【摘要】 为挖掘智能水表的水量数据价值，以福州市7个独立计量区（DMA）的居民小区窄带物联网（NB-IoT）水表的远传数据为基础样本，采用统计学方法，探究户均用水量和户均夜间用水量之间的规律，并据此科学预测福州市居民的夜间用水量。结果表明：当采用样本总体平均值法预测居民夜间用水量时，部分小区的估计值误差较大；当采用线性回归法时，户均合法夜间用水量与户均合法用水量、户均有效夜间用水量与户均有效用水量均存在显著的线性相关性。抽样其他3个户数规模不同的居住小区进行验证，证明线性回归法具有普适性且误差小等特点。据此，未装智能远传水表的DMA小区居民夜间用水量能够由此预测推算，从而为小区的漏损量化控制提供科学依据。更多还原

【Abstract】 To exploit the potential value of water consumption data monitored by smart meters, this paper analyzed the remote transmission data of narrow band Internet of Things(NB-IoT) water meters(basic samples) in 7 district metered area(DMA) residential districts in Fuzhou by using statistical methods to explore the relationship between average household water consumption and average household water consumption at night, so as to scientifically forecast the residential water consumption in Fuzhou at night. When sample population mean method was used to predict residential water consumption at night,the estimation errors of some districts were relatively large. When linear regression method was used to forecast, there was a significant linear correlation between average legitimate household water consumption at night and average legitimate household water consumption, and average effective household water consumption at night and average effective household water consumption. Three other residential areas with different household scales were sampled for verification, which proved that the linear regression method had the characteristics such as universality and small error. Therefore,residential water consumption at night can be predicted in DMA community without intelligent remote water meter, which provides scientific basis for the quantitative control of leakage in DMA community.更多还原