郑州环境空气中粉尘浓度的统计分析及预测

对郑州市2005年1月—2009年12月份的可吸入颗粒物(PM10)指数数据进行统计分析,并利用Matlab软件建立了利用气象要素预测PM10的BP神经网络模型。结果表明:2005—2009年郑州市PM10指数的年均值和空气质量超标天数逐步下降,且趋于稳定;年内各月的PM10指数浓度差异很大,冬季PM10指数显著高于夏季,8月最低,而12月最高;采暖期PM10指数显著高于非采暖期,而节假日对于PM10指数的影响不明显;通过平均风速、平均气温、平均气压和平均相对湿度预测PM10浓度可以达到最高精度86.85%。

Statistical Analysis and Prediction for Dust Concentration of Zhengzhou Environmental Air

This paper conduct analyses on Zhengzhou 5 year of PM10 index data(from Jan.2005 to Dec.2009) and Matlab software is applied to set up BP neural network model,which uses meteorological factors to forecast PM10.The results show that the yearly average value of five years and air quality standards-exceeding days gradually decreases and inclines to be stable;PM10 index concentration difference of every month in a year is great and PM10 index in winter is obviously higher than that in summer,lowest in Aug.,but highest in Dec.;PM10 index in heating period is obviously higher than that in non-heating period and it won’t be affected by holidays;4 kinds of meteorological factors,air pressure,temperature,humidity and wind speed greatly affect PM10 index forecast and the precise value can be reached to 86.85% based on average wind speed,temperature,air pressure and relative humidity.