Applying Evolutionary Algorithms to Optimize Active Sensing for Structural Health Monitoring Applications

Evolutionary algorithms are optimization schemes that mimic mechanisms of biological evolution by using the principles of natural selection and survival of the fittest to "evolve" candidate solutions to a given problem and seek out an optimum. Within the field of structural health monitoring, active sensing involves using an actuator/sensor network to interrogate the system with some form of excitation, measure the system's response to the excitation, and mine the response data for features correlating to system health. In this work, we explore optimizing how this active sensing process, or aspects of the process, may be optimized in terms of damage detection. We show how the input waveform may be tailored or shaped via such an algorithm to provide greatly enhanced damage detection sensitivity in the response features, and we discuss how the choice of features affect the tailored input. We present results in a computational and experimental system.