4.
0 Results and Discussion
In this study, the aluminum beam was analyze in finite element method to determine the dynamic
behavior of the structure such as natural frequency and mode shapes. The results in finite element
FE of aluminum beam will be compared with the experimental modal analysis EMA. The EMA
results was provided from the faculty which is doing experimental of aluminum beam. There are
several design that have been optimized on aluminum beam using finite element method to find
the highest frequency compared to the EMA. Figure 1 show the best design that we have chosen
in our study to determine the dynamic behavior of the aluminum beam. The mass of the aluminum
beam was reduced from 1.5kg to 1.08kg. Figure 1 show the detailed of our optimized design.
Figure 1: Optimized aluminum beam
Figure 1 show the optimized aluminum beam which is the center of the beam have been removed.
The design was model and modified using CATIA V5 software. Then, the design was upload in
the Hypermesh to analyze the dynamic behavior of the aluminum beam. The frequency and mode
shape was obtained after analyze the structure.
Table 1: Comparisons of results of the natural frequencies of the aluminum beam between the
experimental and optimize finite element
mode EMA, Hz Optimize FE, Hz Errors, %
1 103.81 123.16 -18.64
2 243.11 279.14 -14.82
3 280.93 338.11 -20.35
4 546.66 658.03 -20.37
5 661.04 756.33 -14.42
6 898.43 1075.53 -19.71
7 1164.48 1225.14 -5.21
8 1247.05 1448.71 -16.17
9 1333.83 1587.39 -19.01
10 1852.03 2190.34 -18.27
Total Percentage Errors -166.97
Measured mode shape Finite Element mode shape
Mode 1 = 103.81Hz Mode 1 = 123.16Hz
Mode 2 = 279.14Hz
Mode 2 = 243.11Hz
Mode 3 = 280.93Hz Mode 3 = 338.11Hz
Mode 4 = 546.66Hz Mode 4 = 658.03Hz
Mode 5 = 661.04Hz Mode 5 = 756.33Hz
Mode 6 = 898.43Hz Mode 6 = 1075.53Hz
Mode 7 = 1164.48Hz Mode 7 = 1225.54Hz
Mode 8 = 1247.05Hz Mode 8 = 1448.71Hz
Mode 9 = 1333.83Hz Mode 9 = 1587.39Hz
Mode 10 = 1852.03Hz Mode 10 = 2190.34Hz
Table 2: Comparison between measure mode shape and finite element mode shape
Table 1 above show the result obtained after run using the Hypermesh. The result show all
the frequency that obtained in finite element are more than experimental modal analysis. . As can
be seen from Table 1, all ten modes obtained from the initial FE model were successfully adjusted
in the light of the measured data. We have recorded that the natural frequencies of the optimized
model met the requirement as all the mode shape has high value compared to Experiment Modal
Analysis (EMA). The achievement in the adjustment can be clearly seen in the error columns
where the errors were in negative sign means there was a good value of the natural frequency. The
modal based updating method recorded the total percentages errors for our optimized design of
aluminum beam is -166.97% in comparison with the measured result. The greatest error was
contributed by the fourth mode with the errors of 20.37%. The third mode is slightly below than
fourth mode which is 20.35%. The lowest errors was contributed by the mode 7 with the errors of
5. 21%. This percentage error is acceptable for the optimized design because the design structure
is stronger than before the optimized design. Even though, the mass of the aluminum beam was
reduced the stiffness of the structure is high and more rigid.
Comparison natural frequency between FE and EMA
2500
2000
Natural frequency,Hz
1500
1000
500
0
1 2 3 4 5 6 7 8 9 10
EMA, Hz 103.81 243.11 280.93 546.66 661.04 898.43 1164.48 1247.05 1333.83 1852.03
Optimize FE, Hz 123.16 279.14 338.11 658.03 756.33 1075.53 1225.14 1448.71 1587.39 2190.34
Mode shape
EMA, Hz Optimize FE, Hz
Figure 2: Graph Comparison natural frequency between FE and EMA
Based on the graph plotted, the different of percentage error between EMA vs initial FE and EMA
vs optimize model can be seen that there was a significant difference in gap. Thus, this shows that
optimize design is better with higher value of natural frequencies in mode shapes.
