针对目前南水北调中线工程高填方渠道渗漏监测设备综合误差大、不能监测渠道断面间渗漏等问题, 设计了 可用于高填方渠道渗漏的可移动无损监测系统, 建立了高填方渠道渗漏状态监测的 Kalman2BP 融合模型。首先构 建一种基于无线传感网的多区域渗漏信息检测平台, 将传感器设计成便携式可移动的锥形设备, 对渗漏区域的温湿 度、土壤含水率、GPS 位置信息以及渗流等信息进行实时采集, 再通过 ZigBee 和 GPRS 将多传感器信息进行无线传 输; 并结合流场渗漏检测方法, 通过试验模型筛选出与高填方渠道渗流相关的特征变量; 使用卡尔曼( kalman) 算法 对关联的物理变量进行滤波和估值; 最后将多传感器数据通过 BP 神经网络进行渗漏状态模式识别, 实现渗漏的状 态预测, 确定坡面渗漏安全级别。试验结果表明, 基于 Kalman2BP 融合模型的高填方渠道渗漏监测模型识别误差 较小, 达到能在整体上实时监测高填方渠段的渗流状态, 可实现南水北调中线工程高填方渠道断面间的坡面渗流非 破坏性在线监测功能。

To so lv e the pro blems of the leakag e monitoring equipment fo r hig h2filled canals in pro ducing lar ge comprehensive er2 r or and being unable to mo nitor the seepage between canal sect ions, w e desig ned a mov able and non2dest ruct ive leakage monito2 ring sy stem for the hig h2filled canal o f the Middle Ro ut e of the So uth2to2No rth Water Diversion Pro ject and established a fusio n model based on Kalman2BP for leakage monito ring o f hig h2filled canals. Firstly, we const ruct ed a mult i2zo ne leakag e info rmatio n detectio n platform based o n wireless sensor netw ork, and w e desig ned the sensors as po rtable and movable cone devices t hat can be inserted into the soil. T he info rmation of temperatur e, humidity, soil w ater content, seepag e, and GPS lo cation w as collected in real t ime and then w as transmitted wirelessly thro ug h t he ZigBee and GPRS. Using the flow field leakag e detectio n method, we selected the characteristic variables that w ere relevant t o high2filled canal leakag e t hr ough the ex perimental model. Then, w e used Kalman algo rithm t o filter and valuate the associated phy sical variables. Finally, we submitted the multi2senso r data t o the BP neural netw or k for leakag e state patter n r eco gnition and realized the predictio n of slo pe leakag e state and determined the safety lev el o f slope leakag e. The experimental r esults show ed that the Kalman2BP fusion mo del has smaller erro r in r eco gnizing the leakag e o f the hig h2f ill canal, and can monito r in real t ime the leakage state betw een the canal sections. It can r ealize no n2de2 str uctive online mo nitoring of t he slope seepage of t he Middle Ro ut e o f the So ut h2to2Nort h Water Diversion Pro ject.

河南省高等学校重点科研项目计划( 15A510003) ; 河南省科技攻关计划( 172102210050 ) ; 水利部黄河泥沙重点实验室开放课题 基金( 2017001) ; 国家科技重大专项课题( 2014ZX03005001) ; 河南省高等学校重点科研项目计划( 14B170012)