Koh C.G., Perry M.J. Structural Identification and Damage Detection using Genetic Algorithms

CRC Press, 2010. — 151.Structural health monitoring has become a growing R&D area, as witnessed by the increasing number of relevant journal and conference papers. Rapid advances in instrumentation and computational capabilities have led to a new generation of sensors, data communication devices and signal processing software for structural health monitoring. To this end, a crucial challenge is the development of robust and efficient structural identification methods that can be used to identify key parameters and hence, cause change of structural state. There are currently many competing methods of structural identification, both classical and non-classical. Based on our research efforts for over more than a decade, the genetic algorithms (GA) have been found to possess many desired characteristics and offer a very promising way to tackle real systems. It is the intention of this book, believed to be the first on this topic, to provide readers with the background and recent developments on GA-based methods for parameter identification, model updating and damage detection of structural dynamic systems.
Of significance, a novel identification strategy is developed which contains many advantageous features compared to previous studies. The application of the strategy focuses on structural identification problems with limited and noise contaminated measurements. Identification of systems with known mass is first presented to provide physical insight into the effects of various numerical parameters on the identification accuracy. Generalisation is then made to systems with unknown mass, stiffness and damping properties – a much tougher problem rarely considered in many other identification methods, due to the limitation of formulation in separating the effects of mass and stiffness properties.
The GA identification strategy is extended to structural damage detection whereby the undamaged state of the structure is first identified and used to direct the search for parameters of the damaged structure. Furthermore, another rarely studied problem of structural identification without measurement of input forces, i.e. output-only identification, is addressed which will be useful in cases where force measurement is difficult or impossible. It is our strong belief that any research attempt on structural identification and damage detection should be tested not only numerically but also experimentally, and hence a relatively long chapter on experimental study to validate the GA-based identification strategy. Finally, a practical divide-and-conquer approach of substructuring is presented to tackle large structural systems and also to illustrate the power and versatility of the GA-based strategy. The findings presented signify a quantum leap forward from research and practical viewpoints, and this book should therefore be useful to researchers, engineers and graduate students with interests in model updates, parameter identification and damage detection of structural and mechanical systems.A Primer to Genetic Algorithms
An Improved GA Strategy
Structural Identification by GA
Output-Only Structural Identification
Structural Damage Detection
Experimental Verification Study
Substructure Methods of Identification