基于贝叶斯受体模型的排水管网外水来源解析

Source Apportionment of Extraneous Water in Drainage Networks Based on Bayesian Receptor Modeling

排水管网中外水是影响提质增效的重要因素,为了分析排水管网中外水的来源及占比,将FA-NNC法计算得到的外水构成模式,作为约束施加到贝叶斯受体模型上,提出了一种排水管网外水来源解析的组合模型,并应用在深圳龙岗某污水干管的外水解析中。结果表明:FA-NNC和组合模型均能有效识别出生活污水、地下水和工业污水,组合模型的外水构成比FA-NNC更加接近真实的外水构成;组合模型计算得到的外水占比分别为88.6%、9.6%、1.9%,与现场调查的结果相比,可将FA-NNC的计算误差由31%减小到13%;虽然精度不及CBM法,但可节省污染源的检测费用。该方法可为排查项目中排水管网外水来源解析提供参考。

The extraneous water is an important factor affecting the improvement of quality and efficiency of drainage systems.In order to analyze the source proportion of extraneous water in drainage systems,a combined source apportionment model of extraneous water is proposed by applying a pattern constraint on Bayesian receptor model which is obtained by the FA-NNC method and applied to a sewage main pipe in Shenzhen.The results show that both FA-NNC and the model can effectively identify domestic sewage,groundwater and industrial wastewater.The extraneous water profile of the proposed model is closer to the real profile than FA-NNC.The proportion of extraneous water of the combined model is 88.6%,9.6% and 1.9%.Compared with FA-NNC,the error can be reduced from 31%to 13%.Although the accuracy is worse than that of CBM method,it can save cost of monitoring and.This method can provide a reference for the extraneous water source apportionment of drainage networks in the investigation project.