基于微分方程和时间序列的PM2.5预测模型
|
摘要
为了准确描述空气污染物PM2.5的浓度变化规律,运用一阶微分方程和时间序列理论,建立PM2.5浓度预测模型。通过OLS线性回归试解模型参数避免陷入局部最优陷阱,最后求解得到固有削减率r=0.7554,理论最大浓度xm=91.4804μg/m~3。并利用北京市2012～2017年的PM2.5年份浓度数据进行检验和预测。预测结果显示,拟合最大误差率为1.194%,模型准确可靠,北京市2021年PM2.5浓度将会达到7.19μg/m~3,达到世界卫生组织(WHO)公布的PM2.5小于10的安全值标准。
In order to accurately describe the variation of PM2.5 concentration of air pollutants, a prediction model of PM2.5 concentration was established by using first-order differential equation and time series theory. OLS linear regression was used to solve the model parameters to avoid falling into the trap of local optimum. Finally, the inherent reduction rate r = 0.7554 and theoretical maximum concentration xm = 91.4804 μg/m~3 were obtained. The concentration data of PM2.5 from 2012 to 2017 in Beijing were used to test and forecast. The prediction results show that the maximum fitting error rate is1.194%. The model is accurate and reliable. The PM2.5 concentration in Beijing in 2021 will reach 7.19μg/m~3, which meets the safety value standard of PM2.5 less than 10 published by the World Health Organization(WHO).
In order to accurately describe the variation of PM2.5 concentration of air pollutants, a prediction model of PM2.5 concentration was established by using first-order differential equation and time series theory. OLS linear regression was used to solve the model parameters to avoid falling into the trap of local optimum. Finally, the inherent reduction rate r = 0.7554 and theoretical maximum concentration xm = 91.4804 μg/m~3 were obtained. The concentration data of PM2.5 from 2012 to 2017 in Beijing were used to test and forecast. The prediction results show that the maximum fitting error rate is1.194%. The model is accurate and reliable. The PM2.5 concentration in Beijing in 2021 will reach 7.19μg/m~3, which meets the safety value standard of PM2.5 less than 10 published by the World Health Organization(WHO).
引文
[1]刘小生,李胜,赵相博.基于基因表达式编程的PM_(2.5)浓度预测模型研究[J].江西理工大学学报,2013,34(5):1-5.
[2]王媛.PM2.5的危害及其防治措施分析[J].山东工业技术,2017(6):37.
[3]孙柏峰.吉林省某市空气质量评价及预测[D].长春:吉林大学,2013.
[4]陈丽.基于时空数据模型的PM2.5预测[J].仪器仪表学报,2015,36(增刊):4.
[5]张怡文,胡静宜,王冉.基于神经网络的PM_(2.5)预测模型[J].江苏师范大学学报:自然科学版,2015,33(1):63-65.
[6]王凯.二元函数的可微性与一元函数可微性的关系[J].岳阳师范学院学报:自然科学版,2001(2):14-15.
[7]李培志.PM2.5浓度预测模型的研究与应用--基于人群搜索算法优化后支持向量机模型[J].调研世界,2016(2):19-24.
[8]王振波,方创琳,许光,等.2014年中国城市PM_(2.5)浓度的时空变化规律[J].地理学报,2015,70(11):1720-1734.
[9]强琳,刘凯,张瑜,等.线性回归法建立城市空气PM_(2.5)预报模型--以宝鸡市为例[J].能源与节能,2014(2):125-127.
[10]司志娟,孙宝盛,李小芳.基于改进型灰色神经网络组合模型的空气质量预测[J].环境工程学报,2013,7(9):3543-3547.
[11]宋宇辰,甄莎.BP神经网络和时间序列模型在包头市空气质量预测中的应用[J].干旱区资源与环境,2013,27(7):65-70.
[12]苏键,陈军,何洁.主成分分析法及其应用[J].轻工科技,2012,28(9):12-13,16.
[13]付强.主成分分析在环境空气质量评价中的应用.中国环境科学学会.2011中国环境科学学会学术年会论文集(第二卷)[C].中国环境科学学会,2011:3.
[2]王媛.PM2.5的危害及其防治措施分析[J].山东工业技术,2017(6):37.
[3]孙柏峰.吉林省某市空气质量评价及预测[D].长春:吉林大学,2013.
[4]陈丽.基于时空数据模型的PM2.5预测[J].仪器仪表学报,2015,36(增刊):4.
[5]张怡文,胡静宜,王冉.基于神经网络的PM_(2.5)预测模型[J].江苏师范大学学报:自然科学版,2015,33(1):63-65.
[6]王凯.二元函数的可微性与一元函数可微性的关系[J].岳阳师范学院学报:自然科学版,2001(2):14-15.
[7]李培志.PM2.5浓度预测模型的研究与应用--基于人群搜索算法优化后支持向量机模型[J].调研世界,2016(2):19-24.
[8]王振波,方创琳,许光,等.2014年中国城市PM_(2.5)浓度的时空变化规律[J].地理学报,2015,70(11):1720-1734.
[9]强琳,刘凯,张瑜,等.线性回归法建立城市空气PM_(2.5)预报模型--以宝鸡市为例[J].能源与节能,2014(2):125-127.
[10]司志娟,孙宝盛,李小芳.基于改进型灰色神经网络组合模型的空气质量预测[J].环境工程学报,2013,7(9):3543-3547.
[11]宋宇辰,甄莎.BP神经网络和时间序列模型在包头市空气质量预测中的应用[J].干旱区资源与环境,2013,27(7):65-70.
[12]苏键,陈军,何洁.主成分分析法及其应用[J].轻工科技,2012,28(9):12-13,16.
[13]付强.主成分分析在环境空气质量评价中的应用.中国环境科学学会.2011中国环境科学学会学术年会论文集(第二卷)[C].中国环境科学学会,2011:3.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |