寒区再生水湿地水质风险因子筛查模型的构建与应用  

The construction and application of risk factors screening model for wetland quality from recycled water in cold regions

作　　者：周继元[1] 史建勋[1] 李大鹏[1] 高艳清 王宪国[1] Zhou Jiyuan;Shi Jianxun;Li Dapeng;Gao Yanqing;Wang Xianguo(Jilin Water Conservacy Research Institude,Changchun 130022,China;Yitong Manchu Autonomous County Water Conservacy Project Ouality Service Center,Yitong 130700,China)

机构地区：[1]吉林.省水利科学研究院,吉林长春130022 [2]伊通满族自治县水利工程质量服务中心,吉林伊通130700

出　　处：《吉林水利》2024年第3期36-42,共7页Jilin Water Resources

基　　金：2021年吉林省科技厅发展计划项目“再生水在寒区湿地中风险因子筛查及治理研究”,项目编号20210203020SF。

摘　　要：筛查再生水水质变异较大的风险因子,存在监测数据量大、评价指标重叠等分析困难问题。本文用多元统计方法构建再生水水质评价的风险因子筛查模型:首先将监测数据转化成无量纲的随机变量,形成协方差矩阵。由于监测数据矩阵一般不是方阵,依据统计模型原理进行矩阵奇异分解降维(SVD)结合最小二乘法拟合观察数据模型,实现了用较少变量再现监测数据的相关关系,简化了多因子系统的综合评价过程。其次,按照构建的数学模型编制了计算程序,用多个基于因子分析的SPSS计算与本模型该段计算进行对比验证,结果没有差异。由于寒区的再生水湿地水生植物冬季休眠,除污水处理厂排水外,湿地基本无其它补水,湿地生态作用降低。因此,项目组利用该程序对不同季节时段、不同位置再生水流过湿地后的监测数据进行了水质风险因子筛查,效果良好。The screening of risk factors for variations in recycled water quality poses a challenge due to the extensive amount of monitoring data and the overlapping evaluation indices.In this study,we developed a risk factor screening model using multivariate statistical methods to evaluate recycled water quality.Firstly,we transformed the monitoring data into dimensionless random variables to create a covariance matrix.Since the monitoring data matrix is highly unlikely non-square,the matrix singular decomposition dimension reduction(SVD)combined with least square method based on statistical principles can be used to fit the monitoring data model.This approach allows for reproducing correlations within the monitoring data using fewer variables and simplifies the comprehensive evaluation process of the multi-factor system.We compiled a calculation program based on this constructed mathematical model and compared its results against SPSS calculations utilizing multiple factor analysis as well as calculations from this model itself,finding no differences between these results.Due to the recycled water wetlands in cold areas,aquatic plants are dormant in winter and there is the limited water replenishment except for sewage treatment plants,and the ecological function of wetlands is reduced.Therefore,the project team used this program to screen the water quality risk factors of the monitoring data after the regenerated water flows through the wetland in different seasons and different locations,and the effect was satisfactory.

关 键 词：再生水 风险因子筛查 因子分析算法模型 

分 类 号：X52[环境科学与工程—环境工程] X820.4