Studies On Structural Damage Detection And Optimal Sensor Placement

It is very significant to prevent catastrophic structural failures through timely and accurate health monitoring and damage detection, furthermore, the maintenance expense of the structure can be sharply decreased by early detecting structural damages and effective measure. So, structural health monitoring and damage detection have become research hot spot in civil engineering field in recent years.For the reason of self advantages of dynamic techniques for damage detection, it has attracted much attention of domestic and overseas researchers. They have done a lot of work in this field and bear fruit. In this paper, studies focused on structural damage diagnosis indictors including flexibility method,change of strain mode shape and method of dynamic and static combining. The main contents are as follow:①In order to enhance the extents and precision of flexibility method applied to structural damage detection. A new sturctural nondestructive detection indictor is prsented. The indictor is based on difference of flexibility-curvature-matrix. Its availability is verified through simple beam and continuous beam, and by compared with other three kinds of detection indictors. The result shows that to detect the location of damage in the beam mentioned in the paper, only the first mode data is needed. It can also be used to detect a small extent structural damage. Compared with others three indictors, new indictor is more favorable in precision and universality.②In recently years, domestic and foreign scholars have done a lot of work in study of flexibility method. However, most of their study is limited to use flexibility method for locate the damaged location in a structure. In this paper, based on previous chapter the author defined intact flexibility-curvature curve of undamaged structure and used least squares method to fit this curve. On the base of the above work, degree of structural damage has been studied and damaged effective area been presented to assess damaged degree. At last, a numerical example of simple beam with single damage and two damages have been used to verify this method. It will provide a new way to assess degree of structural damage.③Compared with displacement mode shape, strain mode shpe is better to locate local damage in a structure. A new structural nondestructive detection indictor, a dynamic and static combining indictor is presented, which is based on dynamic indictor and static indictor. The new indictor retains advantages of method of dynamic and static. It can filter much effect of non-damaged elements. It is much exact than used dynamic and static indictor for damage detection individual. Then its availability is verified through a steel truss numerical simulation, and by compared with dynamic and static detection indictors. The result shows that new indictor is the best. Furthermore, second-detection and repeated-detection methods is presented for solve shortcoming of little strain is difficult to detect in different kinds of components'damage in numerical example.④A new dynamic and static combining indictor is presented based on dynamic and static structural damage detection in this paper. Structural displacement-difference and change of modal force between intact and damaged structure are used to construct balance equation in the dynamic and static method. Then method of Moore-Penrose generalized inverse is used to solve damage coefficient. Dynamic and static combining method takes advantages and disadvantages of dynamic and static method into account. Take a four-storey steel frame structural model as an example, dynamic and static response information is gained from experiment. The three above method is used for damage detection, the result shows that new method is superior to others methods.⑤The first and second order sensitivity matrixes of modal parameters relative to the damage member were derived, and the modal truncation errors arising from the derivation of modal sensitivity was improved. The first and second order sensitivity equations were established respectively on the basis of principle of Taylor series expansion. And the solving method of these sensitivity equations was studied. The generalized inverse technique was used to solve the first order equation, and the trust-region reflective Newton method was used to solve the second order sensitivity equation. Considering the fast speed in solving the first order sensitivity equation and the high precision in solving the second order equation, a mixed iteration algorithm for structural damage detection was put forward. This algorithm took up the second order nonliner analytical solution as the first substituting value, and then the first substituting value was modified on the basis of the Taylor series bias error using the solution of the first order sensitivity equation. It shows that the mixed iteration algorithm in this paper has better convergence and faster iteration speed because the higher precision second order nonliner analytical solution is adopted.⑥In order to improve the effective independence method , capacity factor is considered. Three optimal sensor placement approaches, the effective independence method,kinetic energy method and modified effective independence method, which were compared in this paper. With the background of a steel truss, the above optimal approaches were used for sensor placement and their effects were evaluated according to the three estimating criteria. Last, the results show that the modified effective independence method is better than the others methods above, it can reflect structural character better.