Pipeline is an important technique to ensure gas pipeline safe operation. However

because of city gas pipeline network is a complex system

there is currently no effective leak detection method for its security monitoring and leak detection. Although pattern recognition technology is a development trend of future leak detection technology

it has a huge disadvantage of large dimension feature vectors when used in pipeline network. To solve this problem

network partition detection thought is proposed and its detailed application procedure is described for pipeline network leak detection. Finally

an example verifies this method feasibility for leak detection in gas pipeline network.