A dynamic laboratory test was carried out on a single-storey frame model placed on a shaking table. Measures were collected by a traditional scheme based on accelerometers installed on the storey and the basis of the frame structure, and by an innovative device, the SOFO dynamic system, which is based on long-gauge fiber-optic sensors installed in two different locations along one of the columns of the frame structure. The use of long-gauge strain sensors allows the detection of local damage that is visible in the global frequency response of the structure. In particular the fiber-optic sensors show a very high sensitivity and extend the frequency range (1mHz-1KHz).
The peculiar contribution of this paper is to conduct an elaboration of the measured data through the damage detection scheme recently proposed by the first author. It is based on the idea of using a response surface model of the measure correlation for damage detection and localization.
Two goals are pursued in this paper. The first goal consists of comparing the performance of the innovative SOFO dynamic system, which uses long-gauge fiber-optic sensors, with the traditional monitoring method based on accelerometers. For this purpose, a dynamic laboratory test was carried out, and measurements were taken from a single-storey three-dimensional steel frame model excited at the base by a shaking table. The SOFO dynamic system was installed on one column of the frame structure, while two accelerometers were mounted on the base and on the frame storey, respectively, for comparison. The use of fiber-optic sensors allows to overcome the difficulties associated with the traditional dynamic measurement methods, such as the limitations in the number and in the locations of the monitoring devices. Furthermore, the long-gauge fiber-optic strain sensors show a very high sensitivity and extend the frequency range (1mHz-1KHz).
The second goal is to investigate the sensitivity to local damage of a recently proposed method for damage detection and localization. Indeed, the use of better performing long-gauge strain sensors allows the detection of local damage that is hardly visible in the global response of the structure. Damages of increasing intensities are therefore gradually introduced in the structure, and the measurements acquisition is repeated for each of the damaged cases. The SHM-RSM method, which is based on the idea of using a response surface model to approximate the relationship between the measurements collected by different sensors during the same test, is finally applied to the collected data to detect and locate the damages of different intensities.