CASES STUDIED CASE 1: DISPLACEMENT RESPONSE There are three different sensor location configurations studied which compare displacements responses at three different locations in the BU structure. All the sensor configurations use 7 out of the 8 locations shown in Figure 6. A. Sensor set-up A – uses locations 2, 3, 4, 5, 6, 7 and 8 for expansion. B. Sensor set-up B – uses locations 1, 2, 3, 5, 6, 7 and 8 for expansion. C. Sensor set-up C – uses locations 1, 2, 3, 4, 5, 6 and 8 for expansion. For the first sensor configuration (Sensor set-up A), 2, 3, 4, 5, 6, 7 and 8 sensor displacement responses were used in expansion process to expand up to 3112 locations. The sensor 1 displacement response recorded experimentally was used for comparison with expanded solution. Note that the impact was made at the location 1 on the Upright structure, so location 1 will see the highest amplitude of response. The comparison of the expanded and the experimentally obtained response is shown in Figure 8a. The response was recorded for 4 seconds as shown. The responses predicted by the proposed technique compares well with the experimentally recorded response at the top-right location. For comparing the time responses, a TRAC was performed between the expanded solution and experimentally recorded response. A TRAC value of 98% was observed which show very good agreement between the two solutions. The transient response comparison for initial time steps (where highest amplitude of response is observed) is shown in Figure 8b which is the first 0.5 sec of the response. As seen, the solution predicted by the proposed technique compares well with the experimentally recorded response. Note that the TRAC value increases from 98% to 100% because good correlation is observed for high Signal-to-Noise-Ratio (SNR) time period. The second sensor configuration (Sensor set-up B) studied was using 1, 2, 3, 5, 6, 7 and 8 sensors excluding the 4th location. The comparison between expanded displacement response (which used the 7 sensors to expand to up to 3112 points in the Upright) and experimentally recorded displacement response at location 4 is made and shown in Figure 8c. The mode configuration used previously in the expansion process is used again in this case. Good correlation between the expanded and experimentally recorded displacement data set is observed. TRAC of up to 97 % is seen which means that the predicted solution is very similar to the experimentally recorded data set even though the comparison location at Location 4 was half-way to the fixed edge of the structure (where more noise in experimental data set is expected). Comparison for initial time period as seen in Figure 8d shows that even for mid-range SNR, the proposed technique does a good job of predicting accurate response. A TRAC of 99 % further substantiates the argument of accurate response prediction. A third sensor configuration (Sensor set-up C) was studied where sensors 1, 2, 3, 4, 5, 6, and 8 were used excluding sensor 7 which is very close to the fixed edge of the structure. Higher noise and low SNR is expected. Figure 8e shows comparison between the predicted solution and experimentally recorded displacement response at location 7. As expected the experimental response is very noisy, especially for the time period from 2 to 4 seconds where the response dies out. A TRAC of 76% is observed for the time response comparison between the predicted and the experimental responses over the full range of 4 seconds. However, a TRAC value of 87% was obtained for the time responses at initial time steps where higher response was expected. A fairly good correlation is observed in Figure 8f between the experimental and the predicted responses. 208
RkJQdWJsaXNoZXIy MTMzNzEzMQ==