環境資源報告成果查詢系統

112年高風險空氣污染事業智慧勾稽實務應用暨建置科技執法教育培訓場域計畫

中文摘要 近年來環境部積極地執行深度稽查,運用現有污染源管理系統勾稽比對及科技工具輔助蒐證後,發現有不肖業者在空污費有不實申報、偽造數據之情況,減少所應繳納之空氣污染防制費用。本計畫持續執行深度稽查,針對PU皮製造程序、膠帶製造程序及凹版印刷程序之事業單位,進行各污染管制系統間之勾稽比對分析。結果顯示,本計畫篩選出的25家事業單位之異常風險值(筆數)較高,其中有15家疑異筆數≥30筆以上,另有6家為空氣污染防制費申報量大於許可核定量1.2倍之情形及8家為單位產品控制前揮發性有機物排放量比值差異大之建議優先查核名單。 本計畫亦執行「產出依製程特性區分產品排放揮發性有機物關鍵因子評估報告」,分析需投入之揮發性有機物排放量之比值,結果指出,凹版印刷程序單位產品控制前,揮發性有機物排放量比值約在160~418公斤VOC/公噸產品,平均值約在290公斤VOC/公噸產品;另PU皮製造程序因查緝數量較少,故其單位產品控制前揮發性有機物排放量比值為119.4公斤VOC/公噸產品;此外,曾遭檢警環查緝之凹版印刷程序及PU皮製造程序事業,於查緝前與查緝後於事業廢棄物申報系統之申報比值,結果發現為單位產品之廢溶劑產生量範圍約為9.14~146公斤廢溶劑/公噸產品,建議設定150公斤廢溶劑/公噸產品作為事業是否有短漏報空污費快速評析之關鍵參數;另、單位產出之廢塑膠混合物範圍約為4%~19%,建議設定20%為快速評析之關鍵參數,俾利後續執法人員利用上述參數快速勾稽高風險事業,並進行查核以減少事業之污染行為。 為精進督察人員執法專業能力,提升打擊環保犯罪績效,及有效掌握工業區或特定地點之工廠煙囪排放影像監控,作為空污自動影像辨識示範,本計畫依指定區域內架設人工智能(AI)影像辨識系統,監控可疑煙囪排放監控,本計畫於112年05月26日完成AI影像辨識架設點之環境現勘作業,於112年06月14日完成設備架設及監控畫面上線,開始蒐集影像畫面進行煙流排放異常模組訓練。本次監控範圍共分為三個區域,彙整統計7月~10月AI影像辨識告警頻率。結果發現符合異常煙流告警頻率較高為A區及B區,建議執法人員得針對該區域之工廠進行現場稽查,確認實際現場運作及防制設備是否依規定操作。 另外,為提升環境執法人員之稽查專業能力,積極追查及遏止環境污染及環保犯罪案件,本計畫辦理了5場次精進環境執法策略相關會議,亦協助建置環境執法訓練數位教材一式,採用360度環景拍攝技術還原稽查現場,本計畫於112年05月05日於臺中市梧棲區OO上膠廠,對該廠空污防制設備汰舊換新及製程改善設施進行3D環景拍攝現勘,並於112年06月02日完成工廠3D環景拍攝,並生成3D模型,透過雲端平台可瀏覽工廠3D虛擬實境之製程及防制設備等畫面,且於112年07月04日邀請專家委員協助進行現場稽查重點教學影片拍攝,透過此方式輔助過往教學時限制於以文字、流程圖及簡報等方式,讓執法人員能透過數位教材進行模擬場景之稽查訓練,以強化督查人員實務經驗,提升教育訓練效能。 為能完整、有效的紀錄稽查、蒐證過程,本計畫已協助機關執行3場次重點污染區域之環境執法相關影像拍攝,並配合製作宣傳影片3支,拍攝主題分別是:第1場輔導製程改善及增設高效能空氣污染防制設備之執行成果,第2場介紹非法棄置智慧圍籬系統及其應用,第3場為檢警環結盟合作共同查緝與偵辦各類環保犯罪案件之歷程。
中文關鍵字 空氣污染防制費,污染源勾稽比對,影像監控辨識

基本資訊

專案計畫編號 經費年度 112 計畫經費 4220 千元
專案開始日期 2023/03/08 專案結束日期 2023/11/30 專案主持人 麥成瑋
主辦單位 環管署環境執法組 承辦人 藍威麟 執行單位 玖進股份有限公司

