基于源信号相关特征的供水管网漏损流量研究

The Choice of the Modeling Mechanism Based on Source Signal Distribution Characteristics and Correlation

在供水管网中,无法直接统计物理漏损流量,通过样本数据和概率密度函数研究物理漏损流量统计分布均不可行,而供水总流量的数据易于测量并保留。通过对分析供水总流量,推断出物理漏损流量与用水流量的统计分布特征,并根据二者与水压关系确定相关性。由此,根据分布特征及相关性,可准确从众多盲源分离理论对应的多种不同算法中,选择适用的算法,与Kalman滤波算法结合,实现对物理漏损流量的分离。

In water supply network,cannot direct statistical physics leakage flow,through sample data and research physical leakage flow statistical distribution probability density function are not feasible,and the data is easy to measure and retention of water traffic.This article infers the statistical distribution characteristics of physical loss flow and water flow by analyzing the characteristics of water delivery total flow,as well as determining the relevance according to the relationship between the two flows and water pressure.Thus,according to the distribution characteristics and correlation,the can choose the suitable algorithm form variety blind source separation theory,combined with Kalman filter algorithm,in order to realize the separation of physical loss flow.