News, Blogs and Updates

Even with an academic and experimental background in fatigue analysis, it is daunting to provide a hard, no-nonsense life-cycle prediction.  It becomes especially daunting when your fatigue prediction can cost or save your client millions of dollars.  Plus, there are tons of computer programs that promise “instant fatigue nirvana” at the press of a button; which leads one to ask: “What is a poor engineer supposed to do?”  Over the years, we have learned that there are three critical components to a quality fatigue analysis: i.) accurate FEA stress results, ii.) accurate FEA stress results and iii.) accurate FEA stress results.  Okay, sad, old, real-estate joke about location, location, and location; but let us just imagine that your stress numbers are good, then what?  Fatigue analysis is all about the protection of structures and systems against failure from cyclic loading.  This is where the ASME Boiler & Pressure Vessel Code (BPVC) provides a tried and true standard that, if your stress numbers are good, then you can be assured that your fatigue prediction will be conservative.

NASA 5020A Requirements for Threaded Fastening Systems in Spaceflight Hardware

Over the years, we have done a number of satellite analysis projects for commercial and those other government agencies.  Looking back, I’m sort of surprised how close we got with what we thought were FEA best practices for linear dynamics (i.e., normal modes and PSD analysis).  The big advance over the last couple of years has been in our approach to fastener modeling.  In prior work, fasteners (bolts, screws, what-ever) were idealized using beams and rigid links (Nastran RBE2) while nowadays, our preference is to use six DOF springs (Nastran C-Bush) in combination with rigid links.  While a bit messy, it provides an efficient methodology to meet the NASA 5020A technical specification.

  The gist of this specification is how to calculate, whether or not, the fastener will fail given: bolt preload, with and without shear pins and joint slippage.  It is a tall order and the specification is a algebraic joy to the mathematically inclined simulation engineer.  Fastener failure is dominated by the designer’s choice of bolt preload.  Interesting enough, the NASA specification favors low bolt preload.  It sounds odd, but if pushed, one can avoid NASA 5020A fastener failure by lowering the bolt preload.  The reason for this is due to the relationship between bolt preload and the applied tensile load.  There is no free lunch and regardless of the initial bolt preload, the applied load adds to the overall bolt tensile load.  The specification favors hand calculation but with some FEA modeling, one can improved upon the hand calculations and eke out a bit more headroom.  If you would like to read more, the NASA 5020A specification can be downloaded here.

We are generalists at Predictive Engineering and it has its pros and cons.  We cross-pollinate from medical (orthopedic to endoscopic), rail (transit to heavy locomotives), automotive (electric to Class 8 trucks to school buses), aviation (commercial, supersonic and military), space (hypervelocity missiles to satellites), marine (ships and submarines), civil (hydroelectric turbines to fish ladders to water treatment tanks) and, not to bore you too much, ASME Section VIII, Division 2, “Design-by-Analysis” pressure vessel work (beer kegs to nuclear waste processing vessels under seismic and fatigue).  It is a long list and it just continues to grow.

Okay, why all this pre-amble?  Our clients’ intellectual property (IP) often represents their “crown jewels” and protection of this data is something we take very seriously.  Although Predictive is ITAR-Registered, we treat all our clients’ data as if it were ITAR data.  What does this really mean at the end of the day?  It means we use best practices to ensure that no harm comes to any data, for example, using the “Dutch Reach” method when opening a car door to prevent an accident (i.e., data loss) and likewise, being “Safe” with our data and having backups that are under “lock and key”.  It sounds a bit silly, but good data security is just about best practices and thinking of others, since how would you feel if someone infected you with COVID by not wearing a mask?