In recent years, the Ministry of Environment, the predecessor of the Environmental Protection Administration (EPA) of the Executive Yuan has been at the forefront of deploying the Internet of Things (IoT) for air quality, with nearly 10,000 IoT sensors already deployed across the island. This project aims to utilize advanced smart inspection methods for approximately 3,700 IoT sensors in the central region's five counties and cities. It integrates and analyzes meteorological monitoring, pollution source data, and air IoT data to instantly grasp the potential risks and information about pollution sources, effectively executing pollution inspection operations. The three main objectives of the project include implementing smart inspections, conducting on-site verification using scientific instruments, and widely promoting education and awareness. During the execution, the project team first established the TrackAIR smart inspection platform, integrating crucial air quality and meteorological data from various sources. These data include IoT sensor data from five counties and cities, such as Miaoli County, Taichung City, Changhua County, Nantou County, and Yunlin County, as well as wind field data at various scales, crucial for high-precision meteorological model analysis. Through this platform, the team can instantly identify pollution event hotspots in the region, conduct traceability analysis for each pollution event, and propose a list of 100 high-pollution potential factories for further investigation, considering both the platform-ranked highest emission factories and several important industries or processes. This list is cross-referenced with various permits and declaration data for subsequent on-site investigations. Throughout the project, the team conducted 101 on-site inspections, using infrared thermal imaging thermometers for aerial shots in 21 instances, Gas-FindIR in 33 instances, enhancing on-site inspection efficiency and accuracy. Among the operational factories, 80 were verified, with 4 showing emissions and 20 exhibiting odor issues. The team suggested targets for in-depth factory inspections. Ultimately, the Central Region Management Center conducted in-depth inspections on 11 factories, including testing total hydrocarbons (THC) and non-methane total hydrocarbons (NMHC) in emission pipelines for 7 sessions, gas-phase organic compound sampling and testing (NIEA A722) for 4 sessions, and odor pollutants functional determination for 7 sessions. All 10 companies were successfully reported, achieving a reporting rate of 91%. Among the tested factories, 6 out of 8 did not meet emission standards, resulting in a non-compliance rate of 75%. This validates that the intelligent platform model combining IoT big data algorithms can significantly improve the success rate of inspections and enhance the ability to control pollution sources. In terms of education and promotion, the project team dedicated efforts to raise public awareness of air pollution issues and smart inspection technology. They produced a professional video, wrote 5 press releases, and organized an expert meeting and a seminar, effectively increasing societal interest in the project's outcomes and support for improving air quality.

Internet of Things, Big data, Volatile Organic Compounds, Geographic Information System