A new damage detection and tracking method using smart sensor network

Large sensor networks have been proposed as part of structural health monitoring (SHM) systems for infrastructure safety. To apply SHM on large structures such as pipelines, a considerably large number of sensors are required. By increasing the number of sensors, the amount of data to communicate and analyze becomes a burden due to the required computational overhead, power and communication cost. In this paper, a new methodology for detection and tracking capable of minimizing the necessary collected data without compromising damage detection and tracking is presented. Our novel approach combines damage feature correlation and a probabilistic on/off scheme to minimize the required number of activated sensors for damage detection. The amount of preprocessing data to select the on sensors compared to the overall processing is considerably small. Consequently, the new approach minimizes power demand for limiting the amount of data being communicated and further promoting the use of wireless technologies. The randomness of the process leads to an efficient damage tracking method due to minimizing the overall cost of the system. A case study of corrosion damage detection and tracking in a steel pipeline is presented and discussed. It is shown that the proposed method enables successful damage detection and tracking with less than 25 % of the total installed sensors at any time of operation.