Abstract
In this paper, we propose a novel circuit with piezoelectric energy harvesters that we developed to power accelerometer sensor nodes placed at optimal locations to maximize energy harvesting, and timely and accurately detect leaks in the pipeline. To reduce energy consumption in sensing and computations, strategies such as duty cycling and a reduction in the number of samples have been incorporated. Due to the conflicting relationship between leak detection accuracy, delay in detecting the leak, and energy consumed by the sensor node, a novel optimization problem is formulated to address this trade-off by linking some crucial design parameters, namely the number of samples per cycle, node sleep time, delay in leak detection, required leak detection accuracy, and remaining sensor node energy. The resulting optimization problem is solved using graphical method. Experimental data is gathered for the harvested energy from a home-grown lab testbed consisting of a Wall-Mounted Pipeline (WMP) system, and various techniques are suggested to increase energy generation. Subsequently, the experimental data is utilized to solve the optimization problem by providing optimal node parameters for a selected remaining node energy after a certain specified number of sampling cycles, as well as a desired leak detection delay and accuracy. The results obtained corroborate our selection of the values of the various parameters used, and the theory underlying our novel optimization problem. The results obtained also provide ample encouragement to pursue different alternative solutions discussed in this paper, as part of our future work in this important practical area.