VTechWorks staff will be away for the Thanksgiving holiday beginning at noon on Wednesday, November 27, through Friday, November 29. We will resume normal operations on Monday, December 2. Thank you for your patience.

Browsing by Author "Marthinuss, Samuel Joseph"

Now showing 1 - 1 of 1
  • Thumbnail Image
    A Computer Vision Approach to Stress Determination in Blisters, and a Fatigue-Based Method Framework for Testing Defect Development
    Marthinuss, Samuel Joseph (Virginia Tech, 2020-11-24)
    With the development of hydrogen fuel cell technology continuing to advance, rapid characterization of membranes is increasingly important for design purposes. Pressurized blister testing has been suggested as an accelerated characterization alternative to traditional relative humidity (RH) cycling tests, and is the focus of this project. Prior efforts to determine the stress state present in the pressurized membrane blister test, however, have required constitutive properties of the membrane (Young's modulus and Poisson's ratio), along with Hencky's classic model for circular membrane stresses. Herein we describe an analysis method and computer vision imaging technique that are capable of determining the stress state in a pressurized circular membrane based solely on simple equilibrium equations and geometric considerations. This analysis method is applied to an image of the blister during testing, and the only additional required data is the pressure at the time the image was taken. By pressurizing circular blisters, an equi-biaxial, mechanical stress state is induced, simulating membrane stresses experienced during fuel cell operation as humidity levels fluctuate. The analysis leverages membrane theory and the axisymmetric geometry to determine the stress state from a profile image of the inflated blister. As a check for the method, an elastomer with known constitutive properties was analyzed using both the previous Hencky's solution method, as well as the new computer vision imaging method. The comparison of stress calculation results show that the two methods agree within 5 percent. A primary mechanism of membrane failure through mechanical stressors is the growth of local defects (usually chemically induced) due to the cyclic equi-biaxial stress state. In order to better understand and characterize the effect of disparate initial defects on CCM, two primary methods to defect membranes were introduced. The first was a compression against sandpaper method meant to simulate GDL compression, and the second was a targeted method using a hypodermic needle to initiate a defect at a central location on the membrane prior to pressurization. Observing the pressure decay in these defected blisters as compared to undefected tests showed that, while undefected samples did not experience pressure decay until failure, defected samples began showing signs of leaking through pressurization cycle profiles and steady state pressures achieved. Pressure data showed that samples tended to lose pressure more quickly with increasing initial defect severity. Undefected samples exhibited no pressure loss until the moment of failure, which was often catastrophic and instantaneous. Sandpaper defected samples exhibited a slow decay in cycle steady state pressure throughout tests, with no increase in cycle pressurization time. Needle samples showed a slow decay in cycle steady state pressure as well as an increase in time for the cycles to reach steady state. The needle defects were the most locally severe and thus the pressure decay indicators were most significant out of all the samples tested. The blister test method rapidly cycles mechanical stresses in a CCM, and elucidates signs of leaking that correlate to flaw development in recorded pressure data. With further development, it might serve as a robust method to quickly test flaw growth rate and development in CCM samples.