基于线性回归法预测南方居民夜间合法用水量研究  

Prediction of nighttime water consumption by residents in southern China based on linear regression method

作　　者：马啸 李露 

机构地区：[1]湖北师范大学城市与环境学院,黄石435002 [2]湖北师范大学资源枯竭城市转型与发展研究中心,黄石435002 [3]湖北理工学院土木建筑工程学院,黄石435005

出　　处：《给水排水》2018年第2期87-93,共7页Water & Wastewater Engineering

基　　金：湖北省教育厅科学技术研究项目(Q20172502);资源枯竭城市转型与发展研究中心开放基金(kf2016y06)

摘　　要：DMA小区内无收益水量类型较多,分析困难,对其类型准确分离的关键是确定居民夜间合法用水量。以南方某市8个居民小区17 000余户智能水表实时远传数据为基础样本,采用统计学方法对居民夜间合法用水量进行了研究。户均夜间合法用水量Y与户均日用水量X之间存在较好的线性相关性,其回归方程为Y=7.876 X+0.446(R^2=0.760)。并在6个新安装的智能小区加以验证,整体准确性较高。采用年度平均值实时估计夜间合法用水量时,其误差范围宽度与用户夜间合法用水总量的年度平均值存在显著的线性关系。应用线性回归方程,同时考虑误差范围,可计算得到未装智能远传水表的DMA小区居民夜间合法用水量,从而为小区漏损控制奠定基础。There are many types of non-revenue water consumption in the district metered area （DMA）, which cannot be analyzed easily. The key to accurately analyze the non-revenue water consumption is to determine the residential nighttime water consumption（RNWC）. Based on the real-time remote data samples from more than 17000 households of eight residential quarters in southern cities, the residential nighttime water consumption has been studied by statistical method. The results indicate that the linear correlativity between the average RNWC per family （Y） and the average residential daily water consumption per family （X） is very prominent and the equation of linear regression is Y=7. 876Xq-0. 446 （R2：0. 760）. This equation was verified to be fairly accu- rate on six new residential quarters. In addition, there is significant linear relationship between the annual average RNWC and the estimation error range which will be produced when estimating the daily RNWC using the annual average RNWC. In a word, linear regression prediction method is a kind of RNWC forecasting method, considering the error range at the same time, which can be used to calculate the RNWC in the DMA where intelligent water meters have not been installed. There- {ore, leakage control in DMA can be carried out more easily.

关 键 词：居民夜间合法用水量 线性回归 DMA 漏损控制 

分 类 号：TU991.31[建筑科学—市政工程]