Optimization of placement of piezoelectric wafers based on a hybrid model using pitch-catch and pulse-echo configurations

The development of Structural Health Monitoring (SHM) systems and integration in our structures is a necessity. It has proven to provide a robust and low-cost solution for monitoring structural integrity and can predict the remaining life of our structures. One of the most important aspects of SHM systems is the design and implementation of sensor networks. This study proposes a new hybrid model for optimizing sensor placement on convex and non-convex structures. We propose a novel framework in which two detection mechanisms are considered: pitch-catch and pulse-echo to provide coverage for a given surface. These two mechanisms will complement each other to minimize the number of sensors used while maintaining a high coverage. This combination also allows for better coverage of the corners and regions in the proximity of geometrical discontinuity (such as holes and openings). The monitored area is discretized into a set of control points. For a control point to be covered, it should satisfy the user-defined coverage level which is the number of sensing paths crossing that point. These sensing paths are provided by two modes of communications (pitch-catch and pulse-echo) between the actuator-sensor pairs. The model, which is solved using a genetic algorithm (GA), provides flexibility by allowing the user to input different parameters such as the attenuation distance of the propagating waves and the sensing path limits of both coverage configurations that can be determined through experimentation. The efficiency of the proposed model is then demonstrated by simulating different geometrical shapes. Significant improvement in the coverage of the monitored area, reaching 34.6%, was achieved when compared to the coverage provided by some preliminary solutions such as uniformly placing the sensors on the plate under study. Also, the advantage of combining both configurations (pitch-catch and pulse-echo) in the same model was investigated. It was shown that the latter highly impacted the coverage in the blind zones (corners and edges) where a single configuration is not effective. Afterward, experimental validation was carried out to evaluate the model’s accuracy in damage localization within the optimized sensor networks. The results demonstrated the proficiency of the model developed in distributing the sensors on the tested specimens.