This project uses the Internet of Things (IoT) technology combined with the miniaturization of sensors to achieve a wide range of intelligent monitoring sensors, and then transmits on-site information to the cloud system through sensing equipment, integrates water quality monitoring big data for analysis, and conducts real-time data through the information system. Achieve real-time pollution warning work to reduce the harm of external water pollutants to the river water environment, reduce the overall impact level, and create an "intelligent river" with innovative thinking. The "109 Yunlin County Water Quality Sensor Joint Test Application Project" is specially formulated to promote and implement related work. During the project period, 9 sensors were sent back for repairs due to equipment failures for a total of 15 times. The reasons for the repairs include abnormal temperature, equipment interruption, equipment or pH cannot be calibrated, and battery charging cable connector burned out. In addition, 42 abnormal equipment events occurred, including: abnormal temperature, abnormal dissolved oxygen (often 0), constant values, equipment interruption, equipment tilted and suspended, being picked up by nearby residents, stranded, stuck under rocks, water in the equipment, equipment was propped up by aquatic plants...etc. During the monitoring period, two water quality abnormalities were discovered, including the high conductivity of the upper reaches of Yunlin Creek and the high temperature of Huilaicuo drainage. The upper reaches of Yunlin Creek were confirmed to be inconsistent with the location of the animal husbandry discharge outlet and the license application location, which violated Article 14 Paragraph 1 of the Water Pollution Prevention and Control Law and was punished. The Huilaicuo drainage did not show any high temperature afterwards, so continue to observe. According to the analysis of on-site inspection and calibration results, the dissolved oxygen value is obviously affected by the growth and coating of the biofilm on the electrode. The correlation coefficient (R2) shows that both cleaning and calibration can improve the accuracy of dissolved oxygen. The R2 of dissolved oxygen before cleaning is 0.4824, but it can reach 0.9712 after calibration. The results show that dissolved oxygen needs to increase its accuracy by increasing the calibration frequency. R2 >0.97 before and after the conductivity sensors cleaning and calibration, its accuracy is high. The pH value is within the error range. Although R2 shows that both cleaning and calibration can improve the pH accuracy, the correlation is not obvious. After calibration, R2 only reaches 0.8391. The temperature cannot be corrected, the temperature value shows low error and high accuracy, R2 >0.99 before and after cleaning. In addition, this plan uses temperature as a variable to perform multiple linear regression analysis of the pH value correlation before cleaning. R2 can be increased from 0.46 to 0.69. The overall data reception rate was 88%, and the data integrity rate was 98%. The work progress of the project has reached 100%, which meets the requirements of the contract.

Water quality sensor, Internet of Things (IoT), Environmental monitoring