利，智慧無線感測網路系統之研製並應用於室內空氣品質監控，碩士論文，國立臺北科技大學電腦與通訊研究所，台北，2010

[43] 陳筠錚，應用藍芽協定之行動導覽服務，碩士論文，大同大學資訊經營研究所，台北，2005。

[9] 李彥頤，辦公空間室內空氣品質管制策略之研究，博士論文，國立成功大學建築研究所，台南，2004。