Abstract
The multi-material structure in which steel members support high load and the other members are made of light metal such as aluminum alloy is becoming mainstream in automobile industries. Though the mechanical joints such as bolts and nuts are often applied to join the members, the mechanical joints are required to apply proper jointed load. This study has proposed a novel method to estimate the surface texture parameters based on the data set of the jointed load and the natural frequency of the target by using the data assimilation. Moreover, applying the proposed method, the jointed load can be estimated from the arbitrary natural frequency by using a few sample of the data set. In order to investigate the behavior of the proposed method subjected to the measurement error in the sample data, a numerical simulation has been performed based on the Monte Carlo method. The correct surface texture parameters were given in advance, and the direct analysis was applied to obtain the simulation data of the natural frequency. Then the proposed method was applied to the simulation data with artificial disturbances. Most of the results show that the estimation error of the natural frequency under the low jointed load was less than 20 Hz when the noise of the natural frequency under high jointed load was less than 10 Hz. In the estimation of the surface texture parameters, the sensitivity of the homogenized elastic modulus was lower than that of the standard deviation of the asperity peak height.
