FE analysis is widely used to analyze dynamic characteristics of mechanical structures. But its credibility is not always guaranteed. Thus mostly FE analysis credibility is decided by comparing the dynamic characteristics predicted from the FE model with those from experiment. Then model updating or model improvement is followed to match the differences between analysis and test. But FE model updating is not easy when the structure is complex and its degree of freedom is huge. So it is important to judge the confidence of initial FE model before analyzing the structure. Modal data and/or FRF are commonly used to judge the confidence of initial FE model. In this work, mode shape was selected as the criterion of confidence judgement. There are many indexes to compare the similarity of mode shapes. All the indexes, which are available in the published papers, are studies and compared with simulated data. A new index, coordinate XOR via GDOP which is based and modified from XOR via GDOP, is proposed and verified its credibility. From comparison study, the characteristics of correlation and model reduction methods were identified. When confidence judgement is performed, it is general to assume experimental data are correct. But frequently, the assumption is not adequate for various reasons (experimental noise, etc.). And the condition of experiment is usually restricted in time and cost. So considering these points, it is important to select the optimum sensor and actuator position to get better experimental data under the restricted experimental conditions. In this work, A simple simulation about the selection of sensor and actuator position was performed on a clamped plate and its effect of distribution of sensor placements was analyzed in the view of model reduction method. The results of this work can be used to determine whether initial FE model needs the performance of FE model updating or not. Also it can be used to select the best modeled FE model when many initial FE models exist for an identical structure. After selection of the best FE model, FE model updating can be followed effectively with the selected FE model.
