基于水力瞬变与扩展卡尔曼滤波的管道流体监测与泄漏定位

Pipeline fluid monitoring and leak location based on hydraulic transient and extended kalman filter

引入一种长输管道流体监测与泄漏定位的新方法,将管道流动的瞬变流模型转化为状态空间模型的描述,以管线沿程流量、压强水头为状态变量,管道进口流量和出口压力视做非线性动态系统的控制输入,出口流量和进口压力观测序列构成系统的测量向量。基于小信号原理首先线性化处理非线性模型,然后用扩展的卡尔曼滤波器结合传统的双曲方程特征线解法估计泄漏尺寸与位置,并实时模拟出管道流体的压力流量过程及其沿管道的分布。试验和仿真算例表明此法模拟的管道流动状态能较快收敛到稳定状态,并且泄漏尺寸估计与定位的结果与给定值比较吻合。因此引入扩展的卡尔曼滤波能够提高瞬变流模拟管道非定常流动的准确性和跟踪能力。

In the present paper a new technology of long-distance pipe fluid monitoring and leak location is introduced which can transfer the hydraulic transient of pipe flow to the state model based on state space theory. The process of fluid transfer in the pipeline can be regarded as a nonlinear dynamics system with entry pressure and outlet flow rate considred as stochastic input and output of system respectively. And the measuring model is constructed by observed entry flow rate and outlet pressure series. The nonlinear model was treated with linearization on small signal principle, and then the leak size and leak location can be estimated by the hydraulic fluid model combined with the Extended Kalman Filter （EKF） with traditional characteristic method （CM）. The EKF is strong tracking filter with the fading factor. And the distribution of fluid transient pressure and flow rate along the pipeline is simulated. The results from test and simulated example given out in the present paper shows that this new method can quickly be converged to the steady state during monitoring process and the result of leak location is well couparably consistent with the experimental data. The conclusion is drawn as the introduction of EKF could increase the leak detection degree and improve its robustness in application.