In the summer of 2023, numerous aircraft engines were recalled due to impurities in the metal powder used to 3D print the turbine blades. These impurities meant that the behavior of the turbine blades could not be reliably predicted, posing a risk of failure. Further information can be found in these articles:
https://www.popularmechanics.com/flight/airlines/a44656285/1200-airbus-jet-engines-recalled/
https://www.aerotelegraph.com/mtu-rechnet-mit-milliardenkosten-wegen-problemen-bei-pratt-and-whitney
This incident illustrates how important stochastics is for predicting component behavior. Using Time-separated Stochastic Mechanics, we have made our own calculations to illustrate the problem.
Figure 1a) shows the expected value of the stress in a turbine blade. This is essentially equivalent to the calculation with fixed, i.e., deterministic, material values.
The loading of the turbine blade results in high stresses at the edges of the upper part. An elasto-viscoplastic material behavior was assumed, which is realistic under the high temperatures.
However, if variable material parameters are taken into account, the standard deviation also becomes relevant. This indicates where particularly large deviations in the stresses are to be expected. The result is shown in Figure 1b). It is quite clear that the standard deviation is distributed differently locally than the expected value. In the lower part of the turbine blade in particular, a high standard deviation but a low expected value of the stresses is visible. This is exactly where some turbine blades can be damaged.
To eliminate the risk of damage, the affected engines were serviced ahead of schedule and the affected turbine blades were replaced. The costs amount to several billion US dollars. Including the stochastics in the design would have prevented this.