成果下載

類型 檔名 檔案大小 說明
期末報告 112AA076-成果報告(完整版).pdf 24MB

112-year practical application of smart inspections for high-risk air pollution undertakings and a plan to build a science and technology law enforcement education and training ground

英文摘要 In recent years, the Ministry of Environment (MOE) has actively carried out in-depth audits and applied existing pollution source management systems to assist in evidence collection technologically. It has been exposed that those illegal companies with falsely reported and fake data in reducing the amount of air pollution fees. In 2023, this project continues to implement auditing and conducts comparative analysis with three selected manufacturing processes including tape manufacturing processes, PU leather manufacturing processes, and gravure printing processes. The results indicated that the abnormal risk values (i.e., number of transactions) of the 25 companies screened were relatively high. Among them, 15 companies had more than 30 suspected abnormal transactions, and another 6 companies had air pollution fee 1.2 times greater than the approved amount. There are also 8 companies on the recommended priority lists that had large differences from the volatile organic compound emission ratios. This project also implements the "Assessment Report on Key Factors of Volatile Organic Compound Emissions from Products Diversified by Process Features" to analyze the ratio of volatile organic compound emissions required for input-output balance. The results showed that prior to unit product control of the PU leather manufacturing processes, the volatile organic compound emission ratio was about 160~418 kg VOC per metric ton of product, with average value at about 290 kg VOC per metric ton of product. Due to the lack of inspections in the PU leather manufacturing process, the volatile organic compound emission ratio per unit product before control is 119.4 kg VOC per metric ton of product. This project analyzed and compared those gravure printing processes and PU leather manufacturing processes before and after inspection by the prosecutors and police. The results showed that the range of waste solvent production per unit product is approximately 9.14~146 kg of waste solvent/metric ton of product. It is recommended to set 150 kg as a key parameter for the waste solvent per metric ton of product, if a rapid assessment of shortage in air pollution fees is required. In addition, this project found that the range of waste plastic mixture per unit output was from 4% to 19%. It is thus recommended to set 20% as a key parameter for rapid evaluation, so that subsequent law enforcement personnel can quickly check high-risk companies and conduct inspections in reducing the risk of polluting behavior. This project also set up an Artificial Intelligence (AI) image recognition system in designated areas. The goals are to improve the professional capabilities of inspectors in law enforcement officers, to improve the performance of combating environmental crimes, and to effectively monitor factory chimney emission images in industrial areas or specific locations. To monitor suspicious chimney emissions, this project completed the environmental survey of the AI image recognition installation point on May 26, 2023. On June 14, 2023, the equipment installation and monitoring screen were on-line, and started to collect images and smoke flow. The monitoring scope were divided into three areas (A, B, and C), and statistics of AI image recognition alarm frequency from July 2023 to October 2023 are compiled. The results showed that areas A and B had higher frequency of abnormal smoke flow alarms. It is recommended that law enforcement officers might conduct on-site inspections of factories in those areas to confirm whether the real on-site operations and control equipment are working in accordance with regulations. This project organized five meetings that were related to improving environmental law strategies, and assisted in the construction of a digital teaching material for environmental law enforcement training using 360 panoramic shooting technology in restoring the inspection sites. One of them was OO glue factory in Wuqi District, Taichung City on May 5, 2023. The factory replaced the out-dated control equipment to improve air pollution of the factory. The on-site shooting survey was completed on June 2, 2023, and the 3D environmental shooting of the factory was completed with a 3D model was generated. Through the cloud platform, the 3D virtual reality of the factory's process and control equipment can be browsed, and on July 4, 2023, the experts were invited to assist in the filming of key teaching videos. Through this way, together with flowcharts and presentations can be used as teaching aids, allowing law enforcement officers to conduct inspection training in simulated scenarios, thereby strengthening the practical experience and improving the effectiveness of education and training. To record the inspection and evidence collection process completely and effectively, this project has assisted the MOE to carry out three environmental law enforcement-related video shootings in key polluted areas. The shooting themes are: (1) The implementation results of the training session on process improvement and the high-efficiency air pollution control equipment, (2) The introduction of the illegally abandoned smart fence system and its application, and (3) The coalition of the prosecutors, police and environmental officers in joint investigate on various environmental crime cases.
英文關鍵字 Air pollution control fees,pollution source articulation comparison,Video Surveillance Recognition