A Project Sponsored by BASF
Vibration welding is a well-established technique that is widely used in high-strength and high-temperature applications, such as in the production of automotive intake manifolds, chain saw housings, and hedge trimmer housings. The main advantage of vibration welding is its fast process time and low cost. Currently, most of the research focus is on the short-term strength of vibration-welded parts, which is typically measured through tensile, hydrostatic burst, or dynamic burst testing. To the best of our knowledge, there is very limited research available on the long-term strength and failure modes under sustained and/or cyclic loading.
In our observations under sustained loading, we notice that failure occurs in a brittle manner at a very small strain. This phenomenon is accompanied by the formation of microcracks in the matrix and fiber separation from the bulk material. The fracture surface also displays well-defined dark and light domains, as visible in the upper optical micrograph in the following optical and SEM micrographs.
Due to the large number of random shapes and sizes of defects with unknown statistical distribution, modeling the accumulation of defects (damage) using a statistical approach is impractical. Hence, we adopt a phenomenological approach, as offered by continuum damage mechanics to describe the damage growth in the composite material. We introduce a phenomenological damage parameter as an amplification factor for engineering stress due to micro-damage, and express the kinetic equation of damage growth in the form of Eyring equation, taking into account the thermo-activated nature of the defects in the composite (short fiber reinforced Nylon 66). The criterion for local failure is defined as the damage growth instability, or unlimited rate of growth.
The microstructural variations within the weld joint material are observed to exist on various scales, with the largest being in the glass fiber reinforcement. Despite controlled and reproducible processing conditions, a noticeable variation in fiber distribution is still present within the weld joint. This variation affects the mechanical properties, particularly the strength and toughness, and results in the identification of the weakest point as the initiation site for fracture. It causes the randomness of time to fracture initiation. This randomness is manifested in the large scatter of the failure time observed in creep experiments. Given that fracture initiation occurs at the weakest point, the actual fracture initiation time is the minimum value among the virtual times to failure elsewhere along the weld joint, i.e., the time to failure tf should be within the following interval 0 ≤ tmin ≤ tf <∞, here, the value tmin is the absolute minimum of time to failure, below which the failure does not occur. This is a crucial consideration for probabilistic modeling, and only the Weibull distribution meets the requirement among the three known types of minimum value probability distributions.
We formulate a deterministic model using continuum damage mechanics and Eyring kinetic equation to estimate key parameters such as activation energy and volume at which fracture occurs, based on data from accelerated life tests. Furthermore, we build on that foundation and develop a probabilistic model to predict the lifetime reliability of weld joint under creep and fatigue loading. This model can be used for plastic part reliability-based design. Future experimental investigations are necessary to validate the statistical model prediction, particularly at lower stresses and temperatures. Additionally, further testing is needed to determine the effect of R-ratio and frequency on statistical characteristics of time to failure.
