不同三维模型对铅酸蓄电池污染场地土壤Pb空间分布预测的影响

为准确界定污染场地土壤中污染物三维分布范围和受污染土壤土方量并提供相关技术方法和思路,以某典型铅酸蓄电池污染场地为例,对比研究了不同三维插值模型(三维克里格、反距离加权、最邻近点)在不同水平垂直向异性系数设置,对土壤铅污染评价和污染边界分布的影响.研究结果显示,不同模型预测的精度不同,随着水平垂直比值增大,平均误差和均方根误差有增大趋势;通过对比修复目标表明,反距离加权统计的受污染土方量最大,其次是三维克里格和最邻近点,统计值相差约为13%;从污染评价结果和预测精度来看,三维克里格模型和反距离加权要优于最邻近点,选择适宜的模型和合理的参数设置对准确界定污染边界和降低不确定性具有重要影响.

The influence of different three dimensional models on Pb distributing prediction in lead-acid battery contaminated sites

In order to define the three dimensional distribution of pollutant and contaminated earth volume in contaminated sites, and provide the relative technical methods and ideas, a typical lead-acid battery contaminated site was studied as an example. The influence of different three dimensional interpolation models (Krig-3D, Inverse Distance Weighted (IDW), Nearest Neighbor (NN)) in different horizontal and vertical anisotropy coefficient setting conditions on Pb pollution assessment and boundary distribution was examined. Different models had different predictive accuracy, the mean error and root mean square error raised, with the increasing of vertical and horizontal ratio. Furthermore, comparison between the results and the remediation target indicated that the contaminated earth volume was largest by IDW statistical model, followed by Krig-3D and NN models, the statistical values differed by 13%. In terms of the contaminated evaluation results and prediction accuracy, the Krig-3D and IDW were better than NN models. Thus, the selections of the appropriate models and rational parameters had very important effects on the accuracy definition of pollution boundary and reducing uncertainty. This result provided important basis for guiding the remediation and management of the contaminated sites.