I think weight reduction is the main reason. Reciprocating masses upset dynamic balance, thump thump, and accelerating excess weight is wasted work. Best to reduce both effects by removing metal that doesn't contribute to strength.

A double acting steam engine conrod in tension need only be thick enough to take pull forces because there's no tendency to buckle. In compression, pushing, the same conrod has to be strong enough to prevent buckling. Compressed rods are weakest at the centre point and least likely to buckle at the ends. So it makes sense to size the ends small for tension and the middle big to resist buckling forces.

Simple tapering seems most evident on big stationary steam engines probably because their conrods are relatively long thin affairs. Locomotive conrods are comparatively stubby but heavy, and their geometry may favour flatter conrods, in which case girder forms and weight reducing holes may be easier to make and stronger than a taper. IC engines are even stubbier. As smooth curves are good at reducing fatigue, I'd guess conrod design uses several ways of reducing weight whilst managing stress. Be interesting to FEM model a few different designs to see what the pretty colours say.

Dave

Edited By SillyOldDuffer on 14/07/2020 15:02:48

Two loads are of interest: compression due to steam pressure and accelerating the piston, etc. And transverse loads due to the mass of the con rod being accelerated up and down by the crankpin. So a classical pin ended strut with compression and transverse load.

The transverse load gives a bending moment in the rod, zero at both ends and maximum, not in the middle, but offset more towards the crank pin. Hence con rods normally being an I-beam (or H-beam on some car engines), often tapering towards the little end. For the loco rods, I dare say that asthetics and ease of machining dominate the design.

Dave

Considering railway locomotives, the connecting rods are long and thin compared with most IC engines (that is really stating the obvious). They are subjected to a number of loads, compression and tension along the rod and their own weight. At speed this weight (remember, equals mass x acceleration and acceleration increases with speed) may be very high. The rod would be a simply supported beam with a continuous load with the maximum deflection would half way along it. The deflection can be reduced by increasing its second moment of area (that is B x D^3 where D is the thickness of the rod in plane of motion and B is the thickness across the rod). Considering the mass of the rod is B x D (per unit length) thickening the rod where it is going to deflect the most has a major effect on the stiffness.

IC engine rods are rather different, because they are short, but I believe most fail in bending close to the little end.

I am sorry if the above is rambling.

JA

Edited By JA on 14/07/2020 17:40:53