Buried Wireless Sensor Network for Monitoring Pipeline Joint Leakage Caused by Large Ground Movements

Abstract

This paper proposes an innovative buried wireless sensor network (B-WSN) system for detecting leakage from pipeline joints caused by large ground movements such as earthquakes. The key challenge to any such system is that electromagnetic (EM) signal strength becomes significantly attenuated over short distances when wireless devices are buried in certain materials—notably soil, this paper’s focus. After simulation results indicated that the EM radio frequency was a key factor influencing the depth through which a signal can propagate in soil, the B-WSN system was developed, which includes a high-performance sub-1-GHz transceiver that utilizes a low-power band frequency at 433 MHz. Field testing indicated that the BWSN can achieve a penetration depth of 2.13 m. The system configuration includes a radio link budget of 120 dB, transmit power of 26 dBm, receive sensitivity of −125 dBm, and omnidirectional antenna gain of 1.5 dBi. The system works on multihop topology, meaning that each sensing node also acts as a relay node to assist other nodes buried deeper in the ground with data communication. For purposes of this paper, four hops were used, and this made wireless communication possible at an overall burial depth of 8 m. As such, the proposed B-WSN system would be compatible with most buried utility pipelines. The conducted full-scale pipeline-rupture experiment results further verified that the system can, in close to real time, pinpoint locations and subsequent patterns of water leakage caused by severe ground deformation. The findings also exemplify how the B-WSN system could aid structural evaluation of pipelines that are likely to experience large ground deformation. The average packet-loss rate was less than 0.1% during the experiment, and in terms of average power consumption, each sensing node used less than 26.5 mA per 30 s data-reporting period. Thus, the sensing nodes can be expected to function continuously for 27 days if powered by four standard industrial D-cell batteries, or for more than 2 years if the data-reporting period is changed to 1 h.