Multi-Sensor Data Fusion and Target Location in Pipeline Monitoring and a Pre-Warning System Based on Multi-Seismic Sensors

As a safe and economical method the pipeline is widely used to transport the oil and gas. At present, most of serious pipeline accidents are the artificial damages. The traditional leakage detection methods only infer the leakages by analyzing the parameters. The sensitivities are low and the leakage warning exclusively arrives after the leakage. The urgent problem is to detect the intrusions and locate the threat position. For this reason, we studied and developed a novel pipeline security monitoring and pre-warning system based on multi-seismic sensors. The system consists of many seismic geophones and data acquisition modules which are laid along the pipeline. The events such as footstep, car moving and the digging generate the seismic waves. These sensors and modules acquire the seismic signals. The empirical mode decomposition is used to acquire the IMFs (intrinsic mode functions). Then the IMF kurtosis features are normalized and used to recognize the different targets. In order to improve the individual target recognition accuracy, it proposes to fuse the identification results by using the Dempster-Shafer evidence theory and obtain the final judgment. If the target belongs to threat, the location will be finished. The time-frequency analysis method is used to process the seismic signals. The EMD outcomes are transformed by Hilbert-Huang Transform and then it can acquire the signal's instantaneous properties such as instantaneous frequency, phase and amplitude. In this paper we also discuss the localization procedures and put forward to a new passive localization method based on HHT and TDOA. The filed trails show the proposed methods work well.