Power-efficient piezoelectric fatigue measurement using long-range wireless sensor networks

In this paper we describe the design of a proof-of-concept wireless embedded sensor system for continuous strain cycle monitoring as a method for fatigue life assessment on civil structures. Monitoring of strain cycles is energy demanding, and therefore not suited to energy-constrained devices, as it requires continuous acquisition of strain data with a high sampling rate, followed by data processing using algorithms for peak-trough detection and cycle counting. To overcome this drawback, at the core of our proposed design is a piezoelectric-based analogue sensor system that can achieve as much as a factor of 9 increase in energy efficiency compared with the conventional approach. The key component is an analogue peak-trough detector that offloads the computation in peak-trough detection from the microcontroller, thus eliminating the need for continuous sampling. The function of the detector is coupled with an energy-efficient interrupt-driven software design for acquisition and strain cycles calculation, which is carried out by using a standard form of the rainflow cycle counting algorithm. For wireless communication and networking, LoRa and LoRaWAN are adopted as core modules. We illustrate the performance of our proposed solution by way of simulation and laboratory experiments. Results show a good agreement in measurement of strain cycles between our proposed system and the conventional approach. Thus, our solution proves to be promising for real fatigue measurement applications. (paper)