A PROBABILISTIC DIAGNOSIS METHOD FOR STRUCTURE MONITORING

Experimental data for monitoring structural condition contain noise. It causes concern with the effectiveness and the reliability of diagnosis, especially for diagnosing small structural damage. A probabilistic method is presented here for more reliable diagnosis. This method is applied in the laboratory for a bridge model using high resolution CCD camera for deflection data acquisition. Results show that this approach can clearly identify both the existence and the location of damage with 3 percent stiffness loss