Wireless sensor and embedded active diagnosis for structural health monitoring

Wireless smart sensor technologies, which integrate sensors and microprocessors with wireless communication, have become increasingly vital in structural health monitoring (SHM). Despite the fact that researches on wireless sensor for a variety of applications have made significant strides in these years, unique demands for structural health monitoring, in particular for active diagnosis, are not fulfilled. Followed by the discussion on the issues and challenges, e.g., contradiction between limited bandwidth and limited energy, this research investigates power aware solutions in wireless sensor platform and embedded damage localization algorithms.;For wireless sensors, improving sampling bandwidth with minimum power consumption is on focus in the first priority. Based on preliminary researches, a dual-processor based architecture is proposed and implemented to remedy inefficiencies caused by the traditional centralized architecture and serial operations. With parallel controlling, the architecture gains mega-hertz range sampling capability without unnecessary power consumption during data acquisition as the traditional architecture has. The implementation of hardware and software is also discussed in detail. The completion of the wireless sensor platform enables feasibility for active diagnosis where high frequency waves (small wavelengths) should be excited to detect damages in the size of millimeter range.;For embedded SHM algorithms, the imposition of distributed (decentralized) collaborative algorithms is required to cope with the large flux of sensory data, the limited processing capability and energy for wireless sensors. The later part of this work focuses on establishing a power aware damage localization algorithm for isotropic plates.;Higher-order theory is used to formulate the dispersion relation of elastic wave propagation in isotropic plates. Upon investigations on dispersion phenomena, a linear mapping algorithm is developed to remove dispersion, thereby enhancing signal resolution. The dispersion removal further facilitates traditional damage localization methods, such as cross-correlation, to be used in dispersive medium.;An energy decay model based on asymptotic expansion of the wave field is derived for damage localization. With the preclusion of noises, the proposed model provides a simple, yet effective method to obtain propagation distance by measuring the intensity of the diagnostic signal. To successfully embed damage localization into resources-constraint wireless sensors, computational complexity is lessened to expedite calculations and reduce energy consumption by providing exact solutions of the cross-correlation method and the energy decay method for damage localization in isotropic plates.;By embedding the proposed damage localization algorithm into the wireless sensor, experiments are conducted to validate the power efficiency of the proposed wireless sensor and algorithm. To compare the performance, three damage localization scenarios are tested, which are full transmission, embedded cross-correlation and embedded energy decay method. It is shown that the wireless sensor and the embedded energy decay method can construct a laboratory based SHM system with significant less power consumption to localize damage in isotropic plates and the combination holds promise for practical SHM systems.