Dr Benjamin Shank of Thermotron Industries tells Jonathan Newell about the dangers of over-vibrating during accelerated life tests.
Accelerated life testing on products has moved on significantly from the “educated guess” profiles that samples were exposed to in the early days of testing. Today, testing includes more scientifically derived representations of the vibration stresses that products are likely to experience during their lifecycle.
Dr Benjamin Shank, a Stanford educated expert in vibration test system control and associated applications at Thermotron Industries, explains how today’s computer derived stress cycles need careful assessment to prevent hidden dangers from invalidating the test programme.
According to Shank, modern computing power provides the ability to capture a sample of acceleration traces from a product’s service environment. It is then possible to calculate stress cycles that would be experienced by the product’s components with a wide range of resonant frequencies.
“Using the Miner-Palmgren Rule (damage cycles are independent and each does damage 1/Ni, where N¬i is the number of cycles of that stress the product could survive), it is possible to assemble a Fatigue Damage Spectrum (FDS), which gives a relative measure of the total damage that the product needs to survive in its lifetime. It is then straight-forward to calculate a Power Spectral Density (PSD) for a Random vibration test which will deliver that FDS in a tiny fraction of the product’s expected life,” he says.
This is the principle of accelerated life testing that enables whole-life evaluations to be performed on a vibration table in the laboratory in a fraction of the time.
Explaining how that time is saved, Shank says that, assuming the product lifetime is more than a few uses, most of the fatigue damage is likely to be due to microscopic crack propagation, the so-called High Cycle Fatigue region. With this, the damage caused by each stress cycle is proportional to the stress cycle raised to the power of the “fatigue exponent”, a property of the material.
To illustrate the power of this exponent, for a product composed of mild steel, a 10% increase in stress levels corresponds to more than a three-fold increase in damage rate, provided the top of the High Cycle Fatigue region is not reached. It’s important therefore that the levels of the product stress cycles don’t exceed its yield strength, resulting in different failure modes than would be expected in its normal life cycle.
For this reason, it is very important for an FDS calculator to keep track of the damage-equivalence sum at each resonant frequency as well as the highest individual cycle. The spectrum of highest stress cycle at each resonant frequency is called the Extreme Response Spectrum (ERS).
Without an ERS limit, the Grms level of any Random vibration test derived from FDS could be turned up arbitrarily to decrease the test duration. This method can quickly change the failure modes encountered from fatigue-related damage to plastic deformations (Low Cycle Fatigue) or even ultimate stress failures.
Unfortunately, some of the Fatigue Damage Spectrum calculators on the market today make unrestricted use of the High Cycle Fatigue approximation without monitoring the Extreme Response Spectrum of the field data fed into them.
However, any programme that allows the user to select an arbitrarily short duration while increasing the Grms level should be viewed with deep suspicion.
According to Shank, when considering accelerated lifetime, it is important to remember that the stresses involved are not merely a group with combined statistics, but also individual events. Any one of these might be hazardous to the product.
“That extreme event determines how much the rest of the lifetime can be accelerated so ignore it at your own peril,” he concludes.
Thermotron Industries has been supplying customers with Accelerated Stress Test equipment for two decades and has a deep understanding of the principles and real challenges in the field. The company also practices what it preaches, exposing their own critical components to a wide range of temperature and vibration conditions to improve robustness and reliability.