The Prediction Of Soil Heavy Metal Content Based On Neural Network And Study On The Pollution Risk

The soil heavy metal pollution can be a threat to human health,the combination of urban and rural ecological environment is complex,farmland scattered and vulnerable to pollution,this paper is mainly about the study on soil heavy metals contamination risk in Shanghai city of Fengxian District in a farmland.41 samples were collected,after initial treatment,detection of heavy metal Chromium(Cr),Arsenic(As),Nickel(Ni),Lead(Pb),Zinc(Zn),Cobalt(Co)and antimony(Sb)content of 7 kinds of elements,and then predict the contents of 7 kinds of heavy metals in 11 groups pre study area the sampling site was not detected by RBF neural network and BP neural network model two.The first 35 sets of data are taken as the training data,and the latter 6 groups of data are used as the verification.The results show that the RBF neural network prediction is better than the BP neural network prediction model.Multivariate statistical data showed that the 52 groups except Co,Sb no national standard reference value,the average value of Cr,As,Ni,Pb and Zn were not more than two national standard value,but the maximum value of As and Ni were higher than the national standard level two;As,Ni,Zn,average Co and Sb5 elements than the city of Shanghai soil environmental background values of 3 elements,Sb,As,Co in the soil have obvious enrichment.As and Sb reached a high degree of variation,Pb,Zn,Ni,Co reached moderate variability.In the geography of statistical analysis,the content of soil heavy metal elements Cr,As,Ni,Pb,Zn,Co and Sb of the ordinary Kriging interpolation method.The results showed that:in general,in the spatial distribution of heavy metals in soil,the southwest of the study area was mostly the high value area of the elements,but the accumulation of heavy metals was not obvious in the middle east.Correlation analysis,heavy metals Cr-Ni,Cr-Co and Co-Ni 22 has reached a high degree of correlation.The cumulative variance contribution rate of the first 3 factors in the principal component analysis was about 89.044%.The first factor,rotating element of the Ni load is highest,after the rotating load element of Cr is the highest,the highest after the rotating element of the Pb maximum load load before rotating elements As second factors,third factors,before rotating element of the Pb load is highest,after the rotating element of the As maximum load.In the study area,Cr and Pb were not polluted by the cumulative index method,and Sb was in the pollution free and moderate pollution as a whole.The study of pollution load index showed that the study area was in the middle pollution.The potential ecological risk index method study found that the average value of Sb reached 43.61,the moderate risk level has been on the whole,the 50.29 is the maximum value of the As element,achieved the moderate risk level,the maximum value of the remaining elements are more than 40,at a mild level of risk.The average value of RI was 80.29,which was less than 150,indicating that the study area was in a mild ecological risk level.