摘要
为评估天津市TD校区配水管网（简称目标管网）中消毒副产物的风险,以水中三氯甲烷（CHCl3）为代表性指标,对该管网进行了为期1年的监测.通过联合使用逻辑回归（Logistic回归）和主成分分析（PCA）方法,建立了以耗氯量、总有机碳（TOC）质量浓度、水温、pH值、SUVA和氨氮（NH3-N）质量浓度为自变量的配水管网CHCl3风险预测模型;以不同风险概率下CHCl3质量浓度的分布情况,定义了管网水体的3级CHCl3风险等级,并根据风险概率和所制定的风险等级提出了分级控制标准.通过分析3级风险下模型各自变量的分布情况,确定了影响管网水质CHCl3风险的关键指标.实验结果表明：水中TOC质量浓度、SUVA与目标管网水体的CHCl3风险有显著的正相关性,而氨氮却对CHCl3的生成有抑制作用.在一定范围内,CHCl3生成量随着温度的升高而增加;耗氯量和pH值则对结果的影响不大.统计结果发现,当水中TOC质量浓度、SUVA和氨氮质量浓度分别控制在0~3.8,mg/L、0~1.5,L/（mg·m）和0.25~0.50,mg/L时,可将CHCl3控制在低等风险水平（风险概率为0~0.2）.
In order to assess the disinfection by-product risk of the water distribution network in Tianjin TD campus, the network in this research was monitored for a year with chloroform（CHCl3）as the representative indicator. Through logistic regression and principal component analysis（PCA）,CHCl3 risk prediction model was established with chlorine consumption,total organic carbon（TOC） mass concentration,water temperature,pH value,SUVA and ammonia（NH3-N） mass concentration as independent variables. According to the CHCl3 mass concentration distribution of different risk probabilities,three levels of risk were defined. The hierarchical control standards were put forward based on risk probabilities and levels. The key impact factors of water quality risk were defined by analyzing the distribution of model variables at three risk levels. The results indicate that TOC mass concentration and SUVA have significant positive correlations with CHCl3 risk in this research,while ammonia inhibits the formation of CHCl3. CHCl3 increases with increasing temperature within a certain range. Chlorine consumption and pH value have little influence the CHCl3 formation. The statistical results show that CHCl3 can be controlled at the low risk level（risk probability： 0—0.2）,when TOC mass concentration,SUVA and ammonia mass concentration are in the range of 0—3.8 mg/L, 0—1.5 L/（mg·m）and 0.25—0.50 mg/L,respectively.
出处
《天津大学学报（自然科学与工程技术版）》
EI
CAS
CSCD
北大核心
2014年第4期343-348,共6页
Journal of Tianjin University：Science and Technology
基金
国家自然科学基金资助项目(51178303)
天津市科委社会发展科技计划资助项目(033113111)
关键词
配水管网
逻辑回归
三氯甲烷风险概率
影响因素
water distribution network
logistic regression
chloroform risk probability
impact factor
