Many industries have faced a long standing problem obtaining critical replacement parts when the original manufacturer is either out of business or no longer providing them. This has been a particularly vexing issue in the aircraft business but similar problems are encountered for highly complex, one of a kind systems in ships, industrial equipment and related markets.
These critically needed parts are generally highly engineered and/or manufactured to tight tolerance controls. The product development effort that originally went into their legacy production was extensive and to redo that now, years later, would be time consuming and exceedingly expensive. In the case of aircraft parts, for example, the development time could have ranged from 2 to 5 years with an extensive amount of performance and durability testing before they were originally placed in-service.
The geometric design isn’t the problem because that can be easily recreated by extracting measurements of the legacy part. Knowing the exact recipe used for production – how the materials were formulated, heated, cooled and at what rates to achieve the grain arrangements creating the achieved level of durability – the secret sauce, so to speak – is the issue. In all likelihood, extensive research, analysis, testing, and revising occurred the first time around, and that can’t be detected visually. At least, that used to be the case.
Probabilistic, material microstructure simulation techniques are now being used to reverse engineer legacy parts in a 3 to 6 month’s time. The investigation process starts by dissecting the legacy parts to catalog the grain arrangements, microstructural arrangements and variability, and processing defects. Applied surface treatment effects are revealed during the microstructure level investigatory work. Some lab-level testing may be required to identify the active failure mechanisms and residual stress affects on samples of the legacy part materials.
All the previous data are used to set up the probabilistic, micromaterial simulation of the part. Since the application is known, the finite element stress analysis can be recreated from scratch. Even if the exact stress loading is unknown – that’s not a huge issue to get around in order to produce a first cut at the analysis – a simulation can be conducted for several loading scenarios (or “missions,” in the case of aerospace). This will effectively “bound” the solution and identify the areas of highest concern for further analysis.
Five years ago, this type of computational processing would not have been possible or even practical. But today it is. And it opens up a whole new approach to resolving the legacy part manufacturing issue.