The new press-together crankshaft shown below may be of topical interest.
(To see how it works directly, scroll vid across to ~2:05 mark):
Here's how R/R engineers did it for their V12 cranks - back in the day - via mental computation,Mudflap wrote: ↑Tue Oct 06, 2020 10:24 pmIt's funny how technology has evolved. When I started in the late 90's we used spreadsheets to calculate crankshaft bending loads as you say i.e. treating the crank as a statically determinate beam. We had simple frequency domain in-house codes to work out torsional loads and the FEA was done one HALF throw at a time.Greg Locock wrote: ↑Tue Oct 06, 2020 9:40 pmThat Ferrari F430 one is interesting. Notice how the counterweights are symmetrical for each throw? That's an old, and not especially efficient trick, to avoid having to do FEA and stuff. Basically treat each throw as a single cylinder engine. For a production engine your aim is usually trying to minimise the maximum load in the mains, and you can normally get a solution that achieves that with less mass for a given load by considering the crank as a flexible whole. FWIW it was about 2 kg on a straight 6.
Nowadays we just flush an entire crank through commercial multi body dynamics codes and fatigue solvers and it spits out factors of strength. It's probably twice as fast but when it goes wrong it can take days to debug.