Optimized State Estimation Method for Structural Health Monitoring Using Heterogeneous Measurements Under Uncertainty
AuthorSaleem, Muhammad Mazhar
KeywordsHeterogeneous data fusion
Impact Force Identification
Sensor configuration Optimization
Structural excitations' Identification
Structural Health Monitoring
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
EmbargoRelease after 01/10/2022
AbstractAn optimized state estimation method that can effectively incorporate uncertainties associated with structural models and measurements is proposed for structural health monitoring. Heterogeneous structural measurements consisting of strain and acceleration time-histories are fused through a state space framework to obtain good quality state estimates over a broader range of frequencies. The unknown states are estimated through an augmented Kalman filter (AKF) which can identify all the states including structural response as well as excitations while effectively incorporating model error and measurement noise. Genetic algorithm-based optimization strategies are adopted to address the uncertainty issues and improve the efficacy of the state estimation process. Optimization is performed at two stages. First, sensor configuration is optimized; second, error covariance values on model, structural measurements and excitations, involved in the Kalman filter process, are optimized. Four studies are carried out for the development and validation of the proposed method. The first two studies are related to optimization of sensor configuration and error covariance values while the other two are focused on the performance validation of the proposed method through case studies, especially related to input force identification. In the first study, sensor numbers, locations and types are optimized using a genetic algorithm with a single objective function whereas in the second study, sensor configuration is optimized for spatially-varying dynamic loading of bridges using a multi-objective genetic algorithm. The third study pertains to spatiotemporal identification of impact force and corresponding structural responses. The concept of redundant sensors is utilized to locate the impact force, and then exact time-histories of force and structural responses are generated by optimizing the error covariance values. The fourth study is about the identification of traffic induced structural excitations. The study shows that once the state estimation method is optimized by selecting suitable error covariance values it remains stable against any traffic regime. A truss bridge example is utilized to demonstrate the effectiveness of the proposed method on identifying the traffic induced structural excitations. The results from all the studies show that the proposed method can effectively identify the structural response and excitations while successfully dealing with the uncertainties in structural models and measurements. Hence, it has a great potential for practical application to real-life structures.
Degree ProgramGraduate College
Civil Engineering and Engineering Mechanics