管道泄漏检测及定位算法的改进和应用

为了提高现有输油管道负压波检测系统的灵敏度和检测性能，在不增加硬件的基础上，采用分段差值的方法确定是否发生泄漏，基于相似性原理计算信号间的欧式距离，由此确定压力信号至上、下游的时差，并通过现场泄油试验进行验证。结果表明，原有算法只能检测到突发泄漏，但对于泄漏量小于1.1%的缓慢泄漏工况未能有效检测；改进算法后可检测微小缓慢泄漏，当采样点数移动步长为5时，定位的绝对误差为0.52～2.47 km，标准差为0.95 km，定位结果误差较小。该方法适合资源受限的嵌入式管道泄漏检测方法的改进。

Improvement and Application of Pipeline Leak Detection and Location Algorithm

In order to improve the existing pipeline negative pressure wave detecting sensitivity and detection performance of the system, without any increase in hardware, on the basis of using segmented difference method to determine whether the leaking, based on the similarity theory to calculate the Euclidean distance between signals, determine signals to upload first, downstream of the jet lag, and is verified by field drainage experiment.The results show that the original algorithm can only detect the burst leakage, and can not detect the slow leakage condition with the leakage volume less than 1.1%. The improved algorithm can monitor small and slow leakage. When the moving step of sampling points is 5, the absolute error of location is 0.52~2.47 km, and the standard deviation is 0.95 km, and the error of location result is small. This method is suitable for the improvement of embedded pipeline leakage method with limited resources.