Sound and Vibration

Research Experience for Undergraduates

CSU Summer 2003

Made Possible by a grant from the NSF and ARO

Dynamic Analysis of Viscoelastic Properties

Charles Schuler

Senior in Civil Engineering at Colorado State University

Purpose

To study the viscoelastic properties of PEEK/IM7 in an attempt to calibrate a Seiko TMA/SS testing machine at the Colorado State University Composite Materials Lab. Previous research using this equiptment was conducted by Daniel Melo, a PhD student. During his work, it was found that the TMA/SS did not produce usable data. Testing was completed offsite, providing data that will be used in the calibration. The focus of this research was in the damping properities of small beams (about 18 x 1 x 0.2 mm). Beams with fibers oriented 0-, 30-, 45-, and 90- degrees to the beam axis were prepared. Beams including 2 and 3 layers (plys) of material were also prepared for the 0-degree orientation. Damping properties of a material can be described by E' the storage modulus and E" the loss modulus Tangent of delta describes the relationship. Description of these properties was desired

Equiptment

The machine used in the testing is a Seiko 5200c TMA/SS. The machine is capable of producing loads up to 500g and accuratly measuring displacements. This machine was designed to accuratly measure the coefficents of thermal expansion and do some dynamic mechanical analysis for tensile samples. Using a 3-point pending of a beam as the geometry meant that new mathmetical procedures would have to be developed. This is done under precise temperature control allowing investigation of properties with respect to temperature.

Analysis

The output file from the TMA/SS is a text file containing temperature, load, and displacement readings. This was analyized in MATLAB to develop tangent of delta and E' and E". MATLAB was chosen because the iterative nature of the required calculations lends itself to a mathmetics program. This was also consistent with the program used in the previous research. A program were created that use equivalent length and area equations that allow calculation of tangent of delta and E' simply, as would be done with tensile specimens. Also developed was a program that used beam theory to calculate tangent of delta and E'.

Results

Both methods produced results that were comparable to the offsite data, though not as close as expected. Errors were especially high in regard to E' calculations.

Discussion

One explanation of the high errors is the sensetivity of the calculations to the thickness of the sample. The thickness is carried as a cubed term in the equations, so calculations were performed to determine what influence it has on the results. Calculations showed that a change of 0.05mm, the accuracy that measurements were conducted to, produced an average of a 5% change in the tangent of delta calculation and a 40% change in the E' calculation. This sensetivity shows that accurate measurements are required. Overall, the problem does not seem to be a calibration problem, as error was distributed around the known data points. Thus, the data is considered extremly vairable and sensestive to sample size. Future work in this area may include investigation of the effects of differences in sample preperation and handling which were possible contributions to error in this test.

Conclusion

Testing conducted using a Seiko 5200c TMA/SS unit has produced results that are considered reasonable. Large errors found in calculating E' show the necesity of accurate sample measurement and careful sample preperation. This research was successful by showing that the equiptment is capable of making these tests in future studies.