光散射原理的大气PM2.5小型传感器监测性能评估研究

随着大气污染治理的不断深入，精细化、精准化和智能化的环境管理需求不断增加，迫切需要在特定区域开展高密度的监测，以弥补现有传统监测的不足.在此背景下，PM_2.5传感器监测方法在国内迅速发展，为研究光散射原理的大气PM_2.5小型传感器监测性能，于2018年7月—2019年7月在北京市建立了比对测试平台，对不同原理、不同品牌的PM_2.5传感器设备进行为期1年的测试与分析.结果表明：(1)激光粒子计数法PM_2.5传感器性能优于红外法PM_2.5传感器，激光粒子计数法PM_2.5传感器设备与自动标准设备比对的相关系数(R²)均大于0.45，红外法PM_2.5传感器设备与自动标准设备比对的R²均小于0.40.采用激光粒子计数法的工业级PM_2.5传感器数据有效率均在95%以上，更适用于业务化PM_2.5监测.(2)多数激光粒子计数法的工业级传感器设备与标准设备有较好的一...

Performance Evaluation of Atmospheric PM2.5 Light Scattering Sensor

The need for better air pollution control has spurred the need for refined, precise, and intelligent environmental management. Therefore, it is essential to perform high-density monitoring in specific areas to make up for the shortcomings of the existing traditional monitoring. In this context, researchers have adopted different PM_2.5 sensor monitoring methods in China. To evaluate the monitoring performance of small atmospheric PM_2.5 sensors using the principle of light scattering, a comparative test platform was established to test and analyze PM_2.5 sensor monitoring equipment with different principles and brands in Beijing from July 2018 to July 2019. The results are summarized below. First, the PM_2.5 sensor performance of laser particle counting method (R2>0.45) is better than that of infrared method (R2<0.40). The data efficiency of the industrial-grade PM_2.5 sensor is above 95%, which is more suitable for commercial PM_2.5 monitoring. Second, most of the industrial sensors based on the laser particle counting method are consistent with the standard equipment (R2>0.70, standard deviation (SD)<10 μg/m3, coefficient of variation (CV)<15%). The comparison results between PM_2.5 sensors of the same brand and model (R2 higher than 0.97, root mean square error (RMSE) lower than 6.0 μg/m3) are better than the comparison results between sensors of different brands (R2 range: 0.67-0.79, RMSE range: 14.1-23.1 μg/m3). Third, relative humidity has a huge influence on PM_2.5 sensor. With the increase of ambient relative humidity, the absolute error between PM_2.5 sensor and the standard equipment tends to increase. In the high relative humidity range, the median absolute error of each brand sensor exceeds 19 μg/m3. Fourth, under the influence of various factors (such as the chemical composition of particulate matter and environment), the maximum relative deviation between the monitoring results of sensor equipment and the standard equipment in different pollution scenarios is ?22.7%-67.0%. Fifth, the stable operation period of a sensor is generally 6 months for the commercial application of PM_2.5 sensors. It is necessary to establish a suitable quality control calibration system and regularly carry out quality control work based on the localized performance assessment results. This study shows that the laser particle counting method PM_2.5 sensor is suitable for large-scale and high-density grid-based monitoring, but there is a certain gap in data quality compared with standard equipment. In the future, continuous optimization research on hardware and quality control algorithms should be conducted to ensure that the data quality meets the needs of environmental management.