Fusion Method Combining Ground-Level Observations with Chemical Transport Model Predictions Using an Ensemble Deep Learning Framework: Application in China to Estimate Spatiotemporally-Resolved PM2.5 Exposure Fields in 2014-2017.

Atmospheric chemical transport models (CTMs) have been widely used to simulate spatiotemporally resolved PM2.5 concentrations. However, CTM results are usually prone to bias and errors. In this study, we improved the accuracy of PM2.5 predictions by developing an ensemble deep learning framework to fuse model simulations with ground-level observations. The framework encompasses four machine-learning models, i.e., general linear model, fully connected neural network, random forest, and gradient boosting machine, and combines them by stacking approach. This framework is applied to PM2.5 concentrations simulated by the Community Multiscale Air Quality (CMAQ) model for China from 2014 to 2017, which has complete spatial coverage over the entirety of China at a 12-km resolution, with no sampling biases. The fused PM2.5 concentration fields were evaluated by comparing with an independent network of observations. The R2 values increased from 0.39 to 0.64, and the RMSE values decreased from 33.7 µg/m3 to 24.8 µg/m3. According to the fused data, the percentage of Chinese population residing under the level II National Ambient Air Quality Standards of 35 µg/m3 for PM2.5 has increased from 46.5% in 2014 to 61.7% in 2017. The method is readily adapted to utilize near-real-time observations for operational analyses and forecasting of pollutant concentrations and can be extended to provide source apportionment forecasts as well.

Performance improvement of a low-frequency vibration generator by using iterative learning control.

This paper presents the performance improvement of a low-frequency vibration generator by using iterative learning control (ILC). A linear motor is designed as a low-frequency vibration generator to calibrate accelerometers. The traditional three-loop control model is first established. The Luenberger observer control method and the closed-loop ILC method are then proposed to improve the performance. Finally, the prototype of this low-frequency vibration system is set up. An accelerometer is calibrated to verify the accuracy of ILC. Subsequently, the total harmonic distortion, amplitude accuracy, and transverse motion of this linear motor vibration generator are tested. Compared with results obtained from the Luenberger observer control, the results derived by the ILC reveal better performance.

A novel method for testing accelerometer transverse sensitivity.

This paper presents a flexure-based compliant mechanism for testing accelerometer transverse sensitivity. The definition of transverse sensitivity is first described. Subsequently, the detailed structure of the developed mechanism is introduced. The principle of this method and the corresponding theoretical model are analyzed. Based on the principle, a prototype is manufactured and an experimental platform is set up. The circular trajectory tests are carried out to verify the feasibility of this method. It shows that the precision of the circular trajectory can be guaranteed. Finally, a three-axis piezoelectric accelerometer is tested. The maximum transverse sensitivity is below 5%, and its maximum measurement uncertainty is 0.15%, while the maximum measurement uncertainty of the corresponding direction angle is 0.79°. It demonstrates that the proposed method is reasonable and accurate.

[Pollution Characteristics of Inorganic Water-soluble Ions in Atmospheric Particulate Matter in Chengdu Plain].

To study the pollution characteristics of water-soluble ions in atmospheric particulate matter in Chengdu Plain, and identify the composition, distribution, time and spatial variation, achieve targeted control of heavy pollution and haze days, 1476 samples were collected at five monitoring sites during August 2013-July 2014, in which eight kinds of inorganic water-soluble ions (SO42-, NO3-, NH4+, K+, Na+, Ca2+, Mg2+, Cl-) were determined by ion chromatography. The results showed that the total mass concentrations of 8 ions in PM2.5-10 and PM2.5 were 11.35 and 36.93µg·m-3, accounting for 37.8% and 46.6% respectively, and SNA (SO42-, NO3- and NH4+) in PM2.5-10 and PM2.5 contributed 81.1% and 89.9% to the total ions, respectively. The concentration of water-soluble ions was highest in winter and lowest in summer. ρ(SO42-)/ρ(PM2.5) was highest in summer and autumn, while ρ(NO3-)/ρ(PM2.5) was highest in winter and lowest in summer. SNA, Cl-, K+mostly distributed in PM2.5, Ca2+ and Mg2+ in PM2.5-10. PM2.5 was generally neutral, the water-soluble ions in which existed as (NH4)2SO4, NH4NO3, KNO3, NaCl, KCl and so on. ρ(NO3-)/ρ(SO42-) revealed that the main source of PM2.5 was given priority to fixed sources. Sulfur oxidation ratio (SOR) and nitrogen oxidation ratio (NOR) were 0.31 and 0.13, respectively, which had opposite changing trend with a highest SOR in summer and NOR in winter. PM2.5 had the characteristics of regional pollution complex, and SNA was the dominant factor causing the increase of ρ(PM2.5).