Sensor Placement Optimization Based On Time Domain Structural Damage Detection

To avoid catastrophic structure failures with a huge loss of properties and human lives, structural damage detection becomes an essential phase of work in structural health monitoring for researchers in civil engineering in recent years. The time domain structural damage detection method is a category of effective method, and it has attracted considerable attention of researchers. Time domain structural damage detection method has the advantages which need fewer sensors and provide more detection equations than frequency domain damage detection method. With these advantages the time domain method will be more applicable in civil engineering, but the damage detection results are also affected by sensor locations because it is a vibration-based method and response data collection is always the first step in in time domain method, which makes the sensor number and sensor locations become key issues and affect the engineering application of the time domain method.Although the time domain method are more economic because it needs less sensors than the frequency method, the time domain method has less engineering application than the frequency method due to the lack of the time domain sensor placement optimization method. To promote the development of the time domain method the time domain single sensor placement optimization method is proposed based on the minimum ill-posedness rule, and then the multiple sensor placement method is proposed. In the end the fast algorithm for sensor placement is proposed. The outline of this dissertation is given as follows:(1) Time domain sensitivity method and corresponding solution method have been summarized through literature reviewing. It finds that the ill-posedness of the damage detection equation is the key factor that influences the accuracy of the damage detection results. Condition number of the sensitivity matrix is the criterion for evaluation of this ill-posed problem, so a sensor placement index is proposed to replace the calculation of the condition number, in which the sensor combinations can be judged by the magnitude of the proposed index. The proposed method has high computational efficiency, and the effectiveness of the proposed method is illustrated with a planar truss structure and a 3D frame structure. The numerical results show that the optimized sensor placement can give accurate damage detection result with measurement noise.(2) The structural damage detection method is studied based on multiple time domain responses, in which the sensitivity matrix is combined by every single normalized sensitivity matrix of each response, and the solution method is also given. With the proposed method the cheaper strain sensors can replace the acceleration sensors, so it will reduce the cost of the structural damage detection. It finds that the ill-posedness of normalized sensitivity matrix is also the key factor that influences the accuracy of the damage detection results, so the minimum ill-posedness rule is also applied to optimize the multiple sensor placements. A similar sensor placement index is proposed to judge the multiple sensor combinations. The proposed methods are studied with numerical simulations of a 2D beam structure, which shows that the proposed methods give satisfactory damage detection result.(3) To improve the calculation efficiency of the sensor placement index, the fast algorithm is proposed to meet the requirement of the engineering application. The fast algorithm referring to the search algorithm has two steps, which can enhance the calculation efficiency and give the number of the optimized sensor combinations. The simulation studies mentioned above are re-studied with the proposed fast algorithm, which validates the effectiveness and high efficiency of the proposed fast algorithm.