基于盲系统辨识的供水管网泄漏定位系统研究

管道长度作为已知参量是传统时延估计泄漏定位方法应用的前提,针对实际应用中长度较难测量或不能获得的情况,采用盲系统辨识方法估计泄漏信号传播信道响应函数,并从中提取源信号从漏点传播到采集点的时间信息,从而不依赖管道长度定位漏点.盲辨识中,采用overlap-save和相关函数配准原理构建代价函数,解决高阶信道估计和信道间的病态问题;采用遗传算法对多目标函数进行全局最优化,避免梯度算法收敛陷入局部最小点.基于虚拟仪器技术构建了整个定位系统,实现信号采集、存储、处理及检测信息显示等功能.应用表明,系统解决了管道长度未知情况下的漏点定位问题.

Research on leak location system for water distribution pipelines based on blind system identification

In conventional leak location application that uses acoustic signals in buried water distribution pipelines, the most widely used technique is correlation method. The prerequisite of the method is that the distance between two detection points is known. However, in practice, this prerequisite is not always satisfied. The blind identification strategy is applied to estimate the transmission performances of the two acoustic paths associated with two sensors and the leak source. Then the time for the leak source signal to travel from the leak point to either sensor is also determined. Consequently, for leak location, the above mentioned prerequisite is no longer needed. For blind system identification, the overlap-save and cross-relation fitting techniques are utilized to identify the long impulse response under a built constraint without the prior knowledge of the channel order. In order to avoid converging to the local minima, genetic algorithm is used to minimize the cost functions. An instrument system based on virtual instrumentation implements the functionalities of information acquisition, storage, processing and presentation. The application results show the validness of the proposed scheme.