摘要
以PM2.5扩散、衰减模式为研究对象,分析探究了PM2.5的扩散规律、危机治理及其后5年的治理问题.首先通过主成分分析法,建立了PM2.5与其它污染物之间的多元非线性对数模型.同时引入相对湿度的影响因素对模型进行再度优化,提高了模型的拟合优度.运用统计学原理,得出采集点之间的PM2.5具有较高的协同性.另外分析了静态下PM2.5污染物颗粒的受力和漂移模式和从点源、面源两方面分析了PM2.5动态扩散模式,建立了PM2.5的扩散偏微分方程模型.根据建立的扩散模型,对突变的污染物浓度确定安全区域的范围.最后建立综合费用和专项费用的多目标优化模型,利用贝叶斯支持向量机方法对PM2.5进行宏观预测,并运用系统动力学理论对目标值进一步优化,并对不同治理模式进行对比分析.
This paper is focused on the PM2.5 diffusion and decay mode, analysis itsdiffusion rule, crisis governance and governance after 5 years. We first established the multivariate nonlinear logarithmic model between PM2.5 and other pollutants by principal component analysis. Meanwhile, introduce the relative humidity factors to re-optimization the model, improves the goodness of fit of the model. Then apply the statistical principles and come to the conclusion that the PM2.5 collection point has a high synergistic. In addition, we analysis the force and drift mode of PM2.5 particulate pollutants under static condition, and the dynamic diffusion mode of PM2.5 from point and non-point source. Then, we established a diffusion partial differential equation model of PM2.5. Based on the diffusion model, we determine the safety range of the mutationalpollutant concentration. Last, we build the multi-objective optimization model about comprehensive cost and special cost, make a macroscopic prediction of PM2.5 by using Bayes support vector machines method, further optimize the target value by system dynamics theory andcompare and analyze the different governance models.
出处
《数学的实践与认识》
CSCD
北大核心
2014年第15期37-46,共10页
Mathematics in Practice and Theory
关键词
多元非线性对数
主成分分析法
贝叶斯支持向量机
系统动力学理论
multiple linear logarithmic
principal component analysis
bayes support vector machines
system dynamics theory
