基于神经网络的土壤重金属含量预测及污染风险研究

发布时间：2018-05-02 07:50

  本文选题：土壤重金属 + 神经网络　； 参考：《昆明理工大学》2017年硕士论文

【摘要】：土壤重金属污染可威胁到人类的健康,城乡结合处生态环境复杂,农田分散且容易受到污染,本文主要是对上海市奉贤区某农田中的土壤重金属污染风险研究。采集41份样品,经初步处理后,检测重金属铬(Cr)、砷(As)、镍(Ni)、铅(Pb)、锌(Zn)、钴(Co)和锑(Sb)7种元素的含量,然后利用RBF神经网络与BP神经网络两种模型预测出研究区域未检测的11组预采样位点的7种重金属的含量。取出前35组数据作为训练数据,后6组数据作为验证,结果显示,RBF神经网络预测优于BP神经网络预测模型。多元统计结果显示52组数据中除Co、Sb没有国家标准的参考值之外,Cr、As、Ni、Pb、Zn的平均值均未超过国家二级标准值,但As、Ni的最大值则超过了环境土壤质量标准中国家二级标准;As、Ni、Zn、Co和Sb5种元素的均值超过了上海市土壤环境背景值,Sb、As、Co 3种元素在土壤中存在明显富集。As和Sb达到高度变异,Pb、Zn、Ni、Co达到中度变异。在地统计分析中,研究区域土壤重金属元素Cr、As、Ni、Pb、Zn、Co和Sb的含量采用普通克里格插值法。结果显示:总体上看,在土壤重金属的空间分布中,研究区域的西南部多为元素的高值区,而中东部地区土壤重金属积累不明显。相关性分析中,重金属元素Cr-Ni、Cr-Co和Co-Ni两两之间已经达到高度相关。主成份分析中前3个因子的累积方差贡献率为89.044%。第1因子中,旋转元素Ni的载荷最高,旋转后元素Cr的载荷最高,第2因子中旋转前元素As的载荷最高旋转后元素Pb的载荷最高,在第三因子中,旋转前元素Pb的载荷最高,旋转后元素As的载荷最高。地累积指数法发现研究区域采集及预测样品中,Cr、Pb未受到污染,Sb整体上处于于无污染-中度污染,研究区域试验样品中污染较为严重。污染负荷指数法研究发现,研究区域整体上处于中等污染。潜在生态危害指数法研究发现Sb的平均值达到43.61,整体上已处于中度风险水平,As元素的最大值是50.29,达到了中度风险水平,其余元素的最大值均为超过40,处在轻度风险水平。研究区综合潜在生态风险指数RI的平均值为80.29,小于150,说明研究区整体上处于轻度生态风险水平。
[Abstract]:Heavy metal pollution in soil can threaten human health. The ecological environment in the combination of urban and rural areas is complex and farmland is scattered and vulnerable to pollution. This paper mainly studies the risk of heavy metal pollution in a farmland in Fengxian District of Shanghai. 41 samples were collected. After preliminary treatment, the contents of 7 kinds of elements, such as Cr, Cr, as, Ni, Pb, Zn, Co) and SB ~ (3 +), were determined. Then, RBF neural network and BP neural network were used to predict the contents of 7 heavy metals in 11 groups of presampled sites in the study area. The first 35 groups of data are taken as training data and the last 6 groups of data are used as verification data. The results show that RBF neural network prediction is better than BP neural network prediction model. The results of multivariate statistics show that the average value of Cr ~ (2 +) -As-Ni ~ (2 +) Pb ~ (2 +) in 52 groups of data does not exceed the national second class standard value except that there is no reference value of the national standard for CoSb. However, the maximum value of As-Ni is higher than that of the national secondary standard, As-NiNiZn-Co and Sb5, in the environmental soil quality standard, which is higher than the soil environmental background value of Shanghai, where there are significant enrichment of. As and SB in the soil. The variation was moderate. In the geostatistical analysis, the contents of heavy metal elements in the soil of the region were studied by the ordinary Kriging interpolation method. The results showed that, in the spatial distribution of soil heavy metals, the southwest of the study area was mostly the high value area of elements, but the accumulation of heavy metals in the middle and eastern regions was not obvious. In the correlation analysis, there is a high correlation between the heavy metal elements Cr-NiCr-Co and Co-Ni. The cumulative variance contribution rate of the first three factors in principal component analysis was 89.04444. In the first factor, the load of rotating element Ni is the highest, the load of rotating element Cr is the highest, the load of element as before rotation is the highest in factor 2, the load of element Pb after rotation is the highest, and the load of element Pb before rotation is the highest in the third factor. The load of as is the highest after rotation. It was found by the method of geoaccumulation index that the samples collected and predicted in the study area were not polluted and the SB was generally in the range of no pollution and moderate pollution, and the pollution in the samples of the study area was more serious. The study of pollution load index method found that the study area is in the middle pollution as a whole. The study of potential ecological hazard index showed that the average value of SB reached 43.61, and the maximum value of element as was 50.29, which reached the level of moderate risk, and the maximum value of other elements was over 40, which was at the level of mild risk. The average value of the comprehensive potential ecological risk index RI is 80.29, which is less than 150, which indicates that the study area is at the level of light ecological risk as a whole.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X53

【参考文献】

中国期刊全文数据库 前10条

1 宁翠萍;李国琛;王颜红;李波;田莉;王世成;;细河流域农田土壤重金属污染评价及来源解析[J];农业环境科学学报;2017年03期

2 闫慧;沈宁娟;;磁化率对城郊耕地土壤重金属污染的指示研究——以许昌市为例[J];地球与环境;2016年06期

3 李晓艳;吴超;;某铅锌矿区公路两侧土壤重金属污染分布研究[J];环境工程;2017年01期

4 张静;邹滨;陈思萱;王泽斌;柴立元;;土壤重金属污染风险时空变化模拟与分析[J];测绘科学;2016年10期

5 虞敏达;张慧;何小松;张媛;马丽娜;檀文炳;高如泰;;典型农业活动区土壤重金属污染特征及生态风险评价[J];环境工程学报;2016年03期

6 张连科;李海鹏;黄学敏;李玉梅;焦坤灵;孙鹏;王维大;;包头某铝厂周边土壤重金属的空间分布及来源解析[J];环境科学;2016年03期

7 段宁;杨思言;魏婉婷;;基于BP神经网络的铅酸蓄电池厂地下水重金属浓度预测[J];环境科学与技术;2016年01期

8 刘硕;吴泉源;曹学江;王集宁;张龙龙;蔡东全;周历媛;刘娜;;龙口煤矿区土壤重金属污染评价与空间分布特征[J];环境科学;2016年01期

9 李宝磊;张榆锋;施心陵;章克信;张俊华;;一种用于山区土壤重金属评估的集成RBF空间插值算法(英文)[J];Journal of Southeast University(English Edition);2015年01期

10 戴彬;吕建树;战金成;张祖陆;刘洋;周汝佳;;山东省典型工业城市土壤重金属来源、空间分布及潜在生态风险评价[J];环境科学;2015年02期

中国博士学位论文全文数据库 前2条

1 卢金娜;基于优化算法的径向基神经网络模型的改进及应用[D];中北大学;2015年

2 孙崇玉;吉林省典型黑土区农田土壤重金属环境风险研究[D];中国科学院研究生院(东北地理与农业生态研究所);2013年

中国硕士学位论文全文数据库 前1条

1 丛艳静;基于GIS技术的兴化平原耕地土壤重金属污染空间分异与环境风险评价[D];福建农林大学;2006年

，

本文编号：1833014