State Key Laboratory of Environmental Criteria and Risk Assessment, Chinese Research Academy of Environmental Sciences, Beijing 100012, China.
Jinan Ecological Environment Monitoring Center, Shandong Province, Jinan 250013, China.
Sci Total Environ. 2020 May 1;715:136791. doi: 10.1016/j.scitotenv.2020.136791. Epub 2020 Jan 20.
With the development of the air pollution control, the low-cost sensors are widely used in air quality monitoring, while the data quality of these sensors is always the most concern for users. In this study, data from nine air monitoring stations with standard PM instruments were used as reference and compared with the data of mobile and fixed PM sensors in Jinan, the capital city of Shandong Province, China. Data quality of PM sensors was checked by the cross-comparison among standard method, fixed and mobile sensors. And the impacts of relative humidity and size distribution (PM/PM) on the performance of PM sensors were evaluated as well. To optimize the calibration method for both fixed and mobile PM sensors, a two-step model was designed, in which the RH and PM/PM ratio were both used as input parameters. We firstly calibrated the sensors with five independent models, and then all the calibrated data were linearly fitted by the LR-final model. In comparison with standard instruments, the LR-final model increased the R values of the PM and PM measured by fixed sensors from 0.89 and 0.79 to 0.98 and 0.97, respectively. The R values of PM and PM measured by the mobile sensors both increased to 0.99 from 0.79 and 0.62. Overall, the two-step calibration model appeared to be a promising approach to solve the poor performance of low-cost sensors.
随着空气污染控制的发展，低成本传感器在空气质量监测中得到了广泛应用，而这些传感器的数据质量一直是用户最关心的问题。本研究以中国山东省省会济南市的九个具有标准 PM 仪器的空气质量监测站的数据作为参考，与移动和固定 PM 传感器的数据进行了比较。通过标准方法、固定传感器和移动传感器之间的交叉比较来检查 PM 传感器的数据质量。并评估了相对湿度和粒径分布（PM/PM）对 PM 传感器性能的影响。为了优化固定和移动 PM 传感器的校准方法，设计了两步模型，其中 RH 和 PM/PM 比均作为输入参数。我们首先用五个独立模型对传感器进行校准，然后用 LR-最终模型对所有校准数据进行线性拟合。与标准仪器相比，LR-最终模型将固定传感器测量的 PM 和 PM 的 R 值从 0.89 和 0.79 分别提高到 0.98 和 0.97。移动传感器测量的 PM 和 PM 的 R 值均从 0.79 和 0.62 分别提高到 0.99。总体而言，两步校准模型似乎是解决低成本传感器性能不佳的一种有前途的方法。
