基于LM-BP的供水管网爆管研究  

STDDY ON PIPE BURST DETECTION OF WATER DISTRIBUTION SYSTEM BASED ON LM-BD

作　　者：李诚[1] 周明[1] 刘年东 唐秀秀[1] 

机构地区：[1]昆明理工大学建筑工程学院,昆明650500

出　　处：《低温建筑技术》2017年第2期128-131,共4页Low Temperature Architecture Technology

基　　金：昆明理工大学2016年大学生课外学术科技创新基金课题(2015YC009)

摘　　要：供水管网正常运行是人们生产生活的基本保障,针对供水管网爆管检测问题,采用Water GEMS对实例管网进行模拟仿真,并对管网进行压力监测点的优化布置。以监测点的压力数据作为输入,爆管程度作为输出,证明了经LM法优化的BP神经网络具有更好的收敛速度和稳定性,最后利用LM-BP对实例管网进行爆管检测,给出了实例管网中各管段可检测的最小爆管流量。The normal operation of the water distribution system is the basic demand for people＇ s living and production. For the purpose of pipe burst detection in the water distribution system, the net is simulated with WaterGEMS and the pressure monitoring points is optimized. The BP neural network optimized by LM method is proved to have better convergence speed and stability by using the pressure data of the monitoring points as the input and the burst level as the output. Finally, the LM-BP is used to detect the pipe burst and the minimum burst flow corresponding to each diameter is obtained.

关 键 词：供水管网 爆管 压力监测 LM-BP神经网络 

分 类 号：TU991.32[建筑科学—市政工程]