湿度对PM_(2.5)/PM_(10)比值影响及MLP神经网络大气污染预测

The Effect of Humidity on PM_(2.5)/PM_(10)and MLP Neural Network Prediction Regarding Air Pollution

颗粒物污染严重影响大气环境质量和公众身体健康.基于郑州、太原、成都2022—2023年秋冬季大气污染和气象数据,讨论了PM_(2.5)、PM_(10)污染特征,通过计算斯皮尔曼系数并进行偏相关分析,得出PM_(2.5)/PM_(10)与相对湿度之间的相关性,并利用MLP神经网络建立关于PM_(2.5)和PM_(2.5)/PM_(10)的预测模型.结果表明,郑州、太原、成都2022—2023年秋冬季PM_(2.5)均值分别为63.52、52.68、50.95μg/m^(3),PM_(2.5)/PM_(10)比值都超过0.5.郑州、太原PM_(2.5)/PM_(10)与相对湿度呈现较强的正相关性.利用MLP神经网络对PM_(2.5)和PM_(2.5)/PM_(10)建立预测模型,误差分别为6.75μg/m^(3)和0.076,模型精度较高.建议未来从秋冬供暖、推广工业水蒸气回收装置、合理开展道路洒水等方面巩固颗粒物污染治理成效,探索和优化机器学习模型用于PM_(2.5)/PM_(10)预测,为大气污染防治提供数据参考,以求发挥大数据在做到精准治污中的作用.

Particle pollution seriously affects atmospheric environment quality and public health.Based on air pollution and meteorological data of Zhengzhou,Taiyuan,and Chengdu from 2022 to 2023(autumn and winter),the characteristics of PM_(2.5)and PM_(10)pollution were discussed.By calculating Spearman correlation coefficients and conducting partial correlation analysis,the relationship between PM_(2.5)/PM_(10)and relative humidity(RH)was investigated.MLP neural network was adopted to establish prediction models for PM_(2.5)and PM_(2.5)/PM_(10).Findings indicated that average PM_(2.5)during autumn and winter of Zhengzhou,Taiyuan,and Chengdu was 63.52μg/m^(3),52.68μg/m^(3)and 50.95μg/m^(3),respectively.The ratios of PM_(2.5)/PM_(10)of the three cities all exceeded 0.5.PM_(2.5)/PM_(10)and RH of Zhengzhou and Taiyuan demonstrated strong correlations.As for MLP neural network,model prediction errors for PM_(2.5)and PM_(2.5)/PM_(10)were 6.75μg/m^(3)and 0.076,respectively,indicating high level of accuracy.Suggestions for enhancing the effectiveness of particulate matter pollution control were promoted in terms of heating in autumn and winter,industrial steam recovery devices,and carrying out reasonable road watering.Besides,machine learning models of PM_(2.5)/PM_(10)prediction should be further explored to provide reference for precise air pollution control.