Well, hydrogen embrittlement is a long known issue that has played havoc with metals at nuclear power plants and has lead to fuel rod cadding failures along with reduction of the rating of some containment vessels.
@TJ McDermott: I would strongly recommend against derating hydrogen embrittled parts, if by derating you mean using the affected parts in a lower-stress application. The problem is that it's extremely difficult to predict what level of stress a hydrogen-embrittled part will fail at. Even residual stresses from the forming process may be enough to produce microcracks, which may propagate later in fatigue. If I knew that a part was likely to be hydrogen embrittled, I wouldn't recommend using it in any application.
The same line of argument applies to baking parts as a re-work method after a problem has been found. As I mentioned, it's important to bake parts as soon as possible after plating or welding. Some aerospace specifications require that this be done within one hour. The more time passes, the less effective baking will be. This is not so much due to the hydrogen being any more difficult to remove (although you will hear this claim), so much as the fact that microcracks may have already formed as a result of residual stress. Obviously, once cracks have formed, no amount of baking will heal them.
As naperlou pointed out, prevention of process-induced hydrogen embrittlement depends on having a good quality system in place. If you're plating or welding high-strength parts, you need to ensure that they are always baked at the proper temperature within the specified length of time.
This is an interesting example of a failure mode that is not going to be easy to predict. It seems that, short of testing a sample of the parts after each process, one can only track the parts after production and remedy situations as they occur. This requires a detailed tracking of the products after delivery and detailed reports of problems. In general, quality control systems do this. By linking those databases with design data in a PLM system, another big theme of Design News lately, one can avoid the problem in the future.
Are they robots or androids? We're not exactly sure. Each talking, gesturing Geminoid looks exactly like a real individual, starting with their creator, professor Hiroshi Ishiguro of Osaka University in Japan.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.