자동측정망 자료를 활용한 도시철도 공기질 실/내외 영향 분석과 농도 변화 예측 (Prediction of Concentration Change and Influence Analysis of Indoor/Outdoor in Subway Station using Indoor Air Quality Monitoring Data)

서지정보

· 발행기관 : 한국대기환경학회
· 수록지 정보 : 한국대기환경학회지 / 38권 / 1호 / 30 ~ 45페이지
· 저자명 : 박덕신, 김종범, 박세찬, 이용일, 이선엽, 김정호

초록

Concern and caution for the atmospheric environment has been increasing recently; the air quality in urban subways, a major means of transportation in large cities, is a major concern. To manage this situation, the Korean government has installed real-time monitoring devices at all subway stations and its data is made available to the public. In this study, we carried out an influence analysis between the indoor and outdoor environment, and future concentration prediction (1 hour later) using machine learning; real-time data was measured at Suyu station. PM10 concentration on a platform at Suyu station was 146.0 μg/m3, exceeding the indoor air quality standards. The annual average concentration of CO2 was 530 ppm, which was below the indoor air quality standards. The correlation analysis between pollutants and measurement points showed that PM10 had a high correlation coefficient for train passing number (TPN), tunnel, concourse, and platform. NO showed high correlation for concourse, platform, and ambient air. The prediction results (R2) for big data obtained using machine learning was 0.69. We confirmed that it is possible to predict indoor air quality of subway stations by employing machine learning and real-time monitoring data. In future, the results of this study can be used as basic information for establishing an indoor air quality management plan for subway stations.

영어초록

Concern and caution for the atmospheric environment has been increasing recently; the air quality in urban subways, a major means of transportation in large cities, is a major concern. To manage this situation, the Korean government has installed real-time monitoring devices at all subway stations and its data is made available to the public. In this study, we carried out an influence analysis between the indoor and outdoor environment, and future concentration prediction (1 hour later) using machine learning; real-time data was measured at Suyu station. PM10 concentration on a platform at Suyu station was 146.0 μg/m3, exceeding the indoor air quality standards. The annual average concentration of CO2 was 530 ppm, which was below the indoor air quality standards. The correlation analysis between pollutants and measurement points showed that PM10 had a high correlation coefficient for train passing number (TPN), tunnel, concourse, and platform. NO showed high correlation for concourse, platform, and ambient air. The prediction results (R2) for big data obtained using machine learning was 0.69. We confirmed that it is possible to predict indoor air quality of subway stations by employing machine learning and real-time monitoring data. In future, the results of this study can be used as basic information for establishing an indoor air quality management plan for subway stations.

참고자료

· 없음