基于时频分析和改进差分进化的多点泄漏定位方法

当管道同时发生多个泄漏点时,多个泄漏点的声波信号出现混叠,从而影响泄漏声信号的传播和衰减规律。在多点泄漏声波信号混叠的情况下,传统的泄漏定位方法无法对多个泄漏点进行有效定位。为此,提出了一种基于时频分析和改进差分进化的多点泄漏定位方法。由于变分模态分解(variational mode decomposing, VMD)的分解模态个数对特征提取有影响,因此采用能量函数估计分解模态个数,从而对VMD改进。然后采用改进VMD对多点泄漏声波信号进行处理。另外,通过时频分析(time-frequency analysis, TFA)估计多个泄漏点数量,从而得到多点泄漏的定位函数。差分进化(differential evolution, DE)算法易陷入局部最优,在进化后期收敛速度较慢,采用粒子群优化(particle swarm optimization, PSO)算法提高DE的全局收敛速度,并通过改进的差分进化算法估计多个泄漏点的位置。在管道现场进行多点泄漏的模拟试验,试验结果表明该方法能够准确地估计多点泄漏位置,泄漏定位的最小误差为18 m。

Multi-point leakage locating method based on TFA and improved DE

When multiple leakage points occur in a pipeline at the same time,their acoustic signals have aliasing to affect propagation and attenuation laws of these leakage acoustic signals.In aliasing case of multi-point leakage acoustic signals,the traditional leakage locating method can’t effectively locate multiple leakage points.Here,a multi-point leakage locating method based on time-frequency analysis(TFA)and improved differential evolution(DE)was proposed.Due to number of decomposed modes of the variational mode decomposition(VMD)affecting feature extraction,the energy function was used to estimate the number of decomposed modes to improve VMD.Then,the improved VMD was used to process multi-point leakage acoustic signals.In addition,the number of multiple leakage points was estimated using TFA to obtain the locating function of multi-point leakage.Due to DE being easy to fall into local optimization and having a slow convergence speed in the later stage of evolution,the particle swarm optimization(PSO)algorithm was used to improve the global convergence speed of DE,and positions of multiple leakage points were estimated with the improved DE algorithm.Simulation tests of multi-point leakage were conducted on site of pipeline.Test results showed that the proposed method can correctly estimate positions of multi-point leakage;the minimum error of leakage locating is 18 m.