基于时间序列的重庆市PM2.5演变规律分析

发布时间：2018-05-28 12:43

本文选题：PM2.5 + 时间序列　；参考：《重庆理工大学》2015年硕士论文

【摘要】：雾霾已成为中国日常空气污染的突出问题。而PM2.5是雾霾的主要构成成分,也是衡量空气质量的重要指标,其浓度的变化直接反映了空气质量好坏的变化。因此研究PM2.5的形成机理和构成成分就显得极其重要。然而,到目前为止,人们对PM2.5的形成机理和构成成分还没有形成一致认识。因此,从不同的角度运用不同的方法对PM2.5的发生和演变规律进一步研究是必要的。本文运用时间序列分析的方法,对重庆市PM2.5浓度以日期为单位的时序变化规律进行分析。首先根据相关文献资料筛选出与PM2.5相关性较大的变量,如温度、CO、PM10、NO2和SO2等,并通过网络渠道收集相关数据并对数据进行整理和预处理;其次分别研究各个变量序列自身的变化规律,得出各自的适应性模型;再次将PM2.5作为输出变量序列,分别将最高温度、最低温度、CO、PM10、NO2和SO2等作为输入变量序列,构建单输入变量传递函数模型;再逐步构建多输入变量的综合性传递函数模型,用来研究PM2.5与温度、CO、PM10、NO2和SO2之间的内在关系。最后,为了评价上述综合性传递函数模型,分别构建PM2.5与温度、CO、PM10、NO2和SO2之间的普通多元线性回归模型及协整模型,通过三种模型的对比进一步明确了PM2.5与其影响因素之间的关系。结果表明:每一个变量序列本身都不平稳,有其自身的变化规律;PM2.5分别与各个输入变量有显著的相关关系;而且PM2.5与各个因素间的多输入变量传递函数模型的拟合效果更好,充分说明了PM2.5浓度变化规律明显受到各个因素的综合影响。另外,通过多元线性回归模型和协整模型的对比分析可知,这种综合影响不是简单的线性相关关系,PM2.5浓度值不仅受到当期相关因素数值的影响以及随机误差的干扰,而且还要受到其自身及各因素前期数值的显著影响。
[Abstract]:Haze has become a prominent problem of daily air pollution in China. PM2.5 is the main component of haze and an important index to measure air quality. The change of its concentration directly reflects the change of air quality. Therefore, it is very important to study the formation mechanism and composition of PM2.5. However, up to now, there is no consensus on the formation mechanism and composition of PM2.5. Therefore, it is necessary to further study the occurrence and evolution of PM2.5 by using different methods from different angles. In this paper, time series analysis is used to analyze the time series of PM2.5 concentration in Chongqing. Firstly, according to the relevant literature, the variables which have a great correlation with PM2.5, such as temperature, PM10, NO2 and SO2, are selected, and the relevant data are collected through the network channel, and the data are sorted out and preprocessed. Secondly, the variation law of each variable sequence itself is studied, and their adaptive models are obtained. Thirdly, PM2.5 is taken as the output variable sequence, and the highest and lowest temperatures, such as COP10, NO2 and SO2, are taken as input variable sequences, respectively. The transfer function model of single input variable is constructed, and the comprehensive transfer function model of multiple input variables is constructed step by step, which is used to study the relationship between PM2.5 and temperature COP10 PM10NO2 and SO2. Finally, in order to evaluate the comprehensive transfer function model mentioned above, the general multivariate linear regression model and cointegration model between PM2.5 and COP10 / NO2 and SO2 are constructed, respectively. The relationship between PM2.5 and its influencing factors is further clarified by comparing the three models. The results show that each variable sequence itself is not stable and has its own variation law. PM2.5 has a significant correlation with each input variable, and the model of transfer function of multiple input variables between PM2.5 and each factor has a better fitting effect. It is fully explained that the variation of PM2.5 concentration is obviously influenced by various factors. In addition, through the comparative analysis of multivariate linear regression model and cointegration model, it can be seen that this kind of comprehensive influence is not a simple linear correlation relation. The concentration of PM2.5 is not only affected by the value of relevant factors in the current period, but also interfered by random errors. It is also subject to its own and various factors of the value of the previous significant impact.
【学位授予单位】：重庆理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X513

【参考文献】