We report a procedure by which structural parameters and input ground motion are identified from measured responses only. We have assumed that the coda of the response time history represents the free vibration response of the structural system. Because the coda is not effected by the input ground motion, we can first identify such structural parameters as the masses, damping coefficients and spring constants from this part of the record. Input ground motion then is estimated from the full record and the identified parameters. The identification and estimation are made with the Kalman filter. To verify the effectiveness of this procedure, we have simulated the responses of a linear, three-degree-of-freedom system for different earthquake inputs and made estimations using the simulated responses as observed records. The estimated accelerograms, the identification of which usually more difficult than the identifications of velocitigrams and displacementgrams, are in good agreement with the recorded ones for the actual earthquakes.

We propose a dynamic analysismethod – a refined version of the DEM- that can simulate three-dimensional elastic, failure and collapse behaviors of structures. A structure is modeled as an assembly of rigid elements. Interaction between elements is modeled using multiple springs and multiple dashpots attached to surfaces of the elements. The elements are assumed to be rigid, but the method allows the simulation of structural deformation by permitting penetration between elements. There are two types of springs: one is a restoring spring to simulate elastic behavior before failure and the other is a contact spring for simulating contact and recontact between elements. A contact dashpot is also used to dissipate the energy of contact. Structural failure is modeled by replacing restoring springs with contact springs and dashpots. A method for determining spring constants is also proposed. The validity of the method is confirmed by the numerical simulation of masonry wall models. First, the elastic behavior induced by an impact force is calculated. It is found that the elastic behavior determined using the proposed method is in good agreement with that determined using the finite element method. Second, the seismic behaviors of masonry wall models with different laying patterns and a wall model with reinforcement are analyzed. It is found that the proposed method allows expression of the difference in behavior due to different laying patterns and reinforcement. The validity of the proposed method is thus confirmed. The proposed method is suitable for simulating seismic behavior of masonry structures.

No abstract

The process of failure propagation of masonry buildings during earthquakes is simulated using a refined version of the distinct element method that simulates three-dimensional elastic, failure, and collapse behaviors of structures. Models with a flat roof and models with a vault roof are considered, and their failure propagation mechanisms are examined. The influence of the direction of the input ground motion on failure propagation is also investigated. Moreover, the effectiveness of three reinforcement measures is compared. One measure is increasing the mortar strength, the second is increasing the thickness of the bearing walls, and the third is introducing wooden columns and beams. Among these measures, increasing the mortar strength is found to be the most effective. Increasing the thickness of the bearing walls and introducing wooden columns and beams are found to be effective only if the roof has sufficient integrity since the collapse of the roof depends on the integrity of the bearing walls and the roof itself.