|
| 引用本文: | 张丹宁，张猛，张博.2020.基于NARX神经网络的PM2.5/10浓度值预测模型——以咸阳市两寺渡监测站为例[J].地球环境学报,11(2):161-168 |
| ZHANG Danning, ZHANG Meng, ZHANG Bo.2020.The model to predict PM2.5/10 concentrations based on NARX neural network — taking Liangsidu monitoring station in Xianyang as an example[J].Journal of Earth Environment,11(2):161-168 |
|
| |
|
|
| 本文已被：浏览 2739次 下载 1963次 |
码上扫一扫！ |
|
|
| 基于NARX神经网络的PM2.5/10浓度值预测模型——以咸阳市两寺渡监测站为例 |
|
张丹宁，张猛，张博
|
|
西安交通大学 人居环境与建筑工程学院，西安 710049
|
|
| 摘要: |
| PM2.5和PM10（记为PM2.5/10）对空气质量和人类健康有着严重威胁，日益引起国内外的关注，并成为大气污染控制工程中最重要的部分。基于陕西省咸阳市两寺渡监测站的污染物（PM2.5、PM10、NO2、NO、NOx、CO）和相关气象参数的监测数据，建立起基于非线性有源自回归神经网络的预测模型，并分别针对不同预测时间段确定最优网络结构，从而实现了对未来6小时、12小时以及24小时PM2.5/10浓度的有效预测。实验结果表明：（1）NARX神经网络模型可对未来24小时内的PM2.5/10污染物浓度进行较为准确的预测；（2）对于PM2.5/10未来6小时的预测能力优于对12小时、24小时的预测；（3）预测值偏高或偏低的结果与前后时间段内的气象因素及其他污染物浓度变化情况也具有相关性。 |
| 关键词: PM2.5 PM10 空气质量 NARX 递归神经网络 大气污染预测 |
| DOI：10.7515/JEE192013 |
| CSTR：32259.14.JEE192013 |
| 分类号: |
| 基金项目:国家自然科学基金项目（41871315） |
| 英文基金项目:National Natural Science Foundation of China (41871315) |
|
| The model to predict PM2.5/10 concentrations based on NARX neural network — taking Liangsidu monitoring station in Xianyang as an example |
|
ZHANG Danning, ZHANG Meng, ZHANG Bo
|
|
School of Human Settlements and Civil Engineering, Xi’an Jiaotong University, Xi’an 710049, China
|
| Abstract: |
| Background, aim, and scope As known, both PM2.5 and PM10 (denoted as PM2.5/10) are typical fine particles matters, which have serious threats to air quality and human health. To better understand the temporal variations of air quality, this research is dedicated to the prediction of PM2.5/10 concentrations, and we regard Xianyang, Shaanxi, China as study area, taking Liangsidu Monitoring Station as example. Materials and methods In this research, the RNN (recurrent neural network) model based on the NARX (nonlinear autoregressive with external input) method has been proposed by using the hourly monitoring data of pollutants (incl. PM2.5, PM10, NO2, NO, NO x and CO) and meteorology (incl. wind direction, wind speed, temperature, humidity, pressure, etc.). Results Six optimal structures of the neurons in hidden layer have been determined to predict the PM2.5 and PM10 concentrations in the following 6 h, 12 h and 24 h, respectively. The conducted experiments shown that (1) for the PM2.5 prediction of 6 h, the performance becomes the best when the neuron number of hidden layers has been settled as 8, while for the PM2.5 prediction of 12 h and 24 h, the neuron number of hidden layers should be turned to 12 and 7 for the best prediction accuracy; and (2) for the prediction PM2.5 prediction of 6 h, 12 h and 24 h, the optimal settlements of the neurons number in hidden layers are 12, 10 and 13 respectively. Discussion In general, the proposed model has revealed satisfactory performance for the predictions of PM2.5/10 concentrations and the prediction accuracy for the next 6 h is slightly better than that for 12 h and 24 h. Some uncertain predictions, however, still exist especially when unusual meteorological situation occurs. In addition, the data used for the neural-network training are not quite enough. Conclusions It has been demonstrated that the established RNN model based on the NARX network can be implemented to effectively predict the concentration of PM2.5/10: the R values for the PM2.5 predictions of the following 6 h, 12 h and 24 h reaches 0.929281, 0.906767 and 0.889691, respectively; the corresponding RMSE values are 0.0008, 0.0010 and 0.0012; for the prediction of PM10, the R reaches 0.929867, 0.921972 and 0.917757, and the corresponding RMSE are 0.0013, 0.0014 and 0.0017 for the following 6 h, 12 h and 24 h respectively. Recommendations and perspectives In order to further improve the prediction performance of the PM2.5/10 concentrations, the effect of unusual methodological conditions should be considered by the proposed RNN model; moreover, the sensitivity analysis of the different input parameters need to be further investigated. |
| Key words: PM2.5 PM10 air quality NARX recurrent neural network air pollution prediction |
|
|
|
|
|
